TRAINING INTENSITY, VOLUME

AND RECOVERY DISTRIBUTION

AMONG ELITE AND RECREATIONAL
ENDURANCE ATHLETES

EDITED BY : Thomas L. Stöggl and Billy Sperlich

PUBLISHED IN: Frontiers in Physiology
 Frontiers Copyright Statement
About Frontiers
© Copyright 2007-2019 Frontiers
Media SA. All rights reserved. Frontiers is more than just an open-access publisher of scholarly articles: it is a
All content included on this site,
pioneering approach to the world of academia, radically improving the way scholarly
such as text, graphics, logos, button
icons, images, video/audio clips, research is managed. The grand vision of Frontiers is a world where all people have
downloads, data compilations and
an equal opportunity to seek, share and generate knowledge. Frontiers provides
software, is the property of or is
licensed to Frontiers Media SA immediate and permanent online open access to all its publications, but this alone
(“Frontiers”) or its licensees and/or
subcontractors. The copyright in the
is not enough to realize our grand goals.
text of individual articles is the property
of their respective authors, subject to a
license granted to Frontiers.
Frontiers Journal Series
The compilation of articles constituting The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access,
this e-book, wherever published,
as well as the compilation of all other
online journals, promising a paradigm shift from the current review, selection and
content on this site, is the exclusive dissemination processes in academic publishing. All Frontiers journals are driven
property of Frontiers. For the
conditions for downloading and
by researchers for researchers; therefore, they constitute a service to the scholarly
copying of e-books from Frontiers’ community. At the same time, the Frontiers Journal Series operates on a revolutionary
website, please see the Terms for
Website Use. If purchasing Frontiers invention, the tiered publishing system, initially addressing specific communities of
e-books from other websites scholars, and gradually climbing up to broader public understanding, thus serving
or sources, the conditions of the
website concerned apply. the interests of the lay society, too.
Images and graphics not forming part
of user-contributed materials may
Dedication to Quality
not be downloaded or copied
without permission. Each Frontiers article is a landmark of the highest quality, thanks to genuinely
Individual articles may be downloaded
and reproduced in accordance
collaborative interactions between authors and review editors, who include some
with the principles of the CC-BY of the world’s best academicians. Research must be certified by peers before entering
licence subject to any copyright or
other notices. They may not be re-sold
a stream of knowledge that may eventually reach the public - and shape society;
as an e-book. therefore, Frontiers only applies the most rigorous and unbiased reviews.
As author or other contributor you Frontiers revolutionizes research publishing by freely delivering the most outstanding
grant a CC-BY licence to others to
reproduce your articles, including any research, evaluated with no bias from both the academic and social point of view.
graphics and third-party materials By applying the most advanced information technologies, Frontiers is catapulting
supplied by you, in accordance with
the Conditions for Website Use and scholarly publishing into a new generation.
subject to any copyright notices which
you include in connection with your
articles and materials. What are Frontiers Research Topics?
All copyright, and all rights therein,
are protected by national and
Frontiers Research Topics are very popular trademarks of the Frontiers Journals
international copyright laws. Series: they are collections of at least ten articles, all centered on a particular subject.
The above represents a summary only. With their unique mix of varied contributions from Original Research to Review
For the full conditions see the
Conditions for Authors and the Articles, Frontiers Research Topics unify the most influential researchers, the latest
Conditions for Website Use. key findings and historical advances in a hot research area! Find out more on how
ISSN 1664-8714
ISBN 978-2-88945-966-7
to host your own Frontiers Research Topic or contribute to one as an author by
DOI 10.3389/978-2-88945-966-7 contacting the Frontiers Editorial Office: [email protected]

Frontiers in Physiology 1 August 2019 | Polarization-Index

 TRAINING INTENSITY, VOLUME
AND RECOVERY DISTRIBUTION
AMONG ELITE AND RECREATIONAL
ENDURANCE ATHLETES

Topic Editors:
Thomas L. Stöggl, University of Salzburg, Austria
Billy Sperlich, University of Würzburg, Germany

Citation: Stöggl, T. L., Sperlich, B., eds. (2019). Training Intensity, Volume and Recovery
Distribution Among Elite and Recreational Endurance Athletes. Lausanne: Frontiers
Media. doi: 10.3389/978-2-88945-966-7

Frontiers in Physiology 2 August 2019 | Polarization-Index

 Table of Contents
05 The Polarization-Index: A Simple Calculation to Distinguish Polarized
From Non-polarized Training Intensity Distributions
Gunnar Treff, Kay Winkert, Mahdi Sareban, Jürgen M. Steinacker and
Billy Sperlich
11 Editorial: Training Intensity, Volume and Recovery Distribution Among
Elite and Recreational Endurance Athletes
Thomas L. Stöggl and Billy Sperlich
15 Relation Between Training Load and Recovery-Stress State in
High-Performance Swimming
Robert Collette, Michael Kellmann, Alexander Ferrauti, Tim Meyer and
Mark Pfeiffer
29 Effects of Initial Performance, Gross Efficiency and VO2peak
Characteristics on Subsequent Adaptations to Endurance Training
in Competitive Cyclists
Knut Skovereng, Øystein Sylta, Espen Tønnessen, Daniel Hammarström,
Jørgen Danielsen, Stephen Seiler, Bent R. Rønnestad and Øyvind Sandbakk
38 Adaptive Changes After 2 Weeks of 10-s Sprint Interval Training With
Various Recovery Times
Robert A. Olek, Sylwester Kujach, Ewa Ziemann, Wieslaw Ziolkowski,
Piotr Waz and Radoslaw Laskowski
46 Impact of Short-Term Training Camp on Aortic Blood Pressure in
Collegiate Endurance Runners
Tsubasa Tomoto, Jun Sugawara, Ai Hirasawa, Tomoko Imai, Seiji Maeda and
Shigehiko Ogoh
52 The Training Characteristics of the World’s Most Successful Female
Cross-Country Skier
Guro S. Solli, Espen Tønnessen and Øyvind Sandbakk
66 Massage Alleviates Delayed Onset Muscle Soreness After Strenuous
Exercise: A Systematic Review and Meta-Analysis
Jianmin Guo, Linjin Li, Yuxiang Gong, Rong Zhu, Jiake Xu, Jun Zou and
Xi Chen
78 Upper-Body Muscular Endurance Training Improves Performance
Following 50 min of Double Poling in Well-Trained Cross-Country Skiers
Jørgen Børve, Steffen N. Jevne, Bjarne Rud and Thomas Losnegard
88 Sedentary Behavior Among National Elite Rowers During
Off-Training—A Pilot Study
Billy Sperlich, Martin Becker, Andreas Hotho, Birgit Wallmann-Sperlich,
Mahdi Sareban, Kay Winkert, Jürgen M. Steinacker and Gunnar Treff
94 The Responses of Elite Athletes to Exercise: An All-Day, 24-h Integrative
View is Required!
Billy Sperlich and Hans-Christer Holmberg

Frontiers in Physiology 3 August 2019 | Polarization-Index

 97 High Intensity Interval Training Leads to Greater Improvements in Acute
Heart Rate Recovery and Anaerobic Power as High Volume Low Intensity
Training
Thomas L. Stöggl and Glenn Björklund
105 Eleven-Week Preparation Involving Polarized Intensity Distribution is not
Superior to Pyramidal Distribution in National Elite Rowers
Gunnar Treff, Kay Winkert, Mahdi Sareban, Jürgen M. Steinacker,
Martin Becker and Billy Sperlich
116 The Mucosal Immune Function is not Compromised During a Period
of High-Intensity Interval Training. Is it Time to Reconsider an old
Assumption?
Dennis-Peter Born, Christoph Zinner and Billy Sperlich
125 No Additional Benefits of Block- Over Evenly-Distributed High-Intensity
Interval Training Within a Polarized Microcycle
Kerry McGawley, Elisabeth Juudas, Zuzanna Kazior, Kristoffer Ström,
Eva Blomstrand, Ola Hansson and Hans-Christer Holmberg
137 Is Marathon Training Harder Than the Ironman Training? An ECO-method
Comparison
Jonathan Esteve-Lanao, Diego Moreno-Pérez, Claudia A. Cardona,
Eneko Larumbe-Zabala, Iker Muñoz, Sergio Sellés and Roberto Cejuela
145 Intensity- and Duration-Based Options to Regulate Endurance Training
Peter Hofmann and Gerhard Tschakert
154 Training Intensity Distribution and Changes in Performance and
Physiology of a 2nd Place Finisher Team of the Race Across America
Over a 6 Month Preparation
Christian Manunzio, Joachim Mester, Walter Kaiser and Patrick Wahl

Frontiers in Physiology 4 August 2019 | Polarization-Index

 HYPOTHESIS AND THEORY
published: 12 June 2019
doi: 10.3389/fphys.2019.00707

The Polarization-Index: A Simple

Calculation to Distinguish Polarized
From Non-polarized Training
Intensity Distributions
Gunnar Treff 1* , Kay Winkert 1 , Mahdi Sareban 2 , Jürgen M. Steinacker 1 and Billy Sperlich 3
1
Division of Sports and Rehabilitation Medicine, Ulm University Hospital, Ulm, Germany, 2 Institute of Sports Medicine,
Prevention and Rehabilitation, Paracelsus Medical University, Salzburg, Austria, 3 Integrative and Experimental Exercise
Science and Training, Institute of Sport Science, University of Würzburg, Würzburg, Germany

The training intensity distribution (TID) of endurance athletes has retrieved substantial
scientific interest since it reflects a vital component of training prescription: (i) the intensity
of exercise and its distribution over time are essential components for adaptation to
endurance training and (ii) the training volume (at least for most endurance disciplines)
is already near or at maximum, so optimization of training procedures including TID
have become paramount for success. This paper aims to elaborate the polarization-
index (PI) which is calculated as log10 (Zone 1/Zone 2∗ Zone 3∗ 100), where Zones 1–3
refer to aggregated volume (time or distance) spent with low, mid, or high intensity
Edited by:
training. PI allows to distinguish between non-polarized and polarized TID using a cut-
François Billaut, off > 2.00 a.U. and to quantify the level of a polarized TID. Within this hypothesis paper,
Laval University, Canada
examples from the literature illustrating the usefulness of PI-calculation are discussed
Reviewed by:
as well as its limitations. Further it is elucidated how the PI may contribute to a more
Pantelis Theodoros Nikolaidis,
University of West Attica, Greece precise definition of TID descriptors.
Kai Roecker,
Hochschule Furtwangen University, Keywords: high-intensity training, high-performance sports, lactate threshold training, endurance training, elite
Germany
*Correspondence:
Gunnar Treff
INTRODUCTION
[email protected]
The training intensity distribution (TID) of endurance athletes has become an important
Specialty section: component of training prescription since (i) the intensity of exercise and its distribution over time
This article was submitted to are essential components of adaptation to endurance training and (ii) the training volume (at least
Exercise Physiology, for most disciplines) is already near or at maximum. Therefore, several optimization procedures
a section of the journal have gained scientific interest, including the manipulation of TID, and other components of
Frontiers in Physiology exercise prescription including exercise duration, volume, frequency, or mode.
Received: 29 October 2018 To quantify TID, the intensity of exercise is commonly defined according to physiological
Accepted: 20 May 2019 thresholds and distributed into an “intensity zone-model,” of which a three-zone model is
Published: 12 June 2019
predominantly employed for scientific evaluation. Briefly, the intensity of Zone 1 incorporates low-
Citation: intensity exercise greater than or equal to 50% of maximal oxygen uptake (V̇O2 max ) and lower than
Treff G, Winkert K, Sareban M, the intensity corresponding to the first lactate or ventilatory threshold. Exercise prescription in
Steinacker JM and Sperlich B (2019)
Zone 1 is often termed “basic-endurance” or “low-intensity” exercise. The first and second lactate or
The Polarization-Index: A Simple
Calculation to Distinguish Polarized
ventilatory thresholds define the lower and upper limits of Zone 2, an exercise intensity that is often
From Non-polarized Training Intensity termed “threshold intensity,” or “lactate threshold training.” Zone 3 is usually defined as an exercise
Distributions. Front. Physiol. 10:707. intensity greater than the second lactate or ventilatory threshold and established as high-intensity
doi: 10.3389/fphys.2019.00707 interval training near or at maximum V̇O2 max . These training intensities may also be defined

Frontiers in Physiology | www.frontiersin.org 5 June 2019 | Volume 10 | Article 707

Treff et al. Polarization-Index

by other variables based on blood lactate concentration, descriptor for pyramidal TIDs (Plews and Laursen, 2017) which
percentage of maximal heart rate or V̇O2 max , or subjective ratings are clearly characterized by decreasing proportions of Zone 1,
like the “Session-RPE”. For details see Seiler and Kjerland (2006), 2, and 3, or for TIDs that do not differentiate between Zone
Seiler (2010). However, the physiological transitions between 2 and Zone 3 (Fiskerstrand and Seiler, 2004), thereby violating
training intensities are fluent and the targeted adaptions depend the aforementioned TID classification. Therefore, the definition
on multiple factors including training volume, TID, health-, and of polarized vs. other non-polarized TID is often unclear and
training-status. For detailed reviews see Gibala et al. (2012), sometimes misleading.
Milanović et al. (2015), and MacInnis and Gibala (2016). For this reason, we would like to present an elaborated
Among several TID patterns, four main distributions have concept of our previously published polarization-index (PI)
been reported and investigated so far, namely the “polarized”-, (Treff et al., 2017), which is based on the assumption of
“high-intensity”-, “pyramidal”-, and “lactate threshold”-TID, two necessary conditions for a polarized TID. (i) a polarized
which are based on previous definitions (Seiler and Kjerland, structure, where Zone 1 > Zone 3 and Zone 3 > Zone 2
2006; Stöggl and Sperlich, 2015): (and consequently Zone 1 > Zone 2) and (ii) a relatively small
proportion of Zone 2. The PI aims to distinguish between
• Polarized TID consists of elevated percentages of time or polarized and non-polarized TID and to quantify the level of a
distance spent in both high- (Zone 3) and low-intensity polarized TID. Further, we aim to highlight the PI’s usefulness
exercise (Zone 1) and only a small proportion of training and limitations, thereby contributing to a more precise TID
in Zone 2. The polarized TID with its fractions of training terminology within the scientific literature. Based on studies
volume spent at low-, threshold-, and high-intensity often published between 2009 and 2018 and reported in our previous
consists of e.g., 80% of training volume spent in Zone 1, 5% paper, we want to highlight examples illustrating why we believe
in Zone 2, and 15% in Zone 3 (80-5-15), or 75-5-20, (i.e., the polarization index may be a valuable tool for practical and
75% within Zone 1, 5% in Zone 2, and 20% in Zone 3), scientific purposes.
with percentages of Zone 1 greater than Zone 3 and Zone
3 always greater than Zone 2.
• Pyramidal TID consists of high percentage of training CALCULATION OF THE
volume spent in Zone 1 and less proportions in Zone 2 POLARIZATION-INDEX
and 3. As an example, a pyramidal TID may be quantified
as 70-20-10, i.e., 70% within Zone 1, 20% in Zone 2, The formula for calculation of the PI is based on a three-zone
and 10% in Zone 3. TID-model:
• Threshold TID consists of training volume emphasizing
Zone 2. This distribution is frequently established by Polarization − index (a.U.)
longer intervals with an intensity between first and second = log 10 (Zone 1/Zone 2 × Zone 3 ∗ 100) (1)
lactate or ventilatory threshold or by continuous exercise
intermixed with higher intensities and without a distinct where Zone is the fraction (given percentage/100) of the training
recovery interval. As an example, a threshold TID could be volume in Zone 1, 2, and 3.
designed as 40-50-10 (i.e., 40% within Zone 1, 50% in Zone The PI increases if a high ratio of Zone 1 to Zone 2 is
2, and 10% in Zone 3). Notably, a threshold TID, e.g., 50-45- combined with a high percentage of training in Zone 3. The log-
5 (i.e., 50% within Zone 1, 45% in Zone 2, and 5% in Zone transformation of the raw-data establishes a quasi linear function.
3), may but not necessarily has to be pyramidal (i.e., with If Zone 2 = 0, Eq. 2 avoids zero in the denominator:
decreasing proportions of Zone 2 and Zone 3).
• High Intensity TID is a TID with training predominantly Polarization − index (a.U.)
performed in Zone 3 and mainly involving interval training. = log 10 (Zone 1/0.01 × Zone 3 − 0.01 ∗ 100) (2)
A typical high-intensity TID could be designed as 20-10-70
(i.e., 20% within Zone 1, 10% in Zone 2, and 70% in Zone 3). If Zone 3 = 0, PI is zero per definition.
If Zone 3 > Zone 1 the PI is not valid and must not be
Notably, the classification of TIDs to one of the four patterns calculated (please see discussion for details below).
shown in Figure 1 maybe ambiguous. Especially the term If PI > 2.00 a.U., the TID is defined as “polarized,” with
“polarized” differs substantially between publications (Stöggl increasing values indicating a higher level of polarization. If PI
and Sperlich, 2015; Plews and Laursen, 2017), nevertheless is ≤ 2.00 a.U., the TID is defined as non-polarized.
the polarized TID has received increasing scientific interest
since retrospective analysis (Seiler and Kjerland, 2006), and
prospective randomized-controlled trials have documented equal JUSTIFICATION OF THE PI-CONCEPT
(Ingham et al., 2008; Treff et al., 2017) or superior gains in
endurance performance (Neal et al., 2013; Stöggl and Sperlich, For a polarized TID we assume (i) a polarized structure and (ii) a
2014; Tønnessen et al., 2014) when compared to the pyramidal, relatively small proportion of Zone 2.
threshold, or high-intensity TIDs. “Polarized” TID comprises a Ad (i) For the polarized structure, we agree on the
variety of fractions of Zone 1-3 and is sometimes even used as a following necessary conditions: a: Zone 1 + Zone 2 +

Frontiers in Physiology | www.frontiersin.org 6 June 2019 | Volume 10 | Article 707

Treff et al. Polarization-Index

FIGURE 1 | Various training intensity distributions (TID), their schematic proportions, and key characteristics (indicated by black bars). Zones refer to following
intensities: Zone 1 (basic endurance), ≥ 50% V̇O2 max and ≤ first lactate or ventilatory threshold; Zone 2 (lactate threshold), ≥ first and ≤ second lactate or ventilatory
threshold; Zone 3 (high intensity), > second lactate or ventilatory threshold.

Zone 3 = 1, b: Zone 3 > Zone 2, c: Zone 1 > Zone 3, and d: Zone ∧ Zone 3 > Zone 2 will be met in each case for Z1 ≤ 100%
1 > Zone 2. Figure 2 visualizes these conditions over the range (Figure 2, green area).
0–1 (or 0 – 100%) and the colored areas represent the range of Ad (ii) At the same time the 2.00-threshold can be employed
fractions where all of these conditions are fulfilled. to identify the fulfillment of the second necessary condition for a
Due to these conditions, a PI > 2.00 a.U. is inevitably polarized TID, i.e., comprising of “a relatively small” proportion
associated with a polarized structure, because if Zone 2 equals of Zone 2: If, for example, the percentage of Zone 1 is as low as
Zone 3 (a TID which does precisely not stand for a polarized 60%, a TID of 60-19-21 will result in a PI of 1.82 a.U., thereby
structure), the result of Zone 1/Zone 2 × Zone 3 will equal the clearly below the cut-off and violating the definition of a polarized
value of Zone 1, since Zone 2 and Zone 3 will shorten each other, TID, even though Zone 1 > Zone 3 ∧ Zone 3 > Zone 2 (i.e., the
e.g., 0.8/0.2 × 0.2 ∗ 100 = 80. Zone 1 approaches the maximal first necessary condition for a polarized TID is true). However,
value of 100%, therefore the raw, i.e., not log-transformed PI if Zone 2 is lower than ∼ 15% in this example, the PI will
will approach a value of 100. Since log10 (100) equals 2.00, the be > 2.00 a.U. and consequently, PI calculates 2.05 a.U. in a
PI approaches 2.00 a.U. and cannot exceed a value > 2.00 a.U. 60-14-26 distribution.
if Zone 2 equals Zone 3. Consequently, the percentage of Zone 3 Of note, with higher percentages of Zone 1 training (e.g.,
must be higher than Zone 2 to result in a PI > 2.00 a.U. Or, vice 80%), a Zone 2 percentage lower than 9% will already allow for
versa: If PI > 2.00 a.U., the necessary condition Zone1 > Zone3 a polarized TID (e.g., 80-8-12) resulting in a PI of 2.08 a.U.

Frontiers in Physiology | www.frontiersin.org 7 June 2019 | Volume 10 | Article 707

Treff et al. Polarization-Index

(Neal et al., 2013) reports a 80-0-20 polarized TID evident with

a high PI of 3.18 a.U. In one study (Ingham et al., 2008) the
index is even higher, due to lower and greater fractions of
Zones 1 and 3, respectively. In three studies the PI was lower
compared to the TID of the aforementioned studies (i.e., the level
of “polarization” was lessened), due to the greater percentage of
Zone 2 (Bourgois et al., 2013; Stöggl and Sperlich, 2014) or lower
absolute percentage spent in Zone 3 (Treff et al., 2017).
In two studies of Table 1 (Neal et al., 2013; Stöggl and Sperlich,
2014) the percentage of training spent in Zone 3 equaled zero.
According to the aforementioned definition, the PI amounts to
zero and consequently the TID of both studies are not polarized.
In five studies of Table 1 (Guellich et al., 2009; Plews et al.,
2014; Stöggl and Sperlich, 2014; Plews and Laursen, 2017; Treff
et al., 2017) the PI varied from 0.71 to 1.80 a.U. representing
non-polarized TIDs due to a PI ≤ 2.0 a.U. However, the PI varies
reasonably according to the respective contributions of Zone 1,
Zone 2, and Zone 3. In detail, the PI is very similar in two
examples (Stöggl and Sperlich, 2014; Plews and Laursen, 2017)
but clearly higher compared to other studies (Plews et al., 2014;
Treff et al., 2017) with a similar fraction of Zone 2, but a nearly
threefold higher percentage of Zone 3 at the expense of a lower
FIGURE 2 | Limits and conditions of the polarization index. Line a indicates
percentage in Zone 1.
the upper limit of Zone 3 for a given (x-axis) fraction of Zone 1 (condition a:
Zone 1 + Zone 2 + Zone 3 = 1), line b indicates the lower limit of Zone 3 for a
Table 1 also illustrates a special TID variant, as the study
given fraction of Zone 1 (condition b: Zone 3 > Zone 2), line c indicates the by Carnes and Mahoney (2018) represents a TID (74-11-15) in
upper limit of Zone 3 for a given fraction of Zone 1 (condition c: Zone 1 > Zone which the percentage of Zone 3 is higher compared to Zone 2,
3), and line d indicates the lower limit of Zone 3 for a given fraction of Zone 1 indicating a polarized structure, however the fraction of time
(condition d: Zone 1 > Zone 2). The colored areas represent the range of
spent in Zone 2 is considerably high, thereby not fulfilling the
values where conditions a-d are fulfilled. The gray broken line indicates the
limit of Zone 3 for a given fraction of Zone 1 resulting in a polarization-index of second necessary criterion for a polarized TID (Zone 2 being
2.00 a.U., i.e., if Zone 3 is higher or lower, polarization-index will result in relatively small), which is mirrored by the PI of 2.00 a.U.,
values higher (green area) or lower than 2.00 a.U. (yellow area), respectively. indicating a non-, but “nearly”-polarized TID.
The broken blue line indicates the upper limit of Zone 2 for a given fraction of
Zone 1 to allow for a polarization-index ≥ 2.00 a.U.

CLASSIFICATION AND DETAILED

This behavior is visualized by the broken blue line in Figure 2, QUANTIFICATION OF TRAINING
indicating the upper limit of Zone 2 allowing for a PI ≥ 2.00 a.U. INTENSITY DISTRIBUTIONS IN THE
and the approximation of the broken lines to line b with
increasing fractions of Zone 1.
LITERATURE
The cut-off > 2.00 a.U. is therefore not arbitrary and clearly As already mentioned, the term “polarized” superficially
providing a benchmark for a polarized structure and an objective describes the TIDs within retrospective training analysis and
(but not physiological) definition of “relatively small” percentages prospective experiments. For example, the “polarized” TIDs of
of Zone 2 training in polarized TIDs. the studies summarized in Table 1 (Ingham et al., 2008; Bourgois
et al., 2013; Neal et al., 2013; Stöggl and Sperlich, 2014) report
substantial differences in the fraction of Zone 3 ranging from 6
APPLICATION OF THE POLARIZATION to 28%. Taking into account that a large body of evidence has
INDEX IN TRAINING INTENSITY revealed that training emphasizing Zone 3 promotes substantial
STUDIES differences in oxygen transport and utilization (Hickson et al.,
1977; Milanović et al., 2015), it seems important to know how
Table 1 shows several original investigations published between “polarized” an experiment was in order to judge the level of
2009 and 2018 which were discussed in one of our previous adaptation in connection with various TIDs, and to allow better
papers (Treff et al., 2017), their PI and the employed descriptor. comparisons between groups and studies.
Five studies of Table 1 (Ingham et al., 2008; Bourgois Table 1 also provides examples of studies reporting a polarized
et al., 2013; Neal et al., 2013; Stöggl and Sperlich, 2014; Treff TID, even though the percentage of Zone 2 is considerably high
et al., 2017) report polarized TIDs according to the PI and (≥17%) and proportions of Zone 1 to 3 continuously decrease,
the aforementioned definitions i.e., Zone 1 > Zone 2 ∧ Zone thereby clearly indicating a pyramidal TID (Plews et al., 2014).
3 > Zone 2 as well as low percentage of Zone 2. One study The borderline TID by Carnes and Mahoney (2018) also claims

Frontiers in Physiology | www.frontiersin.org 8 June 2019 | Volume 10 | Article 707

Treff et al. Polarization-Index

TABLE 1 | Selected studies in the area training intensity distribution, percentages of three training intensity zones, the resulting polarization-index, and their classification
according to the polarization-index and the original publication.

Authors Time or distance spent in Polarization-Index Classification according to

(a.U.)
Zone 1 (%) Zone 2 (%) Zone 3 (%) Polarization-Index Authors

1 Neal et al., 2013 80.0 0.0 20.0 3.18 Polarized Polarized

2 Ingham et al., 2008 72.0 0.0 28.0 3.29 Polarized Polarized
3 Stöggl and Sperlich, 2014 68.0 6.0 26.0 2.47 Polarized Polarized
4 Bourgois et al., 2013 93.1 2.3 4.6 2.27 Polarized Not classified
5 Treff et al., 2017 93.0 1.0 6.0 2.75 Polarized Polarized
6 Neal et al., 2013 57.0 43.0 0.0 0.00 Non-polarized Non-polarized (LT)
7 Stöggl and Sperlich, 2014 46.0 54.0 0.0 0.00 Non-polarized Non-polarized (LT)
8 Plews and Laursen, 2017 67.3 30.2 2.5 0.75 Non-polarized Non-polarized (pyramidal)
9 Stöggl and Sperlich, 2014 83.0 16.0 1.0 0.71 Non-polarized Non-polarized (HVLT Int)
10 Plews and Laursen, 2017 80.4 17.9 1.8 0.91 Non-polarized Polarized
11 Plews et al., 2014 77.3 16.9 5.8 1.42 Non-polarized Polarized
12 Treff et al., 2017 94.0 4.0 2.0 1.67 Non-polarized Non-polarized (pyramidal)
13 Guellich et al., 2009 95.0 3.0 2.0 1.80 Non-polarized Non-polarized
14 Carnes and Mahoney, 2018 74.0 11.0 15.0 2.00 Non-polarized Polarized
15 Stöggl and Sperlich, 2014 43.0 0.0 57.0 n.a. – HIT

LT, lactate threshold training; Hi HVLT, high volume low intensity; HIT, high intensity training. Bold letters indicate that classification according to polarization-index differs
from classification in the original publication.

to have employed a “polarized” TID. In each of these studies, the or training modalities) are clamped, or are at least similar
application of the PI would provide a more precise and objective between subjects or within subject. In this case, we would like to
classification of the respective TIDs. emphasize that the PI as an algorithm assists in discriminating
between various TIDs, but we discourage the interpretation as a
surrogate for training load. For example, a PI of 2.00 a.U. may
APPLICATION OF THE result out of two substantially different TIDs, e.g., 90-5-5 and 74-
POLARIZATION-INDEX IN TRAINING 13-13. As explained above, and from a biological and empirical
MONITORING AND ANALYSIS perspective, these two training regimes, unlike TID, will result
in different central and peripheral adaptations and will affect
Figure 3 is an example retrieved from a previous paper of our performance differently, even if applied in a theoretically perfect
group, illustrating the practical application of the PI in a scientific model, i.e., two identical subjects. It is also worth to mention, that
study simultaneously illustrating the practical application of the the quality of training data and a reliable and valid allocation
PI for training monitoring, and analysis with two groups of of the intensities is crucial for analysis. Finally, successful
rowers performing either a pyramidal (PI = 1.70 a.U.) or a training is not only a quantitatively but also strongly qualitatively
polarized TID (PI = 2.70 a.U.) (Treff et al., 2017). As in similar
training studies, the TIDs differed significantly between groups,
but at the individual level, the training was quite heterogenous.
The PI therefore might allow for a more precise analysis of
training outcomes taking into account the actual individual TID.
Further, the PI can easily be integrated into standard training
monitoring software.
However, some rules should be followed when interpreting
the PI:
As shown before, very small differences between Zone 2 and
Zone 3 will allow for a PI ≥ 2.00 a.U. if percentage of Zone 1 is
high. Therefore, the PI is practically useful within reasonable and
accepted limits for Zone 1 in polarized TIDs, being approximately
70–90% (Stöggl and Sperlich, 2014). Also, a given TID with, for
example, 15 h/week training will affect performance differently
to the same TID with a volume of, for example, 25 h/week. FIGURE 3 | Percentage change of average power in 2000 m rowing
ergometer test (P2000 m ) in internationally competing rowers. Vertical dashed
Therefore, interpretation of changes in performance in relation to
line represents the cut-off between non-polarized (≤ 2.00 a.U.) and polarized
a given TID (as illustrated in Figure 3) is only justified when other (> 2.00 a.U.) TIDs. Figure adapted from Treff et al. (2017).
important variables of training (e.g., training volume, frequency,

Frontiers in Physiology | www.frontiersin.org 9 June 2019 | Volume 10 | Article 707

Treff et al. Polarization-Index

determined intervention and not reflected by the PI or other CONCLUSION

quantitative training variables.
The application of the PI represents an algorithm to
distinguish distinctively between polarized and non-polarized
LIMITATIONS TIDs and to judge the level of polarization. As shown,
the PI has the potential to reduce the ambiguity regarding
Despite the practical usefulness, the PI has some limitations the classification of TIDs in the current literature and
that warrant a brief discussion. Even though the PI provides is easily applicable in any training monitoring software.
an objective cut-off to distinguish polarized from non-polarized Since the PI also allows intra-individual assessment of
distributions, it does not allow the differentiation of sub-types polarization, we aim to stimulate researchers to re-evaluate
of the non-polarized TID structures (for example, lactate- their existing data retrospectively in order to investigate
threshold vs. high-intensity TID) and values between 0 and the response or non-response to various TIDs. In addition,
2.00 must not be interpreted in terms of more or less polarized we encourage future training experiments to state the
distributions. Furthermore, from a theoretical perspective, it level of polarization allowing for more detailed comparison
appears inappropriate to replace 0.00% in Zone 2 by e.g., 1.00% between studies.
to avoid zero in the denominator (Eq. 2). However, from a
practical perspective, it is virtually impossible to achieve high
intensities (i.e., Zone 3), without some fraction of time spent
in Zone 2. Therefore, this limitation appears to be practically AUTHOR CONTRIBUTIONS
irrelevant. Finally, the PI is a statistical measure that increases
data density and thereby - like every index - leads to a loss of GT drafted the manuscript. GT, KW, MS, JS, and BS contributed
detailed information. substantially to the manuscript.

REFERENCES Plews, D. J., and Laursen, P. B. (2017). Training intensity distribution over a four-
year cycle in olympic champion rowers: different roads lead to rio. Int. J. Sports
Bourgois, J., Steyaert, A., and Boone, J. (2013). Physiological and Physiol. Perform. 27, 1–24. doi: 10.1123/ijspp.2017-0343
anthropometric progression in an international oarsman: a 15 year case Plews, D. J., Laursen, P. B., Kilding, A. E., and Buchheit, M. (2014). Heart-
study. Int. J. Sports Physiol. Perform. 9, 723–726. doi: 10.1123/ijspp.2013- rate variability and training-intensity distribution in elite rowers. Int. J. Sports
0267 Physiol. Perform. 9, 1026–1032. doi: 10.1123/ijspp.2013-0497
Carnes, A. J., and Mahoney, S. E. (2018). Polarized vs. high intensity multimodal Seiler, K. S., and Kjerland, G. Ø. (2006). Quantifying training intensity distribution
training in recreational runners. Int. J. Sports Physiol. Perform. 28, 1–28. doi: in elite endurance athletes: is there evidence for an “optimal” distribution?
10.1123/ijspp.2018-0040 Scand. J. Med. Sci. Sports 16, 49–56. doi: 10.1111/j.1600-0838.2004.00418.x
Fiskerstrand, A., and Seiler, K. S. (2004). Training and performance characteristics Seiler, S. (2010). What is best practice for training intensity and duration
among Norwegian international rowers 1970-2001. Scand. J. Med. Sci. Sports distribution in endurance athletes? Int. J. Sports Physiol. Perform. 5, 276–291.
14, 303–310. doi: 10.1046/j.1600-0838.2003.370.x doi: 10.1123/ijspp.5.3.276
Gibala, M. J., Little, J. P., Macdonald, M. J., and Hawley, J. A. (2012). Stöggl, T., and Sperlich, B. (2014). Polarized training has greater impact on key
Physiological adaptations to low-volume, high-intensity interval training in endurance variables than threshold, high intensity, or high volume training.
health and disease. J. Physiol. 590, 1077–1084. doi: 10.1113/jphysiol.2011.22 Front. Physiol. 5:33. doi: 10.3389/fphys.2014.00033
4725 Stöggl, T. L., and Sperlich, B. (2015). The training intensity distribution among
Guellich, A., Seiler, S., and Emrich, E. (2009). Training methods and intensity well-trained and elite endurance athletes. Front. Physiol. 6:295. doi: 10.3389/
distribution of young world-class rowers. Int. J. Sports Physiol. Perform. 4, fphys.2015.00295
448–460. doi: 10.1123/ijspp.4.4.448 Tønnessen, E., Sylta, Ã., Haugen, T. A., Hem, E., Svendsen, I. S., and Seiler, S.
Hickson, R. C., Bomze, H. A., and Holloszy, J. O. (1977). Linear increase in (2014). The road to gold: training and peaking characteristics in the year prior
aerobic power induced by a strenuous program of endurance exercise. J. Appl. to a gold medal endurance performance. PLoS One 9:e101796. doi: 10.1371/
Physiol. Respir. Environ. Exerc. Physiol. 42, 372–376. doi: 10.1152/jappl.1977.42. journal.pone.0101796
3.372 Treff, G., Winkert, K., Sareban, M., Steinacker, J. M., Becker, M., and Sperlich, B.
Ingham, S. A., Carter, H., Whyte, G. P., and Doust, J. H. (2008). Physiological and (2017). Eleven-week preparation involving polarized intensity distribution is
performance effects of low- versus mixed-intensity rowing training. Med. Sci. not superior to pyramidal distribution in national elite rowers. Front. Physiol.
Sports Exerc. 40, 579–584. doi: 10.1249/MSS.0b013e31815ecc6a 8:515. doi: 10.3389/fphys.2017.00515
MacInnis, M. J., and Gibala, M. J. (2016). Physiological adaptations to interval
training and the role of exercise intensity. J. Physiol. 595, 2915–2930. doi: Conflict of Interest Statement: The authors declare that the research was
10.1113/JP273196 conducted in the absence of any commercial or financial relationships that could
Milanović, Z., Sporiš, G., and Weston, M. (2015). Effectiveness of High-intensity be construed as a potential conflict of interest.
interval training (HIT) and continuous endurance training for VO2max
improvements: a systematic review and meta-analysis of controlled trials. Sports Copyright © 2019 Treff, Winkert, Sareban, Steinacker and Sperlich. This is an open-
Med. 45, 1469–1481. doi: 10.1007/s40279-015-0365-0 access article distributed under the terms of the Creative Commons Attribution
Neal, C. M., Hunter, A. M., Brennan, L., O’sullivan, A., Hamilton, D. L., De Vito, License (CC BY). The use, distribution or reproduction in other forums is permitted,
G., et al. (2013). Six weeks of a polarized training-intensity distribution leads to provided the original author(s) and the copyright owner(s) are credited and that the
greater physiological and performance adaptations than a threshold model in original publication in this journal is cited, in accordance with accepted academic
trained cyclists. J. Appl. Physiol. 114, 461–471. doi: 10.1152/japplphysiol.00652. practice. No use, distribution or reproduction is permitted which does not comply
2012 with these terms.

Frontiers in Physiology | www.frontiersin.org 10 June 2019 | Volume 10 | Article 707

 EDITORIAL
published: 21 May 2019
doi: 10.3389/fphys.2019.00592

Editorial: Training Intensity, Volume

and Recovery Distribution Among
Elite and Recreational Endurance
Athletes
Thomas L. Stöggl 1* and Billy Sperlich 2
1
Department of Sport and Exercise Science, University of Salzburg, Salzburg, Austria, 2 Integrative and Experimental Exercise
Science and Training, University of Würzburg, Würzburg, Germany

Keywords: training intensity distribution, endurance diagnostics, training load and responses, high intensity
training (HIT), polarized training, pyramidal training

Editorial on the Research Topic

Training Intensity, Volume and Recovery Distribution Among Elite and Recreational
Endurance Athletes

Stimulated by the debate among endurance athletes and coaches of whether an optimal training
intensity distribution (TID) exists, we recently reviewed the literature of studies dealing with TID in
various sports (Stöggl and Sperlich, 2015; Stöggl, 2018). Our research identified numerous intensity
zone models for quantifying TID. Among all TID models, a three zone-based TID seems widely
applied among the literature. Within this model “Zone 1” represents low-intensity exercise below
first lactate or ventilatory threshold. “Zone 2” exhibits accumulated levels of blood lactate between
the first and second lactate or ventilatory threshold. “Zone 3” represents high-intensity exercise
above the second lactate or ventilatory threshold.
Based on our (Stöggl and Sperlich, 2015) and other findings (e.g., Billat et al., 2001; Seiler and
Edited and reviewed by:
Gary Iwamoto,
Kjerland, 2006; Sandbakk et al., 2011; Tonnessen et al., 2014) various TID exist including a so-
University of Illinois at called polarized (Zone1 > Zone3 > Zone 2) (Fiskerstrand and Seiler, 2004; Seiler and Kjerland,
Urbana-Champaign, United States 2006) or pyramidal TID (Zone1 > Zone 2 > Zone 3) (Stöggl and Sperlich, 2015), depending on
*Correspondence: the discipline.
Thomas L. Stöggl Since controlled and prospective training experiments are time consuming, complex, requiring
[email protected] both the coaches and athletes to adhere to scientific rules and methods we were motivated to
stimulate this research topic to gain broader and deeper insights in TIDs during preparation, pre-
Specialty section: competition, and competition phases in different endurance disciplines and performance levels.
This article was submitted to Ultimately, we wanted to identify TIDs demonstrating greater efficacy than others and highlight
Exercise Physiology, research gaps in an effort to direct future scientific investigations.
a section of the journal
Based on the numerous contributions to this research topic we have learned the following:
Frontiers in Physiology

Received: 23 January 2019

Accepted: 26 April 2019
METHOD FOR TRAINING QUANTIFICATION
Published: 21 May 2019
Based on Manunzio et al.’s analysis of 6-months preparation analysis of a 2nd place Race Across
Citation:
America finisher team (n = 4 athletes) the TID may vary depending on the method employed.
Stöggl TL and Sperlich B (2019)
Editorial: Training Intensity, Volume
Retrospective power data analysis based on the 3-zone model revealed a pyramidal TID (Zone
and Recovery Distribution Among Elite 1: 63%; Zone: 28%; Zone 3: 9%) when including coasting phases (i.e., power output below 50% of
and Recreational Endurance Athletes. power at maximum lactate steady state, MLSS). The same data set without coasting phases reflected
Front. Physiol. 10:592. a threshold TID with a greater Zone 2 emphasis (48/39/13%). The amount of training time <50%
doi: 10.3389/fphys.2019.00592 of MLSS was shown to be remarkably high with 28% (104 h) of total training time.

Frontiers in Physiology | www.frontiersin.org 11 May 2019 | Volume 10 | Article 592

Stöggl and Sperlich Training Intensity Distribution Among Athletes

POWER-DURATION RELATIONSHIP especially with respect to the physical and overall scales, while
the mental and emotional scales were not. Further, the authors
In their concept (Hofmann and Tschakert) describe the recommend using the sRPE method, with respect to volume (km)
relationship between power/velocity and the corresponding rather than training time (h), to monitor the internal training
maximal duration/distance. Based on the power-duration load in swimmers.
relationship their concept may allow coaches or athletes
to individually define the intensity and duration of an
GROSS EFFICIENCY AND TID
exercise session.
Skovereng et al. investigated the possible effects of initial
performance, gross efficiency and VO2peak on subsequent
DISCIPLINE SPECIFIC TID adaptations to a 12-weeks endurance training including HIIT
In their attempt to answer the question of whether the training (24 supervised HIIT sessions, 3/ and 1/weeks during recovery
for marathon is harder than for an Ironman, Esteve-Lanao week in combination with ad libitum low intensity training)
et al. employed the Objective Load Scale (ECOs) training load in competitive cyclists. In general, this training concept led to
quantification method. For Ironman, the highest associations an increase in VO2peak , peak and mean power output during
between performance and training were found with total training a 40 min time trial, while gross efficiency decreased. Initial
time and % of time in zone 1 (<Aerobic Threshold), while no performance demonstrated only small to moderate effects on
association was found for total ECOs or amount of training training response.
in Zone 3 (>anaerobic threshold). Marathon performance was
related to both total training time and total training load. SPRINT INTERVAL TRAINING (SIT) WITH
Further, it was associated with total amount of Zone 1 plus DIFFERENT INTERVAL RECOVERY
the accumulation of any variable related to Zone 3 training.
For both, Ironman and Marathon, the amount of Zone 2 In a 2-weeks experiment (6 sessions) by Olek et al. physically
training was associated with poorer competition performance. active males performed a series of 10-s sprints separated by either
Ironman athletes trained more (more than twice total time, and 1- or 4-min of recovery. The number of sprints progressed from
>1/3 of training load) however less hard (e.g., training load four to six separated by 1–2 d rest. VO2max , citrate synthase
per training hour) than marathon athletes. The differences in activity and Wingate anaerobic test results improved similarly
performance related to TID emphasizes the need for discipline in both groups. Only end power output increased by 10.8%
specific TID analysis. in the group with 1-min recovery. The two SIT protocols
induced metabolic adaptations over a short period of time,
and reduced recovery between SIT bouts may attenuate fatigue
BLOCKED TID during maximal exercise.

The study of McGawley et al. investigated the performance,

stress, recovery and physiological effects to nine session of high
SEVEN-DAY RUNNING BLOCK ON AORTIC
intensity interval training (HIIT) evenly distributed over 3- BLOOD PRESSURE
weeks vs. nine HIIT sessions within 1 week with no further
Tomoto et al. investigated the impact of a condensed 7-day
HIIT in the other 2 weeks. The authors showed that well-
running camp on aortic blood pressure in two groups of
trained junior cross-country skiers are able to complete nine HIT
collegiate endurance runners (i.e., one group accomplishing the
sessions within 1 week without compromising total work done
weekly training target: 31 km/d vs. one group not accomplishing
or experiencing greater stress or reduced recovery over a 3-weeks
the weekly training target 13 km/d). Pulse wave analysis revealed
polarized microcycle. Although, both 3-weeks TIDs improved
elevated aortic blood pressure in the group with 31 km/d while
performance (i.e., 600-m roller skiing time trial) and various
this was not the case in the group with distinctly lower training
hormonal and muscle markers, the findings do not suggest that
volume and intensity.
block-distributed HIIT is superior to evenly distributed HIIT
within a 3-week period.
TID IN ELITE FEMALE CROSS-COUNTRY
SKIERS
RELATION BETWEEN TRAINING LOAD
AND RECOVERY-STRESS STATE The case study by Solli et al. presents the TID of the world’s
most successful female cross-country skier. Following a 12-years
Collette et al. aimed (i) to analyze the individual time-delayed nonlinear increase in training load of approximately 80% (522–
linear effect relationship between training load and recovery- 940 h/years) from the age of 20–35, the annual training volume
stress state with single-case time series methods and (ii) to during the five consecutive most successful years stabilized at
monitor the acute recovery-stress state of high-performance 937 ± 25 h, distributed across 543 ± 9 sessions. She displayed
swimmers over a 17-weeks macro cycle. The Acute Recovery a polarized TID, but to a lesser extent during the latter part
and Stress Scale (ARSS), was shown to be a suitable tool of her career (i.e., 88/22/20% at the age of 20–27 years vs.
for monitoring the acute recovery-stress state in swimming, 92/3/5% at the age of 28–35 years). While the total time in

Frontiers in Physiology | www.frontiersin.org 12 May 2019 | Volume 10 | Article 592

Stöggl and Sperlich Training Intensity Distribution Among Athletes

Zone-1 and 2 reduced from general preparation, to specific improves the neuromuscular status, anaerobic power and the
preparation and competition phase, Zone 3 increased from 4.1 acute HRR in well-trained endurance athletes.
to 9.2 sessions/month.
TID AND IMMUNE FUNCTION
UPPER-BODY EXERCISE
Born et al. hypothesized that nine session of HIIT in
Børve et al. investigated the effects of replacing two HIT sessions 3 weeks would increase levels of salivary cortisol, reduce
with either combined upper-body muscular endurance training Immunoglobin-A secretion rate and impair mood thereby
and running intervals (mixed endurance group) or only running demonstrating marked psycho-immunological stress-response
intervals. Both concepts were performed as pyramidal TID. and compromised mucosal immune function compared to long
The 6-weeks mixed training approach increased both muscular slow distance (LSD) running. Based on their data the authors
endurance and maximal strength in a simulated double poling concluded that the increased Immunoglobin-A secretion rate
exercise and 1,000-m double poling performance following a 50- with HIIT indicates no compromised mucosal immune function
min submaximal trial with no changes in the endurance group. compared to LSD. Further, this shows the functional adaptation
Specific upper-body muscular endurance training thus seems as of the mucosal immune system in response to the increased stress
a promising training model to optimize performance in well- and training load of nine sessions of HIIT.
trained cross-country skiers. Although the research topic broadened our understanding
of various aspects of TID in elite and recreational sports, there
are still various research question subject to future investigation,
OFF-TRAINING ANALYSIS AND TID including the following (the authors are aware that this list is
Not only the training stimulus itself but also other off-training not exhaustive):
stimuli may explain variation in adaptation among individuals i) Currently most TIDs are defined as a certain percentage of
(Sperlich and Holmberg), therefore it seems strange that the training time or session within an intensity zone. However,
aspect of off-training behavior is mostly not considered within this approach does not allow to judge the density of sessions
TID analysis. The study by Sperlich et al. demonstrated that nor the timing between sessions. Furthermore, depending
national elite rowers demonstrate a substantial sedentary off- on the sport strength training maybe included but differs
training behavior of more than 11.5 h/day. The question about substantially in intensity but no TID model, at least to our
the effects of off-training physical activity during recovery knowledge, has implemented this matter into a TID model.
and the long-term performance development is open to ii) It remains unclear why (i) endurance athletes exercise
future research. a large proportion at low-intensity although endurance
competitions are usually executed at higher intensity and (ii)
TID IN ELITE ROWERS why so many different TID may exist.
iii) The long-term effects of TIDs (e.g., inverse polarized or
Treff et al. analyzed different TID (polarized vs. pyramidal HIIT) and potential shifting of TID within a season or
TID with similar amount of Zone 1 training) in national elite between seasons are still not totally understood.
rowers. Based on their analysis both TID showed similar gains iv) We believe that the development of wearable technology
in performance. The polarized compared to the pyramidal TID should further ease and improve our understanding of
seemed not to be superior, possibly due to a very similar different TID (e.g., the interaction of internal-external load,
percentage of Zone 1 training. stress, fatigue and recovery process).
At this point we’d like to thank the authors for their contribution
TID AND ACUTE HEART RATE RECOVERY and we hope this research topic will not only provide new insights
AND ANAEROBIC POWER and viewpoints about the issues of TID in endurance training,
but will also stimulate novel thoughts, experiments, and further
Stöggl and Björklund explored whether four TID (9 weeks advances in this field of research.
of HIIT vs. polarized TID vs. threshold vs. high volume low
intensity) induced different responses on neuromuscular status, AUTHOR CONTRIBUTIONS
anaerobic capacity/power and acute heart rate recovery (HRR)
in well-trained endurance athletes. They concluded that only All authors listed have made a substantial, direct and intellectual
a training regime that includes a significant amount of HIIT contribution to the work, and approved it for publication.

REFERENCES Fiskerstrand, A., and Seiler, K. S. (2004). Training and performance characteristics
among Norwegian international rowers 1970-2001. Scand. J. Med. Sci. Sports
Billat, V. L., Demarle, A., Slawinski, J., Paiva, M., and Koralsztein, J. P. (2001). 14, 303–310. doi: 10.1046/j.1600-0838.2003.370.x
Physical and training characteristics of top-class marathon runners. Med. Sci. Sandbakk, O., Holmberg, H. C., Leirdal, S., and Ettema, G.
Sports Exerc. 33, 2089–2097. doi: 10.1097/00005768-200112000-00018 (2011). The physiology of world-class sprint skiers. Scand.

Frontiers in Physiology | www.frontiersin.org 13 May 2019 | Volume 10 | Article 592

Stöggl and Sperlich Training Intensity Distribution Among Athletes

J. Med. Sci. Sports 21, e9–e16. doi: 10.1111/j.1600-0838.2010. year prior to a gold medal endurance performance. PLoS ONE 9:e101796.
01117.x doi: 10.1371/journal.pone.0101796
Seiler, K. S., and Kjerland, G. O. (2006). Quantifying training intensity distribution
in elite endurance athletes: is there evidence for an “optimal” distribution? Conflict of Interest Statement: The authors declare that the research was
Scand. J. Med. Sci. Sports 16, 49–56. doi: 10.1111/j.1600-0838.2004. conducted in the absence of any commercial or financial relationships that could
00418.x be construed as a potential conflict of interest.
Stöggl, T. L. (2018). What is the best way to train to become a star endurance
athlete? Front. Young Minds 6:17. doi: 10.3389/frym.2018.00017 Copyright © 2019 Stöggl and Sperlich. This is an open-access article distributed
Stöggl, T. L., and Sperlich, B. (2015). The training intensity distribution under the terms of the Creative Commons Attribution License (CC BY). The use,
among well-trained and elite endurance athletes. Front. Physiol. 6:295. distribution or reproduction in other forums is permitted, provided the original
doi: 10.3389/fphys.2015.00295 author(s) and the copyright owner(s) are credited and that the original publication
Tonnessen, E., Sylta, O., Haugen, T. A., Hem, E., Svendsen, I. S., and Seiler, in this journal is cited, in accordance with accepted academic practice. No use,
K. S. (2014). The road to gold: training and peaking characteristics in the distribution or reproduction is permitted which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 14 May 2019 | Volume 10 | Article 592

 ORIGINAL RESEARCH
published: 05 July 2018
doi: 10.3389/fphys.2018.00845

Relation Between Training Load and

Recovery-Stress State in
High-Performance Swimming
Robert Collette 1* , Michael Kellmann 2,3 , Alexander Ferrauti 3,4 , Tim Meyer 5 and
Mark Pfeiffer 1
1
Department Theory and Practice of Sports, Institute of Sport Science, Johannes Gutenberg University Mainz, Mainz,
Germany, 2 Unit of Sport Psychology, Faculty of Sport Science, Ruhr-University Bochum, Bochum, Germany, 3 School of
Human Movement and Nutrition Sciences, The University of Queensland, Brisbane, QLD, Australia, 4 Department of Training
and Exercise Science, Faculty of Sport Science, Ruhr-University Bochum, Bochum, Germany, 5 Institute of Sports and
Preventive Medicine, Saarland University, Saarbrücken, Germany

Background: The relation between training load, especially internal load, and the
recovery-stress state is of central importance for avoiding negative adaptations in high-
performance sports like swimming. The aim of this study was to analyze the individual
time-delayed linear effect relationship between training load and recovery-stress state
with single case time series methods and to monitor the acute recovery-stress
state of high-performance swimmers in an economical and multidimensional manner
over a macro cycle. The Acute Recovery and Stress Scale (ARSS) was used for
Edited by: daily monitoring of the recovery-stress state. The methods session-RPE (sRPE) and
Thomas Leonhard Stöggl,
University of Salzburg, Austria acute:chronic workload-ratio (ACWR) were used to compare different methods for
Reviewed by: quantifying the internal training load with regard to their interrelationship with the
Pedro Jiménez Reyes, recovery-stress state.
Universidad Católica San Antonio
de Murcia, Spain Methods: Internal load and recovery-stress state of five highly trained female swimmers
David Christopher Nieman,
[with a training frequency of 13.6 ± 0.8 sessions per week and specializing in sprint
Appalachian State University,
United States (50 and 100 m), middle-distance (200 and 400 m), or long distance (800 and 1,500
*Correspondence: m) events] were daily documented over 17 weeks. Two different types of sRPE
Robert Collette were applied: RPE∗ duration (sRPEh ) and RPE∗ volume (sRPEkm ). Subsequently, we
[email protected]
calculated the ratios ACWRh and ACWRkm (sRPE last week: 4-week exponentially
Specialty section: weighted moving average). The recovery-stress state was measured by using the ARSS,
This article was submitted to consisting of eight scales, four of which are related to recovery [Physical Performance
Exercise Physiology,
a section of the journal Capability (PPC), Mental Performance Capability (MPC), Emotional Balance (EB), Overall
Frontiers in Physiology Recovery (OR)], and four to stress [Muscular Stress (MS), Lack of Activation (LA),
Received: 30 December 2017 Negative Emotional State (NES), Overall Stress (OS)]. To examine the relation between
Accepted: 14 June 2018
training load and recovery-stress state a cross correlation (CCC) was conducted with
Published: 05 July 2018
sRPEh , sRPEkm , ACWRh , and ACWRkm as lead and the eight ARSS-scales as lag
Citation:
Collette R, Kellmann M, Ferrauti A, variables.
Meyer T and Pfeiffer M (2018)
Relation Between Training Load and Results: A large variation of training load can be observed in the individual week-to-
Recovery-Stress State in week fluctuations whereby the single fluctuations can significantly differ from the overall
High-Performance Swimming.
Front. Physiol. 9:845.
mean of the group. The range also shows that the CCC individually reaches values
doi: 10.3389/fphys.2018.00845 above 0.3, especially with sRPEkm as lead variable. Overall, there is a large range with

Frontiers in Physiology | www.frontiersin.org 15 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

significant differences between the recovery and stress dimensions of the ARSS
and between the training load methods, with sRPEkm having the largest span
(Range = 1.16). High inter-individual differences between the athletes lie in strength and
direction of the correlation | 0.66| ≤ CCC ≥ | −0.50| . The time delayed effects (lags
0–7) are highly individual, however, clear patterns can be observed.
Conclusion: The ARSS, especially the physical and overall-related scales (PPC, OR,
MS, OS), is a suitable tool for monitoring the acute recovery-stress state in swimmers.
MPC, EB, LA, and NES are less affected by training induced changes. Comparably high
CCC and Ranges result from the four internal load methods, whereby sRPE, especially
sRPEkm , shows a stronger relation to recovery-stress state than ACWR. Based on these
results and the individual differences in terms of time delay in training response, we
recommend for swimming to use sRPE to monitor the internal training load and to use
the ARSS, with a focus at the physical and overall-scales, to monitor the recovery-stress
state.
Keywords: monitoring, training, recovery-stress state, internal load, session RPE, ACWR, time series analysis,
individual case

INTRODUCTION instead of the duration for the calculation. It is assumed that

in endurance sports such as swimming, volume has a greater
Analyzing internal and external training loads has become a impact on the recovery-stress state than duration. A validation
critical issue in elite sport practice and research. In this regard, study in which both methods are compared is not yet available.
monitoring the athletes internal training load is essential for The ACWR is a simplification of the fitness-fatigue model of
understanding whether athletes are positively adapting to their Banister et al. (1975) and it was recently reported to provide valid
training program. This implements an understanding of the information regarding injury risk in team sports (Hulin et al.,
individual’s responses to training, assessing fatigue and the 2016; Gabbett, 2016; Bowen et al., 2017; Murray et al., 2017b).
associated need for recovery, in order to minimize the risk of non- It is therefore reasonable to conclude that ACWR also provides
functional overreaching, injury and illness (Bourdon et al., 2017; valid information regarding the impact of the internal load on
Kellmann and Beckmann, 2018; Kellmann et al., 2018). In high- recovery-stress state.
performance sports like swimming an individual monitoring of The recovery-stress state is based on the individual’s ability to
the training load and the recovery-stress state for the prevention utilize resources necessary for recovery in order to compensate
of negative adaptations is recommended (Foster et al., 1999; stressful situations and activities (Nässi et al., 2017). Stress
Smith, 2003; Lambert and Mujika, 2013; Collette, 2016; Soligard and recovery appear to be complex, intertwined processes that
et al., 2016; Crowcroft et al., 2017). should be viewed from different perspectives such as time
Referring to Bourdon et al. (2017) the measures of training frames and/or contexts, and even multiple processes (Kenttä
load can be categorized as either internal or external, where and Hassmén, 1998; Kellmann and Kallus, 2001; Kellmann
external training loads are objective measures of the work and Beckmann, 2018). In high-performance sport self-report
performed by the athlete (e.g., speed, acceleration, volume, . . .). measures via questionnaires represent the most common form
On the other hand, internal training load is defined as the relative for monitoring the athlete’s recovery-stress state (Nässi et al.,
physiological and psychological stressors imposed on the athlete 2017) and for this purpose, several valid and reliable instruments
during training or competition. Various methods for measuring are available, e.g., the Profile of Mood States (POMS) (McNair
internal load exist, such as rating of perceived exertion (RPE), et al., 1971) or the Recovery-Stress-Questionnaire for Sport
session rating of perceived exertion (sRPE), training impulse (RESTQ-Sport) (Kellmann and Kallus, 2001, 2016) are the most
(TRIMP), heart-rate indices, blood lactate, oxygen uptake and/or frequently used instruments (Saw et al., 2016, 2017). However,
psychological scales and questionnaires (Bourdon et al., 2017). both instruments do not examine the acute recovery-stress state
At present, especially the ‘sRPE’ (Foster et al., 1999) as well in a multidimensional manner. The main criterion for POMS
as the ‘acute:chronic-workload ratio’ (ACWR) (Gabbett, 2016; is the predominantly negative orientation of the questionnaire
Hulin et al., 2016) methods are being discussed, whereas sRPE (Martin et al., 2000; Ziemainz and Peters, 2010) and that it
has been extensively investigated and seems to be a valid tool has been developed for the assessment of mood (Hitzschke
for measuring internal training load in a variety of sport (Foster et al., 2016). However, recovery-stress state and mood are
et al., 2001; Herman et al., 2006; Seiler and Kjerland, 2006; regarded as separate psychological constructs (Mäestu et al.,
Borresen and Lambert, 2008), especially in swimming (Wallace 2005). The RESTQ-Sport includes 76 items and is therefore not
et al., 2008, 2009; Toubekis et al., 2013). Nagle et al. (2015) suitable for a weekly or daily use (Kellmann, 2000). Additionally,
modified the method of Foster by using the volume (km) recovery and stress state are evaluated over the last 3 days and

Frontiers in Physiology | www.frontiersin.org 16 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

therefore the questionnaire does not indicate the acute (‘here (c) to compare different methods for quantifying the internal
right now’) condition of the athlete (Kölling et al., 2015). Meeusen training load with regard to their interrelationship with the
et al. (2013) point out that for effective load monitoring a recovery-stress state,
shorter questionnaire or instrument is needed which responds (d) to detect differences in the relationship between internal
sensitively to the current state of recovery and stress. As a result, load and the two states ‘recovery’ and ‘stress’ determined
Kellmann et al. (2016) developed the Acute Recovery and Stress by using the ARSS.
Scale (ARSS) to assess and monitor the acute multidimensional
recovery and stress state, that considers not only the physical,
but also the emotional and psychological recovery or stress. The MATERIALS AND METHODS
ARSS can be applied on a daily basis for training monitoring in
elite sports (Hitzschke et al., 2017). Several laboratory and field Participants
studies in swimming (Collette, 2016), cycling (Hammes et al., Five female high-performance swimmers (S1–S5, mean ± SD:
2016), rowing (Kölling et al., 2016), tennis (Wiewelhove et al., age: 21 ± 2.8 years, body mass: 60.1 ± 6.5 kg, height: 1.72 ± 0.1
2016), football (Pelka et al., 2017) or strength and high-intensity m, best Fédération Internationale de Natation (2014) points in
interval training (Raeder et al., 2016) showed the practicability main event as percentage of world record 72.8 ± 7.9%) monitored
and suitability to changes of the training stimuli of the ARSS. daily over 17 weeks. All participants were well-trained athletes,
These studies showed that daily changes as well as indications for accustomed to a training frequency of more than thirteen sessions
a general trend of the recovery-stress state in different training per week (13.6 ± 0.8), including pool and athletic sessions.
phases will be displayed. Specialized in sprint (50 and 100 m), middle-distance (200 and
Several studies have investigated the relationship between 400 m) or long distance (800 and 1,500 m) events.
training load and well-being or recovery-stress state, with an The study received approval from the Ruhr-Universität
increase in ‘stress’ and reduction in ‘recovery’ after intensive Bochum, Faculty of Psychology, Ethics Committee. All
training, respectively, higher training load in comparison to participants gave their written informed consent to participate
normal training load. In addition, a reduction in stress scales in the study which was conducted in accordance with the
and increase in recovery scales were also observed following a Declaration of Helsinki.
taper phase. Furthermore, the results show a high variability
and indicate a high degree of individuality (Morgan et al., Design
1987; Berglund and Safstrom, 1994; Adams and Kirkby, 2001; Athletes were monitored over a 17-week macrocycle (tw) which
O’Connor and Puetz, 2005; Coutts et al., 2007; Kellmann, included different mesocycles and periodization phases, as well
2010; Bresciani et al., 2011; Brink et al., 2012; Laux et al., as one or two main competitions [German Championships
2015). (tw 9)/German Youth Championships (tw 16)]. Within these
In high-performance sport, there is a high degree of macrocycle a 16 days training camp was included at end of
individuality in terms of training loads and training adaptation week three. The recovery-stress state was recorded every morning
(Collette, 2016; Julian et al., 2017). In addition, the recovery- using the ARSS. Additionally, internal training load using sRPE
stress structure is characterized by high individuality (Bouchard was documented after every training session. In order to provide
and Rankinen, 2001; Hautala et al., 2006; Hecksteden et al., 2015, as much error-free documentation as possible, the training load
2017; Schimpchen et al., 2017). was recorded by both the athletes themselves and the coach.
However, the above-mentioned studies show a number of The athletes were instructed to complete the ARSS questionnaire
shortcomings: (a) mostly the focus was to compare group values, every morning before the first training session. The questionnaire
no single case study performed before; (b) typically pre–post was provided on an online platform1 and was filled out using the
study design was applied so that the process cannot be observed; athlete’s smartphone or tablet. The compliance was high with data
(c) the studies usually comprise short periods or specific training only missing for S1, S2 (1 day each/0.8%) and S5 (6 days/5.0%).
phases (e.g., taper-phase, training camps); (d) the recovery-stress For calculations the following parameters were individually
state was evaluated over a time period (e.g., over 3-days with the collected: duration of every training session (min); volume (km);
RESTQ-Sport), only one-dimensionally (e.g., RPE) or the mood sRPE; ARSS. Figure 1 shows a systematic overview of the study
is measured (e.g., POMS). design as well as the collected or calculated parameters.
Considering these critical points, the aim of the present study
was Training Load
Two different methods of sRPE were used to quantify the internal
(a) to analyze the individual time-delayed linear effect training load and based on the sRPE values the ‘acute:chronic
relationship between training load and recovery-stress state workload ratios’ (ACWR) were calculated (Table 1).
with single case time series methods (bivariate cross-
correlations), Session RPE
(b) to monitor the acute recovery-stress state of high- As described by Foster et al. (2001) within 30 min after every
performance swimmers in an economical and training session participants were given standard instructions for
multidimensional manner over a long period or different
training periods (macro cycle) 1
www.soscisurvey.de

Frontiers in Physiology | www.frontiersin.org 17 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

FIGURE 1 | Systematic overview of the study design and the collected or calculated parameters. Cross-correlations were calculated for every of the four
lead-variables with every scale of the ARSS as lag-variable. sRPEh = RPE∗ duration (min); sRPEkm = RPE∗ volume (km); ACWRh = sRPEh last week/sRPEh 4-week
exponentially weighted moving average; ACWRkm = sRPEkm last week/sRPEkm 4-week exponentially weighted moving average.

overall RPE and were asked to report based on the degree of Mujika et al. (1996) (60-min dryland session equals 2 km of
whole body heaviness and strain experienced during the exercise swimming).
task using a 11-point scale (based on the CR-10 scale by Borg,
1982), with 0 and 10 corresponding to ‘rest’ and ‘maximal,’ Acute:Chronic Workload Ratio
respectively. The training load (sRPE) of 1 week is defined as the acute load and
For the first method, sRPEh , the training load was calculated the chronic training load represent the exponentially weighted
by multiplying the 0–10 rating by the total session duration (in moving average (EWMA) of the load in the four previous weeks
min) and expressed in arbitrary units (Foster et al., 2001). Weekly of training (Nielsen et al., 2014; Gabbett, 2016; Murray et al.,
sRPEh training load was calculated for each athlete individually 2017a).
by summing the sRPEh training loads for all training sessions. As described by Williams et al. (2016) and Murray et al.
The second method, sRPEkm , differs only in a modified (2017a), the EWMA is calculated as follows:
calculation for the sRPE-method by using volume (in km)
EWMAtoday = sRPEtoday × λa + ((1 − λa ) × EWMAyesterday )
instead of duration (in min) (Nagle et al., 2015). Weekly
sRPEkm training load scores were calculated for each athlete and λa is calculated as:
individually by summing the sRPEkm training loads for
all training sessions. To consider dryland workouts the λq = 2/(N + 1)
duration of the session was converted to volume based on
Where N is the time decay constant with 1 week (7 days) for
acute and 4 weeks (28 days) for chronic workloads. For the
TABLE 1 | Overview of the various calculation formulas for the methods of training
ratio the EWMA ACWR value of acute workload was divided
load quantification. by the EWMA ACWR value of chronic workload. To begin the
EWMA calculation, the first observation in the series is arbitrarily
Method Calculation formula Unit recorded as the first workload value in the series (Murray et al.,
sRPEh = RPE∗ duration (min) AU 2017a).
ACWRh = sRPEh last week/sRPEh 4-week exponentially AU Comparing the acute training load to the chronic training load
weighted moving average as a ratio provides an index of athlete preparedness and fatigue
sRPEkm = RPE∗ volume (km) AU (Gabbett, 2016). The ACWR was divided into the following
ACWRkm = sRPEkm last week/sRPEh 4-week AU ranges: very low ≤ 0.49, low 0.5–0.99, moderate 1.0–1.49, high
exponentially weighted moving average 1.50–1.99, and very high 2.0 (Murray et al., 2017a). As such an
sRPE, session rating of perceived exertion; ACWR, acute:chronic workload ratio; ‘acute:chronic workload ratio’ between 0.8 and 1.3 was considered
AU, arbitrary units. the ‘sweet spot,’ while ratios ≥ 1.5 represent the ‘danger zone’

Frontiers in Physiology | www.frontiersin.org 18 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

with an increased risk of injury (Blanch and Gabbett, 2016; time-series analyzes are very sensitive to erroneous data, the few
Gabbett, 2016). The ‘acute:chronic load ratio’ calculated with missing data are estimated. The estimated value was determined
training loads based on the sRPEh - (ACWRh ) and sRPEkm - by calculating the mean value 2 days before and after the data
Method (ACWRkm ) is shown in Table 1. gap. The calculation of the mean values of the cross-correlation
coefficients implemented the Fischer-Z transformations (Bortz
Acute Recovery and Stress Scale and Schuster, 2010).
The ARSS consists of a total of 32 adjectives describing the
physical, emotional, mental, and general aspects of recovery
and stress based on a 7-point Likert scale from 0 (does not RESULTS
apply at all) to 6 (fully applies) (Kellmann et al., 2016). The
adjectives are summarized in eight scales, of which four are Training Load
related to stress, and four to recovery. The recovery-related scales Total mean training volume was 833.7 ± 14.1 km with a mean
are: Physical Performance Capability (PPC), Mental Performance maximum of 89.0 km in tw 4 and a minimum of 30.8 km in tw 7
Capability (MPC), Emotional Balance (EB), and Overall Recovery (Figure 3A).
(OR). The stress-related scales are: Muscular Stress (MS), Lack Total mean sRPEkm -load was 3888.1 ± 72.4 au, with a
of Activation (LA), Negative Emotional State (NES), and Overall maximum (440.5 au) in tw 4 and a minimum (123.6 au) in tw
Stress (OS) (Hitzschke et al., 2017). All scales of the German 7 (Figure 3B).
ARSS showed satisfactory internal consistency (range between The sRPEh values also show a similar distribution as the
α = 0.84 and α = 0.96) and a good model fit for both the sRPEkm values, with the peak loads striking in particular during
recovery (RMSEA = 0.07, CFI = 0.97, SRMR = 0.04) and stress the training weeks with competitions (tw 6, tw 9, tw 16). Mean
(RMSEA = 0.09, CFI = 0.94, SRMR = 0.05) factors (Kellmann sRPEh -training load was 114996 ± 2630 au (Figure 3C).
et al., 2016). Nässi et al. (2017) reported the psychometric Overall, a large variation can be observed in the individual
properties of the English version of the ARSS. week-to-week fluctuations whereby the single fluctuations can
significantly differ from the overall mean plus/minus the standard
Statistical Analyses deviation of the group (Figures 3A–C).
The time-series analysis is designed to investigate the time-lagged Thus, the individual maximum for volume (91.0 km in tw
effects of several variables, in which the correlations between 4) and sRPEh (2110 au in tw 7) can be seen for athlete S1,
the training loads and the recovery-stress state are investigated while athlete S3 has the maximum for sRPEkm (18390 au in tw
using bivariate cross-correlations. In other words, the linear 6). The individual minimum for volume (9.1 km in tw 13) and
dependence on the time delay of two time series is calculated sRPEkm (55.3 au in tw 13) has also athlete S1, while S5 shows the
here (Schmitz, 2000). For this purpose, two time series are minimum for sRPEh (2110 au in tw 7) (Figures 3A–C).
shifted against each other by π-times, resulting in a lead variable Figure 4 shows the ‘ACWR’ based on the sRPEh - (Figure 4A)
and a lag variable. The direction of the displacement, e.g., the and the sRPEkm -values (Figure 4B). Only the ACWRh curves of
variable which is the lead and the lag variable (Schmitz, 1996) is Athletes S1, S4, and S5 are in the ‘very high’ (>2.0) range and
important. As lead variables, the sRPE and ACWR values are set these are only short-term peaks over a day. Except for S1 all
in this study in order to examine the time-delayed effects of the athletes have a peak in the range of 1.5–1.99 (high) or >2.0 in
training on the recovery-stress state in terms of the eight scales training week 9 in which the German Championships took place.
of the ARSS as lag variables. This means that for every athlete In comparison, the values in the training camp (end of tw 3 to tw
32 cross-correlations (Figure 1) with n = 119 data points (days) 5) are all in moderate range (1.0–1.49). In the ACWRkm curves
were calculated. A significant correlation can be interpreted only S1 shows a peak in the very high range and this, as with
in the sense of a co-determination of the lag variable by the ACWRh , in training week 13. In addition, peaks in the high range
lead variable (Schmitz, 2000). According to Maiwald and Rogge can only be seen S1 and S5. Overall, the curves of ACWRkm are
(2005), significant cross-correlations are found for physiological smoother compared to ACWRh and predominantly in the low
variables and ordinally scaled self-estimates when their absolute (0.5–0.99) to moderate (1.0–1.49) range.
value is greater than CCC ≥ 0.2. Due to the method of cross-
correlation, only the maximum significant cross-correlation Cross-Correlation
coefficients (max. CCC) with the associated time delay (lag) are of Only the maximum significant CCC is considered for further
interest for further evaluation. In order to avoid the possibility of analysis due to methodological reasons of the cross-correlation
false inconsistencies, a ‘pre-whitening’ (data transformation into (see above). Table 2 shows the mean (MCCC) and the ranges
‘White Noise’) of the time series is deliberately dispensed with, of the CCC between the sRPE and ACWR values as lead
which is why the strengths of the cross-correlation coefficients variables and the dimensions of the ARSS, separated into
can only be interpreted with extreme caution, especially in recovery and stress, as lag variables. The highest MCCC are
the case of interindividual comparisons (Schmitz, 1996). The observed for MS with sRPEh (MCCC = 0.41), respectively,
IBM© SPSS© Statistics 23 software package was used to perform sRPEkm (MCCC = 0.51) and for OS with sRPEkm (MCCC = 0.39).
the complex statistical calculations. The cross-correlation logs The range in Table 2 also shows that the CCC individually reaches
have been calculated for seven lags (7 days) and show the values above 0.3 in other dimensions, especially with sRPEkm
corresponding 95% confidence intervals (e.g., Figure 2). Since the as lead variable. Overall, there is a large range with significant

Frontiers in Physiology | www.frontiersin.org 19 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

FIGURE 2 | Example of a cross-correlation plot for Athlete S4 with sRPEh as lead-variable and the ARSS-Dimension Emotional Balance as lag-variable.

differences between the recovery and stress dimensions of the 0 to 6 for sRPEh , ACWRkm , ACWRh , and 0 to 7 for
ARSS, and between the training load methods, with sRPEkm sRPEkm .
having the largest span of Range = 1.16. Furthermore, it is To make the time-delayed effects comparable with each other,
noticeable that for sRPEkm , ACWRkm , and ACWRh for the an individual profile using a network diagram was created for
dimensions MPC, EB, LA, and NES ranges from negative to each athlete and for all four lead variables. Figure 7 shows
positive CCC, and thus different effective directions are present. the individual profiles in terms of time-delayed interaction for
Therefore, Figures 5, 6 show the level and effective direction the recovery (PPC, MPC, EB, OR) and stress dimensions (MS,
of the individual CCC, as well as the time delay, based on LA, NES, OS) of the ARSS with sRPEh , sRPEkm , ACWRh ,
their lags. Contrary directions of action show for athletes S1, and ACWRkm as lead variables. If sRPEh and ACWRh are
S3, and S4 but for different lead variables and dimensions, not taken into account only once for S3 and sRPEkm for S4,
whereby only the mental and emotional dimensions (MPC, EB, three basic tread patterns can be distinguished. The athletes
LA, NES) are shown. S3 (MPC CCC = 0.22, EB CCC = 0.22, S1 (sRPEh , ACWRh ), S3 (ACWRkm ), and S5 (sRPEh , sRPEkm ,
NES CCC = −0.26) and S4 (EB CCC = 0.30, NES CCC = −0.24) ACWRh , ACWRkm ) react very quickly or directly with lag 0
show for sRPEkm contrary effective directions, for ACWRh only or lag 1 (or a single maximum lag 3 for S5) in dimensions
S1 (NES CCC = −0.19) and for ACWRkm S1 (EB CCC = 0.24, where significant correlations on the individual training load
LA CCC = −0.18, NES CCC = −0.25), and S4 (MPC CCC = 0.23, exist (Profile 1). Profile 2 shows only a small deviation; one
LA CCC = −0.21). This also shows that there are more significant or two dimensions react with a significantly larger time delay
CCC values with the stress dimensions than with the recovery (lag 4 to 6). For S2 (NES, LA) and S3 (MPC, EB) these are
dimensions and in some cases, there are even considerably only mental or emotional-related scales, for S1 only the overall
higher CCC. scales (OR, OS) and for S4 (PPC, MS, LA, OS) both occurs.
In addition to the magnitude of the relationship between Profile 3 is distinctly different from the other two profiles, as
training load and recovery-stress state, from a training control all dimensions except for MS (lag 0) react explicitly later with
perspective the time-delay of the relation is of interest. lag 5 to 7. It can be noticed that for all athletes and all load
Our findings show high inter-individual as well as intra- methods only for MS, the time-delayed effect is always between
individual differences of the time delayed effects, concerning 0 and 2. The only exception here is athlete S4 for ACWRh
the athletes and each dimension of the ARSS with lag with lag 5.

Frontiers in Physiology | www.frontiersin.org 20 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

FIGURE 3 | Training load (A) volume (km) (B) sRPEkm and (C) sRPEh over 17 training weeks of the athletes S1–S5 as well as the mean values [M (tw)]. To estimate
the changes from week to week, the total mean value [M (total)] as well as plus/minus a standard deviation (M ± SD) for reference are indicated.

DISCUSSION and the dimensions of the ARSS, confirm the theoretical

assumption that the interactions between training load and
The time-delayed linear effect relationship, which has been the recovery-stress state is characterized with high inter- and
calculated according to the time series analysis using bivariate intra-individual differences. It also reaffirms the idea that
cross-correlations between the individual training load recovery and stress should be explored using a multi-level

Frontiers in Physiology | www.frontiersin.org 21 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

FIGURE 4 | The acute:chronic workload-ratio over 17 training weeks of the athletes S1–S5 based on the sRPEh - (A) and the sRPEkm -values (B).

TABLE 2 | Mean (MCCC) and range of significant cross-correlations.

sRPEh sRPEkm ACWRh ACWRkm

MCCC Range MCCC Range MCCC Range MCCC Range

Recovery PPC −0.23 (−0.18/−0.29) −0.27 (−0.21/−0.35) −0.27 (−0.27/−0.27) −0.25 (−0.24/−0.27)
MPC −0.20 (−0.18/−0.21) −0.06 (0.22/−0.32) −0.18 (−0.18/−0.18) −0.21 (0.23/−0.24)
EB −0.26 (−0.25/−0.28) 0.26 (0.30/−0.24) −0.25 (−0.22/−0.30) −0.20 (0.24/−0.20)
OR −0.27 (−0.25/−0.30) −0.36 (−0.24/−0.50) −0.22 (−0.23/−0.41) −0.27 (−0.18/−0.31)
Stress MS 0.41 (0.48/0.33) 0.52 (0.66/0.37) 0.22 (0.26/0.18) 0.29 (0.38/0.20)
LA 0.24 (0.18/0.29) 0.21 (0.29/0.21) 0.23 (0.24/0.22) 0.20 (0.21/−0.21)
NES 0.24 (0.24/0.24) 0.21 (0.25/−0.26) 0.22 (0.22/−0.19) 0.19 (0.23/−0.25)
OS 0.29 (0.37/0.24) 0.39 (0.46/0.21) 0.26 (0.31/0.22) 0.29 (0.33/0.19)

PPC, Physical Performance Capability, MPC, Mental Performance Capability, EB, Emotional Balance, OR, Overall Recovery, MS, Muscular Stress, LA, Lack of Activation,
NES, Negative Emotional State, OS, Overall Stress. MCCC > 0.3 highlighted bold.

approach that takes into account psychological, emotional, |0.66| ≤ CCC ≥ |−0.50| as well as in the time delays from lag 0
cognitive, and social aspects both individually and collectively to lag 7.
(Kellmann, 2002b). High inter-individual differences between Low values in the stress dimensions and high values in the
the athletes lie in strength and direction of the correlation recovery dimensions are generally defined as positive and vice

Frontiers in Physiology | www.frontiersin.org 22 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

FIGURE 5 | Maximum significant cross correlations (max CCC) between the dimensions of ARSS as lag variables and sRPEh and sRPEkm as lead variables with the
associated time delay (lags) for the athletes S1 - S5.

FIGURE 6 | Maximum significant cross correlations (max CCC) between the dimensions of ARSS as lag variables and ACWRh and ACWRkm as lead variables with
the associated time delay (lags) for the athletes S1–S5.

versa as negative. However, in this context terms like “good – action between the training loads and the individual recovery-
bad” and/or “positive – negative” should be used with care and stress-related scales are in contrast to this assumption for
(Kellmann, 2010). The results show that the linear directions of three athletes (S1, S3, and S4). High training loads lead to higher

Frontiers in Physiology | www.frontiersin.org 23 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

FIGURE 7 | Individual profile display of the time delays (lag 0–7) for the examined athletes S1–S5 for the recovery and stress dimensions of the ARSS with sRPEh ,
sRPEkm , ACWRh , and ACWRkm as lead variables on the basis of net diagrams. PPC, Physical Performance Capability; MPC, Mental Performance Capability; EB,
Emotional Balance; OR, Overall Recovery; MS, Muscular Stress; LA, Lack of Activation; NES, Negative Emotional State; OS, Overall Stress.

values in the recovery dimensions and lower values in the stress and Lack of Aviation) and emotional-related scales (Emotional
dimensions, however, in different dimensions and with different Balance and Negative Emotional State). In a study of elite rowers
lead variables. Another aspect is, that the contradictory directions Kellmann (2002a) also concluded, that at high training loads,
of action exist only for the mental- (Mental Performance Capacity individual recovery parameters, in this case the items physical

Frontiers in Physiology | www.frontiersin.org 24 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

recovery and fun of the ‘Recovery Cue’ (Kellmann et al., 2002) that the more suitable method for swimming is sRPEkm , which
can be quite lead to higher values. These results illustrate that suggests that in swimming the influence of the volume on the
the relationship between training load and recovery-stress state perceived exertion is greater than the training time. This has to
cannot be easily generalized, because the interdependencies seem be further investigated. In particular, because the sRPEkm method
to be too complex and highly individual. A possible explanatory was used by Nagle et al. (2015) as a modification of Foster’s sRPE
approach can be provided via the ‘Individual Zones of Optimal method (Foster et al., 2001) without previously performing any
Functioning’ model of Hanin (Hanin, 1980, 2002; Salminen studies on the comparability of the methods.
et al., 1995; Hanin and Hanina, 2009), which supports the
assumption of an individual area or zone in which training Limitations
loads are more likely to have positive effects on the recovery- One limitation of the current study can be seen in the small
stress state. The optimal level of training load does not always sample size and that only female athletes were examined.
occur at the midpoint of the continuum but rather varies from To find a sufficient number of highly trained athletes with
individual to individual. That is, some athletes have a zone of high compliance for studies is a general problem. To reach
optimal functioning at the lower end of the continuum, some the target position beyond the individual case to a typology
in the midrange, and others in the upper end (Weinberg and or group statements may with this group size only pointing
Gould, 2007). In addition, the optimal level is most likely not a the way for further investigation. From a sports practical
specific point but rather includes a particular individual range. point of view, it is clear that single case analysis is sensible
Another step would be to analyze the relationship of performance and necessary for athlete monitoring. On the other hand,
against an appropriate modeling to determine individual profiles for a scientific generalization based on a group statistic,
or individual optimal zones for the dimensions of the ARSS the sample is too small. So further single case studies are
on this basis. In accordance with this, some authors proposed required to investigate whether different types of athletes can be
to conduct repeated measurements to establish an individual differentiated. Even the decision to perform no ‘pre-whitening’
baseline from which changes can be determined (Saw et al., 2016, for the time series analysis that can lead to ‘apparent correlations’
2017; Hecksteden et al., 2017; Hitzschke et al., 2017). in the presence of serial dependence, is to question critically.
The results support the sensitivity of the ARSS, as the overall However, the procedures for how to perform this filtering
and physical-related scales and items showed largest changes are particularly controversial for studies with psychological
in response to the physical stress stimulus. Furthermore, the parameters and are in part rejected, especially as there is a
results underline recent indications from Kölling et al. (2015) risk that the problem will be reversed and ‘fake independence’
and Hitzschke et al. (2017) that the subjective ratings of the will be present. In addition, it may be criticized that, for
physical and overall-related scales Physical Performance Capacity, the analysis of the relationships between training load and
Overall Recovery, Muscular Stress and Overall Stress respond on recovery-stress state, the time-delayed effects were examined
acute load. Nevertheless, mental- and emotional-related scales linear by bivariate cross-correlations and not by multivariate
were affected as well, however, with low cross-correlations. This non-linear models. For future investigations, suitable non-
indicates that these dimensions might be more affected by non- linear models should be used or developed. Furthermore,
training-induced stressors or factors. For the analysis of the an individual profile or individual optimum recovery-stress
relationship between training load and recovery-stress state the zone should be defined, relating to corresponding performance
mental- and emotional-related scales seem not to be suitable for data.
this sample.
The difference in time-delay between the sRPE and the ACWR
method was expected due to the calculation method of the ACWR CONCLUSION
using a ratio 1:4 by means of EWMA. The influence of the
method for recording the internal load on time delay is clearly Based on our data we conclude that the ARSS, especially the
visible. Thus, it might be suspected that the ACWR might not physical and overall-related scales (PPC, OR, MS, OS), is a
be suitable to investigate the time-delayed interdependencies of suitable tool for monitoring the acute recovery-stress state in
load and recovery-stress state. However, ACWR still provides a swimmers. Due to its sensitivity to training loads, the individual
simple and practical method for monitoring load, especially with time-delayed correlations between internal load and recovery-
regards to the control of load intensity over longer periods of time stress state was demonstrated. These interrelationships are
(Hulin et al., 2014; Blanch and Gabbett, 2016; Gabbett et al., 2016; considered in relation with CCC > 0.2 to be significant and in
Soligard et al., 2016). many cases as high, e.g., for Muscular Stress with sRPEkm for
sRPEh and sRPEkm appear to be the more suitable methods for athlete S2 (CCC = 0.66).
monitoring the interdependencies between load and recovery- For Physical Performance Capability, Overall Recovery,
stress state, although significant differences were observed. For Muscular Stress and Overall Stress the effectiveness of training
example, S1 equals profile 1 for sRPEh and profile 3 for sRPEkm , load on the recovery-stress state is in line with the theory. Mental
whereas the inverse applies to S2 (sRPEh profile 3 and sRPEkm Performance Capability, Emotional Balance, Lack of Activation,
profile 1). Also in terms of the number of CCC and the amount and Negative Emotional State appear to be less training-
of significant CCC, the sRPE-methods seem to be better suited induced. Comparably high CCC and Ranges result from the
for monitoring than the ACWR-methods. The results indicate four internal load methods, whereby sRPE, especially sRPEkm ,

Frontiers in Physiology | www.frontiersin.org 25 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

shows a stronger relation to recovery-stress state than ACWR. the data and edited the manuscript. MK, AF, and TM edited the
Based on these results and the individual differences in terms manuscript.
of time delay in training response, we recommend performing
intra individual evaluations as well as repeated measurements
to establish an individual baseline from which changes can FUNDING
be determined in order to enable a training optimization
through individual athletes monitoring. For swimming we also This work was supported by the German Federal Institute of
recommend using session RPE, especially sRPEkm , to monitor Sport Science, and realized within RegMan – Optimization of
the internal training load and to use the ARSS with focus Training and Competition: Management of Regeneration in Elite
on physical and overall-scales to monitor the recovery-stress Sports (IIA1-081901/12-16).
state.

ACKNOWLEDGMENTS
AUTHOR CONTRIBUTIONS
The authors would like to thank the athletes, coaches, and
RC planned and designed the study, conducted measurements, research assistants involved in this study for their participation,
analyzed the data, and prepared the manuscript. MP analyzed enthusiasm, and cooperation.

REFERENCES Crowcroft, S., McCleave, E., Slattery, K., and Coutts, A. J. (2017). Assessing the
measurement sensitivity and diagnostic characteristics of athlete-monitoring
Adams, J., and Kirkby, R. (2001). Exercise dependence and overtraining. The tools in national swimmers. Int. J. Sports Physiol. Perform. 12, S295–S2100.
physiological and psychological consequences of excessive exercise. Sports Med. doi: 10.1123/ijspp.2016-0406
Train. Rehabil. 10, 199–222. doi: 10.1080/10578310210395 Fédération Internationale de Natation (2014). FINA Points Table. Available at:
Banister, E. W., Calvert, I. W., Savage, M. V., and Bach, I. M. (1975). A system http://www.fina.org/content/fina-points
model of training for athletic performance. Aust. J. Sci. Med. Sport 7, 57–61. Foster, C., Florhaug, J. A., Franklin, J., Gottschall, L., Hrovatin, L. A., Parker, S.,
Berglund, B., and Safstrom, H. (1994). Psychological monitoring and modulation et al. (2001). A new approach to monitoring exercise training. J. Strength Cond.
of training load of world-class canoeists. Med. Sci. Sports Exerc. 26, 1036–1040. Res. 15, 109–115.
doi: 10.1249/00005768-199408000-00016 Foster, C., Snyder, A., and Welsh, R. (1999). “Monitoring of training, warm up,
Blanch, P., and Gabbett, T. J. (2016). Has the athlete trained enough to return to and performance in athletes,” in Overload, Performance Incompetence, and
play safely? The acute:chronic workload ratio permits clinicians to quantify a Regeneration in Sport, eds M. Lehmann, C. Foster, U. Gastmann, H. Keizer, and
player’s risk of subsequent injury. Br. J. Sports Med. 50, 471–475. doi: 10.1136/ J. M. Steinacker (New York, NY: Kluwer Academic), 43–51. doi: 10.1007/978-
bjsports-2015-095445 0-585-34048-7_4
Borg, G. (1982). Psychophysical bases of perceived exertion. Med. Sci. Sports Exerc. Gabbett, T. J. (2016). The training-injury prevention paradox: should athletes be
14, 377–381. doi: 10.1249/00005768-198205000-00012 training smarter and harder? Br. J. Sports Med. 50, 273–280. doi: 10.1136/
Borresen, J., and Lambert, M. I. (2008). Quantifying training load: a comparison bjsports-2015-095788
of subjective and objective methods. Int. J. Sports Physiol. Perform. 3, 16–30. Gabbett, T. J., Hulin, B. T., Blanch, P., and Whiteley, R. (2016). High training
doi: 10.1123/ijspp.3.1.16 workloads alone do not cause sports injuries: how you get there is the
Bortz, J., and Schuster, C. (2010). Statistik für Human- und Sozialwissenschaftler real issue. Br. J. Sports Med. 50, 444–445. doi: 10.1136/bjsports-2015-
[Statistics for Human and Social Scientists]. Vollständig Überarbeitete und 095567
Erweiterte Auflage, Vol. 7. Berlin: Springer-Verlag. doi: 10.1007/978-3-642- Hammes, D., Skorski, S., Schwindling, S., Ferrauti, A., Pfeiffer, M., Kellmann, M.,
12770-0 et al. (2016). Can the Lamberts and Lambert submaximal cycle test indicate
Bouchard, C., and Rankinen, T. (2001). Individual differences in response to fatigue and recovery in trained cyclists? Int. J. Sports Physiol. Perform. 11,
regular physical activity. Med. Sci. Sports Exerc. 33, S446–S451. doi: 10.1097/ 328–336. doi: 10.1123/ijspp.2015-0119
00005768-200106001-00013 Hanin, Y. (1980). “A study of anxiety in Sports,” in Sport Psychology: An Analysis
Bourdon, P. C., Cardinale, M., Murray, A., Gastin, P., Kellmann, M., Varley, M. C., of Athlete Behavior, ed. W. Straub (Ithaca, NY: Mouvement Publications),
et al. (2017). Monitoring athlete training loads. Consensus statement. Int. J. 236–249.
Sports Physiol. Perform. 12(Suppl. 2), S2161–S2170. doi: 10.1123/IJSPP.2017- Hanin, Y. (2002). “Individually optimal recovery in sports: an application of the
0208 IZOF model,” in Enhancing Recovery: Preventing Underperformance in Athletes,
Bowen, L., Gross, A. S., Gimpel, M., and Li, F.-X. (2017). Accumulated workloads ed. M. Kellmann (Champaign, IL: Human Kinetics), 199–217.
and the acute:chronic workload ratio relate to injury risk in elite youth football Hanin, Y., and Hanina, M. (2009). Optimization of performance in top-level
players. Br. J. Sports Med. 51, 452–459. doi: 10.1136/bjsports-2015-095820 athletes. An action-focused coping approach. Int. J. Sports Sci. Coach. 4, 47–91.
Bresciani, G., Cuevas, M. J., Molinero, O., Almar, M., Suay, F., Salvador, A., et al. doi: 10.1260/1747-9541.4.1.47
(2011). Signs of overload after an intensified training. Int. J. Sports Med. 32, Hautala, A. J., Kiviniemi, A. M., Mäkikallio, T. H., Kinnunen, H., Nissilä,
338–343. doi: 10.1055/s-0031-1271764 S., Huikuri, H. V., et al. (2006). Individual differences in the responses
Brink, M. S., Visscher, C., Coutts, A. J., and Lemmink, K. A. P. M. (2012). Changes to endurance and resistance training. Eur. J. Appl. Physiol. 96, 535–542.
in perceived stress and recovery in overreached young elite soccer players. doi: 10.1007/s00421-005-0116-2
Scand. J. Med. Sci. Sports 22, 285–292. doi: 10.1111/j.1600-0838.2010.01237.x Hecksteden, A., Kraushaar, J., Scharhag-Rosenberger, F., Theisen, D., Senn, S.,
Collette, R. (2016). Trainings- und Erholungsmonitoring im Leistungssport and Meyer, T. (2015). Individual response to exercise training - a statistical
Schwimmen: [Monitoring Training and Recovery in Competitive Swimming], perspective. J. Appl. Physiol. 118, 1450–1459. doi: 10.1152/japplphysiol.00714.
Vol. 1. Hamburg: Kovac. 2014
Coutts, A. J., Wallace, L. K., and Slattery, K. M. (2007). Monitoring changes in Hecksteden, A., Pitsch, W., Julian, R., Pfeiffer, M., Kellmann, M., Ferrauti, A.,
performance, physiology, biochemistry, and psychology during overreaching et al. (2017). A new method to individualize monitoring of muscle recovery in
and recovery in triathletes. Int. J. Sports Med. 28, 125–134. doi: 10.1055/s-2006- athletes. Int. J. Sports Physiol. Perform. 12, 1137–1142. doi: 10.1123/ijspp.2016-
924146 0120

Frontiers in Physiology | www.frontiersin.org 26 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

Herman, L., Foster, C., Maher, M. A., Mikat, R. P., and Porcari, J. P. (2006). Laux, P., Krumm, B., Diers, M., and Flor, H. (2015). Recovery-stress balance and
Validity and reliability of the session RPE method for monitoring exercise injury risk in professional football players. A prospective study. J. Sports Sci. 33,
training intensity. S. Afr. J. Sports Med. 18, 14–17. doi: 10.17159/2078-516X/ 2140–2148. doi: 10.1080/02640414.2015.1064538
2006/v18i1a247 Mäestu, J., Jürimäe, J., and Jürimäe, T. (2005). Monitoring of performance
Hitzschke, B., Holst, T., Ferrauti, A., Meyer, T., Pfeiffer, M., and Kellmann, M. and training in rowing. Sports Med. 35, 597–617. doi: 10.2165/00007256-
(2016). Entwicklung des Akutmaßes zur erfassung von erholung und 200535070-00005
beanspruchung im Sport. [Development of the acute recovery and stress scale]. Maiwald, M., and Rogge, K. (2005). “Zeitreihenanalyse von psychologischen und
Diagnostica 62, 212–226. doi: 10.1026/0012-1924/a000155 physiologischen variablen: eine langzeit-einzelfalluntersuchung. [Time series
Hitzschke, B., Wiewelhove, T., Raeder, C., Ferrauti, A., Meyer, T., Pfeiffer, M., analysis of psychological and physiological variables: a long-term case-by-case
et al. (2017). Evaluation of psychological measures for the assessment of study],” in Zeitreihenanalysen: Mit Beispielen aus der Psychologie, ed. J. Werner
recovery and stress during a shock-microcycle in strength and high-intensity (Berlin: Logos-Verl), 289–302.
interval training. Perform. Enhanc. Health 5, 147–157. doi: 10.1016/j.peh.2017. Martin, D. T., Andersen, M. B., and Gates, W. (2000). Using Profile of Mood
08.001 States (POMS) to monitor high-intensity training in cyclists: group versus case
Hulin, B. T., Gabbett, T. J., Blanch, P., Chapman, P., Bailey, D., and Orchard, J. W. studies. Sport Psychol. 14, 138–156. doi: 10.1123/tsp.14.2.138
(2014). Spikes in acute workload are associated with increased injury risk in elite McNair, D., Lorr, M., and Droppeleman, L. F. (1971). Manual for the Profile of Mood
cricket fast bowlers. Br. J. Sports Med. 48, 708–712. doi: 10.1136/bjsports-2013- States, 1st Edn. San Diego, CA: Educational and Industrial Testing Service.
092524 Meeusen, R., Duclos, M., Foster, C., Fry, A., Gleeson, M., Nieman, D., et al.
Hulin, B. T., Gabbett, T. J., Lawson, D. W., Caputi, P., and Sampson, J. A. (2016). (2013). Prevention, diagnosis and treatment of the overtraining syndrome. Joint
The acute:chronic workload ratio predicts injury: high chronic workload may consensus statement of the European College of Sport Science (ECSS) and
decrease injury risk in elite rugby league players. Br. J. Sports Med. 50, 231–236. the American College of Sports Medicine (ACSM). Eur. J. Sport Sci. 13, 1–24.
doi: 10.1136/bjsports-2015-094817 doi: 10.1080/17461391.2012.730061
Julian, R., Meyer, T., Fullagar, H. H. K., Skorski, S., Pfeiffer, M., Kellmann, M., Morgan, W. P., Brown, D. R., Raglin, J. S., O’Connor, P. J., and Ellickson, K. A.
et al. (2017). Individual patterns in blood-borne indicators of fatigue-trait (1987). Psychological monitoring of overtraining and staleness. Br. J. Sports
or chance. J. Strength Cond. Res. 31, 608–619. doi: 10.1519/JSC.00000000000 Med. 21, 107–114. doi: 10.1136/bjsm.21.3.107
01390 Mujika, I., Busso, T., Lacoste, L., Barale, F., Geyssant, A., and Chatard, J. C.
Kellmann, M. (2000). Psychologische methoden der erholungs-beanspruchungs- (1996). Modeled responses to training and taper in competitive swimmers.
diagnostik. [Psychological methods for the assessment of recovery and stress]. Med. Sci. Sports Exerc. 28, 251–258. doi: 10.1097/00005768-199602000-
Deutsche Z. Sportmed. 51, 253–258. 00015
Kellmann, M. (2002a). Psychologische erholungs- und beanspruchungssteuerung Murray, N. B., Gabbett, T. J., Townshend, A. D., and Blanch, P. (2017a). Calculating
im ruder- und radsport. [Psychological monitoring of stress and recovery in acute:chronic workload ratios using exponentially weighted moving averages
rowing and cycling]. Leistungssport 32, 23–26. provides a more sensitive indicator of injury likelihood than rolling averages.
Kellmann, M. (2002b). “Underrecovery and overtraining: different concepts- Br. J. Sports Med. 51, 749–754. doi: 10.1136/bjsports-2016-097152
similar impact,” in Enhancing Recovery: Preventing Underperformance in Murray, N. B., Gabbett, T. J., Townshend, A. D., Hulin, B. T., and McLellan, C. P.
Athletes, ed. M. Kellmann (Champaign, IL: Human Kinetics), 3–24. (2017b). Individual and combined effects of acute and chronic running loads on
Kellmann, M. (2010). Preventing overtraining in athletes in high-intensity sports injury risk in elite Australian footballers. Scand. J. Med. Sci. Sports 27, 990–998.
and stress/recovery monitoring. Scand. J. Med. Sci. Sports 20(Suppl. 2), 95–102. doi: 10.1111/sms.12719
doi: 10.1111/j.1600-0838.2010.01192.x Nagle, J., McMillan, J., Munkasy, B., Joyner, B., Roorda, A., Scott, M., et al. (2015).
Kellmann, M., and Beckmann, J. (eds). (2018). Sport, Recovery, and Performance: Changes in swim performance and perceived stress and recovery in female
Interdisciplinary Insights. New York, NY: Routledge. collegiate swimmers across a competitive season. J. Swim. Res. 23, 44–53.
Kellmann, M., Bertollo, M., Bosquet, L., Brink, M. S., Coutts, A., Duffield, R., et al. Nässi, A., Ferrauti, A., Meyer, T., Pfeiffer, M., and Kellmann, M. (2017).
(2018). Recovery and performance in sport: consensus statement. Int. J. Sports Development of two short measures for recovery and stress in sport. Eur. J.
Physiol. Perform. 13, 240–245. doi: 10.1123/ijspp.2017-0759 Sport Sci. 17, 894–903. doi: 10.1080/17461391.2017.1318180
Kellmann, M., and Kallus, K. W. (2001). Recovery Stress Questionnaire for Athletes: Nielsen, R. O., Parner, E. T., Nohr, E. A., Sorensen, H., Lind, M., and Rasmussen, S.
User Manual; [CD-ROM Included]. Champaign, IL: Human Kinetics. (2014). Excessive progression in weekly running distance and risk of running-
Kellmann, M., and Kallus, K. W. (2016). “Recovery-stress questionnaire for related injuries: an association which varies according to type of injury.
athletes,” in The Recovery-Stress Questionnaires: User Manual, eds K. W. Kallus J. Orthop. Sports Phys. Ther. 44, 739–747. doi: 10.2519/jospt.2014.5164
and M. Kellmann (Frankfurt: Pearson), 86–134. O’Connor, P. J., and Puetz, T. W. (2005). Chronic physical activity and feelings of
Kellmann, M., Kölling, S., and Hitzschke, B. (2016). Das Akutmaß und die energy and fatigue. Med. Sci. Sports Exerc. 37, 299–305. doi: 10.1249/01.MSS.
Kurzskala zur Erfassung von Erholung und Beanspruchung im Sport - Manual: 0000152802.89770.CF
[The Acute Measure and the Short Scale of Recovery and Stress for Sports - Pelka, M., Schneider, P., and Kellmann, M. (2017). Development of pre- and post-
Manual]. Auflage, Stand, Vol. 1. Hellenthal: Sportverlag Strauß. match morning recovery-stress states during in-season weeks in elite youth
Kellmann, M., Patrick, T., Botterill, C., and Wilson, C. (2002). “The recovery-cue football. Sci. Med. Football 34, 1–6. doi: 10.1080/24733938.2017.1384560
and its use in applied settings: practical suggestions regarding assessment and Raeder, C., Wiewelhove, T., Simola, R. A. D. P., Kellmann, M., Meyer, T.,
monitoring of recovery,” in Enhancing Recovery: Preventing Underperformance Pfeiffer, M., et al. (2016). Assessment of fatigue and recovery in male and female
in Athletes, ed. M. Kellmann (Champaign, IL: Human Kinetics), 219–229. athletes after 6 days of intensified strength training. J. Strength Cond. Res. 30,
Kenttä, G., and Hassmén, P. (1998). Overtraining and recovery. A conceptual 3412–3427. doi: 10.1519/JSC.0000000000001427
model. Sports Med. 26, 1–16. doi: 10.2165/00007256-199826010-00001 Salminen, S., Liukkonen, J., Hanin, Y., and Hyvönen, A. (1995). Anxiety and
Kölling, S., Hitzschke, B., Holst, T., Ferrauti, A., Meyer, T., Pfeiffer, M., et al. (2015). athletic performance of finnish athletes: application of the zone of optimal
Validity of the acute recovery and stress scale. Training monitoring of the functioning model. Pers. Individ. Dif. 19, 725–729. doi: 10.1016/0191-8869(95)
German junior national field hockey team. Int. J. Sports Sci. Coach. 10, 529–542. 00096-O
doi: 10.1260/1747-9541.10.2-3.529 Saw, A. E., Kellmann, M., Main, L. C., and Gastin, P. B. (2017). Athlete self-report
Kölling, S., Steinacker, J. M., Endler, S., Ferrauti, A., Meyer, T., and Kellmann, M. measures in research and practice: considerations for the discerning reader
(2016). The longer the better. Sleep-wake patterns during preparation of the and fastidious practitioner. Int. J. Sports Physiol. Perform. 12, S2127–S2135.
World Rowing Junior Championships. Chronobiol. Int. 33, 73–84. doi: 10.3109/ doi: 10.1123/ijspp.2016-0395
07420528.2015.1118384 Saw, A. E., Main, L. C., and Gastin, P. B. (2016). Monitoring the athlete training
Lambert, M. I., and Mujika, I. (2013). “Overtraining prevention,” in Recovery for response: subjective self-reported measures trump commonly used objective
Performance in Sport, eds C. Hausswirth and I. Mujika (Champaign, IL: Human measures: a systematic review. Br. J. Sports Med. 50, 281–291. doi: 10.1136/
Kinetics), 23–28. bjsports-2015-094758

Frontiers in Physiology | www.frontiersin.org 27 July 2018 | Volume 9 | Article 845

Collette et al. Training Load and Recovery-Stress State

Schimpchen, J., Wagner, M., Ferrauti, A., Kellmann, M., Pfeiffer, M., and Wallace, L., Coutts, A., Bell, J., Simpson, N., and Slattery, K. (2008). Using session-
Meyer, T. (2017). Can cold water immersion enhance recovery in elite RPE to monitor training load in swimmers. Strength Cond. J. 30, 72–76.
Olympic weightlifters? An individualized perspective. J. Strength Cond. Res. 31, doi: 10.1519/SSC.0b013e31818eed5f
1569–1576. doi: 10.1519/JSC.0000000000001591 Wallace, L. K., Slattery, K. M., and Coutts, A. J. (2009). The ecological
Schmitz, A. (2000). “Erkennung von Nichtlinearitäten und Wechselseitigen validity and application of the session-RPE method for quantifying training
Abhängigkeiten in Zeitreihen: [Detection of Nonlinearities and Interdependencies loads in swimming. J. Strength Cond. Res. 23, 33–38. doi: 10.1519/JSC.
in Time Series]. Doctoral dissertation, Bergische Universität Wuppertal, 0b013e3181874512
Wuppertal. Weinberg, R. S., and Gould, D. (2007). Foundations of Sport and Exercise
Schmitz, B. (1996). “Dynamische interaktion in der trainingswissenschaft: Psychology, 4th Edn. Champaign, IL: Human Kinetics.
multivariate ARIMA-modelle: [Dynamic interaction in exercise Wiewelhove, T., Raeder, C., Meyer, T., Kellmann, M., Pfeiffer, M., and Ferrauti, A.
science: multivariate ARIMA models],” in Proceedings of the Workshop (2016). Effect of repeated active recovery during a high-intensity interval-
Zeitreihenanalyse und “Multiple Statistische Verfahren” in der training shock microcycle on markers of fatigue. Int. J. Sports Physiol. Perform.
Trainingswissenschaft: 16 und 17 Juni 1995 in der Sportwissenschaftlichen 11, 1060–1066. doi: 10.1123/ijspp.2015-0494
Fakultät der Universität Leipzig und im Institut für Angewandte Williams, S., West, S., Cross, M. J., and Stokes, K. A. (2016). Better way to
Trainingswissenschaft e.V, Vol. 1, ed. J. Krug (Cologne: Sport und Buch determine the acute:chronic workload ratio? Br. J. Sports Med. 51, 209–210.
Strauss), 11–44. doi: 10.1136/bjsports-2016-096589
Seiler, K. S., and Kjerland, G. O. (2006). Quantifying training intensity Ziemainz, H., and Peters, S. (2010). Die Messung aktuellen wohlbefindens im
distribution in elite endurance athletes: is there evidence for an “optimal” gesundheitssport. Ein kritisches review [Measurement of actual well-being in
distribution? Scand. J. Med. Sci. Sports 16, 49–56. doi: 10.1111/j.1600-0838.2004. wellness sports. A critical review]. Sportwissenschaft 40, 174–181. doi: 10.1007/
00418.x s12662-010-0130-3
Smith, D. J. (2003). A framework for understanding the training process leading
to elite performance. Sports Med. 33, 1103–1126. doi: 10.2165/00007256- Conflict of Interest Statement: The authors declare that the research was
200333150-00003 conducted in the absence of any commercial or financial relationships that could
Soligard, T., Schwellnus, M., Alonso, J.-M., Bahr, R., Clarsen, B., Dijkstra, be construed as a potential conflict of interest.
H. P., et al. (2016). How much is too much? (Part 1) International
Olympic Committee consensus statement on load in sport and risk Copyright © 2018 Collette, Kellmann, Ferrauti, Meyer and Pfeiffer. This is an open-
of injury. Br. J. Sports Med. 50, 1030–1041. doi: 10.1136/bjsports-2016- access article distributed under the terms of the Creative Commons Attribution
096581 License (CC BY). The use, distribution or reproduction in other forums is permitted,
Toubekis, A. G., Drosou, E., Gourgoulis, V., Thomaidis, S., Douda, H., and provided the original author(s) and the copyright owner(s) are credited and that the
Tokmakidis, S. P. (2013). Competitive performance, training load and original publication in this journal is cited, in accordance with accepted academic
physiological responses during tapering in young swimmers. J. Hum. Kinet. 38, practice. No use, distribution or reproduction is permitted which does not comply
125–134. doi: 10.2478/hukin-2013-0052 with these terms.

Frontiers in Physiology | www.frontiersin.org 28 July 2018 | Volume 9 | Article 845

 ORIGINAL RESEARCH
published: 14 June 2018
doi: 10.3389/fphys.2018.00713

Effects of Initial Performance, Gross

Efficiency and V̇O2peak
Characteristics on Subsequent
Adaptations to Endurance Training in
Competitive Cyclists
Knut Skovereng 1* , Øystein Sylta 2 , Espen Tønnessen 3 , Daniel Hammarström 4 ,
Jørgen Danielsen 1 , Stephen Seiler 2 , Bent R. Rønnestad 4 and Øyvind Sandbakk 1
1
Centre for Elite Sports Research, Department of Neuroscience, Norwegian University of Science and Technology,
Trondheim, Norway, 2 Faculty of Health and Sport Sciences, University of Agder, Kristiansand, Norway, 3 The Norwegian
Olympic Federation, Oslo, Norway, 4 Section for Sport Science, Lillehammer University College, Lillehammer, Norway

Edited by:
The present study investigated the effects of initial levels of cycling performance,
Thomas Leonhard Stöggl, peak oxygen uptake (V̇O2peak ) and gross efficiency (GE) on the subsequent
University of Salzburg, Austria adaptations of these variables and their relationship following high-intensity
Reviewed by: training (HIT) designed to increase V̇O2peak in competitive cyclists. Sixty cyclists
Jeremiah John Peiffer,
Murdoch University, Australia (V̇O2peak = 61 ± 6 mL kg−1 min−1 ) were assigned a 12-week training program
Beat Knechtle, consisting of twenty-four supervised high-intensity interval training sessions and
University Hospital Zurich, Switzerland
ad libitum low intensity training. GE was calculated at 125, 175, and 225 W and
*Correspondence:
Knut Skovereng
performance was determined by mean power during a 40-min time-trial (Power40 min ).
[email protected] In addition to correlation analyses between initial level and pre- to post-intervention
changes of the different variables, we compared these changes between four groups
Specialty section:
This article was submitted to
where participants were categorized with either low and/or high initial levels of V̇O2peak
Exercise Physiology, and GE. Average volume of high- and low-intensity training during the 12-week
a section of the journal
intervention was 1.5 ± 0.3 and 8.3 ± 2.7 h. week−1 , respectively. Following the 12-week
Frontiers in Physiology
training period, there was a significant increase in absolute and body mass normalized
Received: 02 June 2017
Accepted: 23 May 2018 V̇O2peak and Power40 min (p < 0.05) and a significant decrease in GE (p < 0.05) for
Published: 14 June 2018 all athletes pooled. There was no change in body mass following the 12-week training
Citation: period. We found a moderate negative correlation between initial level of V̇O2peak and
Skovereng K, Sylta Ø, Tønnessen E,
Hammarström D, Danielsen J,
the change in V̇O2peak following the training period (r = −0.32; p < 0.05). A small
Seiler S, Rønnestad BR and negative correlation was also found between initial Power40 min and its change following
Sandbakk Ø (2018) Effects of Initial
training both when expressed in absolute power and power normalized for body mass
Performance, Gross Efficiency
and V̇O2peak Characteristics on (r = −0.27 and −0.28; both p < 0.05). A moderate negative correlation was also found
Subsequent Adaptations between initial levels for GE and its change following training (r = −0.44; p < 0.01).
to Endurance Training in Competitive
Cyclists. Front. Physiol. 9:713.
There were no differences between the four groups based on initial levels of V̇O2peak
doi: 10.3389/fphys.2018.00713 and GE in the response to training on V̇O2peak , GE, or Power40 min (all p > 0.12).

Frontiers in Physiology | www.frontiersin.org 29 June 2018 | Volume 9 | Article 713

Skovereng et al. Physiological Characteristics and Training Adaptations

In conclusion, the present findings suggest that there are statistically significant effects
of initial levels of cycling performance and V̇O2peak and on the subsequent adaptations
following a 12-week HIT program, but the small and moderate effects indicate limited
influence on training practice.
Keywords: cycling, performance, maximal oxygen consumption, gross efficiency, high intensity training, interval
training

INTRODUCTION Therefore, the purpose of the present study was to investigate

how baseline characteristics of cycling performance, V̇O2peak and
Cycling performance requires high aerobic energy turnover GE influence subsequent adaptations of these parameters and
and effective transfer of that energy to external power. Hence, their interplay following a 12-week high-intensity intervention
the peak oxygen uptake (V̇O2peak ) and gross efficiency (GE), designed to increase V̇O2peak in competitive cyclists. We
defined as the ratio of work rate to metabolic rate, are two key hypothesized that a large training load would lead to increased
determinants of performance (Joyner and Coyle, 2008). Exercise V̇O2peak and reduced GE, but there would be no influence of
training interventions focusing on high-intensity training (HIT) initial levels of V̇O2peak and GE.
in cyclists have repeatedly shown enhanced V̇O2peak as the main
physiological adaptation (Lucia et al., 2000; Bacon et al., 2013)
whereas GE has been reported to both be unaffected (Ronnestad MATERIALS AND METHODS
et al., 2015) and improved (Hintzy et al., 2005; Hopker et al.,
2009, 2010). However, the study of Ronnestad et al. (2015), where Sixty-three male competitive cyclists (38 ± 8 year, V̇O2peak :
GE was unaffected, used an effort-based approach to control 61 ± 6 mL kg−1 min−1 ) were recruited to take part in the
intensity, which likely induced higher intensity compared to the current multicentre study, involving three test centers completing
studies by Hopker et al. (2009, 2010) who regulated intensity by the same experimental trial. All participants were categorized
set blood lactate levels. The study by Hintzy et al. (2005) used a as well-trained (Jeukendrup et al., 2000) with 9 ± 3 h of
combination of moderate- and high-intensity, but the untrained weekly training in the year prior to participation. All participants
participants in the study are a likely reason for the increased completed the intervention, however, three participants were
GE. Furthermore, sprint and strength training led to reduced excluded from the final analyses due to absence from post-testing.
oxygen cost of submaximal cycling (i.e., GE likely increased) The study was approved by the ethics committee of the Faculty
(Paton and Hopkins, 2005), whereas low- and moderate-intensity for Health and Sport Science, University of Agder, and registered
training led to unchanged (Kristoffersen et al., 2014) or increased with the Norwegian Social Science Data Services (NSD). All
(Hopker et al., 2012) GE in cycling. Overall, the current literature athletes gave their verbal and written informed consent prior to
indicates that both training intensity and fitness level influence study participation. The present study is part of a larger research
the responses on GE. project and thus, the intervention period, testing procedures and
It is generally believed that greater training loads, achieved instrument are described in brief. A detailed description of the
through both volume of training and sufficient intensity, are intervention period, testing procedures, and instruments can be
required to trigger V̇O2peak or performance adaptations in found in Sylta et al. (2016).
individuals with high compared to low initial levels of V̇O2peak
or performance. In previous studies, there has been reported Intervention Period
an effect of initial level of V̇O2peak on the response to In brief, after a 6-week preparation and familiarization period,
exercise (Saltin et al., 1969), but more recently no significant the training intervention consisted of three 4-week mesocycles.
correlation between initial V̇O2peak and the response to a training During the last week in each mesocycle, participants were advised
intervention was reported (Kohrt et al., 1991; Skinner et al., to reduce low-intensity training (LIT) volume by 50% compared
2001). However, these studies (Kohrt et al., 1991; Skinner to previous weeks. In addition, HIT session frequency was
et al., 2001) do not include any measures of performance, reduced from 3 to 1 during the last week of each mesocycle.
and the participants had a relatively low initial level of fitness In total, each participant was prescribed twenty-four supervised
(reported V̇O2peak < 31.8 mL kg−1 min−1 ). As such, inference HIT sessions in addition to testing and self-organized ad libitum
to performance and/or more highly trained populations cannot LIT. All training was recorded using an online training diary and
be made based on these studies. a heart rate monitor was worn for all exercise training. Three
The interaction between V̇O2peak and GE adaptations in different HIT session models were utilized and all included a
response to a training intervention requires further elucidation. self-selected warm up of 20–30 min of LIT, followed by four
An inverse relationship between V̇O2peak and GE has been found high intensity interval efforts of 4, 8, or 16 min at a self-selected
(Lucia et al., 2002) and also an inverse relationship between the cadence, separated by 2 min rest, followed by 10–20 min of LIT
change in GE and V̇O2peak following a training intervention as a cool-down. All HIT training was performed while supervised
(Santalla et al., 2009; Hopker et al., 2012). However, none of these on Computrainer cycling trainers (RacerMate Inc., Seattle, WA,
interventions led to increased V̇O2peak at the group level. United States) using the participants’ own bikes and, additionally,

Frontiers in Physiology | www.frontiersin.org 30 June 2018 | Volume 9 | Article 713

Skovereng et al. Physiological Characteristics and Training Adaptations

blood lactate measurements were taken from a selection of the United States). HR was measured using Polar V800 (Polar Electro
participants on each session. The participants were instructed to Oy, Kempele, Finland) and blood lactate was measured using
cycle at their maximal sustainable intensity for the entire session, capillary blood samples taken from the fingertip (Biosen C-Sport,
and provided with continuous feedback regarding cadence, heart EKF diagnostics, Cardiff, United Kingdom).
rate (HR), and power output. All participants were prescribed the
same number of the three different HIT session models.
Data Analysis
Four subgroups (each with N = 8) were identified from the
Testing Procedures complete cohort, characterized by high or low V̇O2peak relative
In brief, on test day 1, a submaximal, incremental exercise test to body mass and high or low GE. V̇O2peak and GE rank within
consisting of 5-min steps was performed on a bicycle ergometer the cohort from the pre-test was used as criteria for selecting
at work rates of 125, 175, and 225 W. V̇O2 and the respiratory participants to the respective groups (Figure 1). The HH (high
exchange ratio (RER) was used to calculate the metabolic rate GE and high V̇O2peak ) and LL (low GE and low V̇O2peak )
during the three work rates. The work rate was then divided by groups were selected to yield the highest and lowest average rank
the metabolic rate to calculate GE. The 125, 175, and 225 W work for V̇O2peak and GE, respectively. The HL (high GE and low
rates used to calculate GE corresponded to 34 ± 4, 48 ± 5, and V̇O2peak ) group was selected to have the highest V̇O2peak rank
61 ± 6% of the participants’ peak power output (PPO) and to while maintaining an average rank as close to the mean as possible
43 ± 4, 53 ± 5, and 65 ± 6% of their V̇O2peak achieved during and LH (low GE and high V̇O2peak ) group was selected to have
the incremental test. V̇O2 , RER, and HR were measured during the highest possible GE score while maintaining an average rank
the last 2.5 min of each step when a steady state condition had as close to the mean as possible.
occurred. Blood lactate was measured after 4.5 min of each step.
After 10 min recovery, an incremental test to exhaustion was
performed starting with 1 min of cycling at 3 W kg−1 (rounded Statistical Analyses
down to nearest 50 W) and subsequent increases of 25 W every Pearsons correlation coefficients were calculated to determine
minute. Strong verbal encouragement was provided throughout relationships between cycling performance, V̇O2peak and GE and
the test. V̇O2peak was calculated as the average of the two highest a descriptors of effect sizes of the correlations were calculated
consecutive 30-s V̇O2 measurements and PPO was calculated as according to http://www.sportsci.org/resource/stats/effectmag.
the mean power output during the final 60 s the participants html. Differences between PRE and POST conditions were
were able to maintain power output during the incremental test. evaluated using a one-way ANOVA. Differences in responses
HRpeak was observed during the final 5 s before exhaustion and among the four different groups were evaluated using a
blood lactate was measured 60 s post-exhaustion. two-way repeated measures ANOVA. We calculated the smallest
On test day 2, participants performed a 40-min time trial worthwhile change in performance as 0.20 multiplied with the
(Power40min ) after a 30-min warm-up at a self-selected power standard deviation (SD) at the pre-test (Hopkins et al., 2009). All
output. The Power40min test was conducted under supervision in data analyses were conducted using SPSS 22.0 (SPSS Inc, Chicago,
a well-ventilated room. The temperature and relative humidity
were similar at the pre- and post-tests and on both occasions
ad libitum water intake was allowed. Participants were blinded
to all feedback except for elapsed time and the participants were
instructed to cycle at the highest possible mean power output for
40 min.

Instruments and Materials

In brief, all physiological tests were performed on a cycling
ergometer [Velotron (RacerMate, Seattle, WA, United States)
or Lode Excalibur Sport (Lode B. V., Groningen, Netherlands)]
adjusted to the participant’s preference. The type of ergometer
was consistent at pre- and post-tests. Participants were instructed
to remain seated during all tests, with self-selected cadence.
Performance tests and all HIT sessions were performed using
each participant’s personal road bike mounted on Computrainer
FIGURE 1 | Participant distribution of V̇O2peak relative to body mass and GE
LabTM trainers (RacerMate, Seattle, WA, United States), at the baseline measurement. Open circles indicate participants in the HH
calibrated according to the manufacturer’s specifications. V̇O2 (high V̇O2peak and high efficiency) group. Filled circles indicate participants in
was measured using Oxycon ProTM with mixing chamber the HL (high V̇O2peak and low efficiency) group. Filled squares indicate
(Oxycon, Jaeger GmbH, Hoechberg, Germany) calibrated using participants in the LH (low V̇O2peak and high efficiency) group. Open squares
indicate participants in the LL (low V̇O2peak and low efficiency) group.
gases of known concentrations before every test. The flow
X indicates subjects not included in any of the sub groups (N = 28). Dotted
turbine (Triple V, Erich Jaeger) was calibrated using a 3L lines indicate the mean value for V̇O2peak and gross efficiency.
calibration syringe (5530 series; Hans Rudolph, Kansas, MO,

Frontiers in Physiology | www.frontiersin.org 31 June 2018 | Volume 9 | Article 713

Skovereng et al. Physiological Characteristics and Training Adaptations

IL, United States) and are presented as mean ± SD with statistical

significance accepted as α ≤ 0.05.

RESULTS
V̇O2peak ranged from 48 to 76 mL. kg−1 min−1 and GE from
16.5 to 20.8%. The average Power40min ranged from 194 to
342 W and 2.3 to 4.6 W kg−1 when expressed as absolute
values and normalized to body mass, respectively. A plateau in
V̇O2 occurred in 54 and 58 out of the 60 pre- and post-tests,
respectively. Following the 12-week training intervention, there
was a significant increase in V̇O2peak , PPO, and Power40min but
there was a decrease in GE (Table 1; all p < 0.05). The weekly
training volume during the intervention was 10 ± 3 h, of which
97 ± 4% was endurance training. The mean power output during FIGURE 2 | Intensity distribution from heart rate monitored training throughout
all HIT intervals was 310 ± 40 W and average blood lactate taken the study. Zone 1: <75% of HRpeak , zone 2: 75–85% HRpeak , zone 3:
during the sessions was 8.7 ± 4.0 mmol L−1 . The overall intensity 85–90% of HRpeak , zone 4: 90–95% of HRpeak , and zone 5: >95% of HRpeak .

distribution for all training based on heart rate data is presented

in Figure 2.
Initial performance during the Power40min displayed a small
Initial Characteristics negative correlation with the change in performance following
Initial absolute V̇O2peak did not correlate significantly with the training intervention both when expressed as absolute values
the change in V̇O2peak (Figure 3A) following the 12-week (Figure 5A; r = −0.28; p < 0.05) and relative to body mass
intervention (Figure 2A; r = −0.18; p = 0.17). However, for (Figure 5B; r = −0.27; p < 0.05). Initial PPO showed small and
V̇O2peak expressed relative to body mass, there was a moderate moderate correlations with the change following the intervention
significant negative correlation between the initial values and both for absolute power output (r = −0.23; p = 0.08) and values
the change in V̇O2peak following the intervention (Figure 3B; relative to body mass (r = −0.42; p < 0.01), respectively.
r = −0.32; p < 0.05). Smallest worthwhile changes for body
mass normalized V̇O2peak , GE, and body mass normalized GE and V̇O2peak Relationship and Their
performance during the Power40min were 1.2 mL kg−1 min−1 , Relationship With Performance
0.18%, and 0.08 W kg−1 , respectively. There was a moderate negative correlation between GE and both
There was a significant negative moderate correlation between absolute V̇O2peak (Figure 6A; r = −0.36, p < 0.01) and V̇O2peak
initial GE and percentage change in GE following the 12-week relative to body mass (Figure 6B; r = −0.27, p < 0.05).
intervention (Figure 4; r = −0.44; p < 0.01). There was no change There were significant moderate correlations between change
in cadence during the stages used for GE calculation from the pre- in performance during the Power40min and change in absolute
to the post-test (p = 0.35). V̇O2peak (r = 0.30, p < 0.05) and V̇O2peak relative to body
mass (Figure 7A; r = 0.38, p < 0.01). There was no
significant correlation between change in performance during the
TABLE 1 | Pre- and post-test values of performance and physiological factors Power40min and change in GE (Figure 7B; r = −0.06, p = 0.63).
following the 12-week high-intensity training intervention in 60 well-trained cyclists.

Pre Post Sub-group Analyses

Characteristics of the participants included in the four
V̇O2peak (mL min−1 ) 4859 ± 462 5078 ± 484∗
sub-groups are presented in Table 2. The HH group showed the
V̇O2peak (mL kg−1 min−1 ) 61 ± 6 65 ± 6∗
best performance on both the incremental and the time-trial test,
GE (%) 19.0 ± 0.9 18.6 ± 0.9∗
but only when the results where normalized for body mass. Total
PPO (W) 372 ± 40 384 ± 29∗
training volume, LIT, and HIT did not differ between the four
PPO (W kg−1 ) 4.7 ± 0.5 4.9 ± 0.4∗
subgroups (all p > 0.36).
Power40min (W) 282 ± 30 301 ± 31∗
There was a significant main effect of the training intervention
Power40min (W kg−1 ) 3.6 ± 0.4 3.9 ± 0.4∗
on V̇O2peak , Power40min (Figure 8; all p < 0.05) and PPO
Peak blood lactate (mmol L−1 ) 12.1 ± 2.3 12.7 ± 2.1
(not shown) during the incremental test. However, there was
RER 1.15 ± 0.1 1.15 ± 0.1
no difference between the four groups in the change in
Body mass (kg) 80 ± 8 78 ± 8
V̇O2peak , PPO, Power40min or GE (all p > 0.12) and all four
Peak oxygen uptake (V̇O2peak ), gross efficiency (GE) averaged from 125, 175, and groups maintained the selection characteristics following the
225 W, peak power output from the incremental test (PPO), average power output
intervention (i.e., the HH and HL had higher V̇O2peak compared
during the 40-min time trial (Power40min ), peak blood lactate, and RER-values from
the incremental test and body mass are presented as arithmetic mean (±SD). to the LH and LL and the HH and LH had higher GE compared
Asterisks indicate a significant change from the pre- to post-test (p < 0.05). to the HL and LL) (all p < 0.05).

Frontiers in Physiology | www.frontiersin.org 32 June 2018 | Volume 9 | Article 713

Skovereng et al. Physiological Characteristics and Training Adaptations

FIGURE 3 | Correlation between the change in V̇O2peak during the training intervention and initial V̇O2peak in absolute values (A) and normalized for body mass (B).

the relationship in groups based on initial GE and V̇O2peak

characteristics in combination.
While the present study shows significant effects of initial
performance and V̇O2peak on the subsequent adaptation to
HIT in highly trained cyclists, the effect is only moderate. An
effect of initial fitness on the V̇O2peak responses has also been
reported previously in less trained individuals (Saltin et al., 1969),
although other studies (Kohrt et al., 1991; Skinner et al., 2001)
showed no effect. The HIT element in the present study had an
overall positive effect on V̇O2peak and endurance performance,
which differs from the previous studies that utilized a lower
training intensity (e.g., 75% of V̇O2peak (Skinner et al., 2001)).
Furthermore, the total training load added in our study seemed
to be sufficient to elicit an overload stimulus and thereby trigger
FIGURE 4 | Correlation between the percentage change in gross efficiency
endurance adaptions in most of the athletes. This is exemplified
during the training intervention and initial gross efficiency. by a recent study where further increasing the training load
effectively elicited training adaptations in participants who were
unresponsive to 180 min of moderate to high intensity exercise
per week (Montero and Lundby, 2017). However, even in our
DISCUSSION study, the observed improvements were relatively modest, which
may partly be due to differences in how optimal the added
The purpose of the present study was to investigate how the training load was for each individual participant. In addition, the
initial characteristics of cycling performance, V̇O2peak , and GE, as subsequent recovery phase is of importance for adaptations, and
well as their interplay, influences subsequent adaptations in these e.g., subjects need to be healthy and avoid other factors which
parameters following a 12-week HIT intervention in well-trained might negatively influence adaptations.
cyclists. In contrast to the positive effects of the 12-week high intensity
The training intervention led to an increase in performance training intervention on V̇O2peak (∼5%), GE declined slightly
and V̇O2peak but a decrease in GE. Furthermore, we found following the intervention (∼2%), and there was a moderate
significant associations between the athletes’ initial levels of association between initial level of GE and the subsequent
GE, V̇O2peak , and performance and subsequent adaptations of response. However, there were no significant differences in
that same parameter. However, parameter specific initial levels adaptation between the groups with high and low GE. Previous
explained only 4–18% of the variance in adaptations, and when literature examining changes in GE following high intensity
we compared our four groups with combinations of high and/or training reported no effect (Ronnestad et al., 2015) or an increase
low initial levels of GE and V̇O2peak , no differences in adaptations in GE when utilizing an exercise intensity equivalent to five
were found among groups. Overall, our study shows that the heart rate beats above the work rate eliciting 4 mmol·L−1
impact of initial characteristics of cycling performance, V̇O2peak , of blood lactate (Hopker et al., 2010). However, previous
and GE on the subsequent responses to these parameters are studies in participants of a similar performance level to our
relatively small, and the effect disappears if we investigate participants, who utilized low- to moderate-intensity training,

Frontiers in Physiology | www.frontiersin.org 33 June 2018 | Volume 9 | Article 713

Skovereng et al. Physiological Characteristics and Training Adaptations

FIGURE 5 | Correlation between the change in time trial performance during the training intervention and initial time trial performance (Power40min ) in absolute units
(A) and normalized for body mass (B).

FIGURE 6 | Correlation between changes in V̇O2peak relative to body mass (A) and absolute V̇O2peak (B) and changes in gross efficiency (GE) during the training
intervention.

FIGURE 7 | Correlation between change in 40-min time trial performance normalized for body mass and change in V̇O2peak normalized for body mass (A) and
between changes in 40-min time trial performance normalized for body mass and changes in gross efficiency (B) during the training intervention.

Frontiers in Physiology | www.frontiersin.org 34 June 2018 | Volume 9 | Article 713

Skovereng et al. Physiological Characteristics and Training Adaptations

TABLE 2 | Pre-test characteristics and training data during the 12-week intervention period for the four sub-groups.

Hi V̇O2peak /Hi GE Hi V̇O2peak /Low GE Low V̇O2peak /Hi GE Low V̇O2peak /Low GE

Age (years) 32 ± 6cd 36 ± 6d 41 ± 6a 44 ± 4ab

Height (cm) 181 ± 9 184 ± 5 181 ± 4 184 ± 6
Heart rate peak (bpm) 194 ± 7 194 ± 10 184 ± 9 185 ± 13
Body mass (kg) 71.3 ± 7.9bcd 79.6 ± 2.7ad 80.9 ± 3.2ad 88.9 ± 10.3abc
V̇O2peak (mL min−1 ) 4784 ± 509b 5362 ± 268acd 4622 ± 460b 4916 ± 488b
V̇O2peak (mL kg−1 min−1 ) 67.3 ± 5.0cd 67.4 ± 2.8cd 57.2 ± 3.4ab 55.5 ± 4.1ab
GE (%) 19.9 ± 0.4bd 17.9 ± 0.7ac 20.1 ± 0.7bd 18.2 ± 0.7ac
PPO (W) 367 ± 43 394 ± 38 367 ± 43 361 ± 46
PPO (W kg−1 ) 5.2 ± 0.4cd 4.9 ± 0.4cd 4.5 ± 0.5abd 4.1 ± 0.4abc
Power40min (W) 283 ± 22 282 ± 20 274 ± 40 281 ± 39
Power40min (W kg−1 ) 4.0 ± 0.3abc 3.5 ± 0.2a 3.4 ± 0.5a 3.2 ± 0.5a
HIT volume (h week−1 ) 1.4 ± 0.3 1.5 ± 0.3 1.5 ± 0.3 1.6 ± 0.7
LIT volume (h week−1 ) 9.1 ± 3.7 7.3 ± 2.0 7.3 ± 2.4 7.7 ± 2.3

Pre-test values for the HH (high V̇O2peak and high efficiency), HL (high V̇O2peak and low efficiency), LH (low V̇O2peak and high efficiency), and LL (low V̇O2peak and low
efficiency) groups. V̇O2peak , gross efficiency (GE), peak power output from the incremental test (PPO) and average power output during the 40-min time trial (Power40min )
are presented as mean (±SD). a Indicates a difference from HH (p < 0.05). b Indicates a difference from HL (p < 0.05). c Indicates a difference from LH (p < 0.05).
d Indicates a difference from LL (p < 0.05).

reported improved GE and a significant positive relationship higher workloads (i.e., training intensity average interval work
between changes in GE and performance (Hopker et al., 2010, rate of 310 W). Additionally, since 225 W exceed 60% of PPO,
2012). there is a possibility that the V̇O2 slow component influence our
A potential explanation for the different effect of training on measurements slightly, which could have led to an overestimation
GE between our study and others might be the HIT applied in of the decline in GE from pre- to post-test. However, since
our study, which primarily influenced performance via enhanced there was no difference in the change in GE for the three
V̇O2peak (Sylta et al., 2016)). Previous studies (Hopker et al., 2010, work rates used for GE calculation and the corresponding RER
2012) have used a lower intensity compared to the present study, measurements were below 1.0, a possible influence would be
and a lower intensity may be more effective for enhancing GE minor.
at a submaximal work rate. In the present study, training was As expected, we found a positive relationship between change
executed as HIT with the addition of ad libitum LIT. A possible in cycling performance and change in V̇O2peak . In contrast,
influencing factor on the decline in GE in the present study may change in GE was not related to a change in performance. This
be the low amount of training as moderate-intensity exercise, inverse relationship between the change in GE and V̇O2peak
which may be important for maintaining GE (Hopker et al., 2009; (Figure 5) corresponds with previous findings (Lucia et al., 2002;
Kristoffersen et al., 2014). Hopker et al., 2012). However, in contrast to previous studies
Intensity in the present study was controlled utilizing an effort which have shown an increase in GE and small changes in
based approach where the participants are instructed to aim to V̇O2peak (Lucia et al., 2002; Hopker et al., 2012), we show the
achieve the highest possible average power output within each same inverse relationship when the average GE decreases and the
session (Seiler et al., 2013). This is the same approach used by average V̇O2peak increases.
Ronnestad et al. (2015) who demonstrated unchanged GE after Also contrary to previous findings (Hopker et al., 2012) is
HIT training. It is possible that the effort based intensity control an increased performance despite a decrease in GE. Although
leads to higher intensity, as shown through the high blood lactate the 12-week HIT intervention led to a large increase in V̇O2peak
levels during intervals in the present (i.e., average blood lactate of that positively influenced cycling performance, it appears that
8.9 mmol L−1 ) and a previous study using intervals of 4 to 8 min cycling efficiency is slightly reduced, especially in those with
duration (Seiler et al., 2013), compared to the approaches used large V̇O2peak improvements. However, it is important to keep in
by Hopker et al. (2010) who demonstrated increased GE with mind the relatively short duration of the intervention executed
intensity based on absolute blood lactate levels of 4 mmol L−1 + in this study. As demonstrated by the findings in this study,
5 heart rate beats per minute. Furthermore, Hopker et al. (2009, the decrease in GE over a relatively short period where the
2010) proposed that training at moderate intensity, eliciting less objective of training is to increase V̇O2peak was not detrimental to
than 4 mmol L−1 blood lactate, is important for maintaining GE. performance due to the improved V̇O2peak . However, in general,
Hence, training at (very) high-intensity may lead to unchanged long-term decreases in GE can be detrimental to performance.
or declined GE, whereas training at lower intensity may have the Since studies using low- and moderate-intensity training, as
opposite effect. well as sprint and strength training (Paton and Hopkins, 2005;
However, the decline in GE may also be influenced by the fact Hopker et al., 2012) have shown increases in GE, cyclists should
that it was calculated at a moderate intensity (i.e., the average likely use a combination of different training intensities to
of 125, 175, and 225 W) and might not reflect what occurs at optimize their long-term performance development. This also

Frontiers in Physiology | www.frontiersin.org 35 June 2018 | Volume 9 | Article 713

Skovereng et al. Physiological Characteristics and Training Adaptations

FIGURE 8 | Group responses from the pre-test to the post-test for the HH (high V̇O2peak and high efficiency), HL (high V̇O2peak and low efficiency), LH (low V̇O2peak
and high efficiency), and the LL (low V̇O2peak and low efficiency) groups for V̇O2peak , time trial performance and gross efficiency (GE). ∗ Indicates a significant main
effect of the training intervention.

represents a typical training pattern over a season for elite an appropriate training stimulus is administered, the initial
cyclists, where a combination of intensities and periods with physiological characteristics have little effect on the training
different focus is employed (Lucia et al., 2000; Hopker et al., adaptation in a large cohort of well-trained athletes. It
2009). Although different types of training influence GE and demonstrates that an appropriate training stimulus can further
V̇O2peak differently, both high and moderate intensity training enhance performance, independent of the initial physiological
may be necessary for optimal performance increases during a characteristics of even well-trained athletes. A known limitation
competitive season. Additionally, the long-term effect on GE of the present arises from the baseline characteristic and change
and V̇O2peak is potentially different from the effects seen in this following the intervention are not independent measurements,
relatively short 12-week intervention compared to the effect on we violate the assumption of independence in the correlation
performance. analyses. This may lead to an effect known as regression to the
We believe the finding of a small effect of baseline level mean (Altman and Bland, 1994) which ultimately would lead
has important practical implications and demonstrates that if to an overestimation of the correlation. When calculated using

Frontiers in Physiology | www.frontiersin.org 36 June 2018 | Volume 9 | Article 713

Skovereng et al. Physiological Characteristics and Training Adaptations

Oldham’s correction to minimize the effect, the correlations in intervention, GE was reduced and the changes in GE negatively
the present study were weakened. The additional finding that correlated with changes in V̇O2peak . However, when comparing
the groups based on differences in baseline characteristics show adaptations between groups with different levels of V̇O2peak
no difference in the adaptation to the training intervention and/or GE, we found no differences. Overall, this study indicates
also supports our interpretation. Although the groups in the that the effects of initial level of performance and physiological
present study where comprised of only eight athletes which limit capacities on subsequent adaptations are relatively small, and the
statistical power with multiple comparisons, however, the group effect disappears if we investigate the relationship in groups based
size is comparable to similar studies on training adaptation. on GE and V̇O2peak characteristics.

CONCLUSION AUTHOR CONTRIBUTIONS

In conclusion, the present study demonstrates statistically KS, ØSa, ET, JD, BR, and ØSy contributed in conceptualization
significant, but practically trivial effects of initial levels of the study. KS, ØSa, DH, and JD contributed in data collection. KS,
V̇O2peak and performance on subsequent adaptations following ØSa, DH, and ØSy contributed in data handling and statistical
a 12-week HIT intervention in well-trained cyclists. In contrast analysis. KS, ØSa, ET, DH, JD, SS, BR, and ØSy contributed in
to the improvements in performance and V̇O2peak following this preparing the manuscript.

REFERENCES Med. Sci. Sports Exerc. 34, 2079–2084. doi: 10.1249/01.MSS.0000039306.

92778.DF
Altman, D. G., and Bland, J. M. (1994). Statistics notes - regression towards the Montero, D., and Lundby, C. (2017). Refuting the myth of non-response to exercise
mean - reply. Br. Med. J. 309:539. training: ‘non-responders’ do respond to higher dose of training. J. Physiol. 595,
Bacon, A. P., Carter, R. E., Ogle, E. A., and Joyner, M. J. (2013). VO2max trainability 3377–3387. doi: 10.1113/JP273480
and high intensity interval training in humans: a meta-analysis. PLoS One Paton, C. D., and Hopkins, W. G. (2005). Combining explosive and high-resistance
8:e73182. doi: 10.1371/journal.pone.0073182 training improves performance in competitive cyclists. J. Strength Cond. Res. 19,
Hintzy, F., Mourot, L., Perrey, S., and Tordi, N. (2005). Effect of endurance 826–830. doi: 10.1519/R-16334.1
training on different mechanical efficiency indices during submaximal cycling Ronnestad, B. R., Hansen, J., Vegge, G., Tonnessen, E., and Slettalokken, G.
in subjects unaccustomed to cycling. Can. J. Appl. Physiol. 30, 520–528. (2015). Short intervals induce superior training adaptations compared with
doi: 10.1139/h05-138 long intervals in cyclists – an effort-matched approach. Scand. J. Med. Sci. Sports
Hopker, J., Coleman, D., Jobson, S. A., and Passfield, L. (2012). Inverse relationship 25, 143–151. doi: 10.1111/sms.12165
between VO2max and gross efficiency. Int. J. Sports Med. 33, 789–794. Saltin, B., Hartley, L. H., Kilbom, A., and Astrand, I. (1969). Physical training in
doi: 10.1055/s-0032-1304640 sedentary middle-aged and older men. II. Oxygen uptake, heart rate, and blood
Hopker, J., Coleman, D., and Passfield, L. (2009). Changes in cycling efficiency lactate concentration at submaximal and maximal exercise. Scand. J. Clin. Lab.
during a competitive season. Med. Sci. Sports Exerc. 41, 912–919. doi: 10.1249/ Invest. 24, 323–334. doi: 10.3109/00365516909080169
MSS.0b013e31818f2ab2 Santalla, A., Naranjo, J., and Terrados, N. (2009). Muscle efficiency improves over
Hopker, J., Coleman, D., Passfield, L., and Wiles, J. (2010). The effect of training time in world-class cyclists. Med. Sci. Sports Exerc. 41, 1096–1101. doi: 10.1249/
volume and intensity on competitive cyclists’ efficiency. Appl. Physiol. Nutr. MSS.0b013e318191c802
Metab. 35, 17–22. doi: 10.1139/H09-124 Seiler, S., Joranson, K., Olesen, B. V., and Hetlelid, K. J. (2013). Adaptations to
Hopkins, W. G., Marshall, S. W., Batterham, A. M., and Hanin, J. (2009). aerobic interval training: interactive effects of exercise intensity and total work
Progressive statistics for studies in sports medicine and exercise science. Med. duration. Scand. J. Med. Sci. Sports 23, 74–83. doi: 10.1111/j.1600-0838.2011.
Sci. Sports Exerc. 41, 3–13. doi: 10.1249/MSS.0b013e31818cb278 01351.x
Jeukendrup, A. E., Craig, N. P., and Hawley, J. A. (2000). The bioenergetics of world Skinner, J. S., Jaskolski, A., Jaskolska, A., Krasnoff, J., Gagnon, J., Leon, A. S., et al.
class cycling. J. Sci. Med. Sport 3, 414–433. doi: 10.1016/S1440-2440(00)80008-0 (2001). Age, sex, race, initial fitness, and response to training: the HERITAGE
Joyner, M. J., and Coyle, E. F. (2008). Endurance exercise performance: the Family Study. J. Appl. Physiol. 90, 1770–1776. doi: 10.1152/jappl.2001.90.5.1770
physiology of champions. J. Physiol. 586, 35–44. doi: 10.1113/jphysiol.2007. Sylta, O., Tonnessen, E., Hammarstrom, D., Danielsen, J., Skovereng, K., Ravn, T.,
143834 et al. (2016). The effect of different high-intensity periodization models on
Kohrt, W. M., Malley, M. T., Coggan, A. R., Spina, R. J., Ogawa, T., Ehsani, A. A., endurance adaptations. Med. Sci. Sports Exerc. 48, 2165–2174. doi: 10.1249/
et al. (1991). Effects of gender, age, and fitness level on response of VO2max MSS.0000000000001007
to training in 60-71 yr olds. J. Appl. Physiol. 71, 2004–2011. doi: 10.1152/jappl.
1991.71.5.2004 Conflict of Interest Statement: The authors declare that the research was
Kristoffersen, M., Gundersen, H., Leirdal, S., and Iversen, V. V. (2014). Low conducted in the absence of any commercial or financial relationships that could
cadence interval training at moderate intensity does not improve cycling be construed as a potential conflict of interest.
performance in highly trained veteran cyclists. Front. Physiol. 5:34. doi: 10.3389/
Fphys.2014.00034 Copyright © 2018 Skovereng, Sylta, Tønnessen, Hammarström, Danielsen, Seiler,
Lucia, A., Hoyos, J., Pardo, J., and Chicharro, J. L. (2000). Metabolic and Rønnestad and Sandbakk. This is an open-access article distributed under the terms
neuromuscular adaptations to endurance training in professional cyclists: of the Creative Commons Attribution License (CC BY). The use, distribution or
a longitudinal study. Jpn. J. Physiol. 50, 381–388. doi: 10.2170/jjphysiol. reproduction in other forums is permitted, provided the original author(s) and the
50.381 copyright owner are credited and that the original publication in this journal is cited,
Lucia, A., Hoyos, J., Perez, M., Santalla, A., and Chicharro, J. L. (2002). Inverse in accordance with accepted academic practice. No use, distribution or reproduction
relationship between VO2max and economy/efficiency in world-class cyclists. is permitted which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 37 June 2018 | Volume 9 | Article 713

 ORIGINAL RESEARCH
published: 17 April 2018
doi: 10.3389/fphys.2018.00392

Adaptive Changes After 2 Weeks of

10-s Sprint Interval Training With
Various Recovery Times
Robert A. Olek 1*, Sylwester Kujach 2 , Ewa Ziemann 3 , Wieslaw Ziolkowski 1 , Piotr Waz 4 and
Radoslaw Laskowski 2
1
Department of Bioenergetics and Nutrition, Faculty of Rehabilitation and Kinesiology, Gdansk University of Physical
Education and Sport, Gdańsk, Poland, 2 Department of Physiology, Faculty of Physical Education, Gdansk University of
Physical Education and Sport, Gdańsk, Poland, 3 Department of Physiology and Pharmacology, Faculty of Rehabilitation and
Kinesiology, Gdansk University of Physical Education and Sport, Gdańsk, Poland, 4 Department of Nuclear Medicine, Medical
University of Gdańsk, Gdańsk, Poland

Purpose: The aim of this study was to compare the effect of applying two different
rest recovery times in a 10-s sprint interval training session on aerobic and anaerobic
capacities as well as skeletal muscle enzyme activities.
Methods: Fourteen physically active but not highly trained male subjects (mean maximal
oxygen uptake 50.5 ± 1.0 mlO2 ·kg−1 ·min−1 ) participated in the study. The training
Edited by: protocol involved a series of 10-s sprints separated by either 1-min (SIT10:1) or 4-min
Thomas Leonhard Stöggl,
(SIT10:4) of recovery. The number of sprints progressed from four to six over six sessions
University of Salzburg, Austria
separated by 1–2 days rest. Pre and post intervention anthropometric measurements,
Reviewed by:
Antonio Paoli, assessment of aerobic, anaerobic capacity and muscle biopsy were performed. In
Università degli Studi di Padova, Italy the muscle samples maximal activities of citrate synthase (CS), 3-hydroxyacylCoA
Anthony John Blazevich,
Edith Cowan University, Australia
dehydrogenase (HADH), carnitine palmitoyl-transferase (CPT), malate dehydrogenase
*Correspondence:
(MDH), and its mitochondrial form (mMDH), as well as lactate dehydrogenase (LDH)
Robert A. Olek were determined. Analysis of variance was performed to determine changes between
[email protected]
conditions.
Specialty section: Results: Maximal oxygen uptake improved significantly in both training groups, by
This article was submitted to 13.6% in SIT10:1 and 11.9% in SIT10:4, with no difference between groups. Wingate
Exercise Physiology,
a section of the journal
anaerobic test results indicated main effect of time for total work, peak power output
Frontiers in Physiology and mean power output, which increased significantly and similarly in both groups.
Received: 13 November 2017 Significant differences between training groups were observed for end power output,
Accepted: 29 March 2018
which increased by 10.8% in SIT10:1, but remained unchanged in SIT10:4. Both training
Published: 17 April 2018
protocols induced similar increase in CS activity (main effect of time p < 0.05), but no
Citation:
Olek RA, Kujach S, Ziemann E, other enzymes.
Ziolkowski W, Waz P and Laskowski R
(2018) Adaptive Changes After 2
Conclusion: Sprint interval training protocols induce metabolic adaptation over a short
Weeks of 10-s Sprint Interval Training period of time, and the reduced recovery between bouts may attenuate fatigue during
With Various Recovery Times. maximal exercise.
Front. Physiol. 9:392.
doi: 10.3389/fphys.2018.00392 Keywords: Wingate anaerobic test, all-out exercise, skeletal muscle, enzyme activity, recovery

Frontiers in Physiology | www.frontiersin.org 38 April 2018 | Volume 9 | Article 392

Olek et al. Short SIT Adaptive Changes

INTRODUCTION MATERIALS AND METHODS

The study by Parolin et al. (1999) described metabolic Fourteen physically active but not highly trained male subjects
modifications in skeletal muscle during three 30-s bouts of at mean VO2 max 50.5 ± 1.0 mlO2 ·kg−1 ·min−1 volunteered
maximal isokinetic cycling separated by 4-min recovery periods. to participate in the study. The subjects were asked to refrain
This exercise protocol has become a basis for sprint interval from any additional exercise practice during the study. The
training (SIT). The studies investigating the effect of SIT study was approved and then completed in accordance with the
have been conducted using three to seven 30-s supramaximal recommendations of Local Bioethics Committee (http://www.
exercise sprint bouts with 4-min rest periods between (Vollaard komisjabioetyczna.pl). Written informed consent was obtained
et al., 2017). Since the aerobic contribution in ATP resynthesis from all subjects. The study was conducted in accordance with
increases in consecutive exercise bouts (Bogdanis et al., 1996; the Declaration of Helsinki. The overall study protocol is shown
Parolin et al., 1999), SIT is associated with the elevation of in Figure 1. The subjects were assigned to either 1-min interval
muscular oxidative potential as well as central adaptations such recovery (SIT10:1) or 4-min interval recovery (SIT10:4) based on
as increased cardiac output and stroke volume (Sloth et al., the pre-training results.
2013). Consequently, SIT is postulated as a time-efficient exercise
strategy for improving maximal oxygen uptake (VO2 max) Anthropometric Measurements
(Sloth et al., 2013; Vollaard et al., 2017) with biochemical and Body mass and composition were estimated using InBody720
morphological adaptations arising in as little as six sessions (InBody Co., Ltd., Seoul, Korea). Participants were asked to
performed over 2 weeks (Burgomaster et al., 2005, 2006, 2008; arrive at the laboratory fasted, with voided bladders and bowels.
Hazell et al., 2010; Lloyd Jones et al., 2017). The bioelectrical impedance analyses were performed in the
It has been suggested that some of the adaptations to SIT position recommended by the manufacturer guidelines and
are associated with the power generated during the first few the subjects clad only in briefs (Ziemann et al., 2013). The
seconds of each sprint. Therefore, shorter exercise bouts have impedance measured five segments of the body (arms, trunk,
been investigated to determine whether resulting adaptations legs) at frequencies of 1, 5, 50, 250, 500, and 1,000 kHz
are similar to those observed following 30-s SIT (Hazell et al., through the eight electrodes. Based on these impedance values
2010; Lloyd Jones et al., 2017). Two various SIT protocols fat free mass (FFM) and skeletal muscle mass (SMM) were
consisting of repeated 30-s or 10-s “all-out” cycle sprints calculated.
with 4-min rest intervals, performed over 2 weeks indicate
comparable enhancement in aerobic and anaerobic exercise
capacities (Hazell et al., 2010). Moreover, similar improvements Cardiopulmonary Exercise Test
in performance are observed after applying SIT protocols To determine VO2 max participants performed a graded cycle
matched for total sprint time −4 × 30-s or 20 × 6-s (Lloyd ergometry test on an electromagnetically-braked, cycle ergometer
Jones et al., 2017). Another key factor affecting the metabolic (ER 900 Jaeger, Viasys Healthcare GmbH, Germany). After a
modulations may be the work to rest ratio. Reduction in 5-min warm-up at the intensity of 1.5 W·kg−1 with a pedaling
recovery time seems to be less effective in evoking adaptive cadence of 60 rpm, work rate was increased by 25 W·min−1 until
changes (Hazell et al., 2010). The increase in aerobic and volitional exhaustion. Breath by breath pulmonary gas exchange
anaerobic capacities following 10-s SIT with 4-min rest are was measured by Oxycon-Pro analyzer (Viasys Healthcare
double changes in 10-s SIT with 2-min rest (Hazell et al., GmbH, Germany; Olek et al., 2012). Heart rates were monitored
2010). continuously by telemetry (S-625, Polar Electro-Oy, Finland).
Therefore, the aim of the present study was to compare Maximal heart rate (HRmax), maximal ventilation (VEmax),
the effect of applying two different rest intervals in 10-s maximal aerobic power (MAP) were calculated at the VO2 max
SIT. We hypothesized that reducing the recovery time to level (Ziemann et al., 2011). The anaerobic threshold (AT) has
1-min in the SIT protocol consisting of repeated 10-s “all- been determined by three independent members of the research
out” cycle sprints, performed over 2 weeks would be less group using nonlinear increase in ventilation.
effective in evoking aerobic adaptations than 10-s SIT with 4-
min rest, however it would still improve anaerobic performance Anaerobic Power Measurement
measures. After the standard warm-up, all subjects performed a 30-s “all-
out” supramaximal test on a mechanically braked cycle ergometer
(884E Sprint Bike, Monark, Sweden). The test was initiated from
Abbreviations: SIT, sprint interval training; VO2 max, maximal oxygen uptake;
VE, minute ventilation; HR, heart rate; MAP, maximal aerobic power; AT,
a dead stop with the resistance equal to 75 g·kg body mass−1
anaerobic threshold; TW, total work; PPO, peak power output; MPO, mean (corresponding to 7.5% of an individual’s body mass) preset
power output; EPO, end power output; FI, fatigue index; FFM, fat free on the ergometer’s friction belt (Laskowski et al., 2011). The
mass; SMM, skeletal muscle mass; BMI, body mass index; CS, citrate synthase; obtained results were analyzed for total work (TW), peak power
HADH, 3-hydroxyacylCoA dehydrogenase; CPT, carnitine palmitoyl-transferase; output (PPO), mean power output (MPO), end power output
MDH, malate dehydrogenase; mMDH, mitochondrial malate dehydrogenase;
cMDH, cytoplasmic malate dehydrogenase; LDH, lactate dehydrogenase; EDTA,
(EPO), time to PPO (Bar-Or, 1987). Fatigue index (FI) was
ethylenediaminetetraacetic acid; DTNB, 5,5′ -dithiobis (2-nitrobenzoic acid); determined by taking the percentage difference between PPO and
OAA, oxaloacetic acid; NADH, nicotinamide adenine dinucleotide reduced form. EPO (Richardson et al., 2016).

Frontiers in Physiology | www.frontiersin.org 39 April 2018 | Volume 9 | Article 392

Olek et al. Short SIT Adaptive Changes

FIGURE 1 | Scheme of the study protocol. Muscle biopsy (MB), cardiopulmonary exercise test (CPET), Wingate anaerobic test (WAnT).

Muscle Sampling and Analyses 7.2. The substrates NADH and pyruvate were added immediately
Muscle biopsy was made under local anesthesia (2% lidocaine), before the measurement was started. The change in absorbance
in the supine position. The sample was obtained by the sterile was followed in time at 340 nm. HADH, MDH, and LDH
single-use micro biopsy needle (M.D.L. srl, Delebio, Italy). The activities were calculated using molar absorption coefficient of
section of the Vastus Lateralis muscle (∼10 mg) was immediately NADH 6,220 M−1 cm−1 . Protein content was determined by
frozen in liquid nitrogen and stored in −80◦ C until analysis. All using Bradford protein assay. All reagents were obtained from
biopsies were collected by the same person to ensure a standard Sigma-Aldrich.
localization and muscle depth.
Muscles were homogenized in an ice-cold buffer contained Training Intervention
50 mM potassium phosphate, 1 mM ethylenediaminetetraacetic Subjects participated in training sessions on Monday,
acid (EDTA), 1 mM threo-1,4-dimercapto-2,3-butanediol at pH Wednesday, and Friday for 2 weeks. The same research assistant
7.4. The homogenates were then centrifuged at 600 g at 4 ◦ C supervised all sessions, controlled the flywheel resistance, as well
for 10 min. The obtained supernatant was used to determine as the time of exercise and recovery periods. Each training session
enzyme activities as previously (Antosiewicz et al., 1995; Kaczor began with a 5-min warm-up at the intensity approximating 30%
et al., 2005; Olek et al., 2013) with a Super Aquarius CE9200 MAP. All training was completed using a load of 75 g·kg body
spectrophotometer (Cecil Instruments Ltd., Cambridge, UK) at mass−1 .
30◦ C. Citrate synthase (CS) activity was measured by the rate of
1. SIT10:1; performed repeated, 10-s “all-out” efforts separated
SH production as CoASH using the thiol reagent 5,5′ -dithiobis
by 1-min recovery.
(2-nitrobenzoic acid) (DTNB). The reagent cocktail contained
2. SIT10:4; performed repeated, 10-s “all-out” efforts separated
50 mM potassium phosphate, 0.1 mM DTNB, and 0.1 mM
by 4-min recovery.
acetylCoA. The reaction was started by 0.1 mM oxaloacetic
acid (OAA). Carnitine palmitoyl-transferase (CPT) activity was The number of repeats increased from four repetitions during
measured in the reaction mixture composed by 60 mM of Tris the first two training sessions, by five repetitions during the
HCl at pH 8.0, 1.5 mM of EDTA with 0.05% Triton X-100 and middle two training sessions, to six repetitions for the final
0.25 mM DTNB and 1.67 mM of carnitine. The reaction was two training sessions as has been done previously (Hazell et al.,
started by the addition of 0.025 mM palmitoyl-CoA. The kinetics 2010). Both training groups performed a similar training protocol
of change in absorbance were followed at 412 nm, and the molar consisting of: 30-min low intensity warm-up and 5-min high
absorption coefficient 14,150 M−1 cm−1 was used for calculation intensity exercise throughout the training intervention (Stöggl
of CS and CPT activities. 3-hydroxyacylCoA dehydrogenase and Sperlich, 2015). The difference between the groups was the
(HADH) activity was determined in a buffer containing 100 mM time of recovery between the 10-s exercises. Therefore, overall
potassium phosphate and 0.05% Triton at pH 7.4. After addition time of training sessions was equal 59-min for SIT10:1 and
of supernatant and 0.1 mM NADH the cuvette was preincubated 131-min for SIT10:4 (Table 1).
for 3 min. The reaction was started by 0.1 mM acetoacetyl-
CoA. Malate dehydrogenase (MDH) activity was determined in Statistical Analyses
the 50 mM Tris-HCl buffer pH 7.6 containing 5 mM EDTA, Two-way analysis of variance ANOVA was performed to examine
0.1 mM NADH, and 0.2 mM OAA. For the mitochondrial MDH the main effects of group and/or time. In case the ANOVA
(mMDH) assay, the cytoplasmic form of the enzyme (cMDH) yielded a significant effect, a Tukey’s HSD test was used for
was inactivated by 3 min preincubation of homogenate with post hoc comparisons. A probability level p < 0.05 was considered
equal volume of ethanol 99.5% (v/v) at room temperature. Then statistically significant. Due to the small number of subjects, the
mMDH activity was followed by the same procedure as for effect size (η2 ) has been calculated. The values of η2 has been
total MDH measurements. Lactate dehydrogenase (LDH) activity interpreted as follows: 0.1 a small effect, 0.3 a medium effect and
was measured in the assay medium contained 50 mM potassium 0.5 a large effect, as previously (Olek et al., 2014). All data are
phosphate, 1 mM EDTA, 0.1 mM NADH, 2.1 mM pyruvate, at pH expressed as mean ± SEM (standard error of mean).

Frontiers in Physiology | www.frontiersin.org 40 April 2018 | Volume 9 | Article 392

Olek et al. Short SIT Adaptive Changes

TABLE 1 | Training distribution over the 2-weeks intervention. TABLE 2 | Anthropometric characteristics of participants.

SIT10:1 SIT10:4 SIT10:1 SIT10:4

Total time (min) 59 131 Age (years) 20.1 ± 0.3 20.7 ± 0.2
Low intensity (% of total time) 50.8 22.9 Height (cm) 180 ± 1 185 ± 2
High intensity (% of total time) 8.5 3.8 Body mass (kg) 75.9 ± 1.7 83.1 ± 2.7#
Recovery (% of total time) 40.7 73.3 FFM (kg) 66.4 ± 1.5 73.6 ± 2.6#
SMM (kg) 37.8 ± 0.9 42.3 ± 1.5#
The intensity of exercise and its distribution over time have been classified as previously
described (Stöggl and Sperlich, 2015). BMI 23.5 ± 0.5 24.3 ± 0.5

Values are means ± SEM. FFM, fat free mass; SMM, skeletal muscle mass; BMI, body
mass index. # p < 0.05 as compared to SIT10:1.

RESULTS
The mean body mass of SIT10:4 subjects was higher than TABLE 3 | Aerobic capacities before (pre) and after (post) 2 weeks of SIT.
SIT10:1 subjects (Table 2), therefore physiological parameters
SIT10:1 SIT10:4
have been presented as relative values. There were no
significant differences between groups in aerobic and anaerobic Pre Post Pre Post
capacities as well as skeletal muscle enzymatic activities at
MAX
baseline.
VO2 51.4 ± 1.4 58.4 ± 2.2 49.6 ± 1.5 55.5 ± 2.1
The training elevated VO2 max in both groups (Table 3), by
(mlO2 ·min−1 ·kg−1 )a
13.6% in SIT10:1 and 11.9% in SIT10:4 (p < 0.001, η2 = 0.63),
VE (L·min−1 )c 143 ± 6 159 ± 8 152 ± 7 155 ± 6
with no difference between the groups. Moreover, there was
MAP (W·kg−1 ) 3.9 ± 0.2 4.0 ± 0.2 3.8 ± 0.1 3.9 ± 0.1
a significant VEmax increase (p < 0.05, η2 = 0.29) after the
HR (bpm) 194 ± 4 191 ± 4 192 ± 1 192 ± 2
training. No significant differences in either HRmax or MAP
AT
were observed.
VO2 34.0 ± 1.4 39.2 ± 1.2 34.2 ± 1.5 38.2 ± 1.6
Significant changes in oxygen uptake (p < 0.001, η2 = 0.78), (mlO2 ·min−1 ·kg−1 )a
VE (p < 0.001, η2 = 0.74) and workloads (%MAP) (p < 0.001,
VE (L·min−1 )a 67 ± 3 80 ± 3 71 ± 4 77 ± 3
η2 = 0.72) at the exercise intensity corresponding to AT intensity
Power (W·kg−1 )b 2.5 ± 0.1 2.7 ± 0.1 2.3 ± 0.1 2.7 ± 0.2
following 2 weeks of SIT were observed (Table 3). No significant
Power (% MAP)a 63 ± 3 67 ± 3 62 ± 2 69 ± 2
main effects for group and group × time interaction in these
HR (bpm) 164 ± 4 163 ± 5 154 ± 3 161 ± 4
parameters were noted.
Wingate anaerobic test results indicated main effect of time Values are means ± SEM. VO2 , oxygen uptake; VE, minute ventilation; MAP, maximal
for TW (p < 0.001, η2 = 0.79; Figure 2A), MPO (p < 0.001, aerobic power; HR, heart rate. a p < 0.001 main training effect; b p < 0.005 main training
effect; c p < 0.05 main training effect.
η2 = 0.79; Figure 2B) and PPO (p < 0.001, η2 = 0.67; Figure 2D)
which increased significantly in both groups. However, these
changes did not differ between the groups. Significant differences muscle enzyme activities. Moreover, the shorter recovery time
between training groups (p < 0.05, η2 = 0.32), time (p < 0.01, induced higher EPO and lower FI during Wingate anaerobic
η2 = 0.48) and a group x time interaction (p < 0.02, η2 = 0.39) test.
were recorded for EPO (Figure 2C). EPO in SIT10:1 increased The increase in VO2 max is similar to data obtained by
significantly by 10.8% (p < 0.005), whereas in SIT10:4 group McGarr et al. (2014) and Richardson and Gibson (2015), who
remained unchanged (p = 0.9). Changes in EPO affected used 30-s sprints with 4-min recoveries in training protocols
FI, which approached statistical significance after the training performed over 2 weeks, reporting 14.2 and 11.2% improvement,
(p = 0.07, η2 = 0.25; Figure 2E). No differences in time to PPO respectively. Hazell et al. (2010), by applying six SIT sessions
was noted (Figure 2F). consisting of 10-s sprints with 4-min recovery intervals, indicated
There was a main effect of time for CS activity, which elevated an 8.5% enhancement in aerobic capacity, similar to 30-s “all-
significantly in both training groups (p < 0.05, η2 = 0.33), but no out” cycles interspersed with 4-min recoveries (8.3%) reported
for HADH (p = 0.18, η2 = 0.14), mMDH (p = 0.17, η2 = 0.15) in the same study. However, the workload used during SIT
and CPT (p = 0.9, η2 = 0.0). There were no differences in LDH was slightly higher (10% of individual’s body mass) (Hazell
activity and LDH/CS ratio (Table 4). et al., 2010) than in the current protocol (7.5% of individual’s
body mass). On the contrary, two other studies applying 10-
DISCUSSION s maximal cycling SIT reported no change in VO2 max: 2.4%
(Hellsten-Westing et al., 1993) and −1.3% (Skleryk et al.,
The main finding of this study is that both 1-min and 4-min 2013). These discrepancies may be caused by the number of
recovery intervals in repeated 10-s sprint intervention for 2 repetitions applied during SIT (Vollaard et al., 2017). SIT
weeks, resulted in a similar improvements in aerobic (VO2 max, performed by Hellsten-Westing et al. (1993) consisted of 15
AT) and anaerobic (TW, PPO, MPO) capacities, and skeletal 10-s maximal sprints, three times per week for a total of 6

Frontiers in Physiology | www.frontiersin.org 41 April 2018 | Volume 9 | Article 392

Olek et al. Short SIT Adaptive Changes

FIGURE 2 | Anaerobic capacity before and after 2 weeks of SIT. Total work (A), mean power output (B), end power output (C), peak power output (D), fatigue index
(E) and time to peak power output (F). ‡ p < 0.001 main training effect; *p < 0.01 as compared to post-training between the groups; **p < 0.005 as compared to
pre-training within the group; # p = 0.07 main effect of group.

weeks. Skleryk et al. (2013) implemented six SIT sessions of (Rasmussen et al., 2001). SIT-induced VO2 max improvement is
10-s “all-out” cycling, repeated 8–12 times. In the recent meta- associated with changes in mitochondrial bioenergetics (Larsen
analysis, Vollaard et al. (2017) indicated that training protocols et al., 2013). Six SIT sessions of 30-s sprints, performed over
consisting of fewer repetitions during session induced greater 2 weeks caused maximal CS activity elevation (Burgomaster
VO2 max increase. Authors negated the role of total energy et al., 2005, 2006). On the other hand, HADH, the rate-limiting
use, energy turnover or oxygen transfer in improvement of fat oxidation mitochondrial enzyme, does not change after
VO2 max with SIT, because for each of these factors the stimulus 2 weeks (Burgomaster et al., 2006), however, it significantly
should be greater with more sprint repetitions (Vollaard et al., increases after 6 weeks of 30-s SIT (Burgomaster et al., 2008).
2017). Higher HADH has also been observed in legs, but not in
Our results revealed that 1-min or 4-min rest periods did arms, following 72 “all-out sprints” (36 with arm cycling and
not impact the increase in relative values of VO2 max. Both 36 with leg cycling separated by 1-h of recovery) during
training protocols were characterized by supramaximal intensity. the seven training sessions (Zinner et al., 2016). We have
Despite being short-lasting, both had a positive effect on shown that a SIT protocol consisting the same number of
maximal ventilation values, allowing subjects to perform the repetitions as previously reported (Burgomaster et al., 2005,
high volume workload. Moreover, maximal oxygen uptake also 2006), although with lower total work done, induced a significant
strongly correlates with skeletal muscle mitochondrial capacities rise in CS activity (but not in other mitochondrial enzymes).

Frontiers in Physiology | www.frontiersin.org 42 April 2018 | Volume 9 | Article 392

Olek et al. Short SIT Adaptive Changes

TABLE 4 | Maximal activities of skeletal muscle enzymes before (pre) and after reported in a study utilizing 6 sessions of 10-s supramaximal
(post) 2 weeks of SIT. bouts, with 2-min or 4-min recoveries between sprints (Hazell
SIT10:1 SIT10:4
et al., 2010). Consistently, we observed a significant PPO increase
in both SITs. Despite no differences between the two protocols
Pre Post Pre Post in other measured parameters, we observed an improvement in
EPO during Wingate test in the group with a shorter recovery
CPT 1.6 ± 0.1 1.6 ± 0.1 1.7 ± 0.1 1.7 ± 0.1
(mU·mg protein−1 )
time.
CS 161 ± 19 193 ± 13 177 ± 13 202 ± 13
Aerobic energy provision contributes almost half of the
(mU·mg protein−1 )a ATP turnover during the repeated 30-s sprint (after 4-min
HADH 116 ± 11 132 ± 5 114 ± 11 123 ± 12 recovery; Bogdanis et al., 1996). Hence, the adaptive changes
(mU·mg protein−1 ) in oxidative metabolism seem to be reasonable following such
MDH 5.2 ± 0.7 5.4 ± 0.5 4.9 ± 0.5 5.9 ± 0.6 a training protocol. During the 10-s sprint, the PCr availability
(U·mg protein−1 ) is important for high power output (Bogdanis et al., 1998).
cMDH 3.9 ± 0.5 4.0 ± 0.3 3.7 ± 0.4 4.4 ± 0.4 Since PCr resynthesis 2-min after cessation of exercise reaches
(U·mg protein−1 ) about ∼90% of the resting value (Hultman et al., 1967; Bogdanis
mMDH 1.3 ± 0.2 1.4 ± 0.1 1.2 ± 0.1 1.5 ± 0.2 et al., 1998), subjects are able to reproduce PPO after 2-min
(U·mg protein−1 )
recovery (Bogdanis et al., 1998). It seems that the reported
LDH 3.2 ± 0.4 3.8 ± 0.2 3.6 ± 0.5 3.8 ± 0.3
adaptations following such training protocols may be caused
(U·mg protein−1 )
by increased flux in the creatine-PCr energy shuttle (Bessman
LDH/CS 20.3 ± 0.8 20.2 ± 2.1 20.8 ± 2.5 19.6 ± 2.6
and Carpenter, 1985). Bessman and Savabi (1988) suggested
Values are means ± SEM. CPT, carnitine palmitoyl-transferase; CS, citrate synthase; that creatine plays important role in high energy phosphate
HADH, 3-hydroxyacylCoA dehydrogenase; MDH, malate dehydrogenase; cMDH,
transport, necessary for protein synthesis, ion transport and
cytoplasmic malate dehydrogenase; mMDH, mitochondrial malate dehydrogenase; LDH,
lactate dehydrogenase. a p < 0.05 main training effect. muscle contraction. The PCr resynthesis 1-min and 4-min after
the exercise is ∼80 and 90% (Hultman et al., 1967), consequently
the small difference seems to be insufficient to induce various
Since training induced changes in muscle CS activity are adaptive response for two SIT protocols. On the other hand,
matched by changes in whole body oxidative capacity (Vigelsø accumulation of incomplete recoveries in 1-min recovery SIT
et al., 2014; Meinild Lundby et al., 2018), our results confirm may contribute to some adaptations in ionic regulation resulting
the effectiveness of both SIT protocols in training-induced in EPO improvement. However, an increasing number of
adaptations. sprints may accumulate fatigue (Vollaard et al., 2017) and thus
Elevated oxidative ATP synthesis, attenuates the contribution the effectiveness of training protocol in aerobic adaptations
of anaerobic ATP production (Larsen et al., 2014) and delays may be reduced (Hellsten-Westing et al., 1993; Skleryk et al.,
the blood lactate accumulation (Jakeman et al., 2012) following 2013).
SIT. We have not determined lactate production, but we
have observed changes in maximal oxygen uptake and CS
activity, which were associated with higher workload at the CONCLUSION
intensity corresponding to AT. In addition, aerobic/anaerobic
This study found that 2 weeks of SIT comprising either of
energy supply depends on the rate of pyruvate production
1- or 4-min recovery time between exercises, which were
and its mitochondrial oxidative decarboxylation catalyzed by
matched for total sprint time, elicited similar performance
pyruvate dehydrogenase competing to cytosolic reduction to
changes in fit, healthy men. Moreover, reduced time of
lactate via LDH (Spriet et al., 2000). Moreover, malate-
recovery between bouts may be more effective in attenuating
aspartate shuttle appears to be quantitatively important in lactate
fatigue during maximal exercise. Overall, obtained data
production during exercise (Schantz et al., 1986). Although the
complement presently available knowledge about interactions
training protocol applied in our study did not modify LDH
between intensity as well as duration of interval protocols
activity, and the increase in MDH activity was not statistically
and recovery. This is particularly significant in the context
significant, we have observed higher workout at the intensity
of the latest recommendations of American College of
corresponding to AT, suggesting reduced lactate production
Sports Medicine; which cites high/sprint intensity interval
following SIT.
training as the most effective form of exercise (Thompson,
SIT performed daily for 2 weeks increases skeletal muscle
2017).
phosphocreatine (PCr) content and creatine kinase (CK) activity,
but not PPO nor MPO during the Wingate test (Parra et al.,
2000; Rodas et al., 2000). On the contrary, the same 14 sessions ETHICS STATEMENT
applied for 6 weeks (resting for 2 days between each session)
causes improvement in PPO and MPO with no change in PCr This study was carried out in accordance with the
and CK (Parra et al., 2000). PPO increases with no modification recommendations of Local Bioethics Committee with written
in PCr also after completing only 6 SIT sessions over 2 weeks informed consent from all subjects. All subjects gave written
(Burgomaster et al., 2005). Similarly, higher PPO has been informed consent in accordance with the Declaration of Helsinki.

Frontiers in Physiology | www.frontiersin.org 43 April 2018 | Volume 9 | Article 392

Olek et al. Short SIT Adaptive Changes

The protocol (Figure 1) was approved by the Local Bioethics FUNDING

Committee (http://www.komisjabioetyczna.pl/).
This study was supported by the Ministry of Science
and Higher Education DSWF/I/6/2017 and DSRiK/
AUTHOR CONTRIBUTIONS
10/2016.
RO, SK, and RL: Conceived and design the experiment; RO, SK,
EZ, WZ, and RL: Performed the data collection; RO and PW: ACKNOWLEDGMENTS
Performed the statistical analysis and interpretation of data; RO,
SK, EZ, WZ, PW, and RL: Participated in drafting the article or The authors would like to thank a group of subjects
revising it critically for important intellectual content; RO, SK, participating in the study. The excellent technical assistance of
EZ,WZ, PW, and RL: Approved the final manuscript. Dr. Rafal Camilleri is acknowledged.

REFERENCES skeletal muscle from children and adults. Pediatr. Res. 57, 331–335.
doi: 10.1203/01.PDR.0000150799.77094.DE
Antosiewicz, J., Popinigis, J., Ishiguro, H., Hayakawa, T., and Wakabayashi, Larsen, R. G., Befroy, D. E., and Kent-Braun, J. A. (2013). High-intensity interval
T. (1995). Cerulein-induced acute pancreatitis diminished vitamin E training increases in vivo oxidative capacity with no effect on P(i)–>ATP rate
concentration in plasma and increased in the pancreas. Int. J. Pancreatol. 17, in resting human muscle. Am. J. Physiol. Regul. Integr. Comp. Physiol. 304,
231–236. R333–R342. doi: 10.1152/ajpregu.00409.2012
Bar-Or, O. (1987). The Wingate anaerobic test. An update on Larsen, R. G., Maynard, L., and Kent, J. A. (2014). High-intensity interval training
methodology, reliability and validity. Sports Med. 4, 381–394. alters ATP pathway flux during maximal muscle contractions in humans. Acta
doi: 10.2165/00007256-198704060-00001 Physiol. 211, 147–160. doi: 10.1111/apha.12275
Bessman, S. P., and Carpenter, C. L. (1985). The creatine-creatine Laskowski, R., Ziemann, E., Olek, R. A., and Zembron-Lacny, A. (2011).
phosphate energy shuttle. Annu. Rev. Biochem. 54, 831–862. The effect of three days of judo training sessions on the inflammatory
doi: 10.1146/annurev.bi.54.070185.004151 response and oxidative stress markers. J. Hum. Kinet. 30, 65–73.
Bessman, S., and Savabi, F. (1988). “The role of the phosphocreatine energy doi: 10.2478/v10078-011-0074-1
shuttle in exercise and muscle hypertrophy,” in International Series on Sport Lloyd Jones, M. C., Morris, M. G., and Jakeman, J. R. (2017). Impact of
Sciences: Biochemistry of Exercise, VII, eds A. Taylor, P. Gollnick, and H. Green time and work:rest ratio matched sprint interval training programmes on
(Champaign, IL: Human Kinetics), 167–178. performance: a randomised controlled trial. J. Sci. Med. Sport 20, 1034–1038.
Bogdanis, G. C., Nevill, M. E., Boobis, L. H., and Lakomy, H. K. (1996). doi: 10.1016/j.jsams.2017.03.020
Contribution of phosphocreatine and aerobic metabolism to energy McGarr, G. W., Hartley, G. L., and Cheung, S. S. (2014). Neither short-
supply during repeated sprint exercise. J. Appl. Physiol. 80, 876–884. term sprint nor endurance training enhances thermal response to
doi: 10.1152/jappl.1996.80.3.876 exercise in a hot environment. J. Occup. Environ. Hyg. 11, 47–53.
Bogdanis, G. C., Nevill, M. E., Lakomy, H. K., and Boobis, L. H. (1998). Power doi: 10.1080/15459624.2013.816429
output and muscle metabolism during and following recovery from 10 and Meinild Lundby, A. K., Jacobs, R. A., Gehrig, S., De Leur, J., Hauser, M., Bonne,
20 s of maximal sprint exercise in humans. Acta Physiol. Scand. 163, 261–272. T. C., et al. (2018). Exercise training increases skeletal muscle mitochondrial
doi: 10.1046/j.1365-201x.1998.00378.x volume density by enlargement of existing mitochondria and not de novo
Burgomaster, K. A., Heigenhauser, G. J., and Gibala, M. J. (2006). Effect of biogenesis. Acta Physiol. 222:e12905. doi: 10.1111/apha.12905
short-term sprint interval training on human skeletal muscle carbohydrate Olek, R. A., Kujach, S., Wnuk, D., and Laskowski, R. (2014). Single sodium
metabolism during exercise and time-trial performance. J. Appl. Physiol. 100, pyruvate ingestion modifies blood acid-base status and post-exercise lactate
2041–2047. doi: 10.1152/japplphysiol.01220.2005 concentration in humans. Nutrients 6, 1981–1992. doi: 10.3390/nu6051981
Burgomaster, K. A., Howarth, K. R., Phillips, S. M., Rakobowchuk, M., Macdonald, Olek, R. A., Safranow, K., Jakubowska, K., Olszewska, M., Chlubek, D., and
M. J., McGee, S. L., et al. (2008). Similar metabolic adaptations during exercise Laskowski, R. (2012). Allopurinol intake does not modify the slow component
after low volume sprint interval and traditional endurance training in humans. of V(.)O(2) kinetics and oxidative stress induced by severe intensity exercise.
J. Physiol. 586, 151–160. doi: 10.1113/jphysiol.2007.142109 Physiol. Res. 61, 89–96.
Burgomaster, K. A., Hughes, S. C., Heigenhauser, G. J., Bradwell, S. N., and Olek, R. A., Ziolkowski, W., Wierzba, T. H., and Kaczor, J. J. (2013). Effect of ethyl
Gibala, M. J. (2005). Six sessions of sprint interval training increases muscle pyruvate on skeletal muscle metabolism in rats fed on a high fat diet. Nutrients
oxidative potential and cycle endurance capacity in humans. J. Appl. Physiol. 5, 2372–2383. doi: 10.3390/nu5072372
98, 1985–1990. doi: 10.1152/japplphysiol.01095.2004 Parolin, M. L., Chesley, A., Matsos, M. P., Spriet, L. L., Jones, N. L.,
Hazell, T. J., Macpherson, R. E., Gravelle, B. M., and Lemon, P. W. and Heigenhauser, G. J. (1999). Regulation of skeletal muscle glycogen
(2010). 10 or 30-s sprint interval training bouts enhance both aerobic phosphorylase and PDH during maximal intermittent exercise. Am. J. Physiol.
and anaerobic performance. Eur. J. Appl. Physiol. 110, 153–160. 277, E890–E900. doi: 10.1152/ajpendo.1999.277.5.E890
doi: 10.1007/s00421-010-1474-y Parra, J., Cadefau, J. A., Rodas, G., Amigó, N., and Cussó, R. (2000).
Hellsten-Westing, Y., Balsom, P. D., Norman, B., and Sjödin, B. (1993). The effect The distribution of rest periods affects performance and adaptations
of high-intensity training on purine metabolism in man. Acta Physiol. Scand. of energy metabolism induced by high-intensity training in human
149, 405–412. doi: 10.1111/j.1748-1716.1993.tb09636.x muscle. Acta Physiol. Scand. 169, 157–165. doi: 10.1046/j.1365-201x.2000.0
Hultman, E., Bergström, J., and Anderson, N. M. (1967). Breakdown and 0730.x
resynthesis of phosphorylcreatine and adenosine triphosphate in connection Rasmussen, U. F., Rasmussen, H. N., Krustrup, P., Quistorff, B., Saltin, B.,
with muscular work in man. Scand. J. Clin. Lab. Invest. 19, 56–66. and Bangsbo, J. (2001). Aerobic metabolism of human quadriceps muscle:
doi: 10.3109/00365516709093481 in vivo data parallel measurements on isolated mitochondria. Am. J.
Jakeman, J., Adamson, S., and Babraj, J. (2012). Extremely short duration high- Physiol. Endocrinol. Metab. 280, E301–E307. doi: 10.1152/ajpendo.2001.280.
intensity training substantially improves endurance performance in triathletes. 2.E301
Appl. Physiol. Nutr. Metab. 37, 976–981. doi: 10.1139/h2012-083 Richardson, A. J., and Gibson, O. R. (2015). Simulated hypoxia does not further
Kaczor, J. J., Ziolkowski, W., Popinigis, J., and Tarnopolsky, M. improve aerobic capacity during sprint interval training. J. Sports Med. Phys.
A. (2005). Anaerobic and aerobic enzyme activities in human Fitness 55, 1099–1106.

Frontiers in Physiology | www.frontiersin.org 44 April 2018 | Volume 9 | Article 392

Olek et al. Short SIT Adaptive Changes

Richardson, A. J., Relf, R. L., Saunders, A., and Gibson, O. R. (2016). Similar uptake adaptations in response to exercise training. Int. J. Physiol. Pathophysiol.
inflammatory responses following sprint interval training performed in Pharmacol. 6, 84–101.
hypoxia and normoxia. Front. Physiol. 7:332. doi: 10.3389/fphys.2016.00332 Vollaard, N. B. J., Metcalfe, R. S., and Williams, S. (2017). Effect of number of
Rodas, G., Ventura, J. L., Cadefau, J. A., Cussó, R., and Parra, J. (2000). A short sprints in an SIT session on change in VO2 max: a meta-analysis. Med. Sci.
training programme for the rapid improvement of both aerobic and anaerobic Sports Exerc. 49, 1147–1156. doi: 10.1249/MSS.0000000000001204
metabolism. Eur. J. Appl. Physiol. 82, 480–486. doi: 10.1007/s004210000223 Ziemann, E., Grzywacz, T., Łuszczyk, M., Laskowski, R., Olek, R. A., and Gibson,
Schantz, P. G., Sjöberg, B., and Svedenhag, J. (1986). Malate-aspartate and A. L. (2011). Aerobic and anaerobic changes with high-intensity interval
alpha-glycerophosphate shuttle enzyme levels in human skeletal muscle: training in active college-aged men. J. Strength Cond. Res. 25, 1104–1112.
methodological considerations and effect of endurance training. Acta Physiol. doi: 10.1519/JSC.0b013e3181d09ec9
Scand. 128, 397–407. doi: 10.1111/j.1748-1716.1986.tb07993.x Ziemann, E., Olek, R. A., Grzywacz, T., Antosiewicz, J., Kujach, S., Łuszczyk,
Skleryk, J. R., Karagounis, L. G., Hawley, J. A., Sharman, M. J., Laursen, P. B., and M., et al. (2013). Whole-body cryostimulation as an effective method of
Watson, G. (2013). Two weeks of reduced-volume sprint interval or traditional reducing low-grade inflammation in obese men. J. Physiol. Sci. 63, 333–343.
exercise training does not improve metabolic functioning in sedentary obese doi: 10.1007/s12576-013-0269-4
men. Diabetes Obes. Metab. 15, 1146–1153. doi: 10.1111/dom.12150 Zinner, C., Morales-Alamo, D., Ørtenblad, N., Larsen, F. J., Schiffer, T. A., Willis,
Sloth, M., Sloth, D., Overgaard, K., and Dalgas, U. (2013). Effects of sprint interval S. J., et al. (2016). The physiological mechanisms of performance enhancement
training on VO2max and aerobic exercise performance: a systematic review and with sprint interval training differ between the upper and lower extremities in
meta-analysis. Scand. J. Med. Sci. Sports 23, e341–e352. doi: 10.1111/sms.12092 humans. Front. Physiol. 7:426. doi: 10.3389/fphys.2016.00426
Spriet, L. L., Howlett, R. A., and Heigenhauser, G. J. (2000). An enzymatic
approach to lactate production in human skeletal muscle during exercise. Conflict of Interest Statement: The authors declare that the research was
Med. Sci. Sports Exerc. 32, 756–763. doi: 10.1097/00005768-200004000- conducted in the absence of any commercial or financial relationships that could
00007 be construed as a potential conflict of interest.
Stöggl, T. L., and Sperlich, B. (2015). The training intensity distribution
among well-trained and elite endurance athletes. Front. Physiol. 6:295. Copyright © 2018 Olek, Kujach, Ziemann, Ziolkowski, Waz and Laskowski. This
doi: 10.3389/fphys.2015.00295 is an open-access article distributed under the terms of the Creative Commons
Thompson, W. R. (2017). Worldwide survey of fitness trends for Attribution License (CC BY). The use, distribution or reproduction in other forums
2018: the CREP edition. ACSM’s Health Fitness J. 21, 10–19. is permitted, provided the original author(s) and the copyright owner are credited
doi: 10.1249/FIT.0000000000000341 and that the original publication in this journal is cited, in accordance with accepted
Vigelsø, A., Andersen, N. B., and Dela, F. (2014). The relationship between academic practice. No use, distribution or reproduction is permitted which does not
skeletal muscle mitochondrial citrate synthase activity and whole body oxygen comply with these terms.

Frontiers in Physiology | www.frontiersin.org 45 April 2018 | Volume 9 | Article 392

 ORIGINAL RESEARCH
published: 28 March 2018
doi: 10.3389/fphys.2018.00290

Impact of Short-Term Training Camp

on Aortic Blood Pressure in
Collegiate Endurance Runners
Tsubasa Tomoto 1 , Jun Sugawara 1*, Ai Hirasawa 2 , Tomoko Imai 3 , Seiji Maeda 4 and
Shigehiko Ogoh 5
1
Human Informatics Research Institute, National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan,
2
Faculty of Health Sciences, Kyorin University, Mitaka, Japan, 3 Center for General Education, Aichi Institute of Technology,
Toyota, Japan, 4 Faculty of Health and Sports Sciences, University of Tsukuba, Tsukuba, Japan, 5 Department of Biomedical
Engineering, Toyo University, Kawagoe, Japan

To investigate the influence of short-term vigorous endurance training on aortic blood

pressure (BP), pulse wave analysis was performed in 36 highly trained elite collegiate
endurance runners before and after a 7-day intense training camp. Subjects participated
three training sessions per day, which mainly consisted of long distance running and
sprint training to reach the daily target distance of 26 km. After the camp, they were
divided into two groups based on whether the target training was achieved. Aortic systolic
BP, pulse pressure, and tension-time index (TTI, a surrogate index of the myocardial
oxygen demand) were significantly elevated after the camp in the accomplished group
Edited by: but not in the unaccomplished group, whereas the brachial BP remained unchanged in
Thomas Leonhard Stöggl, both groups. The average daily training distance was significantly correlated with the
University of Salzburg, Austria
changes in aortic systolic BP (r = 0.608, p = 0.0002), pulse pressure (r = 0.415,
Reviewed by:
Beat Knechtle,
p = 0.016), and TTI (r = 0.438, p = 0.011). These results suggest that aortic BP is
University of Zurich, Switzerland affected by a short-term vigorous training camp even in highly trained elite endurance
Naoto Fujii,
athletes presumably due to a greater training volume compared to usual.
University of Tsukuba, Japan
*Correspondence: Keywords: aortic blood pressure, pulse wave analysis, endurance training, vigorous training, athletic conditioning
Jun Sugawara
[email protected]
INTRODUCTION
Specialty section:
This article was submitted to There is a widely held notion that the central arterial pressure waveform is synthesized by the
Exercise Physiology, overlapped reflection waves returning from the periphery (mainly lower body) on the incident wave
a section of the journal in phase; therefore, there are disparities between aortic and peripheral BP waveforms (Nichols and
Frontiers in Physiology McDonald, 2011). Importantly, aortic BP is more strongly related to concentric left ventricular
Received: 31 December 2017 geometry than brachial BP (Roman et al., 2007). Furthermore, systolic BP is strongly related to left
Accepted: 12 March 2018 ventricular hypertrophy, and pulse pressure is strongly related to vascular stiffening (Roman et al.,
Published: 28 March 2018 2007, 2010). Thus, assessment of aortic BP could be an indicator of cardiac and vascular condition.
Citation: The beneficial effects of endurance exercise at a moderate intensity on the cardiovascular
Tomoto T, Sugawara J, Hirasawa A, system are well recognized (Thompson et al., 2003; Seals et al., 2008; Rowe et al., 2014).
Imai T, Maeda S and Ogoh S (2018)
On the other hand, the unfavorable effects of cardiovascular system including higher aortic
Impact of Short-Term Training Camp
on Aortic Blood Pressure in Collegiate
stiffness and aortic BP have been confirmed among highly endurance-trained populations
Endurance Runners. who participate in prolonged intense exercise events, especially marathons running (Scharhag
Front. Physiol. 9:290. et al., 2005; Vlachopoulos et al., 2010) and ultramarathons running (Knez et al., 2006; Burr
doi: 10.3389/fphys.2018.00290 et al., 2014) compared with age-matched physically active peers. However, the potential genetic

Frontiers in Physiology | www.frontiersin.org 46 March 2018 | Volume 9 | Article 290

Tomoto et al. Short-Term Training and Aortic Pressure

influences in the aforementioned results cannot be completely (AIx75 ), time to wave reflection (TR ), tension time index (TTI),
ruled out due to a cross-sectional study design (Scharhag et al., diastolic time index (DTI), and sub-endocardial viability ratio
2005; Knez et al., 2006; Vlachopoulos et al., 2010; Burr et al., (SEVR) were computed as previously reported (Vlachopoulos
2014). Vlachopoulos et al. (2010) and Scharhag et al. (2005) et al., 2010).
observed that peripheral and aortic BP and pulse pressure The 7-day camp was conducted in August (summer season).
were significantly reduced 24 h after a marathon running race Although all subject engaged in regular long distance running,
(Scharhag et al., 2005; Vlachopoulos et al., 2010), but this the target training distance during the 7-day camp was longer
response seems to reflect post-exercise hypotension (Halliwill, (26 km/day) than the weekly mandatory training distance.
2001). Thus, the effect of high intense exercise training (e.g., The target running distance the week before the camp was
repetition of intense endurance exercise bouts) on the aortic BP 13 km/day. To accomplish such a long distance, subjects
and pulse pressure is still unknown. participated in three practice sessions per day; 1st session
We previously reported that in well-trained male collegiate (5:30–7:30), athletes mainly performed long distance running
endurance runners, systemic arterial stiffness increased after a 7- (group, individual, tempo running); 2nd session (8:30–12:30)
day intense endurance training camp (Tomoto et al., 2015). Since consisted of jogging, basic core, and lower body strength
arterial stiffening promotes early return of the reflected wave training; 3rd session (14:00–18:30) consisted mainly of speed
from peripheral to the heart and increases aortic BP (Nichols development training such as repetition of 400- and 1,000-
and McDonald, 2011), we hypothesized that the arterial stiffening m sprints. During the camp, all subjects were required to
induced by intense endurance training may cause to amplify participate all practice sessions, and their schedules were
the aortic arterial pressure waveform. As a follow-up study, the strictly controlled, especially practice duration, meal times, and
purpose of this study was to determine the effect of intense sleep.
endurance exercise bouts to aortic BP. To determine the effect of the intervention on the
hemodynamic parameters, repeated measures analysis of
METHODS variance was performed. In the case of a significant F-value,
a post hoc test (the Bonferroni method) was performed to
We recruited the subjects from national ranked Ekiden relay identify significant differences in the mean values of interest.
race team in Japan. We studied a total of 36 well-trained male For simple correlation analysis as well as to identify the
collegiate endurance runners. The average of their official best effect of training distance on aortic and brachial BP, three
times for a 5,000-m race was 14′ 28′′ ± 0′ 17′′ (mean ± SD). All subjects who did not engage any running practice during
of the subjects were healthy, normotensive (<140/90 mmHg), the camp were excluded due to lower leg injuries. These
non-obese (Body mass index, BMI<25 kg/m2 ), nonsmokers, three subjects trained on the same schedule as others and
who were free of medication as well as overt chronic heart performed stationary bike training instead of running. Data
and lung disease as assessed by their medical histories. None were reported as the mean ± SD. All comparisons were based
of the subjects were taking cardiovascular-acting medication. on a 95% confidence limit with P < 0.05 considered statistically
This study was reviewed and approved by the Institutional significant.
Review Board (Toyo University: 2012-R-04). Additionally, all
procedures conformed to the ethical guidelines of the Helsinki RESULTS
Declaration. All subjects provided informed written consent
prior to participation. After the camp, the athletes were divided into two groups: the
All measurements were performed at the same time of accomplished and unaccomplished group. The accomplished
the day on the first day of the 7-day training camp and the group (n = 24) achieved the daily and overall target, while
day after the camp ended. Each subject fasted overnight the unaccomplished group (n = 12) did not complete the
prior to all measurements. Aortic and brachial hemodynamic training menu due to deconditioning during the camp. The
parameters were recorded after at least 15 min supine-position accomplished group constantly completed all training menu,
rest in a quiet air-conditioned room (24–25◦ C). Subjects whereas the unaccomplished group mainly engaged to long
abstained from alcohol for 24 h and caffeine for 12 h prior distance jogging and basic core and lower body strength training
to the experiment. Electrocardiogram (ECG), left carotid without speed development training (e.g. repetition of 400- and
arterial pressure waveforms (via applanation tonometry 1,000-m sprints). During the camp, the accomplished group
sensor), and brachial BP (via oscillometric sensors) were completed a longer training distance than the unaccomplished
simultaneously measured using a vascular function-screening group (31 ± 3 km/day vs. 13 ± 10 km/day, P < 0.001).
device (form PWV/ABI, Omron-Colin, Kyoto, Japan). Carotid The completed training distance in the accomplished group
arterial pressure waveforms were transferred into aortic was approximately 2.5 times longer than 1 week before
pressure waveforms by pulse wave analysis software involving a the camp. Table 1 shows the physical characteristics and
validated generalized transfer function (SphygmoCor software, hemodynamic parameters before and after the camp. There was
AtCor Medical, Sydney, Australia) (Figure 1). The aortic no significant group-difference in either physical characteristics
hemodynamic parameters including aortic systolic BP, pulse or hemodynamic parameters prior to the camp. After the
pressure, augmentation pressure (AP), augmentation index camp, excessive body weight loss, an indicator of dehydration,
(Alx), Alx corrected for heart rate at 75 beats per minutes was not observed. The heart rate, brachial BP, aortic AP,

Frontiers in Physiology | www.frontiersin.org 47 March 2018 | Volume 9 | Article 290

Tomoto et al. Short-Term Training and Aortic Pressure

FIGURE 1 | Aortic pulse wave analysis. Type A waveform: reflecting wave during early systole produces an augmented systolic pressure. Type C waveform: reflecting
wave during late systole produces a longer TR and Pi after Ps. AP, augmentation pressure; DTI, diastolic tension index; Pi, incident pressure from a reflecting pressure
wave; PP, pulse pressure; Ps, systolic pressure; TR , round-trip travel time of the reflecting pressure wave; TTI, tension-time index.

TABLE 1 | Physical characteristics and hemodynamic variables in unaccomplished and accomplished training group before and after the camp.

Unaccomplished group (n = 12) Accomplished group (n = 24)

Before After Before After

Height, cm 171 ± 5 – 171 ± 5 –

Weight, kg 58 ± 5 58 ± 5 58 ± 4 58 ± 4
Heart Rate, bpm 48 ± 5 48 ± 7 47 ± 5 48 ± 5
Brachial systolic BP, mmHg 109 ± 4 109 ± 4 110 ± 8 112 ± 8
Brachial Mean BP, mmHg 76 ± 3 75 ± 3 76 ± 6 78 ± 6
Brachial diastolic BP, mmHg 60 ± 4 58 ± 4 59 ± 6 60 ± 6
Brachial PP, mmHg 49 ± 4 50 ± 5 51 ± 4 52 ± 5
Aortic AP, mmHg −2 ± 6 0±4 0±5 1±6
AIx, % −4 ± 12 0±9 −1 ± 11 3 ± 11
AIx75 , % −17 ± 13 −13 ± 9 −14 ± 11 −11 ± 11
TR , msec 173 ± 10 175 ± 7 172 ± 10 167 ± 11*
TTI, mmHg*ms 1,399 ± 146 1,366 ± 159 1,342 ± 168 1,434 ± 134*
DTI, mmHg*ms 3,096 ± 172 3,129 ± 215 3,115 ± 290 3,186 ± 316
SEVR, % 223 ± 27 233 ± 40 235 ± 31 224 ± 27

Data are mean ± SD. BP, blood pressure; MAP, mean arterial pressure; PP, pulse pressure; AP, augmentation pressure; AIx, augmentation index; AIx75 , Alx corrected for heart rate at
75 beats per minutes; TR , round-trip travel time of the pressure wave from the heart to the peripheral reflected sites; TTI, time tension index; DTI, diastolic tension time integral; SEVR,
subendocarial viability ratio. *P < 0.05 vs. before the camp.

AIx, and AIx75 were unchanged in both groups. The aortic P = 0.929), and AIx75 (r = −0.064, P = 0.725); TTI (r = −0.203,
systolic BP and pulse pressure significantly increased after the P = 0.235); DTI (r = 0.056, P = 0.746); SEVR (r = 0.250,
camp in the accomplished group (P < 0.001) but not in the P = 0.142).
unaccomplished group (Figure 2). In the accomplished group,
TR shortened and TTI increased significantly (P < 0.01), DISCUSSION
whereas DTI and SEVR did not significantly change after the
camp. The average training distance during the camp was not The primary findings from the present study are as flows. First,
significantly correlated with the changes in brachial systolic BP aortic systolic BP, pulse pressure, and tension-time index (TTI,
and pulse pressure, whereas changes in aortic hemodynamic a surrogate index of the myocardial oxygen demand) were
measures were correlated with the average training distance: significantly elevated after the camp in the accomplished group
aortic systolic BP, pulse pressure, TTI, and SEVR (Figure 3). but not in the unaccomplished group, whereas the brachial BP
Individual changes in TR did not correlate with corresponding remained unchanged in both groups. Secondly, the average daily
changes in aortic hemodynamic measures: aortic systolic BP training distance was significantly correlated with the changes in
(r = −0.230, P = 0.166); aortic pulse pressure (r = −0.055, aortic systolic BP, pulse pressure, and TTI. These results suggest
P = 0.750); aortic AP (r = 0.012, P = 0.945), AIx (r = −0.016, that aortic BP is affected by a short-term vigorous training camp

Frontiers in Physiology | www.frontiersin.org 48 March 2018 | Volume 9 | Article 290

Tomoto et al. Short-Term Training and Aortic Pressure

et al., 2010). More importantly, the training camp did not

significantly alter the peripheral BP, and individual changes
in peripheral BP were not associated with the training
distance during the camp. These results suggest that aortic
hemodynamic measures are more sensitive than peripheral
BP to acutely increased training volume. To confirm these
findings, prospective data linking aortic BP response to chronic
overtraining are needed.
The aortic pressure wave may be augmented by overlapping
the early return and high amplitude of the reflected wave
from the periphery (mainly from lower body) to the proximal
aorta on the forward traveling wave generated by the LV
ejection in-phase (Nichols and McDonald, 2011). In the present
study, TR (a surrogate index of aortic pulse wave velocity)
significantly correlated with other aortic hemodynamic measures
such as aortic systolic BP and PP. Therefore, increases in aortic
systolic BP and PP could not be explained by the early return
of the reflected wave from peripheral to the heart. In this
context, contrary to the traditional wave theory (Nichols and
McDonald, 2011), recent studies suggest that the reflected wave
component of arterial hemodynamics does not contribute to
augmented central pressure (Schultz et al., 2015; van Mil et al.,
2016). Furthermore, it has been also reported that the reflected
wave might become diminished until in the proximal aorta,
and it no longer possible to identify contributions to central
pressure augmentation (Davies et al., 2012). Growing evidence
suggests that major determinants of central BP waveform may
be the incident waves arising from left ventricular ejection
and proximal aortic compliance rather than the wave reflection
(Sharman et al., 2009; Davies et al., 2010; Schultz et al.,
2014).
The changes in central hemodynamic parameters in this study
FIGURE 2 | Change in aortic systolic blood pressure (A) and pulse pressure were potentially impacted by repeated mechanical stress from
(B) in the unaccomplished and accomplished groups before and after the
the heart, inflammatory state from muscle damage, and heart
camp. *P < 0.05 vs. before the camp.
accumulation from prolonged exercise training during summer.
First, repeated and particularly excessive stress imposed on the
elastic elements of the aortic wall may be a cause of their
even in highly trained elite endurance athletes partly due to a mechanical fatigue which leads to elevating aortic BP (Nichols
greater training volume compared to usual. and McDonald, 2011). Secondary, aortic BP is affected by
To the best of our knowledge, this is the first study inflammatory state common observed in marathon runner after
demonstrating the acute elevation of aortic systolic BP and the race (Knez et al., 2006, 2007). It is plausible that the vigorous
pulse pressure in elite endurance athletes after short-term training-induced muscle damage and systemic inflammation
endurance training camp characterized by a greater-than- may contribute to elevating aortic BP (Jee et al., 2013). In the
normal training volume. Such changes were associated with present study, the participants of the camp practiced not only on
a proximally greater than twice longer running distance flat track conditions but also on up-and downhill courses; thus, a
compared with 1 week before the camp. In addition, individual large amount of exercise stress—especially eccentric contractions
changes in TTI, a surrogate index of oxygen demand, was during downhill running—was given to the muscle. Furthermore,
positively correlated and SEVR, surrogate indices of myocardial the 7-day training camp was conducted on summer, thus, the
oxygen consumption (Sarnoff et al., 1958) and sub-endocardial prolonged exercise in the heat causes greater hyperthermia that
perfusion (Buckberg et al., 1972), was correlated negatively may yield heat acclimation. Further investigation to clarify the
with the training distance during the camp. These results underlying mechanisms for sustained aortic arterial pressure
suggest that increased training volume with short-term may elevation following vigorous intensity exercise training, especially
lead imbalance of myocardial oxygen demand and supply during heat stress, is needed.
and cause of myocardial fatigue. Our findings could expand Several experimental considerations should be noted. The
the evidence from the cross-sectional investigation that highly primary issue is the lack of recorded training intensity, such
trained marathon runners have a higher aortic BP compared with as heart rate during exercise. For example, by multiplying
age-matched recreationally active control subjects (Vlachopoulos the duration of a training session by the average heart rate

Frontiers in Physiology | www.frontiersin.org 49 March 2018 | Volume 9 | Article 290

Tomoto et al. Short-Term Training and Aortic Pressure

FIGURE 3 | Relationships between the daily training distance and changes in brachial and aortic hemodynamics: brachial and aortic systolic blood pressure (SBP)
(A,C) and pulse pressure (PP) (B,D); aortic tension-time index (TTI) (E); sub-endocardial viability ratio (SEVR) (F).

achieved during a session (i.e., training impulse; Macdougall the repetition of over-reaching and super-compensation. The
et al., 1982), the training stimulus could be quantified in detail. imbalance between training volume/intensity and recovery can
In addition, we did not evaluate the concomitant changes lead to an advanced fatigue state (i.e., overtraining syndrome).
in biochemical parameters (i.e., inflammation biomarker) and Therefore, useful (e.g., sensitive) markers to detect fatigue are
autonomic nervous activity. Measuring these parameters may needed to prevent and manage overtraining for not only athletes
yield insight on the mechanism of the increased aortic systolic but also coaches. Since the strength of this study was a field
BP after the training camp. study with measured hemodynamics parameters in controlled
Training programs for highly trained athletes are planned condition among well-trained endurance athletes before and
with the repetition in the training cycle composed of intense after the camp, the results of this study may provide the
training periods followed by shorter recovery periods, such as consideration of planning short-term summer training camps.

Frontiers in Physiology | www.frontiersin.org 50 March 2018 | Volume 9 | Article 290

Tomoto et al. Short-Term Training and Aortic Pressure

In conclusion, in highly-trained elite endurance athletes, the figures; TT, JS, and SO: Drafted manuscripts; TT, JS, AH, TI, SM,
aortic systolic BP and pulse pressure increases acutely without and SO: Approved final version of manuscripts.
concomitant elevations in brachial BP after a 7-day training camp
characterized by a greater training volume compared with regular FUNDING
training. These alterations might be associated with a greater
training volume. This study was funded by a Grant-in-Aid for Scientific Research
from the Japanese Ministry of Education, Science, and Culture
AUTHOR CONTRIBUTIONS (Grant Numbers JP25702045 and JP26670116).

TT, JS, AH, TI, and SO: Decided conception and design of ACKNOWLEDGMENTS
research; performed experiments; analyzed data; TT, JS, SM, and
SO: Interpreted results of experiments; TT and JS: Prepared We thank all subjects of this study and their team staffs.

REFERENCES Scharhag, J., Herrmann, M., Urhausen, A., Haschke, M., Herrmann, W., and
Kindermann, W. (2005). Independent elevations of N-terminal pro-brain
Buckberg, G. D., Towers, B., Paglia, D. E., Mulder, D. G., and Maloney, J. V. (1972). natriuretic peptide and cardiac troponins in endurance athletes after prolonged
Subendocardial ischemia after cardiopulmonary bypass. J. Thorac. Cardiovasc. strenuous exercise. Am. Heart J. 150, 1128–1134. doi: 10.1016/j.ahj.2005.01.051
Surg. 64, 669–684. Schultz, M. G., Davies, J. E., Hardikar, A., Pitt, S., Moraldo, M., Dhutia, N.,
Burr, J. F., Drury, C. T., Phillips, A. A., Ivey, A., Ku, J., and Warburton, D. E. (2014). et al. (2014). Aortic reservoir pressure corresponds to cyclic changes in aortic
Long-term ultra-marathon running and arterial compliance. J. Sci. Med. Sport volume: physiological validation in humans. Arterioscler. Thromb. Vasc. Biol.
17, 322–325. doi: 10.1016/j.jsams.2013.04.018 34, 1597–1603. doi: 10.1161/ATVBAHA.114.303573
Davies, J. E., Alastruey, J., Francis, D. P., Hadjiloizou, N., Whinnett, Z. I., Manisty, Schultz, M. G., Hughes, A. D., Davies, J. E., and Sharman, J. E. (2015). Associations
C. H., et al. (2012). Attenuation of wave reflection by wave entrapment and clinical relevance of aortic-brachial artery stiffness mismatch, aortic
creates a “horizon effect” in the human aorta. Hypertension 60, 778–785. reservoir function, and central pressure augmentation. Am. J. Physiol. Heart
doi: 10.1161/HYPERTENSIONAHA.111.180604 Circ. Physiol. 309, H1225–H1233. doi: 10.1152/ajpheart.00317.2015
Davies, J. E., Baksi, J., Francis, D. P., Hadjiloizou, N., Whinnett, Z. I., Manisty, C. Seals, D. R., Desouza, C. A., Donato, A. J., and Tanaka, H. (2008).
H., et al. (2010). The arterial reservoir pressure increases with aging and is the Habitual exercise and arterial aging. J. Appl. Physiol. 105, 1323–1332.
major determinant of the aortic augmentation index. Am. J. Physiol. Heart Circ. doi: 10.1152/japplphysiol.90553.2008
Physiol. 298, H580–H586. doi: 10.1152/ajpheart.00875.2009 Sharman, J. E., Davies, J. E., Jenkins, C., and Marwick, T. H. (2009). Augmentation
Halliwill, J. R. (2001). Mechanisms and clinical implications of post- index, left ventricular contractility, and wave reflection. Hypertension 54,
exercise hypotension in humans. Exerc. Sport Sci. Rev. 29, 65–70. 1099–1105. doi: 10.1161/HYPERTENSIONAHA.109.133066
doi: 10.1097/00003677-200104000-00005 Thompson, P. D., Buchner, D., Pina, I. L., Balady, G. J., Williams, M.
Jee, H., Park, J., Oh, J. G., Lee, Y. H., Shin, K. A., and Kim, Y. J. (2013). Effect of A., Marcus, B. H., et al. (2003). Exercise and physical activity in the
a prolonged endurance marathon on vascular endothelial and inflammation prevention and treatment of atherosclerotic cardiovascular disease: a
markers in runners with exercise-induced hypertension. Am. J. Phys. Med. statement from the Council on Clinical Cardiology (Subcommittee
Rehabil. 92, 513–522. doi: 10.1097/PHM.0b013e31829232db on Exercise, Rehabilitation, and Prevention) and the Council on
Knez, W. L., Coombes, J. S., and Jenkins, D. G. (2006). Ultra-endurance exercise Nutrition, physical activity, and metabolism (Subcommittee on Physical
and oxidative damage: implications for cardiovascular health. Sports Med. 36, Activity). Circulation 107, 3109–3116. doi: 10.1161/01.CIR.0000075572.401
429–441. doi: 10.2165/00007256-200636050-00005 58.77
Knez, W. L., Jenkins, D. G., and Coombes, J. S. (2007). Oxidative stress Tomoto, T., Sugawara, J., Hirasawa, A., Imai, T., Maeda, S., and Ogoh, S. (2015).
in half and full Ironman triathletes. Med. Sci. Sports Exerc. 39, 283–288. Impact of short-term training camp on arterial stiffness in endurance runners.
doi: 10.1249/01.mss.0000246999.09718.0c J. Physiol. Sci. 65, 445–449. doi: 10.1007/s12576-015-0383-6
Macdougall, J. D., Wenger, H. A., Green, H. J. Canadian Association of Sports, van Mil, A. C., Pearson, J., Drane, A. L., Cockcroft, J. R., Mcdonnell, B. J., and Stöhr,
Sciences, and Sport Medicine Council of Canada (1982). Physiological Testing E. J. (2016). Interaction between left ventricular twist mechanics and arterial
of the Elite Athlete. Hamilton, ON: Published by the Canadian Association of haemodynamics during localised, non-metabolic hyperaemia with and without
Sport Sciences, in collaboration with the Sport Medicine Council of Canada. blood flow restriction. Exp. Physiol. 101, 509–520. doi: 10.1113/EP085623
Nichols, W. W., and McDonald, D. A. (2011). Mcdonald’s Blood Flow in Arteries Vlachopoulos, C., Kardara, D., Anastasakis, A., Baou, K., Terentes-Printzios, D.,
Theoretical, Experimental and Clinical Principles, 6th Edn. London: Hodder Tousoulis, D., et al. (2010). Arterial stiffness and wave reflections in marathon
Arnold. runners. Am. J. Hypertens. 23, 974–979. doi: 10.1038/ajh.2010.99
Roman, M. J., Devereux, R. B., Kizer, J. R., Lee, E. T., Galloway, J. M., Ali, T., et al.
(2007). Central pressure more strongly relates to vascular disease and outcome Conflict of Interest Statement: The authors declare that the research was
than does brachial pressure: the Strong Heart Study. Hypertension 50, 197–203. conducted in the absence of any commercial or financial relationships that could
doi: 10.1161/HYPERTENSIONAHA.107.089078 be construed as a potential conflict of interest.
Roman, M. J., Okin, P. M., Kizer, J. R., Lee, E. T., Howard, B. V., and Devereux,
R. B. (2010). Relations of central and brachial blood pressure to left ventricular The reviewer NF declared a shared affiliation, with no collaboration, with
hypertrophy and geometry: the Strong Heart Study. J. Hypertens. 28, 384–388. one of the authors, SM, to the handling Editor.
doi: 10.1097/HJH.0b013e328333d228
Rowe, G. C., Safdar, A., and Arany, Z. (2014). Running forward: new Copyright © 2018 Tomoto, Sugawara, Hirasawa, Imai, Maeda and Ogoh. This is an
frontiers in endurance exercise biology. Circulation 129, 798–810. open-access article distributed under the terms of the Creative Commons Attribution
doi: 10.1161/CIRCULATIONAHA.113.001590 License (CC BY). The use, distribution or reproduction in other forums is permitted,
Sarnoff, S. J., Braunwald, E., Jr Welch, G. H., Case, R. B., Stainsby, W. N., and provided the original author(s) and the copyright owner are credited and that the
Macruz, R. (1958). Hemodynamic determinants of oxygen consumption of original publication in this journal is cited, in accordance with accepted academic
the heart with special reference to the tension-time index. Am. J. Physiol. 192, practice. No use, distribution or reproduction is permitted which does not comply
148–156. doi: 10.1152/ajplegacy.1957.192.1.148 with these terms.

Frontiers in Physiology | www.frontiersin.org 51 March 2018 | Volume 9 | Article 290

 ORIGINAL RESEARCH
published: 18 December 2017
doi: 10.3389/fphys.2017.01069

The Training Characteristics of the

World’s Most Successful Female
Cross-Country Skier
Guro S. Solli 1 , Espen Tønnessen 2 and Øyvind Sandbakk 3*
1
Department of Sports Science and Physical Education, Nord University, Bodø, Norway, 2 The Norwegian
Olympic Federation, Oslo, Norway, 3 Department of Neuromedicine and Movement Science, Centre for Elite Sports
Research, Norwegian University of Science and Technology, Trondheim, Norway

The main aim of this study was to investigate the training characteristics of the
most successful female cross-country skier ever during the best period of her career.
The participant won six gold medals at the Olympic Games, 18 gold medals at the
World Championship, and 110 World Cup victories. Day-to-day training diary data,
interviews, and physiological tests were analyzed. Training data was systemized by
training form (endurance, strength, and speed), intensity [low- (LIT), moderate- (MIT),
and high-intensity training (HIT)], and mode (running, cycling, and skiing/roller skiing),
followed by a division into different periodization phases. Specific sessions utilized in the
Edited by: various periodization periods and the day-to-day periodization of training, in connection
Billy Sperlich,
with altitude camps and tapering toward major championships, were also analyzed.
Integrative & Experimentelle
Trainingswissenschaft, Universität Following a 12-year nonlinear increase in training load, the annual training volume during
Würzburg, Germany the five consecutive successful years stabilized at 937 ± 25 h, distributed across 543 ± 9
Reviewed by: sessions. During these 5 years, total training time was distributed as 90.6% endurance-,
Thierry Busso,
University of Saint-Etienne, France 8.0% strength-, and 1.4% speed-training, with endurance-training time consisting of
Jamie Stanley, 92.3 ± 0.3% LIT, 2.9 ± 0.5% MIT, and 4.8 ± 0.5% HIT. Total LIT-time consisted
South Australian Sports Institute,
of 21% warm-up, 14% sessions <90 min, and 65% long-duration sessions >90 min.
Australia
While the total number of LIT sessions remained stable across phases (32 sessions),
*Correspondence:
Øyvind Sandbakk total LIT-time was reduced from GP (76 h/month) to SP (68 h/month) and CP (55
[email protected] h/month). MIT-time decreased from GP (2.8 h/month) to SP (2.2 h/month) and CP
(1 h/month). HIT-time increased from GP (2.8 h/month) to SP (3.2 h/month) and CP
Specialty section:
This article was submitted to (4.7 h/month). Altitude training accounted for 18–25% of annual training volume and
Exercise Physiology, performed across relatively short training camps (≤16 days) with a clear reduction of
a section of the journal
Frontiers in Physiology
HIT training, but increased total and LIT volume compared to sea-level training. Training
Received: 29 September 2017
before international championships included a 2-week increase in LIT and strength
Accepted: 05 December 2017 volume followed by a gradual reduction of training volume and increased HIT during
Published: 18 December 2017
the last week. This study provides unique data on the world’s most successful female
Citation:
cross-country skier’s long-term training process, including novel information about the
Solli GS, Tønnessen E and
Sandbakk Ø (2017) The Training distribution of and interplay between sessions of different forms, intensities, and exercise
Characteristics of the World’s Most modes throughout the annual season.
Successful Female Cross-Country
Skier. Front. Physiol. 8:1069. Keywords: altitude training, endurance training, high-intensity training, performance, periodization, speed
doi: 10.3389/fphys.2017.01069 training, strength training, tapering

Frontiers in Physiology | www.frontiersin.org 52 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

INTRODUCTION (Sandbakk et al., 2011b, 2016; Tonnessen et al., 2014), the design
of specific sessions within the different zones and modes are not
Cross-country (XC) skiers optimize their training to perform well illustrated in the current literature.
in competitions ranging from multiple 3-min sprint races to In terms of periodization, most XC skiers use a traditional
prolonged endurance races lasting up to 2 h. These competitions model, alternating between high- and low-volume weeks while
are performed across varying terrain while changing between keeping the number of MIT and HIT sessions relatively stable
the different sub-techniques in classic and skating (Sandbakk (two to three sessions per week; Tonnessen et al., 2014). However,
and Holmberg, 2017). Although the average aerobic energy some athletes organize the training in blocks with increased focus
contribution is 70–75% in sprint races and 85–95% for longer on developing specific capacities over shorter periods (Sandbakk
distances, the race format is interval-based, with increased effort and Holmberg, 2017). Altitude training represents a significant
in uphill terrain and lower intensities downhill (Norman et al., portion of world-class XC skier training and is an important
1989; Sandbakk et al., 2011a; Sandbakk and Holmberg, 2017). piece of the periodization puzzle (Sandbakk et al., 2016). The
Furthermore, the majority of competitions involve mass-starts main aim of these altitude training camps (living at ∼1,800–
in which sprint ability is critical in determining the final result. 2,000 m above sea-level and training at 1,000–3,000 m) is mainly
Accordingly, high aerobic capacity is of crucial importance in to positively stimulate hematological parameters, and thereby
XC skiing, as reflected by world-class XC skiers’ high VO2max improve performance during the subsequent training and/or
values (>80 and ∼70 mL·kg−1 ·min−1 for men and women, competition period (Millet et al., 2010). However, in XC skiing,
respectively) (Saltin and Astrand, 1967; Ingjer, 1991; Sandbakk altitude training also provides an opportunity for many hours
et al., 2011b, 2016; Tonnessen et al., 2015a). However, XC skiers of skiing on snow throughout the dry-land training period
also need the ability to rapidly elevate their peak oxygen uptake, (Sandbakk and Holmberg, 2017). Although training at altitude
utilize a high fraction of their VO2max in all of the sub-techniques, has been used by endurance athletes for several decades, accurate
and have well-developed skiing efficiency and anaerobic capacity descriptions of the micro periodization of successful athletes is
(Sandbakk and Holmberg, 2017). Unique for XC skiing, both lacking in the literature.
training and competitions involve large fluctuations in speed, To ensure peak performance, a typical tapering approach
work rate, and metabolic intensity, in addition to a varying load has been to perform 2–4 weeks of overload training, followed
on the upper and lower body with the different training modes by 1–3 weeks with decreased load (Hellard et al., 2013).
and sub-techniques utilized. Hence, the training of XC skiers is Bosquet et al. (2007) reported that performance improvements
made up of a sophisticated puzzle of training sessions of different are highly sensitive to reductions in training volume and that
forms, intensities, and exercise modes that has not yet been the optimal range of volume reduction is 41–60%, without
examined in detail. substantial decreases in training frequency, compared to the pre-
World-class XC skiers have previously reported 800–950 tapering training. However, a recent study (Tonnessen et al.,
annual training hours (Tonnessen et al., 2014; Sandbakk 2014) observed that gold medal winning athletes in XC skiing
et al., 2016). Of the total annual training volume, >90% of and biathlon used a more modest and progressive reduction
elite XC skiers’ has been reported to be endurance training, in training volume, with a relatively small reduction during
with the remaining ∼10% performed as strength or speed the last weeks prior to gold-medal performance. The authors
work (Sandbakk et al., 2011b, 2016; Tonnessen et al., 2014). speculate that this progressive tapering strategy could be ideal
These studies showed a pattern of endurance training time in sports with a dense competition schedule, as is the case
distributed as 88–91% low-intensity training (LIT, < VT1 in XC skiing. However, the study reports large individual
aerobic threshold), 3–7% moderate-intensity (MIT, VT1 < >VT2 differences in tapering behavior and does not consider the
anaerobic threshold), and 5–8% high-intensity training (HIT, specific sessions utilized during the final phases of the taper.
>VT2 ). As also shown in other endurance sports (Stoggl and Thus, tapering behavior in world-class endurance athletes is
Sperlich, 2015), a pyramidal training pattern is often found lacking, particularly in athletes who have attained repeated
during the preparation period, whereas more polarized training success in major championships.
is done in the competition phase. The training of world-class XC skiers involves manipulation of
Approximately 60% of the total training time is performed variables such as different training forms (endurance, strength,
during the general preparation period between May and October. and speed), exercise modes (loading of the upper and/or lower
This period typically includes high volumes of LIT and 50–60% body), session organization (continuous or interval), and varying
of the endurance-training conducted as sport-specific exercise terrain, making it more complex than many other endurance
(e.g., roller skiing and skiing), with the remainder mainly sports. This makes it particularly challenging to investigate
performed as running (Tonnessen et al., 2014; Sandbakk et al., at a group level. Case studies allow us to investigate every
2016). The remaining 40% of annual training is performed piece of the training in detail and expand our understanding
during the specific preparation and competition phase, with of champion performance development. Earlier case studies of
decreased total volume, increased amount of HIT (including 30– high-level endurance athletes have focused on physiological test
40 competitions) and higher amount of sport-specific activity data (Jones, 1998, 2006; Bell et al., 2017) or short-term training
forms (Sandbakk and Holmberg, 2017). While the intensity studies (Stellingwerf, 2012; Mujika, 2014; Manunzio et al., 2016;
distribution and use of activity forms during different phases Ronnestad et al., 2017). Only a minority of studies consist
of the year are well covered by several retrospective studies of longitudinal training data spanning several years, and the

Frontiers in Physiology | www.frontiersin.org 53 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

majority of these focus on male subjects (Ingham et al., 2012; 10.5% incline using a graded protocol, including 4–6 periods
Tjelta, 2013; Bourgois et al., 2014; Tjelta et al., 2014; Pinot and of 5-min stages with stepwise 1-km/h increases in workload
Grappe, 2015). (Enoksen et al., 2011). The same treadmill (Woodway Gmbh,
Therefore, the primary aim of this study was to investigate 124 Weil am Rhein, Germany) was used at all tests and lactate
the training characteristics of the most successful female XC skier concentration was measured from the fingertip by an YSI 1500
ever during the best period of her career, including the day-to-day sport lactate analyzer (YSI, Ohio, USA) directly after completion
periodization of her training in connection with altitude camps of each stage. VO2 was recorded between the third and fourth
and tapering toward major international competitions. In order minute at each stage using an Oxycon Pro (Jaeger-Toennis,
to interpret these findings in the perspective of her long-term Wurtzburg, Germany) metabolic test system. AT was determined
development process, the secondary aim was to characterize her at the workload corresponding to 1.5 mmol/1 higher lactate
longitudinal training patterns over 17 years. concentration than the baseline value (averaged over the two
first measurements). Total, lean, and fat mass were analyzed
for the legs, trunk, arms, and head using dual-energy X-ray
METHODS
absorptiometry (DXA) (Encore 2007, Version 11.4, General
Participant Electric Medical Systems, Madison, WI, USA), and presented in
The participant (born in 1980) is the most successful female absolute values (Table 1).
competitor of all time in the winter Olympics and Nordic
skiing World Championships. This includes six Olympic gold Training Monitoring
medals (four individual, two team), 18 gold medals from the The participant recorded her day-to-day training in digital
FIS World Ski Championship (12 individual, six team), 110 diaries designed by the Norwegian Ski Association and the
individual FIS World-Cup victories, and four wins of the overall Norwegian Olympic Federation. The training recorded for
FIS World Cup (FIS, 2017). The study was evaluated by the each session included total training time distributed across
regional ethics committee of mid-Norway, and approved by training form (endurance, strength, and sprint), activity form
the Norwegian Social Science Data Services (NSD). Written (skiing, roller-skiing, running, cycling, etc.), and intensity
informed consent was obtained from the participant for the zone. Specific comments regarding session details were also
publication of this study, which was performed according to the recorded.
Helsinki declarations. To register the endurance-training intensity, the five-zone
intensity scale developed by the Norwegian Olympic Federation
Overall Design was used, which has been reported to provide a valid and accurate
To give both comprehensive understanding of and detailed measurement of the duration and intensity of training by XC
insight into the athlete’s training, the study was divided into two skiers (Sylta et al., 2014a). However, since these zone boundaries
parts (1) investigations of the participant’s longitudinal training, do not clearly correspond with underlying physiological events
performance, and test data spanning 17 years from the age of 20 (Boulay et al., 1997), we used a three zone scale based on the
to 37 years old (2000–2017); and (2) detailed investigations of the ventilatory changes corresponding to the first- and second-lactate
training during five consecutive successful seasons from the age turning point (Boulay et al., 1997; Seiler and Kjerland, 2006).
of 30 to 35 years old (2010–2015), including four international LIT refers to a training intensity below the first lactate threshold
championships with nine individual gold medals. (LT1 ) (<2 mM blood lactate, 60–82% of maximal heart rate;
HRmax ). Moderate-intensity training (MIT) refers to an intensity
Performance Data between LT1 and LT2 (2–4 mM blood lactate, 82–87% of HRmax ).
Performance data for each year was calculated using both race High-intensity training (HIT) refers to an intensity above LT2
results and the International Ski Federation’s (FIS) ranking points (>4 mM blood lactate, >87% of HRmax ) (Seiler and Kjerland,
from all individual distance and sprint competitions, including 2006). Standardized intensity scales do not take into account the
World Cup competitions, the Olympic Games, and the FIS World individual or activity-specific variation, such as the tendency for
championships (FIS, 2017). maximal steady-state concentrations of blood lactate tending to
be higher in activities activating less muscle mass (Beneke and
Physiological Testing von Duvillard, 1996; Beneke et al., 2001). Hence, the participant
The participant underwent regular VO2max and lactate profile in this study tailored her intensity zones in accordance with both
testing (test results presented in Table 1). No physiological tests test results and her own experience. Her self-reported intensity
were performed during the competition period (CP), and the zones are presented in Table 2.
presented results therefore represent tests from May or June
(the start of the general preparation period; GP) and October Registration and Systematization of
or November (the start of the specific preparation period: SP). Training Data
All physiological testing during the period was conducted at the To register training time, the participant used a combination of
Norwegian Olympic Sports Centre, primarily supervised by the the session-goal approach and time in training zone often called
same exercise physiologist. The apparatus and testing procedures a modified session-goal approach, described in detail by (Sylta
used during the lactate profile and VO2max test are previously et al., 2014b). The participant registered endurance training
described (Ingjer, 1991; Tonnessen et al., 2015a). Anaerobic by allocating the time of the different parts of the sessions
threshold (AT) was determined during treadmill running at (e.g., warm-up, intervals, and cool-down) into intensity zones

Frontiers in Physiology | www.frontiersin.org 54 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

TABLE 1 | Physiological characteristics of the world’s most successful female cross-country skier during the successful period from 2010 to 2015.

2010 2011 2012 2013 2014 Mean ± SD

GP1 GP2 GP1 GP2 GP1 GP2 GP1 GP2 GP1 GP2

Age (year) 30 30 31 31 32 32 33 33 34 34 32.0 ± 1.5

Body height (cm) 167 167 167 167 167 167 167 167 167 167 167.0 ± 0.0
Body mass (kg) 65.4 64.6 64.9 64.2 65.7 64.6 65.2 65.2 64.1 64.0 64.8 ± 0.6
Body mass index (kg·m−2 ) 23.5 23.2 23.3 23.0 23.6 23.2 23.4 23.4 23.0 22.9 23.2 ± 0.2
Lean body mass (kg) – – – – – 54.9 54.5 – 54.6 55.0 54.8 ± 0.2
Lean upper body mass (kg) – – – – – 35.0 34.1 – 34.0 34.5 34.4 ± 0.5
Lean lower body mass (kg) – – – – – 18.3 17.4 – 17.5 17.6 17.7 ± 0.4
Total body fat (%) – – – – – 14.8 15.2 – 14.2 12.8 14.3 ± 1.1
VO2max (L·min−1 )* 4.23 4.49 4.31 4.39 4.47 4.52 4.33 4.37 4.42 – 4.39 ± 0.1
VO2max (ml·kg−1 ·min−1 )* 64.7 69.5 66.4 68.4 68.0 70.0 66.4 66.7 69.0 – 67.7 ± 1.7
VO2@AT (ml·kg−1 ·min−1 )* 58.9 61.1 57.0 59.2 60.7 63.6 58.7 61.0 59.4 60.8 60.0 ± 1.8
V@AT (km/h)* 9.8 10.9 10.3 10.7 10.8 10.9 10.6 10.8 10.7 11.2 10.7 ± 0.4

VO2max , maximal oxygen uptake; AT, estimated anaerobic threshold; VO2@AT oxygen uptake at the anaerobic threshold (running); V@AT, velocity at the anaerobic threshold; GP1,
general preparation period one; GP2, general preparation period two; * , gradient of treadmill 10.5%.

TABLE 2 | Self-reported intensity zones presented as maximal, minimal and most commonly used (target) heart rates in the specific training zones, as well as the average
rating of perceived exertion across the different categories of endurance sessions for the world’s most successful female cross-country skier.

Intensity zones HR zones RPE Session categories

Min Target Max

Beat · min−1 (% HRmax ) Beat · min−1 (% HRmax ) Beat · min−1 (% HRmax )

LIT* 115 115–130 149 11 Warm up and cool down**

(67) (67–75) (86) Short-duration session < 50 min
Medium-duration session [50–90 min>
Long-duration session [90–150 min >
Very long-duration session ≥ 150 min
MIT 150 155–160 160 15 Continuous training
(87) (89–92) (92) Intervals with periods from 10 to 15 min
Intervals with periods from 6 to 10 min
HIT 161 161–170 173 19 Continuous training#
(93) (93–98) (100) Intervals with periods from 4 to 7 min
Intervals with periods < 4 min##

LIT, low-intensity training; MIT, moderate-intensity training; HIT, high-intensity training; RPE, Rating of perceived exertion (BORG scale, 6–20).
* When sprints were integrated into LIT sessions, sprint time (including 1-2 min recovery after each sprint) was subtracted from the overall duration of the session. The remaining time
was categorized as LIT.
** The category includes LIT performed as warm up or cool down in connection with MIT, HIT and strength sessions.
# Including distance competitions.
## Including sprint competitions.

based on actual HR registration supported by external load, distributed in line with “the training distribution method”
lactate measurements, and self-perceived exertion. For MIT and previously described (Tonnessen et al., 2014). All endurance
HIT sessions performed as intervals, the time in the MIT/HIT sessions were categorized based on duration and/or design, as
zone was registered from the beginning of the first interval presented in Table 2.
to the end of the last interval, including recovery periods.
Strength and speed training was registered from the start to the Periodization Phases
finish of the specific strength/speed/jump part of the session, General training data are either presented as annual training
including recovery periods. When speed training was integrated characteristics or divided into different periodization phases, as
into LIT sessions, 2 min per sprint was registered as speed presented in Table 3. The day-to-day periodization of training
training. before, during, and after altitude camps is quantified based on
All data from training diaries were systematically analyzed the final 2 weeks prior to altitude training, the first 2 weeks
session-by-session by researchers from the current research of altitude training, and the 2 weeks after the altitude camp
group. Total training time and frequency of sessions were in October. The training during and after the second week

Frontiers in Physiology | www.frontiersin.org 55 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

TABLE 3 | The division of periodization phases across the annual training cycle, RESULTS
including altitude- and peaking phases*.
Longitudinal Training Characteristics
Phase Period in annual training Duration In total, 8,105 training sessions were analyzed during the period
cycle (days)
from 2000 to 2017. These sessions comprised of 7,642 workouts
General preparation Period (GP) May–October 184 and 463 XC skiing competitions. Performance and training data
General preparation Period 1 (GP1) May–July 92 during all 17 years are presented in Figures 1A,B. Total annual
General preparation Period 2 (GP2) August–October 92
training volume increased by 80% (from 522 to 940 h) from the
Pre-altitude phase Day 14-1 before altitude camp 14
age of 20–35 (2000–2015). This a yearly progression of 30 ±
Altitude phase Day 1-14 of the altitude camp 14
53 h and an increase from ∼10 to 18 weekly training hours. The
After-altitude phase Day 1-14 after altitude camp 14
relative distribution of endurance training into LIT/MIT/HIT
was polarized, but to a lesser extent during the latter part of her
Specific preparation period (SP) November–December 61
career, i.e., ∼88/2/10 during the first part of her senior career (20–
Competition period (CP) January–March 90
27 years old) and ∼92/3/5 during the latter part (28–35 years old).
Pre-peaking Phase 1 Day 42-29 before first 14
championship event Subsequently, LIT volume increased from ∼430 h (20 years old)
Pre-peaking Phase 2 Day 28-15 before first 14
to ∼800 h (35 years old). The amount of MIT + HIT was ∼60 h
championship event during both the early (20–23 years old) and latter (29–35 years
Peaking Phase Day 14-1 before first 14 old) stage of her career, but was markedly higher (∼80 h) during
championship event a 5-year period from 23 to 28 caused by the use of extensive HIT
blocks during the general preparation phase.
*April was defined as regeneration period and was not included in any of the other periods.
However, training time in April is included in the calculation of the total annual training.
Training Characteristics of the Five Most
Successful Seasons
of the altitude camp in 2012 was excluded from the analysis During the five consecutive seasons, from May 2010 to April
because of illness. Tapering characteristics are quantified based 2015, the participant achieved 107 individual podium places
on the six final weeks of training prior to the FIS World in international competitions, including Olympic Games, FIS
Championships in 2011, 2013, and 2015 and the Olympic Games World championships and FIS World Cup. This consisted of
in 2014. 63 individual World Cup victories, two gold medals from
the 2014 Olympics and seven gold medals from the three
World Championships. Annual ranking and FIS points in the
Interviews abovementioned international races were 2.5 ± 0.8 (2.0 ± 0.8
To track missing data, ensure compliance with the training in distance races and 3.4 ± 2 in sprint races) and 9.3 ±
diary commentaries, and verify the training intensity of different 3.0 (including distance races and sprint qualifications points),
training sessions, two structured and one semi-structured respectively.
interview with the participant were conducted during the data- Physiological tests (Table 1) showed an average VO2max of
analysis phase of this study. 4.39 ± 0.09 (L·min−1 ) and 67.7 ± 1.7 (ml·kg−1 ·min−1 ) during
the 5 years, with increased values from GP1 to GP2. VO2 at AT
Missing Data was approximately 89% of VO2max .
Training information was lacking for March and April of the A total of 2,713 training sessions, performed in the period
2010/2011 season. Data for these months was calculated based from May 2010 to April 2015 were categorized based on the
on the years in which data was completely documented and detailed design of the session (Table 3). Total annual training
modified based on training plans and an interview with the volume was 937 ± 25 h, distributed across 543 ± 9 sessions. This
athlete. Sessions where information about session design was consisted of 849 ± 18 h (91%) endurance training, 75 ± 21 h (8%)
lacking (10% of MIT sessions and 3% of the HIT sessions) were strength training and 14 ± 2 h (1%) speed training. Monthly and
only used in the time-in zone analyses. weekly training patterns during different phases of the annual
cycle are presented in Figure 2 and Table 4.

Statistical Analyses Endurance Training

All data from the 2010–2015 period is presented as mean ± Using the modified session goal approach, 92.3 ± 0.3% of
standard deviation (SD) of the five years. To calculate the total endurance training time was executed as LIT, 2.9 ± 0.5%
monthly and weekly distribution of training, total training was as MIT and 4.8 ± 0.5% as HIT (including competitions).
divided by duration (days) of the specific phase and multiplied When all endurance sessions were categorized using the
by 30.4 to determine monthly time/frequency (Figures 2, 3A–D, session goal method, the distribution was 76.1 ± 1.1% LIT
4A–C) or by seven to determine the weekly time and frequency sessions, 7.3 ± 1.2% MIT sessions and 16.6 ± 1.2% HIT
(Table 4 and Figures 5A,B). All statistical analyses were carried sessions.
out in Microsoft Office Excel 2013 (Microsoft, Redmond, WA, Annual LIT volume was 784 ± 10 h. Monthly LIT volume
USA). decreased from GP (76 h), to SP (68 h) and further to CP (55 h)

Frontiers in Physiology | www.frontiersin.org 56 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

FIGURE 1 | (A,B) Annual top three performances (A) in international competitions and annual training characteristics (B) distributed into endurance (low-, moderate-,
and high-intensity), strength and speed training during a 17-year period for the world’s most successful female cross-country skier.

while the number of LIT sessions remained relatively stable Annual MIT volume was 24.6 ± 3.6 h. Monthly MIT volume
across all phases (31.8 ± 1.6 sessions/month). Hence, the average decreased from GP (2.8 h) to SP (2.2 h) and further to CP (1.0 h).
duration of LIT sessions was reduced from GP (2.0 h) to SP The monthly number of MIT sessions was relative stable across
(1.5 h) and CP (1.3 h). Total LIT time was distributed as 4% GP1, GP2, and SP (3.5 sessions), but decreased markedly in CP (2
sessions < 50 min, 10% as sessions [50–90> min, 42% as sessions sessions). Average duration of MIT sessions decreased from GP
[90–150> min and 23% as sessions ≥150 min. The remaining (0.8 h) to SP (0.7 h) and CP (0.5 h). The annual number of 35 ± 5
21% of LIT time was performed as warm-up or cool down in MIT sessions consisted of 20% continuous sessions, 48% interval
connection with MIT, HIT, or strength sessions. The number of sessions with interval-durations from 6 to 10> min, and 22% as
LIT sessions in the different categories during the different phases interval sessions with interval-durations from 10 to 15 min. The
are presented in Figure 4A. most common MIT session was an interval session consisting of

Frontiers in Physiology | www.frontiersin.org 57 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

5 × 7–8 min working periods, with 1–2 min rest in between. The
use of specific MIT sessions during different phases of the annual
cycle is presented in Figure 4B.
Annual HIT volume was 40.4 ± 3.6 h. Monthly HIT volume
increased from GP1 (2.0 h) to GP2 (3.7 h), was slightly reduced
in SP (3.2 h) and then increased in CP (4.7 h). The monthly
number of HIT sessions increased from GP1 (4.1 sessions) to
GP2 (7.0 sessions), SP (8.2 sessions) and CP (9.2 sessions).
Average duration of HIT sessions was approximately equal
(0.5 h) across all phases except from SP (0.4 h). The number
of annual competitions was 38.6 ± 6.3 (∼70% distance- and
∼30% sprint-competitions). Competition time increased from
GP (0.5 h/month), to SP (1.6 h/month) and further to CP (3.1
h/month). Competitions accounted for 42 and 49% of total
HIT time and number of HIT sessions, respectively. The annual
number of 79 ± 8 HIT sessions consisted of 45% continuous
sessions (including distance competitions), 38% interval training
with interval-durations from 4 to 7 min and 14% intervals with
FIGURE 2 | Training distribution across phases during the successful interval-durations < 4 min (including sprint competitions). The
2010–2015 period distributed into endurance (low-, moderate-, and most typical HIT interval session was 5 × 4–5 min with 2–3 min
high-intensity), strength and speed training for the world’s most successful
female cross-country skier.
rest in between. The use of specific HIT sessions during different
phases of the annual cycle is presented in Figure 4C.

FIGURE 3 | (A–D) Distribution of specific (skiing classical or skating) and non-specific activity forms (running and cycling) presented as total endurance and speed
training time (A), low—intensity training time (B), moderate—intensity training time (C) and high—intensity training time (D) across phases during the successful
2010–2015 period.

Frontiers in Physiology | www.frontiersin.org 58 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

TABLE 4 | Weekly training distribution (mean ± SD) across the diffrent periodization phases including the different phases of the altitude camp performed in October and
the 6 weeks prior to international championships during the five successfull years from 2010–2015 for the world’s most successful female cross-country skier.

GP1 GP2 SP CP

Overall Pre During After Overall Pre-peaking Pre-peaking Peaking

altitude altitude altitude Phase 1 Phase 2 phase

TOTAL TRAINING
Hours 20.9 ± 1.3 21.7 ± 0.6 19.9 ± 1.0 27.6 ± 1.6 18.3 ± 1.5 18.4 ± 0.5 14.7 ± 0.8 15.8 ± 2.2 19.7 ± 2.7 16.2 ± 0.5
Sessions 10.8 ± 0.3 11.1 ± 0.2 10.9 ± 0.4 11.9 ± 0.6 10.3 ± 0.9 11.7 ± 0.5 10.5 ± 0.6 10.8 ± 1.0 11.3 ± 1.5 11.9 ± 0.3
TRAINING FORMS
Endurance (h) 18.3 ± 0.6 19.4 ± 0.6 16.9 ± 0.4 26.1 ± 1.0 16.2 ± 1.7 16.9 ± 0.4 14.0 ± 0.7 15.2 ± 2.0 17.6 ± 2.1 15.3 ± 0.7
Strength (h) 2.2 ± 0.6 1.9 ± 0.7 2.6 ± 0.6 1.4 ± 0.8 1.8 ± 0.7 1.2 ± 0.3 0.6 ± 0.2 0.5 ± 0.0 1.9 ± 0.9 0.6 ± 0.5
Speed (h) 0.4 ± 0.1 0.4 ± 0.1 0.4 ± 0.2 0.2 ± 0.1 0.3 ± 0.2 0.2 ± 0.0 0.1 ± 0.0 0.1 ± 0.1 0.2 ± 0.1 0.3 ± 0.1
EXERCISE MODE
Specific (h) 9.8 ± 0.5 10.0 ± 0.5 7.1 ± 2.3 19.2 ± 4.8 9.2 ± 1.1 13.3 ± 1.2 12.0 ± 0.8 13.8 ± 2.0 15.3 ± 2.1 13.6 ± 0.7
Unspecific (h) 8.8 ± 0.6 9.7 ± 0.4 10.2 ± 2.9 7.1 ± 2.9 7.3 ± 0.9 3.8 ± 0.9 2.1 ± 0.3 1.5 ± 0.6 2.5 ± 0.6 2.0 ± 0.2
SPE/UNSPE (%) 53/47 51/49 41/59 73/27 56/44 78/22 85/15 90/10 86/14 87/13
INTENSITY DISTRIBUTION
LIT (h) 17.2 ± 0.5 17.9 ± 0.6 15.4 ± 0.5 24.7 ± 1.1 14.8 ± 1.8 15.7 ± 0.4 12.7 ± 0.7 14.0 ± 2.1 16.2 ± 0.9 14.0 ± 1.6
MIT (h) 0.6 ± 0.2 0.7 ± 0.1 0.4 ± 0.4 1.2 ± 0.3 0.7 ± 0.2 0.5 ± 0.1 0.2 ± 0.0 0.3 ± 0.1 0.4 ± 0.3 0.3 ± 0.2
HIT (h) 0.5 ± 0.1 0.8 ± 0.1 1.1 ± 0.2 0.2 ± 0.2 0.7 ± 0.2 0.7 ± 0.1 1.1 ± 0.1 0.9 ± 0.2 1.0 ± 0.1 1.0 ± 0.1
LIT/MIT/HIT (%) 94/3/3 92/4/4 91/2/7 94/5/1 92/4/4 93/3/4 91/1/8 92/2/6 92/2/6 91/2/7
INTENSITY DISTRIBUTION
LIT (sessions) 7.1 ± 0.3 6.9 ± 0.3 6.1 ± 1.2 8.6 ± 0.8 6.6 ± 0.8 7.8 ± 0.5 7.3 ± 0.5 7.1 ± 0.9 7.1 ± 1.0 8.3 ± 0.5
MIT (sessions) 0.8 ± 0.2 0.9 ± 0.1 0.7 ± 0.8 1.6 ± 0.3 0.9 ± 0.3 0.8 ± 0.2 0.5 ± 0.1 0.8 ± 0.5 0.5 ± 0.4 0.6 ± 0.3
HIT (sessions) 0.9 ± 0.2 1.6 ± 0.1 2.2 ± 0.3 0.3 ± 0.3 1.3 ± 0.3 1.9 ± 0.3 2.1 ± 0.2 2.3 ± 0.3 2.0 ± 0.0 2.3 ± 0.3
LIT/MIT/HIT (%) 80/9/11 74/9/17 68/8/24 82/16/2 76/10/14 74/8/18 74/5/21 71/7/22 74/5/21 74/6/20
CATEGORIZATION OF LIT
<50 min (sessions) 0.2 ± 0.2 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 1.9 ± 0.7 2.8 ± 0.5 2.1 ± 0.5 1.1 ± 0.8 3.0 ± 0.4
50–90 min (sessions) 1.0 ± 0.2 1.1 ± 0.3 1.1 ± 1.0 0.5 ± 0.4 1.5 ± 1.1 1.7 ± 0.6 1.3 ± 0.4 1.4 ± 0.8 1.3 ± 0.5 2.4 ± 1.7
90–150 min (sessions) 4.4 ± 0.5 4.0 ± 0.4 4.7 ± 0.8 3.9 ± 0.6 4.8 ± 1.0 3.2 ± 0.2 2.4 ± 0.4 3.0 ± 0.7 3.1 ± 0.8 2.4 ± 0.9
≥150 min (sessions) 1.6 ± 0.5 1.8 ± 0.1 0.5 ± 0.4 4.5 ± 0.7 0.5 ± 0.4 1.0 ± 0.4 0.7 ± 0.2 0.9 ± 0.8 1.8 ± 0.3 1.0 ± 0.7
AVG. SESSION DURATION
LIT (h) 2.0 ± 0.1 2.1 ± 0.1 1.8 ± 0.1 2.6 ± 0.1 1.8 ± 0.1 1.5 ± 0.1 1.3 ± 0.1 1.4 ± 0.1 1.7 ± 0.1 1.3 ± 0.2
MIT (h) 0.8 ± 0.1 0.8 ± 0.1 0.7 ± 0.1 0.8 ± 0.1 0.7 ± 0.0 0.7 ± 0.1 0.5 ± 0.1 0.2 ± 0.1 0.4 ± 0.3 0.3 ± 0.2
HIT (h) 0.5 ± 0.0 0.5 ± 0.0 0.5 ± 0.1 0.7 ± 0.0 0.5 ± 0.0 0.4 ± 0.0 0.5 ± 0.1 0.4 ± 0.1 0.5 ± 0.1 0.5 ± 0.0
COMPETITIONS
Hours 0.1 ± 0.1 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 0.4 ± 0.1 0.7 ± 0.1 0.6 ± 0.3 0.5 ± 0.3 0.3 ± 0.1
Number 0.2 ± 0.1 0.3 ± 0.2 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 1.1 ± 0.2 1.5 ± 0.2 1.6 ± 0.5 1.0 ± 0.6 0.9 ± 0.3

GP1, general preparation period 1; GP2, General preparation period 2; SP, Specific preparation period; CP, Competitions phase; SPE, Specific exercise mode; UNSPE, Unspecific
exercise mode; LIT, Low intensity training; MIT, moderate intensity training; HIT, High intensity training.

Strength and Speed Training of heavy strength training increasing from GP (∼50%), to SP
An important change during the five investigated years was (∼60%) and further to CP (∼65%).
an increase in annual strength training time from ∼51 h (43% A typical strength session consisted of 30–45 min of
core stabilization and 57% heavy strength training) during the core/stabilization exercises followed by 30–45 min of heavy
first 2 years (30–32 years old), to ∼90 h (50% core stabilization strength training. The core stabilization portion included various
and 50% heavy strength training) in the following 3 years. This exercises targeting muscles involved in the force transfer during
increase was due to both an increased number (55–75 sessions) specific ski movements and exercises aiming to stabilize and
and duration (0.9–1.2 h) of sessions. The amount of strength move these segments functionally while skiing. Heavy strength
training increased from 6.0 to 10.9 h/month in GP, from 2.6 sessions consisted of one or two leg exercises (e.g., squats)
to 4.2 h/month in SP and from 1.7 to 3.0 h/month in CP. and three to four upper-body exercises (e.g., seated pull-
The proportion of heavy versus core/stabilization training across down, standing double poling, pull-ups, lying bench-pull, and
phases was relatively similar during all 5 years, with the amount pullover).

Frontiers in Physiology | www.frontiersin.org 59 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

training in the classic and skating techniques were approximately

equal in total training time (48 and 52%), LIT (49 and 51%),
and HIT (49 and 51%), while the proportion of skating was
substantially higher during MIT (61%).

Altitude Training
Total annual days spent at altitude was 61 ± 9, which were mainly
distributed across five altitude camps (12–14 days June/July,
12–14 days August/September, 14–16 days October/November,
10–14 days in December and 10–12 days January/February).
Total training volume at altitude ranged from 170 to 230 h,
accounting for 18–25% of the total annual training volume. The
average weekly training volume decreased from altitude camps
performed in GP (∼26 h) to SP (∼22 h) and further to CP
(∼20 h).
Training during the 2 weeks before, 2 weeks during and
the 2 weeks after the altitude camp in October/November are
presented in Table 4 and Figure 5A. Total training volume was
∼35% higher during altitude than the phases before and after.
The increased training volume occurred due to an increased
number of LIT session’s ≥2.5 h, whereas strength training time
was lower during altitude compared to the phases before and
after. The amount of training in specific modes increased
markedly at altitude, while the total volume of MIT and HIT
remained stable (∼1.5 h/week) across all three phases. However,
the MIT/HIT distribution changed from containing more HIT
before altitude (0.4 h MIT vs. 1.1 h HIT), but more MIT during
altitude (1.2 h MIT vs. 0.2 h HIT) and equal amounts of MIT and
HIT after altitude training (0.7 h MIT vs. 0.7 h HIT).

Tapering toward International Championships

FIGURE 4 | (A–C) Distribution of low- (A), moderate- (B) and high-intensity The distribution of training during the final 6 weeks prior to gold
(C) sessions, categorized after duration and/or organization, across phases
medal performance is presented in Table 4 and Figure 5B. Here,
during the five most successful years (2010–2015).
training volume increased by 25% from pre-peaking phase 1 (day
42–29) to pre-peaking phase 2 (day 28–15), due to increased
amount of long duration LIT sessions and strength training. The
Annually, 14 ± 2 h of speed training, including 11.1 h ski-
training volume was then reduced by 18% to the peaking phase
specific exercises and 2.5 h of jumps/plyometrics, was performed.
(day 14–1), which included a modest reduction in volume (6%)
The amount of speed training decreased from GP (1.7 h/month)
during the first week (day 14–8) and a larger reduction (30%) in
to SP (0.9 h/month) and CP (0.5 h/month). Speed training was
the final week (day 7–1). The training volumes during the three
included 64 ± 9 times/year and typically performed as 6–10 ×
consecutive peaking phases were 73, 91, and 75 % of the average
10–20 s sprints or 5–8 series of 10–15 plyometric jumps using
GP2 training volume, respectively, while the number of sessions
ski specific movements integrated into LIT sessions of 90–120
was stable across all phases (∼11 sessions/week).
minutes or performed before strength sessions.
The amount of MIT+HIT volume remained relatively stable
Exercise Modes across all peaking phases (∼1.3 h/week). Week-by-week analyses
63 ± 3% (545 ± 18 h) of the yearly endurance and sprint training showed a progressive increase in the proportion of HIT vs. MIT
was performed as sport-specific exercise modes (i.e., skating and during the final 3 weeks before the championship start. A detailed
classical on skis or roller skis), with the remaining 37 ± 2% description of the specific sessions performed during the final 14
(318 ± 18 h) performed as non-specific activity forms (34% days before the 2014 Olympic Games is presented in Table 5.
running and 3% cycling). The proportion of specific activity
forms increased from GP (52%) to SP (78%) and further to DISCUSSION
CP (85%). Specific training time also increased from GP (44
h/month) to SP (58 h/month), but then decreased slightly to The present study investigated the training routines during the
CP (52 h/month). Sport-specific training accounted for 62, 83, best-performing period of the most successful female XC skier
and 72% of the annual LIT, MIT, and HIT volume, respectively. of all time, analyzed in the context of her longitudinal training
Figures 3A–D illustrate the distribution of activity forms across patterns. Following a 12-year progressive, non-linear increase in
the different intensities and training phases. The distribution of training load, the annual training volume during these 5 years

Frontiers in Physiology | www.frontiersin.org 60 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

period she also increased the amount of HIT due to extensive

blocks of HIT training during GP. The participant was already
at a high international level and achieved her first international
gold medal at 23 years old, with training volumes slightly in
excess of 700 h. However, before the age of 25 she was not
stable at the top level in distance races and mainly performed
at a world-class level in sprint skiing; after which she performed
equally well in all disciplines and techniques. While these data
highlight the importance of a high training volume to achieve
a top international level in XC-skiing (Tonnessen et al., 2014;
Sandbakk et al., 2016; Sandbakk and Holmberg, 2017), they also
indicate that a progressive increase in training load is beneficial.
Throughout the 12 initial years with increases in training
load, our participant had two major changes in “training
philosophy,” in which rapid performance improvements also
occurred. The first of these periods was in an early stage
of her senior career, where extensive blocks of HIT were
included. This led to rapid performance improvements that
stagnated after a few years, during which the progression in
training load and/or variation in stimulus were limited. The
next major performance improvement coincided with a change
to a more even distribution of training volume/intensity, a
reduction in the amount of HIT, and the implementation of
relatively large amounts of LIT. This occurred directly before
entering the successful 5-year period analyzed here. In this
period, both physiological values and performance improved
over the first year before remaining at a stable high level. The
annual training volume of >910 h consisted of ∼850 h endurance
training distributed as 92% LIT, 3% MIT, and 5% HIT when
using the modified session-goal approach to quantify training.
When quantified by the number of sessions in each zone, ∼475
FIGURE 5 | (A,B) Total weekly training distributions of (A) the 2 weeks before-, endurance sessions were distributed into 76% LIT, 7% MIT, and
2 weeks during- and 2 weeks after the annual altitude camp in October, and
(B) the final 6 weeks before the four international championships performed in
17% HIT. While these amounts of HIT and MIT are similar to
February during the period from 2010–2015, distributed into endurance what was previously reported in world-class XC skiers (Sandbakk
training (low-, moderate-, and high-intensity), strength- and speed training. et al., 2011b, 2016; Tonnessen et al., 2014), the volume of LIT
is remarkably high. This was combined with relatively high
amounts of strength training and regular speed training, which
was ∼940 h, consisting of 91% endurance-, 8% strength-, and 1% may have been beneficial for maintaining muscle mass and sprint
speed training. Endurance training was gradually more polarized, ability. A further progression in stimuli was achieved through
due to reduced LIT and increased HIT, from GP to CP. 18–25% inclusion of more strength training halfway through the 5-
of the annual training time was done at altitude, performed with year period, while the amount of endurance training remained
relatively short training camps (≤16 days) where HIT training relatively stable. One unique feature of our participant is that
is reduced and LIT training increased compared to sea-level she combines a high aerobic capacity with greater muscle mass
training. Training before international championships included a than normally reported among female XC skiers (Hegge et al.,
2-week increase in LIT and strength volume followed by gradual 2016), particularly in the upper-body where women typically
reduction in total training volume and increased HIT during the have the largest difference in body composition and performance
final week. compared to men (Sandbakk et al., 2017).
In XC skiing, not just exercise volume, frequency, and
intensity are of importance. Through the 17-year period, ∼60%
Longitudinal and General Training of annual endurance and sprint training hours were ski specific
Characteristics (skiing on snow or roller skis), while the rest was primarily
In this study, in which 17 years of training were analyzed, our running. This alternation between exercise modes, loading the
participant had a 12-year progressive, non-linear increase in whole body, upper body, and the legs to a different extent, is
training load from the age of 20 until the successful 5-year period unique for XC-skiing compared to other endurance sports. The
analyzed in detail, where annual training volume was ∼940 h, variation between employment of these training modes permits
distributed across ∼540 sessions. The overall progression in high training loads during GP, while the total training load is
training mainly included an increase in LIT, although in a 6-year reduced when less variation and increased sport-specific training

Frontiers in Physiology | www.frontiersin.org 61 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

TABLE 5 | Detailed description of the training performed during the final 14 days before the 2014 Olympic Games in Sochi, including information about the commonalities
during the same period before the World Championships in 2011, 2013, and 2015 for the world’s most successful female cross-country skier.

Training content Commonalities 2011, 2013, 2014, and 2015

14 AM: 2.5 h LIT, ski skating on varying terrain Day 14–8 before first championship event:
PM: Warm-up + 30 min strength training# • Second part of 10–12 days altitude camp at 1,800 m.a.s.l (i.e., the entire
13 Rest day altitude camp was 8–20 days before the first competition)
12 AM: 2.5 h LIT, classical skiing on varying terrain • Training volume 17–20 h
PM: 1.3 h LIT, ski skating on varied terrain, including sprints • 2–3 LIT sessions >2.5 h
11 AM: 5 × 7-min MIT*, ski skating on varied terrain • 2 MIT/HIT sessions performed at 1,000 m.a.s.l
PM: 1.3 h LIT, classical skiing on varied terrain, including sprints • 1–2 strength sessions
• 2–4 LIT sessions whit integrated sprints
10 AM: 2.7 h LIT, classical skiing on varied terrain
• 1 rest day
PM: Warm-up + 30 min strength training#
9 AM: 2.3 h LIT, ski skating on varied terrain, including sprints
PM: Rest
8 AM: 6 min MIT + 5-km HIT*, classical skiing varying terrain
PM: 0.5 h LIT, running

7 AM: 1.3 h LIT, classical skiing on varied terrain, including sprints Day 7–1 before first championship event:
PM: 0.5 h LIT, running • Training at championship elevation
6 MO: 0.5 h LIT, running • Total training volume of 13–16 h
AM: 10-km classic competition* • 3–4 HIT/MIT sessions
PM: 0.5 h LIT, running • Frequent medium and short duration LIT sessions
• Timing of sessions
5 AM: Sprint skating competition*
Day 6–4: 1–3 competitions
PM: 0.5 h LIT, running
Day 3: Easy day with LIT
4 Rest day with traveling Day 2: HIT session or easy training with LIT
3 AM: 1.3 h LIT, ski skating on varied terrain Day 1: Easy training or short duration MIT session
PM: 1.5 h LIT, classical skiing on varying terrain
2 AM: 30 min HIT*, duathlon ski classical and skating varying terrain
PM: 0.5 LIT, running
1 AM: 1.3 h LIT, classical skiing on varying terrain
PM: 0.5 h LIT, running
0 Gold medal, skiathlon Olympic Winter Games Sochi 2014

LIT, low-intensity training: heart rate < 87% max; MIT, moderate-intensity training: heart rate 87–92% max; HIT, high-intensity training: heart rate > 92% max.
*MIT and HIT sessions normally included 30–45 min of LIT as warm up and 15–30 min LIT as cool-down.
# Strength training sessions normally included 30–45 min of LIT as warm u.

is used toward CP. However, variations between exercise modes improved ability to train with relatively high speed and a high
were also employed on the micro periodization level; e.g., by technical quality also during LIT and MIT sessions in all exercise
performing heavy strength training of the upper body in the modes. This is supported by her lactate profiles, where her
morning session followed by lower body endurance training (e.g., speed at various submaximal lactate levels gradually increased
running) in the afternoon. This way of loading the upper and throughout her career. Similar results were shown in the female
lower body may not only increase the tolerable training load, marathon world record holder (Jones, 2006) and a world-class
but could also reduce negative cross-over adaptation effects from rower (Bourgois et al., 2014). This is most likely a result of
concurrent strength and endurance training. In our case, the her long-term progressive increase in endurance training load,
participant confirmed during interviews that she was conscious leading to enhanced peak oxygen uptake, fat utilization and
about the use of terrain, e.g., by combining uphill sessions where improved efficiency in all exercise modes. In this context, it is
the legs are mainly employed, with sessions primarily loading also important to note that the body mass of the participant was
the upper body by using the double poling technique on the very stable throughout her senior career, and measurements of
same day. This is likely an important factor contributing to the body composition during the five successful years showed that
combination of high endurance capacity and a relatively large both her fat percentage and bone mineral density were within
muscle mass obtained by XC skiers. healthy values. We suggest that this is an important reason for
Following a gradual increase in aerobic capacity, the her continuity in training during the 17 years with high loads of
participant’s average VO2max was ∼68 (ml·kg −1 ·min−1 ) during endurance training.
her five most successful years. This is at the same level reported Overall, our data indicate that a progressive increase in
in female champions in running and orienteering (Jones, 2006; training load until the age of 30 may be necessary in order to
Tonnessen et al., 2015b). Her AT increased correspondingly, and optimize the full potential of a top-level XC skier. We hypothesize
both the participant and her coaches highlighted her gradually that this allowed our champion XC skier to tolerate and respond

Frontiers in Physiology | www.frontiersin.org 62 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

positively to the high training volumes utilized in the 5-year sessions was observed as the CP approached. Both exercise mode,
period analyzed, where she used a polarized training pattern with organization of HIT, and use of terrain got more specific closer
a large amount of LIT. to CP. The fact that 42% of annual HIT time was competitions
emphasizes the importance of specific training to achieve success
in XC skiing. The participant also confirmed that competitions
Training Characteristics during Five were an important part of her training, particularly during her
Successful Years tapering phase. The organization of endurance sessions changed
Annual Periodization of Training from longer to gradually prioritizing shorter LIT sessions, while
The total number of LIT sessions remained stable across phases MIT and HIT sessions became more competition-specific.
throughout the training year, while total LIT-time was gradually
reduced from GP to CP. The amount of MIT, speed, and strength Altitude Training
training also decreased from GP to CP, while HIT showed the 18–25% of annual training volume was performed during
opposite pattern, which altogether induced a gradually more relatively short (10–16 days) altitude camps, living at 1,800–
polarized training pattern toward CP. The transition from 2,000 m.a.s.l and training at 1,000–3,000 m.a.s.l., with a clear
a more “pyramidal” to a more polarized endurance training reduction in HIT but an increased volume of LIT compared to
pattern was previously shown in successful athletes (Stoggl sea-level training. This altitude exposure is significantly shorter
and Sperlich, 2015). However, the large amounts of speed and than the 4 weeks recommended to fully stimulate erythropoiesis.
strength training during GP might be an important addition to However, comparable duration of camps is reported to have
concurrently develop endurance and strength capacities during beneficial effects on work economy, muscle buffering capacity,
the preparation period, whereas the subsequently more polarized and ventilatory factors (Millet et al., 2010). Furthermore, the
pattern may facilitate the ability to utilize these capacities at long-term effect of repeated short-duration altitude exposure
competition-specific intensities. over several years is currently unknown.
Simultaneously, the amount of specific training increased The participant experienced marked progress after altitude
from 50% during GP to 85% in CP. This is in line with previous training, although it is not known to what extent altitude-
studies of XC skiers and probably functions as an important facilitated effects and/or the periodization of training occurring
substitute for reduced volume during CP (Tonnessen et al., 2014). in connection with altitude camps influence the experienced
While the sport-specific proportion of LIT and HIT increased progress. Before altitude, training changed toward a more
markedly from GP to CP, the amount of specific MIT was >80% polarized pattern, including lower total volume and more HIT
during all phases. In addition, the MIT sessions were performed and strength training. During altitude, HIT was reduced and
at relative high heart rates (87–92% of HRmax ), which is higher training shifted to a more pyramidal intensity distribution,
than normally reported in elite athletes, although RPE ratings with more LIT sessions ≥2.5 h and increased amounts of MIT.
and lactate values were in the normal range for such sessions. The Training after altitude consisted of some easy days with reduced
participant confirmed that she was able to perform MIT sessions volume and no MIT or HIT during the first 4 days after
at this level, which allowed her to accumulate and tolerate much altitude exposure, followed by increased intensity in training. The
more time at >90% of HRmax than most of her peers. Such participant also highlighted the opportunity to ski on snow, more
training has previously been reported to be highly effective for time to rest, and an increased focus on recovery as possible factors
endurance adaptations and performance in well trained elite contributing to the positive effect of altitude camps.
athletes (Stepto et al., 1999; Sandbakk et al., 2013).
The accumulated LIT-time during GP was very high (76 Tapering toward International Championships
h/month) and reduced in CP (55 h/month). While the The tapering phase prior to international championships
number of LIT sessions remained stable across phases (∼32 included a phase with frequent competitions, followed by
sessions/month), the amount of LIT sessions ≥90 min decreased elevated training volume including more LIT and strength
from GP (25 sessions/month) to CP (14 sessions/month). at altitude. Thereafter, our participant reduced her training
Another pronounced change between phases was the increase volume and increased the amount of HIT during the final
of LIT sessions <50 min from GP (∼0 sessions/month) to CP week before championships. However, the reduction in training
(∼14 sessions/month). This methodological approach is novel, volume during the final 2 weeks (18%) was much lower than
and clearly shows how LIT sessions of different duration are recommendations in the literature (Bosquet et al., 2007). The
distributed differently throughout the year. The effect of duration same observation was made by Tonnessen et al. (2014), where
versus frequency of LIT sessions has not yet been examined, the authors speculated that this might be optimal in sports
although up to 90% of the total training among endurance with a dense competition schedule. As such, top athletes in XC
athletes is LIT. Interestingly, 21% (167 h) of the annual LIT skiing appear to reduce their training volume less than that
volume was warm up or cool down in connection with MIT, recommended by the current literature. Maintenance of training
HIT, or strength sessions. This part of LIT probably functions volume until the final week before the first championship event
as an important contributor to the long-term development by could also be important in order to maintain performance level
enhancing the total training volume. over 5–6 competitions during a championship lasting 9–15 days.
While the majority of MIT and HIT sessions were organized Our participant integrated the competition schedule into
as intervals during GP, an increase in continuous MIT and HIT the tapering strategy and had a relatively similar timing

Frontiers in Physiology | www.frontiersin.org 63 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

of the final competitions in the peaking phase during all training forms, intensities, and exercise modes throughout
5 years. Specifically, a period with frequent competitions, the annual season. By using a single-case approach, where
allowing less training hours, is followed by a competition quantitative data were supported by qualitative interviews, we
break, prioritizing altitude training with more MIT, long were able to present the sophisticated training of a world-class
duration LIT sessions and strength training. Thereafter, three athlete from a macro- to a micro-level, allowing the generation
HIT sessions were performed during the final 7 days, which of new hypotheses that can be tested in future research with
include competitions at day 6–4 before the championship’s larger samples.
start. However, since this analysis is based on the training
conducted prior to the first competition in each championship, AUTHOR CONTRIBUTIONS
it is not certain that this was the day with the best
performance (although gold medals were won already at the first GS, ET, and ØS designed the study; GS performed data collection;
competition). GS, ET, and ØS performed data and statistical- analysis; GS, ET,
and ØS contributed to interpretation of the results; GS and ØS
CONCLUSION wrote the draft manuscript; GS, ET, and ØS contributed to the
final manuscript.
Our study supports previous findings highlighting the
importance of a high training volume, using a polarized ACKNOWLEDGMENTS
training pattern with a large amount of LIT to reach world-class
level in XC skiing. This study provides unique data on the world’s The authors sincerely thank Marit Bjørgen, her manager Guri
most successful XC skier’s long-term training process, including Hetland and her coaches Svein Tore Samdal, Egil Kristiansen,
novel information about the physiological development and and Idar Terje Belsvik for their valuable cooperation and
the distribution of and interplay between sessions of different participation in this study.

REFERENCES Ingjer, F. (1991). Maximal oxygen uptake as a predictor of performance ability in

women and men elite cross-country skiers. Scand. J. Med. Sci. Sports 1, 25–30.
Bell, P. G., Furber, M. J., Van Someran, K. A., Anton-Solanas, A., and Swart, doi: 10.1111/j.1600-0838.1991.tb00267.x
J. (2017). The physiological profile of a multiple tour de france winning Jones, A. M. (1998). A five year physiological case study of an Olympic runner. Br.
cyclist. Med. Sci. Sports Exerc. 49, 115–123. doi: 10.1249/MSS.0000000000 J. Sports Med. 32, 39–43. doi: 10.1136/bjsm.32.1.39
001068 Jones, A. M. (2006). The physiology of the world record holder for
Beneke, R., and von Duvillard, S. P. (1996). Determination of maximal lactate the Women’s Marathon. Int. J. Sports Sci. Coach. 1, 101–116.
steady state response in selected sports events. Med. Sci. Sports Exerc. 28, doi: 10.1260/174795406777641258
241–246. doi: 10.1097/00005768-199602000-00013 Manunzio, C., Mester, J., Kaiser, W., and Wahl, P. (2016). Training intensity
Beneke, R., Leithauser, R. M., and Hutler, M. (2001). Dependence of the maximal distribution and changes in performance and physiology of a 2nd place finisher
lactate steady state on the motor pattern of exercise. Br. J. Sports Med. 35, team of the race across America over a 6 month preparation period. Front.
192–196. doi: 10.1136/bjsm.35.3.192 Physiol. 7:642. doi: 10.3389/fphys.2016.00642
Bosquet, L., Montpetit, J., Arvisais, D., and Mujika, I. (2007). Effects of tapering Millet, G. P., Roels, B., Schmitt, L., Woorons, X., and Richalet, J. P. (2010).
on performance: a meta-analysis. Med. Sci. Sports Exerc. 39, 1358–1365. Combining hypoxic methods for peak performance. Sports Med. 40, 1–25.
doi: 10.1249/mss.0b013e31806010e0 doi: 10.2165/11317920-000000000-00000
Boulay, M. R., Simoneau, J. A., Lortie, G., and Bouchard, C. (1997). Monitoring Mujika, I. (2014). Olympic preparation of a world-class female triathlete. Int. J.
high-intensity endurance exercise with heart rate and thresholds. Med. Sci. Sports Physiol. Perform. 9, 727–731. doi: 10.1123/ijspp.2013-0245
Sports Exerc. 29, 125–132. doi: 10.1097/00005768-199701000-00018 Norman, R. W., Ounpuu, S., Fraser, M., and Mitchell, R. (1989). Mechanical
Bourgois, J., Steyaert, A., and Boone, J. (2014). Physiological and anthropometric power output and estimated metabolic rates Nordic skiers during Olympic
progression in an international oarsman: a 15-year case study. Int. J. Sports competition. Int. J. Sport Biomech. 5, 169–184. doi: 10.1123/ijsb.5.2.169
Physiol. Perform. 9, 723–726. doi: 10.1123/ijspp.2013-0267 Pinot, J., and Grappe, F. (2015). A six-year monitoring case study of
Enoksen, E., Shalfawi, S. A., and Tonnessen, E. (2011). The effect of a top-10 cycling Grand Tour finisher. J. Sports Sci. 33, 907–914.
high- vs. low-intensity training on aerobic capacity in well-trained doi: 10.1080/02640414.2014.969296
male middle-distance runners. J. Strength Cond. Res. 25, 812–818. Ronnestad, B. R., Hansen, J., Vegge, G., and Mujika, I. (2017). Short-term
doi: 10.1519/JSC.0b013e3181cc2291 performance peaking in an elite cross-country mountain biker. J. Sports Sci.
FIS (2017). International Ski Federation World Cup Results [Online]. Avaliable 35, 1392–1395. doi: 10.1080/02640414.2016.1215503
online at: http://www.fis-ski.com/ (Accessed September 1, 2017). Saltin, B., and Astrand, P. O. (1967). Maximal oxygen uptake in athletes. J. Appl.
Hegge, A. M., Bucher, E., Ettema, G., Faude, O., Holmberg, H. C., and Sandbakk, Physiol. 23, 353–358.
O. (2016). Gender differences in power production, energetic capacity and Sandbakk, O., and Holmberg, H. C. (2017). Physiological capacity and
efficiency of elite crosscountry skiers during wholebody, upperbody, and training routines of elite cross-country skiers: approaching the upper
arm poling. Eur. J. Appl. Physiol. 116, 291–300. doi: 10.1007/s00421-015- limits of human endurance. Int. J. Sports Physiol. Perform. 12, 1003–1011.
3281-y doi: 10.1123/ijspp.2016-0749
Hellard, P., Avalos, M., Hausswirth, C., Pyne, D., Toussaint, J. F., and Mujika, I. Sandbakk, O., Ettema, G., Leirdal, S., Jakobsen, V., and Holmberg, H.
(2013). Identifying optimal overload and taper in elite swimmers over time. J. C. (2011a). Analysis of a sprint ski race and associated laboratory
Sports Sci. Med. 12, 668–678. determinants of world-class performance. Eur. J. Appl. Physiol. 111, 947–957.
Ingham, S. A., Fudge, B. W., and Pringle, J. S. (2012). Training distribution, doi: 10.1007/s00421-010-1719-9
physiological profile, and performance for a male international 1500-m runner. Sandbakk, O., Hegge, A. M., Losnegard, T., Skattebo, O., Tonnessen, E., and
Int. J. Sports Physiol. Perform. 7, 193–195. doi: 10.1123/ijspp.7.2.193 Holmberg, H. C. (2016). The physiological capacity of the world’s highest

Frontiers in Physiology | www.frontiersin.org 64 December 2017 | Volume 8 | Article 1069

Solli et al. The World’s Best Cross-Country Skier

ranked female cross-country skiers. Med. Sci. Sports Exerc. 48, 1091–1100. analysis. Int. J. Sports Physiol. Perform. 9, 100–107. doi: 10.1123/ijspp.2013-
doi: 10.1249/MSS.0000000000000862 0298
Sandbakk, O., Holmberg, H. C., Leirdal, S., and Ettema, G. (2011b). The Tjelta, L. I. (2013). A longitudinal case study of the training of the 2012 European
physiology of world-class sprint skiers. Scand. J. Med. Sci. Sports 21, e9–e16. 1500 m track champion. Int. J. Appl. Sport Sci. 25, 11–18. doi: 10.24985/ijass.
doi: 10.1111/j.1600-0838.2010.01117.x 2013.25.1.11
Sandbakk, O., Sandbakk, S. B., Ettema, G., and Welde, B. (2013). Effects of Tjelta, L. I., Tonnessen, E., and Enoksen, E. (2014). A case study of the training of
intensity and duration in aerobic high-intensity interval training in highly nine times New York marathon winner Grete Waitz. Int. J. Sports Sci. Coach 9,
trained junior cross-country skiers. J. Strength Cond. Res. 27, 1974–1980. 139–159. doi: 10.1260/1747-9541.9.1.139
doi: 10.1519/JSC.0b013e3182752f08 Tonnessen, E., Haugen, T. A., Hem, E., Leirstein, S., and Seiler, S. (2015a). Maximal
Sandbakk, O., Solli, G. S., and Holmberg, H. C. (2017). Sex differences in world aerobic capacity in the winter-Olympics endurance disciplines: olympic-medal
record performance: the influence of sport discipline and competition duration. benchmarks for the time period 1990-2013. Int. J. Sports Physiol. Perform. 10,
Int. J. Sports Physiol. Perform. 10, 1–23. doi: 10.1123/ijspp.2017-0196 835–839. doi: 10.1123/ijspp.2014-0431
Seiler, K. S., and Kjerland, G. O. (2006). Quantifying training intensity distribution Tonnessen, E., Svendsen, I. S., Ronnestad, B. R., Hisdal, J., Haugen, T.
in elite endurance athletes: is there evidence for an “optimal” distribution? A., and Seiler, S. (2015b). The annual training periodization of 8 world
Scand. J. Med. Sci. Sports 16, 49–56. doi: 10.1111/j.1600-0838.2004. champions in orienteering. Int. J. Sports Physiol. Perform. 10, 29–38.
00418.x doi: 10.1123/ijspp.2014-0005
Stellingwerf, T. (2012). Case study: nutrition and training periodization in Tonnessen, E., Sylta, O., Haugen, T. A., Hem, E., Svendsen, I. S., and Seiler,
three elite marathon runners. Int. J. Sport Nutr. Exerc. Metab. 22, 392–400. S. (2014). The road to gold: training and peaking characteristics in the
doi: 10.1123/ijsnem.22.5.392 year prior to a gold medal endurance performance. PLoS ONE 9:e101796.
Stepto, N. K., Hawley, J. A., Dennis, S. C., and Hopkins, W. G. (1999). Effects doi: 10.1371/journal.pone.0101796
of different interval-training programs on cycling time-trial performance.
Med. Sci. Sports Exerc. 31, 736–741. doi: 10.1097/00005768-19990500 Conflict of Interest Statement: The authors declare that the research was
0-00018 conducted in the absence of any commercial or financial relationships that could
Stoggl, T. L., and Sperlich, B. (2015). The training intensity distribution be construed as a potential conflict of interest.
among well-trained and elite endurance athletes. Front. Physiol. 6:295.
doi: 10.3389/fphys.2015.00295 Copyright © 2017 Solli, Tønnessen and Sandbakk. This is an open-access article
Sylta, O., Tonnessen, E., and Seiler, S. (2014a). Do elite endurance athletes distributed under the terms of the Creative Commons Attribution License (CC BY).
report their training accurately? Int. J. Sports Physiol. Perform. 9, 85–92. The use, distribution or reproduction in other forums is permitted, provided the
doi: 10.1123/ijspp.2013-0203 original author(s) or licensor are credited and that the original publication in this
Sylta, O., Tonnessen, E., and Seiler, S. (2014b). From heart-rate data to journal is cited, in accordance with accepted academic practice. No use, distribution
training quantification: a comparison of 3 methods of training-intensity or reproduction is permitted which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 65 December 2017 | Volume 8 | Article 1069

 REVIEW
published: 27 September 2017
doi: 10.3389/fphys.2017.00747

Massage Alleviates Delayed Onset

Muscle Soreness after Strenuous
Exercise: A Systematic Review and
Meta-Analysis
Jianmin Guo 1, 2 † , Linjin Li 3 † , Yuxiang Gong 1, 2 , Rong Zhu 2 , Jiake Xu 1, 4 , Jun Zou 1* and
Xi Chen 2*
1
School of Kinesiology, Shanghai University of Sport, Shanghai, China, 2 School of Sports Science, Wenzhou Medical
University, Wenzhou, China, 3 Wenzhou People’s Hospital, The Third Clinical Institute of Wenzhou Medical University,
Wenzhou, China, 4 Molecular Laboratory, School of Pathology and Laboratory Medicine, The University of Western Australia,
Perth, Australia
Edited by:
Billy Sperlich,
Purpose: The purpose of this systematic review and meta-analysis was to evaluate the
University of Würzburg, Germany
effects of massage on alleviating delayed onset of muscle soreness (DOMS) and muscle
Reviewed by:
Estêvão Rios Monteiro, performance after strenuous exercise.
Federal University of Rio de Janeiro,
Brazil
Method: Seven databases consisting of PubMed, Embase, EBSCO, Cochrane Library,
Wigand Poppendieck, Web of Science, CNKI and Wanfang were searched up to December 2016. Randomized
Mannheim University of Applied controlled trials (RCTs) were eligible and the outcomes of muscle soreness, performance
Sciences, Germany
(including muscle maximal isometric force (MIF) and peak torque) and creatine kinase
*Correspondence:
Jun Zou (CK) were used to assess the effectiveness of massage intervention on DOMS.
[email protected]
Results: Eleven articles with a total of 23 data points (involving 504 participants)
Xi Chen
[email protected] satisfied the inclusion criteria and were pooled in the meta-analysis. The findings
†
These authors have contributed demonstrated that muscle soreness rating decreased significantly when the participants
equally to this work. received massage intervention compared with no intervention at 24 h (SMD: –0.61, 95%
CI: –1.17 to –0.05, P = 0.03), 48 h (SMD: –1.51, 95% CI: –2.24 to –0.77, P < 0.001),
Specialty section:
This article was submitted to 72 h (SMD: –1.46, 95% CI: –2.59 to –0.33, P = 0.01) and in total (SMD: –1.16, 95%
Exercise Physiology, CI: –1.60 to –0.72, P < 0.001) after intense exercise. Additionally, massage therapy
a section of the journal
Frontiers in Physiology
improved MIF (SMD: 0.56, 95% CI: 0.21–0.90, P = 0.002) and peak torque (SMD: 0.38,
Received: 11 April 2017
95% CI: 0.04–0.71, P = 0.03) as total effects. Furthermore, the serum CK level was
Accepted: 13 September 2017 reduced when participants received massage intervention (SMD: –0.64, 95% CI: –1.04
Published: 27 September 2017
to –0.25, P = 0.001).
Citation:
Guo J, Li L, Gong Y, Zhu R, Xu J, Conclusion: The current evidence suggests that massage therapy after strenuous
Zou J and Chen X (2017) Massage exercise could be effective for alleviating DOMS and improving muscle performance.
Alleviates Delayed Onset Muscle
Soreness after Strenuous Exercise: A Keywords: massage, physiotherapy, exercise, delayed onset muscle soreness (DOMS), systematic review,
Systematic Review and meta-analysis
Meta-Analysis. Front. Physiol. 8:747.
doi: 10.3389/fphys.2017.00747 Systematic Review Registration: PROSPERO registration number: CRD42016053118.

Frontiers in Physiology | www.frontiersin.org 66 September 2017 | Volume 8 | Article 747

Guo et al. Massage Alleviates DOMS

INTRODUCTION massage on DOMS and muscle performance after strenuous

exercise.
It is well established that keeping an active lifestyle through
exercise benefits human health, especially reducing the risk
of obesity and cardiovascular disease. However, exhaustive
METHOD
or unaccustomed exercise (particularly involving eccentric Search Strategy
contractions) frequently result in temporary muscle damage, Relevant research articles from January 1980 to December 2016
leading to delayed onset muscle soreness (DOMS) (Bleakley et al., were collected with keywords such as “exercise,” “massage,”
2012). “delayed onset muscle soreness,” and “random” from the
DOMS commonly occurs within the first 24 h after exhaustive following databases: PubMed, Embase, EBSCO, Cochrane
or intense exercise, reaching a peak between 24 and 72 h Library, Web of Science, China Knowledge Resource Integrated
(Howatson and van Someren, 2008). It is often accompanied Database (CNKI) and Wanfang Database. The search was
by muscle swelling and reduction in muscle performance imposed with the limitation of randomized controlled
(Kargarfard et al., 2016; De Marchi et al., 2017), as well trials (RCTs), and without language or status limitations.
as a decrease in range of motion (Cheung et al., 2003; Supplementary Data (S1) details all of the search strategies used
Lavender and Nosaka, 2006). Although the exact mechanism in this study. The protocol was registered on the international
of DOMS remains unclear, the most accepted theory suggests prospective register of systematic reviews PROSPERO (http://
primary mechanical damage induced by exercise, followed by www.crd.york.ac.uk/PROSPERO), registration number:
inflammation attributing to the symptoms of DOMS. This is CRD42016053118.
verified by microscopic analysis showing disruption of muscle
fibers. In addition, there is also an increase of intracellular Inclusion Criteria
enzymes such as creatine kinase (CK) and inflammatory markers Trials were included when they met the inclusion criteria as
in blood (Peake et al., 2005a,b; Chatzinikolaou et al., 2010). follows: (a) only the trials designed as RCTs were covered,
In an attempt to prevent or alleviate the symptoms of (b) participants were human with no muscle or bone diseases,
DOMS, a number of conventional physiotherapeutic strategies (3) the trials compared post-exercise massage intervention with
were used such as massage, cold-water immersion, whole body the control group receiving usual care, or no intervention, (4)
vibration, compression garments, and stretching (Weerapong outcomes included the primary outcome of muscle pain or
et al., 2005; Howatson et al., 2009; Bleakley et al., 2012; Costello soreness rating, the secondary outcome: muscle strength (i.e.,
et al., 2012; Hill et al., 2014). Therapeutic massage has been maximal isometric force (MIF) and peak torque), and the level
used for body health for thousands of years worldwide. As of serum CK. Mean and standard deviation (SD) were reported
a physiotherapeutic intervention, massage treatment is widely in the trials.
used to alleviate clinical symptoms of DOMS, and to benefit
the athlete’s recovery after exercise in preparation for the next Exclusion Criteria
event (Poppendieck et al., 2016). The potential effectiveness Trials were excluded when they met the exclusion criteria as the
of massage therapy is proposed to increase skin and muscle following: (a) the articles were conference posters, abstracts, or
temperature, blood and lymphatic flow, and parasympathetic case reports; (b) mean and SD could not be obtained from the
activity. Subsequently, the effects include then relief of muscle articles or authors.
tension and stiffness, reduction of muscle soreness, and increased
joint range of motion (Weerapong et al., 2005). In addition, Studies Selection
changes in the Hoffman reflex after massage also contributes to Two reviewers (Guo JM and Li LJ) independently reviewed the
the reduction of muscle pain (Morelli et al., 1999). Moreover, the titles or abstracts of all studies, and the full contents of the
psychophysiological response to massage, including relaxation, relevant studies were checked carefully to evaluate whether the
leads to mood enhancement and fatigue reduction (Hemmings study could be included. Any disagreements were resolved by
et al., 2000). discussion or consultation with a third author (Chen X) if
Previous systematic reviews have shown benefits from necessary.
massage for symptoms of DOMS after intense exercise. However,
little support for the use of massage was observed to enhance Quality Assessment
muscle performance (Ernst, 1998; Best et al., 2008), including The Cochrane Collaboration tool was used to evaluate the
peak torque and maximal isometric force, factors that are risk of bias of the included trials (Higgins et al., 2011).
important for sporting events. It seems that the psychological Two reviewers (Guo JM and Li LJ) independently evaluated
effect is larger than the physiological effect. Torres et al. seven domain biases as follows: random sequence generation
demonstrated that massage could reduce muscle soreness and (selection bias), allocation concealment (selection bias), blinding
increase muscle recovery 24 h post-exercise based on a meta- of participants and personnel (performance bias), blinding of
analysis, but there were only few trials pooled in the meta-analysis outcome assessment (detection bias), incomplete outcome data
(3 trials) (Torres et al., 2012). Currently, the evidence of the (attrition bias), and selective reporting (reporting bias). Three
effects of massage intervention on DOMS still limited. Therefore, grades of high, low, or unclear bias were labeled for every
the purpose of this study was to evaluate the effectiveness of study included. Disagreements were resolved by discussion or

Frontiers in Physiology | www.frontiersin.org 67 September 2017 | Volume 8 | Article 747

Guo et al. Massage Alleviates DOMS

consulting with a third independent reviewer (Chen X) if

necessary.

Data Extraction
The two reviewers independently extracted the data from every
included eligible trial as the following: study characteristics
(i.e., author and year), participant characteristics (i.e., age and
number of participants), description of interventions, study
period, outcomes, and time points. Any disagreements were
settled by discussion to reach unanimity, and the authors of the
trials were contacted directly to acquire original studies and data
if necessary.

Statistical Analysis
The Review Manager software (RevMan 5.3; Cochrane
Collaboration) was used to perform the meta-analysis.
Heterogeneity among studies was evaluated by I2 statistic
as follows: low heterogeneity was assumed when I2 < 25%;
moderate heterogeneity when I2 < 75% and >25%; high
heterogeneity when I2 ≥ 75% (Higgins et al., 2003). Meta-
analysis was used to combine two or more outcomes from the
studies by random-effects or by a fixed-effect model. The fixed-
effect model was used when I2 <25% and the random-effects
model was used if I2 > 25%. As the trials used different scales
for muscle soreness rating assessment, we used standardized
mean difference (SMD) to analyze the effects. If P < 0.05, it was
considered to be a significant difference. Subgroups were used to
analyze the effectiveness of massage at different time points after
strenuous exercise. The possible publication bias was evaluated
by Funnel plot asymmetry if more than nine trials were included.
Sensitivity analysis was used when I2 > 50%, removing each trial
one by one to evaluate the stability of the results.
If the included articles did not report mean and SD, we would
contact the authors to ask for data, and the articles would be
excluded if there was no reply. When the articles showed the FIGURE 1 | Flow diagram of the study selection.
data as median and interquartile range (IQR), the mean would
be equivalent to the median, and the SD would be calculated as
SD = IQR/1.35 (Wang et al., 2016). If the articles reported the
data in terms of mean and √ standard error (SE), the SD would be Description of Included Studies
calculated as SD = SE× n (n meaning sample size) (Higgins and The characteristics of the included articles were summarized
Green, 2011; Higgins et al., 2011). in Table 1. Eleven articles were included in this meta-analysis.
The distribution of publication countries was as follows: the
United States (n = 5, 45.5%), the People’s Republic of China
RESULTS (n = 3, 27.2%), Canada (n = 1, 9.1%), India (n = 1, 9.1%) and
Search Results Iran (n = 1, 9.1%). Eight articles were published in English, and
From our initial search, 882 records were obtained. After three in Chinese.
reviewing the information on the articles’ title and abstract, 70
potentially eligible articles were identified. After reviewing the Massage Therapy
full content of the studies, 11 articles with 23 data points (504 Most (eight articles) of the trials used western massage
participants) satisfied the inclusion criteria and were pooled in techniques or Swedish massage techniques including effleurage
this meta-analysis (Weber et al., 1994; Lightfoot et al., 1997; and petrissage. Effleurage is a succession of light and stroking
Hilbert et al., 2003; Mancinelli et al., 2006; Yang, 2007; Frey Law massage, and petrissage is deep-tissue kneading of the muscles.
et al., 2008; Xiong et al., 2009; Hu et al., 2010; Andersen et al., Three articles performed traditional Chinese massage including
2013; Imtiyaz et al., 2014; Kargarfard et al., 2016). The details pushing, swing, grasping, vibrating and plucking. According to
of the process of identifying articles from initially searching to the size of the muscle area, the massage intervention duration
inclusion are shown in Figure 1. ranged from 6 to 30 min, but the duration per area approximately

Frontiers in Physiology | www.frontiersin.org 68 September 2017 | Volume 8 | Article 747

 TABLE 1 | Characteristics of included studies.

No Article, year Country/ Subjects’ characteristic, sample size Induce DOMS Intervention Duration Outcomes Time point
Guo et al.

Region

1 Weber et al., 1994 USA 20 healthy female untrained volunteers (G1 = Performed eccentric exercise of G1: no intervention, G2:massage, 8 min Muscle soreness 0, 24, 48 h after
10, G2 = 10). Mean age(SD): G1 = 22.3(4) nondominant elbow flexion until light effleurage for 2 min, petrissage rating, maximal DOMS induced
years, G2 = 25.9(4.5) years fatigued for 5 min, effleurage for 1 min. isometric force, exercise
peak torque at
60◦ /s
2 Lightfoot et al., 1997 USA 21 normally active college-age volunteers [G1 4 sets of 15 repetitions of eccentric G1: no intervention, G2: 10 min of 10 min Muscle soreness 0, 24, 48 h after
= 11 (3 men and 8 women), G2 = 10 (6 men exercise (heel-drop) to induce calf petrissage on calf muscle twice rating, creatine DOMS induced
and 4 women)]. Mean age(SD): G1 = muscle DOMS immediately after eccentric kinase level exercise

Frontiers in Physiology | www.frontiersin.org

23.9(6.4) years, G2 = 26.9(5.6) years exercise, and again at 24 h
post-exercise.
3 Hilbert et al., 2003 USA 18 male and female volunteers (G1 = 9, G2 6 sets of 10 maximal eccentric G1:rest for 20 min while listening to 20 min Muscle soreness 0, 2, 6, 24, 48 h
= 9).mean age (SD) for all the subjects is contractions with hamstring (1 min the same audiotape heard by the rating, peak after DOMS
20.4 (1.0) years of rest between sets), followed by 5 massage group on placebo lotion. eccentric torque induced exercise
more maximal eccentric G2:20 min of massage classical
contractions. Swedish techniques
4 Mancinelli et al., 2006 USA 22 female basketball and volleyball players. Intense strength training and drills G1: no intervention. G2: Western 17 min Muscle soreness 0 h after treatment
G1 = 11, G2 = 11. Mean age (SD) for all massage techniques of effleurage, rating (VAS and (48 h after DOMS
subjects is 20(0.93) years petrissage and vibration were used PPT) induced training)
for 17 min on 48 h after DOMS
induced training.
5 Yang, 2007 China 16 male trained athletic volunteers (G1 = 8, 10 sets of 30 m leapfrog with 1–2 G1: no intervention. G2: Chinese 15 min Muscle soreness 0, 24 h after

69
G2 = 8), mean age (SD) for all subjects is min rest between sets. traditional massage (including rating, creatine DOMS induced
19.2(0.96) years pushing, grasping, vibrating, and kinase level exercise
plucking) for 15 min.
6 Frey Law et al., 2008 USA 27 healthy individuals participanted in this 3 sets of eccentric wrist extensor G1: had a thin layer of massage 6 min Muscle soreness 24–48 h after
study as volunteers, G1 = 11, G2 = 16.mean contractions using a 10 lb hand cream applied but no massage rating (VAS and DOMS induced
age (SD) for all subjects is 23.3 (3.5) years, weight with 1–2 min between sets. received. G2: received a PPT), Peak Torque exercise
range of 19–41 years deep-tissue massage of the
fore-arm (including effleurage for 2
min and petrissage for 4 min)
7 Xiong et al., 2009 China 20 healthy male untrained students (G1 = 10, 2 sets of eccentric nondominant G1: no intervention. G2: received 30 min for Muscle soreness 0, 24, 48, 72 h
G2 = 10). Mean age (SD): G1 = 23.40(2.88) elbow flexion with 60% maximal Chinese traditional massage every day rating, creatine after DOMS
years, G2 = 23.20(2.90) years. isometric force. 25 repetitions for (including pushing, swing, grasping, kinase level induced exercise
each set with 5 min rest between vibrating and plucking) for 30 min
sets. each day with 3 days post-exercise.
8 Hu et al., 2010 China 15 healthy male students (G1= 8, G2 = 7). 7 sets of 50 m leapfrog with 2–3 G1: no intervention. G2: Chinese 20 min for Mucle soreness 0, 2, 24, 48, 72,
Mean age (SD): G1 = 22.13(3.23) years, G2 min rest between sets. traditional massage (including every day rating, creatine 96 h after DOMS
= 22.14(3.08) years. pushing, grasping, vibrating and kinase level, induced exercise
plucking) for 20 min each day with 5 maximal isometric
days post-exercise. force, Peak Torque
at 120◦ /s

(Continued)
Massage Alleviates DOMS

September 2017 | Volume 8 | Article 747

Guo et al. Massage Alleviates DOMS

similar between the trials (i.e., 6 min massage for wrist muscles,

0 h after treatment
(48 h after DOMS

24, 48, 72 h after

after intervention
induced training)
30 min massage for lower extremity muscles).

DOMS induced

0, 24, 48, 72 h
Time point

Control Conditions

exercise
Most (eight articles) of the studies used no intervention (just
sitting for rest). The participants in one study listened to

maximal isometric

maximal isometric
audiotapes while resting, while one study had a thin layer of
Mucle soreness

Mucle soreness

Mucle soreness
rating (VAS and

rating, creatine

rating, creatine
massage cream applied but no massage received. Another study

kinase level,

kinase level,
Outcomes

performed unilateral shoulder shrugs with Thera-Band elastic

tubing.

force

force
PPT)

Risk of Bias of Included Studies

Duration

Every included study was assessed for the risk of bias according to
10 min

15 min

30 min
instructions by Higgins and Green (2011); Higgins et al. (2011).
As shown in Figure 2, all of the 11 articles used a randomization
tubing. G2: 10 min massage on the

petrissage and vibration were used

shrugs with the Thera-Band elastic

massage techniques of effleurage,

method, but none of these reported any information about

petrissage and effleurage) on 48 h
G1: performed unilateral shoulder

G1: no intervention. G2: received

G1: no intervention. G2: Western

therapeutic massage for 15 min

allocation concealment. None of the trials met the requirements

after DOMS induced training.
trapezius muscle (including

for the blinding of participants. However, it seems unfeasible to

use the blinding method in view of the massage intervention.
Only two studies (18.2%) masked their outcome assessors, which
increased the risk of detection bias. Seven studies (63.6%) showed
Intervention

low risk of incomplete outcome bias, and the other four were
for 30 min

unclear. Nine studies (81.8%) showed a low risk bias of selective

reporting, whereas it was unclear if the studies have additional
bias. Funnel plot asymmetry did not show any publication bias
75–77% 1RM until exhaustion, with
30 healthy female non athletic subjects, G1 = 30 repetitions of eccentric exercise
Mean age (SD) for all subjects is 32(11) years. maximal eccentric contractions of

of elbow flexion using a dumbbell

Performed squats or leg press to

of muscle soreness rating, MIF, or the serum levels of CK

upper trapezius, with 1 min rest

90◦ knee flexion for five sets at

10 sets of 10–15 repetitions of

(Supplemental Figures S1–S3).

with weight of 80% maximal

1 min rest between sets

Effects of Massage on Muscle Soreness

Rating
isometric force.
Induce DOMS

between sets.

Four data points (72 participants) reported a muscle soreness

rating immediately after strenuous exercise. No significant
differences were found when the massage and control groups
were compared based on a fixed model (SMD –0.03, 95%CI
0.5–0.44, P = 0.90, I2 = 0%) (Figure 3A). Twenty-three data
15. Mean age (SD): G1 = 28.07 (3.33) years,
experience in bodybuilding, G1 = 15, G2 =
15, G2 = 15. Mean age (SD): G1 = 19.66
20 female volunteers, G1 = 10, G2 = 10.

points (504 participants) showed that a muscle soreness rating

Country/ Subjects’ characteristic, sample size

(1.01) years, G2 = 20.33 (0.97) years.

30 healthy males with at least 2 years

significantly decreased after the participants received massage

intervention when compared with no intervention (SMD –1.16,
95%CI –0.72 to –1.60, P < 0.001, I2 = 79%) (Figure 3B).
Among these, the muscle soreness rating presented a significant
G2 = 29.47 (3.72) years.

decrease compared to the control group when the participants

received massage intervention in subgroups of 24 h (SMD –0.61,
95%CI –1.17 to –0.05, P = 0.03, I2 = 67%), 48 h (SMD –
1.51, 95%CI –2.24 to –0.77, P < 0.0001, I2 = 82%) and 72 h
VAS, Visual analog scale; PPT, pressure pain threshold.

(SMD –1.46, 95%CI –2.59 to –0.33, P = 0.01, I2 = 82%)

post-exercise, respectively (Figure 3B). Based on these results,
massage intervention after exercise showed higher efficacy at 48
Canada
Region

and 72 h after exercise than at 24 h.

India

Iran

Sensitivity analysis revealed that outcomes of the total effect of

muscle soreness rating were stable when trials were removed one
Kargarfard et al., 2016
Andersen et al., 2013

by one.
Imtiyaz et al., 2014
TABLE 1 | Continued

Effects of Massage on MIF

No Article, year

Three data points with 70 participants compared MIF between

massage and control groups immediately after intense exercise.
No significant difference was observed based on the fixed
10

11

model meta-analysis (SMD 0.13, 95%CI –0.34 to 0.60, P =

9

Frontiers in Physiology | www.frontiersin.org 70 September 2017 | Volume 8 | Article 747

Guo et al. Massage Alleviates DOMS

0.73, P = 0.16, I2 = 0%) post-exercise based on each subgroup

meta-analysis.
Sensitivity analysis revealed that outcomes of the total effect
of MIF after strenuous exercise were stable when removing trials
one by one.

Effects of Massage on Peak Torque

Two data points with 38 participants compared peak torque
between massage and control groups immediately after exercise.
No significant difference was observed based on the fixed model
meta-analysis (SMD –0.26, 95%CI –0.89 to 0.38, P = 0.43, I2 =
0%) (Figure 5A). Eight data points (148 participants) as a total
effect showed that massage intervention significantly increased
peak torque after strenuous exercise when compared to the
control group (SMD 0.38, 95%CI 0.04–0.71, P = 0.03, I2 = 24%)
though no significant difference was found in the subgroups of
24 or 48 h (Figure 5B).
Sensitivity analysis demonstrated that the outcome of peak
torque after strenuous exercise was reversed when one of the
trials was removed (Hu et al., 2010).

Effects of Massage on Serum CK Level

Five data points with 102 participants reported the outcome
of serum CK level comparing massage and control groups
immediately after exercise. No significant difference was found
between the two groups (SMD 0.12, 95%CI –0.31 to 0.56, P =
0.58, I2 = 0%) based on fixed meta-analysis (Figure 6A). Seven
articles including twelve data points (including 247 participants)
showed that the serum CK level significantly decreased in the
massage intervention group when compared to the control group
(SMD –0.64, 95%CI –0.25 to –1.04, P = 0.001, I2 = 54%) as a
total effect (Figure 6B). Among these, the subgroups of 48 and
72 h showed massage intervention could significantly decrease
the serum CK level at 48 h (SMD –0.63, 95%CI –1.25 to –0.01,
P = 0.05, I2 = 60%) and 72 h (SMD –1.69, 95%CI –3.27 to –
0.11, P = 0.04, I2 = 66%) post-strenuous exercise. However, no
significant difference was observed in the 24 h subgroup (SMD
–0.33, 95%CI –0.75 to 0.09, P = 0.12, I2 = 9%).
Sensitivity analysis showed that outcomes of the total effects
of the serum CK levels were stable when trials were removed one
by one.

DISCUSSION
FIGURE 2 | Risk of bias summary of included studies.
Massage therapy intervention is often used for reducing muscle
pain or soreness (Field, 2016), and increasing post-exercise
muscle performance (Buttagat et al., 2016; Best and Crawford,
0.60, I2 = 0%) (Figure 4A). However, 16 data points (including 2017). Several lines of evidence contributed to explaining the
265 participants) demonstrated that massage intervention could mechanisms of massage therapy on DOMS: (a) modulation of
significantly enhance the performance of MIF compared to the activity of the parasympathetic nervous system (Weerapong
controls (SMD 0.56, 95%CI 0.21–0.00, P = 0.002, I2 = 46%) et al., 2005), (b) increase in blood and lymphatic flow to
(Figure 4B). Among these, only 72 h post-exercise subgroups rapidly clear the biochemical markers of muscle damage [e.g.,
showed significant positive effects of massage intervention on CK and lactate dehydrogenase (LDH); Bakar et al., 2015], (c)
MIF (SMD 1.11, 95%CI 0.12–2.1, P = 0.03, I2 = 78%). No psychophysiological response also plays an essential role in
significant difference of MIF was observed between massage reducing pain (Arroyo-Morales et al., 2011).
and control groups either at 24 h (SMD 0.34, 95%CI –0.09 to Best et al. reviewed the effectiveness of massage on muscle
0.77, P = 0.12, I2 = 0%) or 48 h (SMD 0.31, 95%CI –0.12 to recovery including 27 studies and 440 participants, showing

Frontiers in Physiology | www.frontiersin.org 71 September 2017 | Volume 8 | Article 747

Guo et al. Massage Alleviates DOMS

FIGURE 3 | Meta-analysis of effects of massage intervention on muscle soreness rating. (A) The time point immediately. (B) Time points after received massage
intervention combined 24, 48, and 72 h after exercise.

little effect of massage on muscle recovery after intense exercise the data could not be extracted for meta-analysis since the
(Best et al., 2008). However, only six RCTs investigated DOMS published papers did not report the mean and SD. Thus, Best
and muscle function post-exercise, and in some of the studies, et al.’s review did not pool the extracted data in a meta-analysis.

Frontiers in Physiology | www.frontiersin.org 72 September 2017 | Volume 8 | Article 747

Guo et al. Massage Alleviates DOMS

FIGURE 4 | Meta-analysis of effects of massage intervention on MIF. (A) The time point immediately. (B) Time points after received massage intervention combined
24, 48, and 72 h after exercise. MIF, maximal isometric force.

Torres et al. showed that massage could alleviate muscle soreness In addition, it is well known that the evidence from RCTs will be
24 h after intense exercise, but few trials were pooled in the much better than from case studies. Therefore, this review pooled
meta-analysis (3 trials) (Torres et al., 2012). Other reviews also only RCTs in the meta-analysis with a total number of 23 data
reported that self-myofascial release (a type of self-massage points (including 504 participants), to be better able to evaluate
using a foam roller) could alleviate muscle pain and enhance the effectiveness of massage on DOMS and muscle performance.
muscle performance after strenuous exercise (Beardsley and This systematic review and meta-analysis showed that
Skarabot, 2015; Cheatham et al., 2015), but no trials were pooled the participants receiving massage intervention post-strenuous
into a meta-analysis to assess the effectiveness. Consistent with exercise experienced a reduction in muscle soreness rating
previous reviews, the present study confirmed that massage was as a total effect. Additionally, our findings demonstrated that
an effective intervention for reducing DOMS after strenuous the SMD of 48 and 72 h subgroups were –1.51 and –1.46,
exercise with the outcomes of muscle soreness rating, muscle respectively. This result was greater than that of 24 h subgroups
performance (MIF and Peak Torque) and the serum CK level. (–0.61), indicating that alleviating muscle pain from massage

Frontiers in Physiology | www.frontiersin.org 73 September 2017 | Volume 8 | Article 747

Guo et al. Massage Alleviates DOMS

FIGURE 5 | Meta-analysis of effects of massage intervention on peak torque. (A) The time point immediately. (B) Time points after received massage intervention
combined 24, 48, and 72 h after exercise.

intervention would be the more efficacious at 48 and 72 h post- not different from those of the trials using Western massage.
exercise compared to 24 h post-exercise. Moreover, the stability In addition, two studies (Xiong et al., 2009; Hu et al., 2010)
of the total effect sensitivity analysis also strengthened the results. used daily massage and one study (Mancinelli et al., 2006)
Previous systematic reviews demonstrated that there was little applied massage 48 h post-exercise, however the effects of these
evidence supporting the use of massage to enhance muscle single trials were not different from the others. Conversely, the
performance after strenuous exercise (Best et al., 2008; Torres effectiveness of single trials is not enough to show significant
et al., 2012). Poppendieck et al. observed that post-exercise benefits of massage intervention on muscle performance post-
manual massage increased recovery of sprint performance strenuous exercise. Additional trials and participants are needed.
(Poppendieck et al., 2016). However, many articles included in Furthermore, contrary to Poppendieck et al.’s review, which
this meta-analysis are not RCTs. in contrast to Best et al.’s and included studies using endurance-type exercise, most of the
Torres et al.’s reviews, the current evidence showed that massage studies included in this review used eccentric strength exercise
intervention increased MIF and peak torque after exercise as to induce DOMS, which may also contribute to the difference
the total effects. These results may reflect more trials included between this meta-analysis and that of previous reviews.
that had positive effects of post-exercise massage on muscle The serum CK level was frequently considered a marker
performance. Three studies (Yang, 2007; Xiong et al., 2009; Hu of inflammation and skeletal muscle damage influencing the
et al., 2010) using Chinese traditional massage were pooled recovery of muscle performance (Clarkson and Sayers, 1999;
in this meta-analysis. The single effects of those trials were Romagnoli et al., 2015). Our findings showed that massage

Frontiers in Physiology | www.frontiersin.org 74 September 2017 | Volume 8 | Article 747

Guo et al. Massage Alleviates DOMS

FIGURE 6 | Meta-analysis of effects of massage intervention on the serum CK level. (A) The time point immediately. (B) Time points after received massage
intervention combined 24, 48, and 72 h after exercise. CK, creatine kinase.

intervention decreased the serum CK levels as a total effect, alleviating DOMS. Compared with previous studies, the present
suggesting that massage decreased inflammation and muscle study included three articles using Chinese traditional massage
damage and promoted muscle performance recovery. This result intervention and more trials were pooled into the meta-analysis
is consistent with the results of muscle soreness rating, MIF, and than in previous systematic reviews. This strategy implied greater
peak torque, supporting evidence of the positive physiological evidence to evaluate the effectiveness of massage intervention
effects of massage therapy on DOMS. Fast clearance of serum CK after exercise. Furthermore, current evidence suggests that
level from the circulatory system was thought to be the reason massage is not only effective in reducing muscle pain after intense
that massage promotes muscle recovery and performance. exercise, but also in increasing muscle performance and reducing
the serum CK level.
Strength and Limitations This review searched a wide variety of database including two
This study is the first meta-analysis that combines both Chinese Chinese electronic databases for relevant articles and included
and Western massage at different time points after strenuous the trials performing Chinese traditional massage that have not
exercise, assessing whether massage intervention was effective in previously been reviewed. Two authors independently searched

Frontiers in Physiology | www.frontiersin.org 75 September 2017 | Volume 8 | Article 747

Guo et al. Massage Alleviates DOMS

and selected the included studies, extracted the data, and assessed a useful and practical therapy for exercise participants or athletes.
the risk of bias of every trial using recommended protocols and Nevertheless, it is necessary to be cautious about the results
methodological schemes. Therefore, the results of this meta- in view of the limitations outlined in the present study. More
analysis are considered a significant contribution. RCTs with large sample sizes are needed for better understanding
However, this meta-analysis has several limitations. First, the the effectiveness of massage intervention on DOMS and muscle
quality of all the included trials in this meta-analysis was low. performance.
None of the studies detailed allocation concealment although
all the trials used a randomization method. None of the studies AUTHOR CONTRIBUTIONS
met the blinding of participants though it seems unfeasible to
use the blinding method in view of the massage intervention. XC and JZ designed the systematic review and supervised the
This factor increased the selection bias and performance bias. entire program; JG and LL reviewed all the studies and extracted
Only two articles (18.2%) masked their outcome assessors, which the information from the eligible trials; YG analyzed the data and
increased the risk of detection bias. Second, the trials varied in prepared the figures and table; XC, JG, and LL wrote the paper;
methodological design (i.e., exercise type, control intervention), XC, RZ, and JX revised the manuscript. All authors reviewed and
massage type, and conditions (i.e., duration). Thus, the outcomes approved the manuscript.
of muscle soreness have high heterogeneity although the total
effects of outcomes were stable. Sensitivity analysis implied that FUNDING
the outcome of peak torque after strenuous exercise was unstable;
therefore, the results should be considered with caution. Third, This work was supported by University student’s Innovation
there may be some publication bias as unpublished articles Project of Zhejiang province (Xinmiao project) Grant No.
could not be searched in this review although the funnel plot 2014R413034 and Wenzhou public welfare science and
asymmetry did not show the bias. Finally, the number of trials technology plan project (Grant No. Y20170225) and the
and participants was relatively small; therefore, larger sample National Natural Science Foundation of China (Grant No.
sizes in future studies are needed to better understand the effects 81702235). The funders had no role in the study design, data
of massage intervention on DOMS and muscle performance. collection and analysis, decision to publish, or preparation of the
manuscript.
CONCLUSION
SUPPLEMENTARY MATERIAL
This systematic review and meta-analysis demonstrated that
massage intervention could be effective for alleviating DOMS, as The Supplementary Material for this article can be found
well as increasing muscle performance after strenuous exercise. online at: http://journal.frontiersin.org/article/10.3389/fphys.
The highest efficacy was achieved at 48 h post-exercise. Massage is 2017.00747/full#supplementary-material

REFERENCES Buttagat, V., Narktro, T., Onsrira, K., and Pobsamai, C. (2016). Short-term effects
of traditional Thai massage on electromyogram, muscle tension and pain
Andersen, L. L., Jay, K., Andersen, C. H., Jakobsen, M. D., Sundstrup, E., Topp, among patients with upper back pain associated with myofascial trigger points.
R., et al. (2013). Acute effects of massage or active exercise in relieving muscle Complement. Ther. Med. 28, 8–12. doi: 10.1016/j.ctim.2016.07.004
soreness: randomized controlled trial. J. Strength Cond. Res. 27, 3352–3359. Chatzinikolaou, A., Fatouros, I. G., Gourgoulis, V., Avloniti, A., Jamurtas, A. Z.,
doi: 10.1519/JSC.0b013e3182908610 Nikolaidis, M. G., et al. (2010). Time course of changes in performance and
Arroyo-Morales, M., Fernandez-Lao, C., Ariza-Garcia, A., Toro-Velasco, C., inflammatory responses after acute plyometric exercise. J. Strength Cond. Res.
Winters, M., Diaz-Rodriguez, L., et al. (2011). Psychophysiological effects of 24, 1389–1398. doi: 10.1519/JSC.0b013e3181d1d318
preperformance massage before isokinetic exercise. J. Strength Cond. Res. 25, Cheatham, S. W., Kolber, M. J., Cain, M., and Lee, M. (2015). The effects of self-
481–488. doi: 10.1519/JSC.0b013e3181e83a47 myofascial release using a foam roll or roller massager on joint range of motion,
Bakar, Y., Coknaz, H., Karli, U., Semsek, O., Serin, E., and Pala, O. O. (2015). muscle recovery, and performance: a systematic review. Int. J. Sports Phys. Ther.
Effect of manual lymph drainage on removal of blood lactate after submaximal 10, 827–838.
exercise. J. Phys. Ther. Sci. 27, 3387–3391. doi: 10.1589/jpts.27.3387 Cheung, K., Hume, P., and Maxwell, L. (2003). Delayed onset muscle soreness:
Beardsley, C., and Skarabot, J. (2015). Effects of self-myofascial release: a systematic treatment strategies and performance factors. Sports Med. 33, 145–164.
review. J. Bodyw. Mov. Ther. 19, 747–758. doi: 10.1016/j.jbmt.2015.08.007 doi: 10.2165/00007256-200333020-00005
Best, T. M., and Crawford, S. K. (2017). Massage and postexercise Clarkson, P. M., and Sayers, S. P. (1999). Etiology of exercise-induced muscle
recovery: the science is emerging. Br. J. Sports Med. 51, 1386–1387. damage. Can. J. Appl. Physiol. 24, 234–248. doi: 10.1139/h99-020
doi: 10.1136/bjsports-2016-096528 Costello, J. T., Algar, L. A., and Donnelly, A. E. (2012). Effects of whole-body
Best, T. M., Hunter, R., Wilcox, A., and Haq, F. (2008). Effectiveness of sports cryotherapy (-110 degrees C) on proprioception and indices of muscle damage.
massage for recovery of skeletal muscle from strenuous exercise. Clin. J. Sport Scand. J. Med. Sci. Sports 22, 190–198. doi: 10.1111/j.1600-0838.2011.01
Med. 18, 446–460. doi: 10.1097/JSM.0b013e31818837a1 292.x
Bleakley, C., McDonough, S., Gardner, E., Baxter, G. D., Hopkins, J. T., and De Marchi, T., Schmitt, V. M., da Silva Fabro, D. C., da Silva, L. L., Sene, J., Tairova,
Davison, G. W. (2012). Cold-water immersion (cryotherapy) for preventing O., et al. (2017). Phototherapy for improvement of performance and exercise
and treating muscle soreness after exercise. Cochrane Database Syst. Rev. recovery: comparison of 3 commercially available devices. J. Athl. Train. 52,
2:CD008262. doi: 10.1002/14651858.CD008262.pub2 429–438. doi: 10.4085/1062-6050-52.2.09

Frontiers in Physiology | www.frontiersin.org 76 September 2017 | Volume 8 | Article 747

Guo et al. Massage Alleviates DOMS

Ernst, E. (1998). Does post-exercise massage treatment reduce delayed onset Mancinelli, C. A., Davis, D. S., Aboulhosn, L., Brady, M., Eisenhofer, J., and
muscle soreness? A systematic review. Br. J. Sports Med. 32, 212–214. Foutty, S. (2006). The effects of massage on delayed onset muscle soreness and
doi: 10.1136/bjsm.32.3.212 physical performance in female collegiate athletes. Phys. Ther. Sport 7, 5–13.
Field, T. (2016). Massage therapy research review. Complement. Ther. Clin. Pract. doi: 10.1016/j.ptsp.2005.10.004
24, 19–31. doi: 10.1016/j.ctcp.2016.04.005 Morelli, M., Chapman, C. E., and Sullivan, S. J. (1999). Do cutaneous receptors
Frey Law, L. A., Evans, S., Knudtson, J., Nus, S., Scholl, K., and Sluka, contribute to the changes in the amplitude of the H-reflex during massage?
K. A. (2008). Massage reduces pain perception and hyperalgesia in Electromyogr. Clin. Neurophysiol. 39, 441–447.
experimental muscle pain: a randomized, controlled trial. J. Pain 9, 714–721. Peake, J. M., Suzuki, K., Hordern, M., Wilson, G., Nosaka, K., and Coombes, J. S.
doi: 10.1016/j.jpain.2008.03.009 (2005a). Plasma cytokine changes in relation to exercise intensity and muscle
Hemmings, B., Smith, M., Graydon, J., and Dyson, R. (2000). Effects of massage on damage. Eur. J. Appl. Physiol. 95, 514–521. doi: 10.1007/s00421-005-0035-2
physiological restoration, perceived recovery, and repeated sports performance. Peake, J. M., Suzuki, K., Wilson, G., Hordern, M., Nosaka, K., Mackinnon,
Br. J. Sports Med. 34, 109–114. discussion: 115. doi: 10.1136/bjsm.34.2.109 L., et al. (2005b). Exercise-induced muscle damage, plasma cytokines, and
Higgins, J. P., Altman, D. G., Gotzsche, P. C., Juni, P., Moher, D., Oxman, A. D., markers of neutrophil activation. Med. Sci. Sports Exerc. 37, 737–745.
et al. (2011). The Cochrane Collaboration’s tool for assessing risk of bias in doi: 10.1249/01.MSS.0000161804.05399.3B
randomised trials. BMJ 343:d5928. doi: 10.1136/bmj.d5928 Poppendieck, W., Wegmann, M., Ferrauti, A., Kellmann, M., Pfeiffer, M., and
Higgins, J. P., Thompson, S. G., Deeks, J. J., and Altman, D. G. Meyer, T. (2016). Massage and performance recovery: a meta-analytical review.
(2003). Measuring inconsistency in meta-analyses. BMJ 327, 557–560. Sports Med. 46, 183–204. doi: 10.1007/s40279-015-0420-x
doi: 10.1136/bmj.327.7414.557 Romagnoli, M., Sanchis-Gomar, F., Alis, R., Risso-Ballester, J., Bosio, A., Graziani,
Higgins, J. P. T., and Green, S. (2011). Cochrane Handbook for Systematic Reviews R. L., et al. (2015). Changes in muscle damage, inflammation, and fatigue-
of Interventions. The Cochrane Collaboration. Available online at: http:// related parameters in young elite soccer players after a match. J. Sports Med.
handbook.cochrane.org Phys. Fitness. 56, 1198–1205.
Hilbert, J. E., Sforzo, G. A., and Swensen, T. (2003). The effects of Torres, R., Ribeiro, F., Alberto Duarte, J., and Cabri, J. M. (2012). Evidence of the
massage on delayed onset muscle soreness. Br. J. Sports Med. 37, 72–75. physiotherapeutic interventions used currently after exercise-induced muscle
doi: 10.1136/bjsm.37.1.72 damage: systematic review and meta-analysis. Phys. Ther. Sport 13, 101–114.
Hill, J., Howatson, G., van Someren, K., Leeder, J., and Pedlar, C. doi: 10.1016/j.ptsp.2011.07.005
(2014). Compression garments and recovery from exercise-induced Wang, X. Q., Pi, Y. L., Chen, P. J., Liu, Y., Wang, R., Li, X., et al. (2016).
muscle damage: a meta-analysis. Br. J. Sports Med. 48, 1340–1346. Traditional chinese exercise for cardiovascular diseases: systematic review and
doi: 10.1136/bjsports-2013-092456 meta-analysis of randomized controlled trials. J. Am. Heart. Assoc. 5:e002562.
Howatson, G., Goodall, S., and van Someren, K. A. (2009). The influence of cold doi: 10.1161/JAHA.115.002562
water immersions on adaptation following a single bout of damaging exercise. Weber, M. D., Servedio, F. J., and Woodall, W. R. (1994). The effects of three
Eur. J. Appl. Physiol. 105, 615–621. doi: 10.1007/s00421-008-0941-1 modalities on delayed onset muscle soreness. J. Orthop. Sports Phys. Ther. 20,
Howatson, G., and van Someren, K. A. (2008). The prevention and 236–242. doi: 10.2519/jospt.1994.20.5.236
treatment of exercise-induced muscle damage. Sports Med. 38, 483–503. Weerapong, P., Hume, P. A., and Kolt, G. S. (2005). The mechanisms of massage
doi: 10.2165/00007256-200838060-00004 and effects on performance, muscle recovery and injury prevention. Sports Med.
Hu, J. C., Ma, L. H., Shi, Q. Z., Zhang, F., Tnag, H., Wang, C. X., et al. (2010). 35, 235–256. doi: 10.2165/00007256-200535030-00004
The effects of different massage interventions on delayed onset muscle soreness. Xiong, Y., Wu, Y. C., Jin, H. Z., and Gu, Y. H. (2009). The effect of massage on
Chin. J. Rehabil. Med. 3, 265–267. delayed onset muscle soreness after eccentric exercise: randomized controlled
Imtiyaz, S., Veqar, Z., and Shareef, M. Y. (2014). To compare the effect of vibration trial. Chin. J. Orthopa Trauma 9, 669–673.
therapy and massage in prevention of delayed onset muscle soreness (DOMS). Yang, Z. X. (2007). Comparing three relax method for reducing delayed onset of
J. Clin. Diagn. Res. 8, 133–136. doi: 10.7860/JCDR/2014/7294.3971 mucle soreness. J. Guangzhou Sport Uni. 4, 119–121.26.
Kargarfard, M., Lam, E. T., Shariat, A., Shaw, I., Shaw, B. S., and Tamrin,
S. B. (2016). Efficacy of massage on muscle soreness, perceived Conflict of Interest Statement: The authors declare that the research was
recovery, physiological restoration and physical performance in male conducted in the absence of any commercial or financial relationships that could
bodybuilders. J. Sports Sci. 34, 959–965. doi: 10.1080/02640414.2015.10 be construed as a potential conflict of interest.
81264
Lavender, A. P., and Nosaka, K. (2006). Changes in fluctuation of isometric force Copyright © 2017 Guo, Li, Gong, Zhu, Xu, Zou and Chen. This is an open-access
following eccentric and concentric exercise of the elbow flexors. Eur. J. Appl. article distributed under the terms of the Creative Commons Attribution License (CC
Physiol. 96, 235–240. doi: 10.1007/s00421-005-0069-5 BY). The use, distribution or reproduction in other forums is permitted, provided the
Lightfoot, J. T., Char, D., McDermott, J., and Goya, C. (1997). Immediate original author(s) or licensor are credited and that the original publication in this
postexercise massage does not attenuate delayed onset muscle soreness. journal is cited, in accordance with accepted academic practice. No use, distribution
J. Strength Cond. Res. 11, 119–124. or reproduction is permitted which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 77 September 2017 | Volume 8 | Article 747

 ORIGINAL RESEARCH
published: 22 September 2017
doi: 10.3389/fphys.2017.00690

Upper-Body Muscular Endurance

Training Improves Performance
Following 50 min of Double Poling in
Well-Trained Cross-Country Skiers
Jørgen Børve † , Steffen N. Jevne † , Bjarne Rud and Thomas Losnegard*

Department of Physical Performance, Norwegian School of Sport Sciences, Oslo, Norway

This study investigated the effect of muscular endurance training on O2 -cost and
performance in double poling (DP) on a rollerski treadmill. Twenty-two well-trained
cross-country skiers (31 ± 4 years, 77 ± 9 kg, 181 ± 8 cm, VO2max running:
64 ± 5 mL·kg−1 ·min−1 ) were counter-balanced to either a combined muscular
endurance and running interval training group [MET; n = 11 (♂ = 9, ♀ = 2)], or an
endurance running interval training group [ET; n = 11 (♂ = 9, ♀ = 2)]. Both groups
Edited by:
Thomas Leonhard Stöggl, continued their normal low-and moderate intensity training, but replaced 2 weekly high
University of Salzburg, Austria intensity-training sessions with two project-specific sessions for 6 weeks. In these
Reviewed by: sessions, MET combined upper-body muscular endurance training (4 × 30 repetitions,
Juha Oksa,
Finnish Institute of Occupational
90 s rest between sets) and running intervals (3 × 4 or 2 × 6 min, 3 min rest),
Health, Finland while ET performed running intervals only (6 × 4 or 4 × 6 min, 3 min rest). The DP
Boye Welde,
test-protocol consisted of 50 min submaximal poling for O2 -cost measurement, followed
UiT The Arctic University of Norway,
Norway by a self-paced 1,000-m performance test. In addition, subjects performed a VO2max
*Correspondence: test in running. MET increased muscular endurance (P < 0.05) and 1RM in simulated DP
Thomas Losnegard (P < 0.01) more than ET. Further, MET reduced the 1,000-m time and O2 -cost compared
[email protected]
†
to baseline values (P < 0.05), and tended to improve the 1,000-m time more than ET
These authors have contributed
equally to this work.
(P = 0.06). There were no changes in VO2max running or VO2peak DP in either MET
or ET. In conclusion, 6 weeks of muscular endurance training increased both muscular
Specialty section: endurance and 1RM in simulated DP. Further, specific upper-body muscular endurance
This article was submitted to
Exercise Physiology,
training improved DP performance and thus, seems as a promising training model to
a section of the journal optimize performance in well-trained cross-country skiers.
Frontiers in Physiology
Keywords: cross-country skiing, high-intensity training, maximal oxygen uptake, O2 -cost, running, training
Received: 23 April 2017 intensity
Accepted: 28 August 2017
Published: 22 September 2017

Citation: INTRODUCTION
Børve J, Jevne SN, Rud B and
Losnegard T (2017) Upper-Body In cross-country (XC) skiing, the classic style double poling (DP) technique has been considerably
Muscular Endurance Training
developed over the last decade and is today the main technique used in races. The DP technique is
Improves Performance Following 50
min of Double Poling in Well-Trained
characterized by a symmetrical DP action, which transfers propulsive forces solely through the
Cross-Country Skiers. poles. This emphasizes the importance of well-developed upper-body power in employing DP
Front. Physiol. 8:690. successfully throughout an entire race (Stöggl et al., 2007; Losnegard et al., 2011). Consequently,
doi: 10.3389/fphys.2017.00690 specific upper-body training, both in research and practical situations, has gained interest as a

Frontiers in Physiology | www.frontiersin.org 78 September 2017 | Volume 8 | Article 690

Børve et al. Upper-Body Training in XC Skiers

training model for improving such abilities (Nilsson et al., 2004; MATERIALS AND METHODS
Terzis et al., 2006; Losnegard et al., 2011; Skattebo et al., 2015).
The introduction of new competition formats such as sprint Subjects
and mass starts has also increased the importance of a high During the intervention, two men withdrew from the project
work-intensity during the closing part of races, which makes it due to unrelated reasons. In total, 22 well-trained XC skiers
vital to conserve power for the final sprint. This requirement (18 males and 4 females) completed the study with the
could potentially be met by working at a lower relative intensity required number of completed project-specific training sessions
during the submaximal part of the competition, resulting in less (minimum adherence 85%; MET = 94 ± 8% adherence; ET = 90
fatigued muscles during the final sprint (Bassett and Howley, ± 5% adherence). Inclusion criteria were completing the long-
2000). However, limited data are available on which type of distance XC ski race Birkebeinerrennet the previous year with a
training is most efficient in improving such abilities in XC skiing finishing time <4 h and < 3 h 30 min for females and males
generally and DP more specifically. respectively. None of the subjects performed specific upper-body
In previous research, the main training model for developing muscular endurance training systematically prior to the start
upper-body power to improve DP performance has been heavy of the study, but all performed weekly aerobic high-intensive
strength training (≤12 repetition maximum). However, these interval training. After completing the pre-test, participants
studies have yielded varying results, displaying both large (Hoff were counter-balanced to either a muscular endurance training
et al., 1999, 2002; Østeras et al., 2002) and trivial effects group (MET: 30 ± 4 years, 76 ± 8 kg, 180 ± 7 cm) or an
(Losnegard et al., 2011; Skattebo et al., 2015). Furthermore, the endurance training group (ET: 32 ± 4 years, 78 ± 11 kg, 182
metabolic response during DP seems to be different for the arms ± 10 cm) based on the following pre-test characteristics: DP
and legs, indicating that intense upper-body endurance training 1,000-m time, VO2max running, 1 repetition maximum (1RM)
may increase the arms’ ability to extract oxygen and thus enhance and gender. All skiers gave their written informed consent
DP performance (Rud et al., 2014). before participating. The study was conducted according to the
The effect of muscular endurance training (20–100 reps/set, Declaration of Helsinki and Norwegian law.
Campos et al., 2002) on endurance performance has been
investigated in various sports including running (Mikkola Design
et al., 2011; Sedano et al., 2013) and rowing (Ebben et al., The investigation was conducted during the pre-competition
2004; Gallagher et al., 2010). Such training seem to target period for XC skiers (August–November), and contained a 6-
different muscular and neurological adaptations compared week intervention period, enclosed by a pre- and post-test
to heavy strength training; e.g., mitochondria and capillary (Table 1). The week before pre-tests, subjects performed one
density, muscle fiber composition and cross-sectional area familiarization session on the rollerski treadmill which consisted
(Campos et al., 2002). Specifically for XC skiing, Nilsson of three different submaximal speeds, followed by a 1,000-m
et al. (2004) showed that 20-s or 180-s interval training performance test (see Prolonged double poling protocol). Then,
in a DP ergometer increased both 30-s and 6-min power a familiarization session with the muscular endurance exercise
output in well-trained XC skiers, while Vandbakk et al. (2017) in standing DP was performed, using the cable-pulley and
demonstrated increased time to exhaustion after 8 weeks consisted of a warm-up and 1RM test (2–3 attempts). Pre-
of 30-s DP intervals. This indicates that upper-body power and post-tests were conducted in the week immediately before
training might have the potential to increase performance in and after the training period and included 2 test days for each
XC skiing, although data on the effect on finishing abilities is participant, separated by at least 48 h. Subjects were not allowed
limited. to perform any strength training the day before testing, and
In terms of training, three variables (intensity, duration, and only a maximum of 90 min endurance training at low intensity.
frequency) together set the training load with the explicit goal Day 1 included a running VO2max test followed by an upper-
of maximizing performance (Seiler, 2010). When investigating body 1RM and a muscular endurance test in a cable-pulling
the effect of a specific training program for recreational apparatus. Day 2 included a submaximal and maximal test in
athletes, often with limited time to execute training, it seems DP on a rollerski treadmill. During the intervention period,
important to keep the three variables similar before and both groups replaced two of their weekly high-intensity interval
during interventions to determine the effect of the intervention training (HIT) sessions with a specific protocol based on their
itself. Otherwise, it could be unclear whether improved respective group. For MET, these sessions consisted of both
performance is a result of adaptations related to the added running intervals and upper-body muscular endurance training,
training, or changed training load in general. The aim of while for ET it involved two sessions with running intervals
the present study was, therefore, to investigate the effect of only. During the other weekly training sessions, the subjects were
replacing parts of high intensity interval training with muscular encouraged to maintain their normal training routines.
endurance training for 6 weeks on performance after completing
50 min of submaximal DP. The main hypotheses of the Running VO2max (Day 1)
present study were; Upper body muscular endurance training Following a 20-min warm up (60–85% of HRpeak ), subjects
would (I) improve performance following 50-min of double performed a VO2max running test at a 10.5% incline with stepwise
poling, and (II) reduce O2 -cost during submaximal double increments of 1 km-h−1 every minute until volitional exhaustion.
poling. Starting speed was set individually based on subjects’ race history

Frontiers in Physiology | www.frontiersin.org 79 September 2017 | Volume 8 | Article 690

Børve et al. Upper-Body Training in XC Skiers

TABLE 1 | The test-battery and time-line of the study.

Familiarization x
Test day 1:
VO2max running test x x
1RM- and muscular endurance test x x
Test day 2:
Double poling protocol x x

Pre-test Intervention Post-test

Week: −2 −1 1 2 3 4 5 6 7

VO2max , maximal oxygen uptake; RM, repetition maximum.

(average: 9.4 ± 0.7 km-h−1 ), but was similar for each subject at
pre- and post-test. All tests lasted 4.5–7.5 min, and the highest
VO2 over a 60 s period was considered as VO2max .

Maximal Strength and Muscular Endurance

(Day 1)
Fifteen minutes after the running VO2max test, the subjects
performed an exercise-specific warm-up in the standing DP,
consisting of four sets (15 reps at 40%, 10 reps at 55%, 5 reps at
75%, 3 reps at 85% of 1RM as estimated during familiarization).
In the 1RM test, the load was set to 95% of estimated 1RM (at
post-test, 95% of pre-test 1RM was used), and increased 2–5% FIGURE 1 | The standing double poling exercise used during testing and
training. The figure illustrate the poling phase of the cycle. The participant gave
after each successful attempt (3 min break) until the subject failed his consent to publish the picture.
on two consecutive lifts. The heaviest successful attempt was
considered 1RM. After a 5 min break, the muscular endurance
test was conducted at 55% of individual 1RM. This relative
workload was set based on pilot testing and previous studies, with (2013). In brief, a fixed individual speed, identical at pre- and
the aim of completing more than 20, but less than 100 repetitions post-test, was set for the initial 200 m to avoid over-pacing (avg.
(Stone and Coulter, 1994; Campos et al., 2002). The test was 3.41 ± 0.5 m·s−1 ). Thereafter, the subjects were free to adjust
performed using a constant DP motion (cycle time; pre 1.4 ± 0.2 their speed based on their position on the treadmill (0.25 m·s−1
s; post 1.4 ± 0.2 s, based on video recording) until exhaustion. increase or decrease). VO2 and HR were measured continuously,
All testing and training with the cable-pulley was done with a and the highest VO2 and HR (avg. 30 s) were considered peak
customized handlebar to simulate a DP grip (Figure 1). values. RPE was reported immediately after finishing the test.
During the prolonged exercise, participants were asked to drink
Prolonged Double Poling Protocol (Day 2) 3 dl of water.
The protocol was initiated by a 5 min trial identical to the
first submaximal speed to ensure a steady state oxygen uptake Apparatus
at the first submaximal load. Further, subjects performed three Oxygen consumption was measured with an automatic
different submaximal speeds (50 min total), followed directly ergospirometry system (Oxycon Pro, Jaeger Instrument,
by a 1,000-m test (Figure 2). All DP tests were conducted at Hoechberg, Germany), as evaluated by Foss and Hallen (2005).
2.5◦ . Submaximal speeds were based on the linear relationship The gas analysers and the flow turbine (Tripel V; Erick Jeager
between workload and O2 -cost, with the aim of matching speed GmbH, Hoechberg, Germany) were calibrated before each test
to 75% of VO2peak DP from 15 to 50 min and adjusted with according to the instruction manual, as described previously
the RPE (rate of perceived exertion, Borg, 1982) with the aim (Losnegard et al., 2011). The same gas analyser was used
of not exceeding 16 on the RPE-scale before the 1,000-m test. during the running VO2max test and DP protocol. Heart rate
Thus, speeds were individually set, but were identical for each was recorded using a Polar V800 (Polar Electro Oy, Kempele,
participant in the pre- and post-test. The average speed for all Finland). A rollerski treadmill with belt dimension 3 × 4.5 m
participants at submaximal speeds one, two and three was 2.3 ± (Rodby, Södertelje, Sweden), was used during the running
0.3, 2.7 ± 0.4 and 3.0 ± 0.5 m·s−1 , respectively. The O2 -cost was VO2max test and the prolonged DP test. Two different pairs
calculated from 8-10, 13-15, 18-20, 33-35 and 48-50 min during of rollerskis (Swenor Fiberglass, Trøsken, Norway), with front
the prolonged submaximal test. Heart rate (HR) was registered wheel type 2 and rear wheel type 3, were used depending on
after 10, 15, 20, 35, and 50 min and RPE after 20, 35, and 45 min. which ski binding system the skiers used (SNS, Salomon, Annecy,
The 1,000-m test was a modified protocol from Losnegard et al. France or NNN, Rottefella, Klokkarstua, Norway). The rollerski

Frontiers in Physiology | www.frontiersin.org 80 September 2017 | Volume 8 | Article 690

Børve et al. Upper-Body Training in XC Skiers

FIGURE 2 | Schematic illustration of prolonged double poling protocol. Steady-state VO2 was measured from 5 to 20, 31 to 35, and 46 to 50 min followed by a
self-paced 1,000-m maximal test. ↓ = start O2 measurements. ↑ = end O2 measurements. All tests conducted at 2.5◦ incline and the speed shown is the average
from both groups. Speed from 0 to 5 min was identical to 5 to 10 min.

TABLE 2 | Training programs for the 2 weekly project sessions in MET (Combined endurance training and muscular endurance training) and ET (endurance training).

Week MET ET

Running interval Muscular endurance Running interval

Duration (sets × min) Intensity (% HRpeak ) Sets Rep Workload (% 1 RM) Duration (sets × min) Intensity (% of HRpeak )

1–2 3 × 4/2 × 6 88–92 4 30 42.5 6 × 4/4 × 6 88–92

3–4 3 × 4/2 × 6 88–92 4 30 45 6 × 4/4 × 6 88–92
5–6 3 × 4/2 × 6 88–92 4 30 47.5 6 × 4/4 × 6 88–92

During running intervals, a 3 min rest was applied.

HRpeak , peak heart rate; RM, repetition maximum.

friction coefficient was 0.026 before, during and after the project, 2013). All subjects where individually supervised during the 1st,
and tested as described in Hoffman et al. (1990). All participants 5th, and 9th training session to ensure proper intensity during
used Swix Triac 1 poles (Swix, Lillehammer, Norway), with intervals. Additionally, MET was observed in simulated DP to
customized tips for rollerski treadmills (self-selected pole length ensure proper technique and the load was adjusted according to
for all tests; 153.3 ± 7.7 cm, corresponding to 82 ± 3% of body Table 2. All training from the 4 weeks prior to the beginning
height). A cable-pulley machine (Technogym Cable Jungle, of the intervention until the end of the intervention period
Gambettola, Italy), with a 45 cm-wide custom-made handlebar were reported. The endurance training intensity was divided into
(Losnegard et al., 2011) was used during training and tests in the three HR zones: low-intensity (LIT; 60–81% of peak HRmax ),
standing DP. Cycle times during submaximal DP (after 21 and moderate-intensity (MIT; 82–87%) and high-intensity training
41 min) and the muscular endurance test were recorded (Sony (HIT 88–100%), based on intensity zones developed by the
DCR-TRV900E; Sony, Tokyo, Japan). Norwegian Olympic Federation.

Training Statistical Analysis

A description of the specific training sessions is provided in All data are presented as mean ± standard deviation (SD).
Table 2. The training sessions lasted approx. 50 min, including Baseline differences between groups were tested with unpaired
a 10 min warm-up. Running intervals lasted 4 or 6 min Student’s t-tests. Within-group and between-group changes were
with 3 min rest periods in-between. During the muscular tested with paired Student’s t-tests and two-way repeated-
endurance training, there was a passive rest of 90 s between measures analysis of variance (ANOVA), respectively. All
sets. Speed during running intervals was based on individual statistical analysis was performed using Excel 2003 (Microsoft
HRmax measured during the VO2max running test, and adjusted Corporation, Redmond, Washington, USA) and IBM SPSS
to match the intended HR-zone within 90 s of trial-start (88– Statistics 20 (International Business Machines, New York, USA).
92% of HRmax ) (Helgerud et al., 2007). The intended intensity The level of confidence was set to 90% and P ≤ 0.05 were
has previously been shown to be advantageous in increasing the considered statistically significant, while 0.05 < P ≤ 0.1 was
maximal oxygen uptake in well-trained athletes (Seiler et al., considered a tendency toward statistical significance.

Frontiers in Physiology | www.frontiersin.org 81 September 2017 | Volume 8 | Article 690

Børve et al. Upper-Body Training in XC Skiers

RESULTS 1,000-m Time

The mean 1,000-m time decreased in MET (−4 ± 2%), but
Training was unchanged in ET (−1 ± 2%, Figure 4), which resulted in
There was a tendency toward differences between groups in total a between-group difference of 4 ± 2% (P = 0.06). Individual
training volume prior to the training intervention (P = 0.10), differences from pre- to post-test are shown in Figure 7A.
mainly due to a higher volume of LIT in MET compared to From pre- to post-test, MET increased speed more than
ET (Table 3). There were no differences in the intensity-zone ET between 500 and 700-m (7 ± 1%), and there was a
distribution between pre-intervention and intervention in MET. tendency to difference between groups at intervals 200–400-m
In ET, there was a small, but significant (0.1 h) increase in (5 ± 1%, P = 0.08) and 700–800-m (5 ± 1%, P = 0.08)
total training volume of HIT per week from pre-intervention to (Figure 5).
intervention.

Muscular Endurance and Maximal Strength Physiological Responses during 1,000-m

MET improved muscular endurance by 21 ± 8% (mean ± Time and VO2max Running
90% confidence interval) while no changes occurred in ET VO2max running and VO2peak DP was unchanged from pre-
(−1 ± 11%), resulting in a 22 ± 8% difference between to post-test in both groups (Table 4). Individual changes in
groups (Figure 3A). Individual differences from pre- to post- VO2peak DP are shown in Figure 7C. The percent differences
test are shown in Figure 7B. 1RM increased significantly in between VO2peak DP and VO2max running were 89 ± 3%
MET (6 ± 2%) and tended to increase in ET (2 ± 3%; in MET and 89 ± 2% in ET at both pre- and post-test.
P = 0.05), with a significant group difference of 4 ± 4% There was a tendency toward increase in HRpeak DP in
(Figure 3B). MET (2 ± 1%; P = 0.06), while ET remained unchanged

TABLE 3 | Training distribution (hours per week) in the 4 weeks prior to pre-test (Pre-intervention) and during the 6-week intervention period for MET (n = 11) and ET
(n = 11).

Variable MET ET

Pre-intervention Intervention Pre-intervention Intervention

LIT (60–81% of HRpeak ) 6.4 ± 3.1 5.6 ± 2.9 4.3 ± 2.2 4.1 ± 2.3
MIT (82–87% of HRpeak ) 1.2 ± 0.6 1.2 ± 0.7 1.0 ± 0.5 1.0 ± 0.6
HIT (88–100% of HRpeak ) 0.5 ± 0.2 0.6 ± 0.3 0.6 ± 0.3 0.7 ± 0.2*
Total endurance training 8.1 ± 3.6 7.5 ± 3.3 5.9 ± 2.1 5.8 ± 2.0
Muscular endurance training – 0.3 ± 0.1 – –
Strength training 0.5 ± 0.7 0.5 ± 0.5 0.3 ± 0.4 0.4 ± 0.5
Total training 8.6 ± 4.6 8.3 ± 4.2 6.2 ± 3.2 6.2 ± 3.1
Double poling 3.3 ± 2.7 2.4 ± 2.4 1.5 ± 1.1 1.2 ± 0.8

LIT, low-intensity training; MIT, moderate intensity training; HIT, high-intensity training; MET, combined muscular endurance and endurance training group; ET, endurance training group.
* Significantin-group difference from pre-intervention to intervention (P < 0.05). Double poling is performed during rollerski.

FIGURE 3 | Muscular endurance (A) and 1RM (B) in the standing double poling exercise before (pre) and after (post) 6 weeks of combined endurance- and muscular
endurance training (MET) or endurance training (ET). Data are expressed as group mean ± standard deviation. *Significant change from pre- to post-test (P < 0.05).
**Significant differences from pre- to post-test between MET and ET (P < 0.05).

Frontiers in Physiology | www.frontiersin.org 82 September 2017 | Volume 8 | Article 690

Børve et al. Upper-Body Training in XC Skiers

(−1 ± 1%), resulting in a significant difference of 3 ± 1% RER in MET after 35 min (−2 ± 1%, Figure 6C), and in
between groups. Rating of perceived exertion (RPE) after the ET after 15 (−2 ± 1%) and 35 min (−2 ± 1%, Figure 6D),
1,000-m test was unchanged from pre- to post-test in both resulting in a larger decrease in ET than in MET after 15
groups. min (2 ± 1%). Heart rate remained unchanged from pre-
to post-test for both groups (Figures 6E,F). Cycle time was
Physiological Response during not different from pre- to post-test both within and between
groups.
Submaximal Double Poling
The average O2 -cost for all five submaximal measurements were
reduced in MET from pre- to post-test, (−2 ± 2%) mainly DISCUSSION
due to reduced O2 -cost after 15 (−3 ± 2%) and 20 min
(−2 ± 1%, Figure 6A). No significant change was found in This study investigated the effect of replacing parts of high-
ET for mean O2 -cost (−1 ± 1%), or at any specific time- intensity interval training with upper-body muscular endurance
point (Figure 6B). For both groups, individual changes in training in well-trained XC skiers. The principal findings
average O2 -cost are shown in Figure 7C. However, there was were: (I) Six weeks of muscular endurance training increased
no significant change between groups in mean O2 -cost or O2 - muscular endurance and maximal strength in a simulated
cost for single time measurements. There was a decrease in DP exercise. (II) MET tended to improve 1,000-m DP
performance after 50 min of submaximal DP compared to
ET. (III) MET reduced the O2 -cost during submaximal DP,
but it was not significant different from ET. (IV) No changes
in VO2peak DP or VO2max running were found in either
group.
A novel finding of the present study was that upper-body
muscular endurance training improved 1,000-m poling time
completed immediately after 50 min of submaximal DP. In a
similar study, but with a slightly different training model, Nilsson
et al. (2004) found a significant improvement in mean power
output during a 6-min all-out test after 6 weeks of 20-s DP
interval training. Moreover, Vandbakk et al. (2017) showed that
30-s DP intervals over ∼8 weeks resulted in improved time
to exhaustion in the interval group, but not in the control
group. Together, these studies indicate that short-term upper-
body endurance training may be a promising training model
and, thus, have direct applications for well-trained skiers aiming
to improve their DP performance. However, whether these
FIGURE 4 | 1,000-m time before (pre) and after (post) 6 weeks of combined findings on an indoor treadmill are valid on snow needs to be
endurance training and muscular endurance training (MET) and endurance examined.
training (ET). Data are expressed as group mean ± standard deviation.
*Tendency to change from pre- to post-test (P = 0.06).
The improved 1,000-m time should be a result of improved
energy turnover and/or reduced O2 -cost (Bassett and Howley,

FIGURE 5 | Relative change in speed during the 1,000-m time trial from pre- (black horizontal line at 0%) to post-test. Data are expressed as group mean ± standard
deviation. The speed was set the first 200 m. All tests conducted at 2.5◦ incline. *Significant differences from pre- to post-test between combined endurance training
and muscular endurance training (MET) and endurance training (ET) (P < 0.05).

Frontiers in Physiology | www.frontiersin.org 83 September 2017 | Volume 8 | Article 690

Børve et al. Upper-Body Training in XC Skiers

2000). No significant changes were measured in aerobic relative change in O2 -cost during prolonged DP was in
energy turnover in either group while the anaerobic capacity favor the group that had trained upper-body muscular
was not estimated (due to methodology). However, the endurance, possibly contributing to the enhanced performance
due to a lower level of fatigue before the 1,000-m time.
These findings are in accordance with Rønnestad et al.
TABLE 4 | Physiological response during the 1,000-m time test (MET: n = 11; ET:
(2011) who showed increased 5 min all-out performance
n = 10) and running VO2max test (MET: n = 11; ET: n = 11). Data are mean ±
standard deviation. following 185 min of cycling after a heavy strength training
intervention. However, mechanisms for increase in work
Variable MET ET economy, when including heavy strength training or muscular
Pre Post Pre Post endurance training, remain unclear. Furthermore, in both
the present study and in Rønnestad et al. (2011) the
1,000-m TIME reduction in O2 -cost should be taken with caution since
VO2peak (mL·kg−1 ·min−1 ) 57.2 ± 6.7 58.0 ± 6.3 56.2 ± 5.8 57.7 ± 6.7 the relative differences in O2 -cost between groups was not
VO2peak (L·min−1 ) 4.5 ± 0.8 4.4 ± 0.8 4.3 ± 0.7 4.4 ± 0.8 significant.
HRpeak (beat ·min−1 ) 184 ± 4 187 ± 6** 190 ± 8 188 ± 5
The ratio VO2peak DP/VO2max running was ∼90% at pre-
RPE (6–20) 18.5 ± 1.0 18.5 ± 1.5 19.0 ± 1.0 18.5 ± 1.0
RUNNING VO2max TEST
test, which is similar to other studies (Rud et al., 2014; Skattebo
VO2max (mL·kg−1 ·min−1 ) 64.1 ± 5.1 64.7 ± 5.6 63.2 ± 5.2 63.9 ± 5.0 et al., 2015). However, no changes were observed in either
VO2max (L·min−1 ) 4.9 ± 0.8 5.0 ± 0.8 4.9 ± 0.9 4.9 ± 0.9 group, a finding consistent with most studies on heavy strength
HRpeak (beat ·min−1 ) 190 ± 5 188 ± 5 190 ± 8 188 ± 8* training, short-term speed training or endurance training in
RPE (6–20) 18.5 ± 1.0 19.0 ± 1.0* 19.0 ± 0.5 19.0 ± 1.0 XC skiing (Hoff et al., 1999; Østeras et al., 2002; Nilsson
MET, Combined muscular endurance and endurance training group; ET, endurance
et al., 2004; Rønnestad et al., 2012; Skattebo et al., 2015).
training group. *Significant in-group difference from pre- to post-test (P < 0.05). Thus, reducing this “gap” has been speculated to be one of the
**Significant difference in pre- to post-test change between groups (P < 0.05).

FIGURE 6 | O2-cost; VO2 (A,B), respiratory exchange ratio, RER (C,D) and heart rate, HR (E,F) during the prolonged 50-min double poling protocol before (pre-test)
and after (post-test) the 6-week intervention period. MET, Combined endurance training and muscular endurance training (left), ET, endurance training (right). Data are
expressed as group mean ± standard deviation. *Different from pre-test (P < 0.05).

Frontiers in Physiology | www.frontiersin.org 84 September 2017 | Volume 8 | Article 690

Børve et al. Upper-Body Training in XC Skiers

FIGURE 7 | Percent individual changes from pre- to post-test: (A) 1,000-m time, (B) muscular endurance, (C) VO2peak double poling (mL·kg−1 ·min−1 ) and (D)
average O2 -cost from the following measuring intervals: 8 to 10, 13 to 15, 18 to 20, 33 to 35, and 48 to 50 min. MET, Combined endurance training and muscular
endurance training (gray columns); ET, endurance training (black columns).

possible training benefits of increased upper-body training in 2011; Rønnestad et al., 2012; Skattebo et al., 2015). This implies
XC skiers (Sandbakk and Holmberg, 2017). However, there is that low resistance strength training could be an alternative
very little information on which type of training is most effective training method to heavy strength training, at least for short-
in stimulating these adaptations, and future studies should term adaptations prior to competitions or during short block
examine possible mechanisms more in detail. Interestingly, both periodization.
groups displayed the same VO2max in running after the training
intervention despite MET replaced half of the high-intensity Methodological Considerations
interval sets with upper-body muscular endurance training. This We counter-balanced the two groups based on 1,000-m time,
may be important information regard tapering strategies where VO2max running, 1RM and gender at pre-test. No significant
reduction in volume, while maintaining intensity and training differences between groups were found in training volume in
frequency, have been proposed to induce an “optimal” strategy any of the training categories. However, since the subjects were
in the final weeks before competition (Bosquet et al., 2007). well-trained and not elite athletes, the range between subjects
The relative increase in muscular endurance per session was large in most categories, causing substantial variation in-
(∼1.8%) is larger than previous reported in comparable studies and between groups. The absolute difference in total training
on specific upper-body training (0.7–1.1%) (Stone and Coulter, volume of 132 min/week [485 min (MET) vs. 353 min (ET)]
1994; Schoenfeld et al., 2015). However, Schoenfeld et al. (2015) prior to the intervention period was mainly caused by a 107 min
and Stone and Coulter (1994) included 24 and 27 training difference in weekly DP training on rollerski. This can indicate
sessions respectively, compared to 12 in the current study. It that MET had less potential for physiological adaptation than
is therefore possible that athletes could gain a relative large ET, which strengthens findings of the improved 1,000-m. On
increase in muscular endurance within a few sessions. This the other hand, MET reduced weekly DP training by 50 min
notion is particularly interesting from a “block periodization” (−27%) from pre-intervention compared to the intervention
perspective, herein shorter training periods (1–4 weeks) are period, and a reduction in training volume is related to tapering
utilized to focus on improving a few selected abilities (Issurin, and potentially improved performance (Bosquet et al., 2007;
2010). Moreover, even though the aim of the training was to Mujika, 2010). Altogether, using recreational but well-trained
increase muscular endurance, strength gain (1RM) per session in skiers with large variations in training load may potentially be
the current study was similar to studies in heavy strength training a limitation in the present study and should therefore be taken
for XC skiers (0.6 vs. 0.5–1.2% per session; Losnegard et al., into consideration when interpreting the results. Another aspect

Frontiers in Physiology | www.frontiersin.org 85 September 2017 | Volume 8 | Article 690

Børve et al. Upper-Body Training in XC Skiers

is that HIT was replaced with muscular endurance, and not that sets with upper-body muscular endurance training had no
training was added to their normal training, as done in most negative effects on the skiers’ VO2peak in DP or VO2max running.
other studies. This was based on the fact that recreational athletes
normally have limited time to execute training (e.g., full time ETHICS STATEMENTS
work) and thereby relative low training volume compared to elite
athletes. Hence, adding training in one group would lead to an This study was carried out in accordance with the
increase in total volume, which could lead to a greater training recommendations of Regional Committee for Medical and
stimuli it selves, and potentially enhanced performance. Finally, Health Research Ethics, Norwegian Research Ethics Act (2006)
one strengthen of the present study is the applied perspective and Act on Medical and Health Research (2008) with written
with direct practical application for coaches and athletes that informed consent from all subjects. The study was conducted
aim on optimizing performance. However, one clear limitation of according to the Declaration of Helsinki and Norwegian law.
the study is the lacking methodology to analyse possible changes
in intrinsic factors (such as muscle fiber types, mitochondria, AUTHOR CONTRIBUTIONS
capillary density and neuromuscular characteristics). Hence,
since this was out of the scope of the present study, further We hereby state that the contributions from the authors
studies should investigate the potential adaptations more in are in line with author guidelines as described below. The
detail. conception or design of the work; or the acquisition, analysis,
or interpretation of data for the work; JB, SNJ, BR, and
CONCLUSION TL Drafting the work or revising it critically for important
intellectual content; JB, SNJ, BR, and TL Final approval of
Six weeks of upper-body muscular endurance training increased the version to be published; JB, SNJ, BR, and TL Questions
muscular endurance and maximal strength in a simulated related to the accuracy or integrity of any part of the work
DP exercise and improved DP performance following 50 min are appropriately investigated and resolved; JB, SNJ, BR,
submaximal trial. Finally, replacing half of the running interval and TL.

REFERENCES Hoffman, M. D., Clifford, P. S., Bota, B., Mandli, M., and Jones, G. M. (1990).
Influence of body mass on energy cost of roller skiing. Int. J. Sport Biomech.
Bassett, D. R., and Howley, E. T. (2000). Limiting factors for maximum oxygen 6, 374–385. doi: 10.1123/ijsb.6.4.374
uptake and determinants of endurance performance. Med. Sci. Sports Exerc. 32, Issurin, V. B. (2010). New horizons for the methodology and
70–84. doi: 10.1097/00005768-200001000-00012 physiology of training periodization. Sports Med. 40, 189–206.
Borg, G. A. (1982). Psychophysical bases of perceived exertion. Med. Sci. Sports doi: 10.2165/11319770-000000000-00000
Exerc. 14, 377–381. doi: 10.1249/00005768-198205000-00012 Losnegard, T., Mikkelsen, K., Rønnestad, B. R., Hallén, J., Rud, B., and Raastad,
Bosquet, L., Montpetit, J., Arvisais, D., and Mujika, I. (2007). Effects of tapering T. (2011). The effect of heavy strength training on muscle mass and physical
on performance: a meta-analysis. Med. Sci. Sports Exerc. 39, 1358–1365. performance in elite cross country skiers. Scand. J. Med. Sci. Sports 21, 389–401.
doi: 10.1249/mss.0b013e31806010e0 doi: 10.1111/j.1600-0838.2009.01074.x
Campos, G. E., Luecke, T. J., Wendeln, H. K., Toma, K., Hagerman, F. C., Losnegard, T., Myklebust, H., Spencer, M., and Hallén, J. (2013). Seasonal
Murray, T. F., et al. (2002). Muscular adaptations in response to three variations in V [Combining Dot Above] O2max , O2 -cost, O2 -deficit, and
different resistance-training regimens: specificity of repetition maximum performance in elite cross-country skiers. J. Strength Cond. Res. 27, 1780–1790.
training zones. Eur. J. Appl. Physiol. 88, 50–60. doi: 10.1007/s00421-002-0 doi: 10.1519/JSC.0b013e31827368f6
681-6 Mikkola, J., Vesterinen, V., Taipale, R., Capostagno, B., Häkkinen, K., and
Ebben, W. P., Kindler, A. G., Chirdon, K. A., Jenkins, N. C., Polichnowski, Nummela, A. (2011). Effect of resistance training regimens on treadmill
A. J., and Ng, A. V. (2004). The effect of high-load vs. high-repetition running and neuromuscular performance in recreational endurance
training on endurance performance. J. Strength Cond. Res. 18, 513–517. runners. J. Sports Sc. 29, 1359–1371. doi: 10.1080/02640414.2011.5
doi: 10.1519/R-12722.1 89467
Foss, Ø., and Hallen, J. (2005). Validity and stability of a computerized Mujika, I. (2010). Intense training: the key to optimal performance
metabolic system with mixing chamber. Int. J. Sports Med. 26, 569–575. before and during the taper. Scand. J. Med. Sci. Sports 20, 24–31.
doi: 10.1055/s-2004-821317 doi: 10.1111/j.1600-0838.2010.01189.x
Gallagher, D., DiPietro, L., Visek, A. J., Bancheri, J. M., and Miller, T. A. (2010). Nilsson, J., Holmberg, H. C., Tveit, P., and Hallén, J. (2004). Effects of 20-s and 180-
The effects of concurrent endurance and resistance training on 2,000-m rowing s double poling interval training in cross-country skiers. Eur. J. Appl. Physiol.
ergometer times in collegiate male rowers. J. Strength Cond. Res. 24, 1208–1214. 92, 121–127. doi: 10.1007/s00421-004-1042-4
doi: 10.1519/JSC.0b013e3181d8331e Østeras, H., Helgerud, J., and Hoff, J. (2002). Maximal strength-training
Helgerud, J., Hoydal, K., Wang, E., Karlsen, T., Berg, P., Bjerkaas, M., et al. effects on force-velocity and force-power relationships explain increases
(2007). Aerobic high-intensity intervals improve VO2max more than moderate in aerobic performance in humans. Eur. J. Appl. Physiol. 88, 255–263.
training. Med. Sci. Sports Exerc. 39, 665–671. doi: 10.1249/mss.0b013e31803 doi: 10.1007/s00421-002-0717-y
04570 Rønnestad, B. R., Hansen, E. A., and Raastad, T. (2011). Strength training
Hoff, J., Gran, A., and Helgerud, J. (2002). Maximal strength training improves improves 5-min all-out performance following 185 min of cycling.
aerobic endurance performance. Scand. J. Med. Sci. Sports 12, 288–295. Scand. J. Med. Sci. Sports 21, 250–259. doi: 10.1111/j.1600-0838.2009.
doi: 10.1034/j.1600-0838.2002.01140.x 01035.x
Hoff, J., Helgerud, J., and Wisloff, U. (1999). Maximal strength training improves Rønnestad, B. R., Kojedal, O., Losnegard, T., Kvamme, B., and Raastad, T.
work economy in trained female cross-country skiers. Med. Sci. Sports Exerc. (2012). Effect of heavy strength training on muscle thickness, strength, jump
31, 870–887. doi: 10.1097/00005768-199906000-00016 performance, and endurance performance in well-trained Nordic Combined

Frontiers in Physiology | www.frontiersin.org 86 September 2017 | Volume 8 | Article 690

Børve et al. Upper-Body Training in XC Skiers

athletes. Eur. J. Appl. Physiol. 112, 2341–2352. doi: 10.1007/s00421-011- Skattebo, Ø., Hallen, J., Rønnestad, B. R., and Losnegard, T. (2015). Upper body
2204-9 heavy strength training does not affect performance in junior female cross-
Rud, B., Secher, N. H., Nilsson, J., Smith, G., and Hallen, J. (2014). country skiers. Scand. J. Med. Sci. Sports 29, 1007–1016. doi: 10.1111/sms.12517
Metabolic and mechanical involvement of arms and legs in simulated Stöggl, T., Lindinger, S., and Müller, E. (2007). Analysis of a simulated sprint
double pole skiing. Scand. J. Med. Sci. Sports 24, 913–919. doi: 10.1111/sms. competition in classical cross country skiing. Scand. J. Med. Sci. Sports 17,
12133 362–372. doi: 10.1111/j.1600-0838.2006.00589.x
Sandbakk, Ø., and Holmberg, H. C. (2017). Physiological capacity and training Stone, W. J., and Coulter, S. P. (1994). Strength/endurance effects from three
routines of elite cross-country skiers: approaching the upper limits of human resistance training protocols with women. J. Strength Cond. Res. 8, 231–234.
endurance. Int. J. Sports Physiol. Perform. doi: 10.1123/ijspp.2016-0749. [Epub Terzis, G., Stattin, B., and Holmberg, H. C. (2006). Upper body training and the
ahead of print]. triceps brachii muscle of elite cross country skiers. Scand. J. Med. Sci. Sports 16,
Schoenfeld, B. J., Peterson, M. D., Ogborn, D., Contreras, B., and Sonmez, G. 121–126. doi: 10.1111/j.1600-0838.2005.00463.x
T. (2015). Effects of low-vs. high-load resistance training on muscle strength Vandbakk, K., Welde, B., Kruken, A. H., Baumgart, J., Ettema, G., Karlsen, T.,
and hypertrophy in well-trained men. J. Strength Cond. Res. 29, 2954–2963. et al. (2017). Effects of upper-body sprint-interval training on strength and
doi: 10.1519/JSC.0000000000000958 endurance capacities in female cross-country skiers. PLoS ONE 12:e0172706.
Sedano, S., Marin, P. J., Cuadrado, G., and Redondo, J. C. (2013). doi: 10.1371/journal.pone.0172706
Concurrent training in elite male runners: the influence of strength
versus muscular endurance training on performance outcomes. J. Conflict of Interest Statement: The authors declare that the research was
Strength Cond. Res. 27, 2433–2443. doi: 10.1519/JSC.0b013e3182 conducted in the absence of any commercial or financial relationships that could
80cc26 be construed as a potential conflict of interest.
Seiler, S. (2010). What is best practice for training intensity and duration
distribution in endurance athletes? Int. J. Sports Physiol. Perform. 5, 276–291. Copyright © 2017 Børve, Jevne, Rud and Losnegard. This is an open-access article
doi: 10.1123/ijspp.5.3.276 distributed under the terms of the Creative Commons Attribution License (CC BY).
Seiler, S., Jøranson, K., Olesen, B. V., and Hetlelid, K. J. (2013). Adaptations The use, distribution or reproduction in other forums is permitted, provided the
to aerobic interval training: interactive effects of exercise intensity original author(s) or licensor are credited and that the original publication in this
and total work duration. Scand. J. Med. Sci. Sports 23, 74–83. journal is cited, in accordance with accepted academic practice. No use, distribution
doi: 10.1111/j.1600-0838.2011.01351.x or reproduction is permitted which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 87 September 2017 | Volume 8 | Article 690

 ORIGINAL RESEARCH
published: 20 September 2017
doi: 10.3389/fphys.2017.00655

Sedentary Behavior among National

Elite Rowers during Off-Training—A
Pilot Study
Billy Sperlich 1*, Martin Becker 2 , Andreas Hotho 2, 3 , Birgit Wallmann-Sperlich 4, 5 ,
Mahdi Sareban 6, 7 , Kay Winkert 6 , Jürgen M. Steinacker 6 and Gunnar Treff 6
1
Integrative and Experimental Exercise Science, Institute for Sport Sciences, University of Wuerzburg, Wuerzburg, Germany,
2
DMIR Research Group, University of Wuerzburg, Wuerzburg, Germany, 3 L3S Research Center, Hanover, Germany,
4
Institute for Sport Sciences, University of Wuerzburg, Wuerzburg, Germany, 5 Institute of Health Promotion and Clinical
Movement Science, German Sport University Cologne, Cologne, Germany, 6 Division of Sports and Rehabilitation Medicine,
Ulm University Hospital, Ulm, Germany, 7 Institute of Sports Medicine, Prevention and Rehabilitation, Paracelsus Medical
University, Salzburg, Austria

The aim of this pilot study was to analyze the off-training physical activity (PA) profile
in national elite German U23 rowers during 31 days of their preparation period. The
hours spent in each PA category (i.e., sedentary: <1.5 metabolic equivalents (MET); light
physical activity: 1.5–3 MET; moderate physical activity: 3–6 MET and vigorous intense
Edited by:
Luca Paolo Ardigò, physical activity: >6 MET) were calculated for every valid day (i.e., >480 min of wear
University of Verona, Italy time). The off-training PA during 21 weekdays and 10 weekend days of the final 11-week
Reviewed by: preparation period was assessed by the wrist-worn multisensory device Microsoft Band
Daniel Aggio,
University College London,
II (MSBII). A total of 11 rowers provided valid data (i.e., >480 min/day) for 11.6 week days
United Kingdom and 4.8 weekend days during the 31 days observation period. The average sedentary
Pantelis Theodoros Nikolaidis,
time was 11.63 ± 1.25 h per day during the week and 12.49 ± 1.10 h per day on
Hellenic Army Academy, Greece
Beat Knechtle, the weekend, with a tendency to be higher on the weekend compared to weekdays
Institute of Primary Care, University of (p = 0.06; d = 0.73). The average time in light, moderate and vigorous PA during the
Zurich, Switzerland
weekdays was 1.27 ± 1.15, 0.76 ± 0.37, 0.51 ± 0.44 h per day, and 0.67 ± 0.43,
*Correspondence:
Billy Sperlich
0.59 ± 0.37, 0.53 ± 0.32 h per weekend day. Light physical activity was higher during
[email protected] weekdays compared to the weekend (p = 0.04; d = 0.69). Based on our pilot study
of 11 national elite rowers we conclude that rowers display a considerable sedentary
Specialty section:
off-training behavior of more than 11.5 h/day.
This article was submitted to
Exercise Physiology, Keywords: accelerometer, microsoft band 2, multi-sensor, recovery, sedentary behavior, wearable
a section of the journal
Frontiers in Physiology

Received: 23 June 2017 INTRODUCTION

Accepted: 17 August 2017
Published: 20 September 2017 Elite rowers invest a considerable amount of time for their training averaging >1,000 h per year
Citation: (Fiskerstrand and Seiler, 2004) i.e., approximately 17% of h per year of waking time. Nevertheless,
Sperlich B, Becker M, Hotho A, a great proportion of available time is not spent for training but for recovery including activities of
Wallmann-Sperlich B, Sareban M, daily living, such as studying, working, traveling etc.
Winkert K, Steinacker JM and Treff G
Past investigations focused on analyzing and optimizing the quality of training (Fiskerstrand
(2017) Sedentary Behavior among
National Elite Rowers during
and Seiler, 2004; Stoggl and Sperlich, 2015), however very little is known about the intensity
Off-Training—A Pilot Study. and volume of physical activity (PA) performed by elite athletes during their off-training time
Front. Physiol. 8:655. which, as mentioned above, accounts for more than 80% of waking time. This is astonishing as we
doi: 10.3389/fphys.2017.00655 know that the rate of adaptation (although not exclusively) is an integral of the training stimulus

Frontiers in Physiology | www.frontiersin.org 88 September 2017 | Volume 8 | Article 655

Sperlich et al. Off-Training Analysis in Elite Rowers

itself (intensity, duration and frequency of stimulus), consent to participate in the study which was conducted in
environmental surrounding, behavior (e.g., nutrition) but accordance with the Declaration of Helsinki. All protocols were
also the type of (acute) recovery strategies (Bishop et al., 2008). pre-approved by the ethical review board of the University
Largely, this “integrative dose” determines one’s individual of Ulm.
biological adaptation as well as health.
Surprisingly, to the best of our knowledge only one study
so far investigated the PA of elite athletes outside their sport-
Assessment of Physical Activity (PA)
Data collection took place during the final 11-week preparation
activity (Weiler et al., 2015) concluding that the elite soccer
period (i.e., calendar week 3–14) before the rowers’ first
players were surprisingly sedentary during off-training, especially
competition of the season. Each rower was instructed to wear
when compared to non-athletic groups. In this context, recent
a wrist-worn multisensory device Microsoft Band II (MSBII),
studies also showed increased prevalence of overweight and obese
for a period of 1 month (31 days, with 21 weekdays, and 10
athletes indicating increased sedentary behavior (Nikolaidis,
weekend days) only removing it for scheduled training sessions
2012, 2013). Sedentary behavior as such is defined as any waking
and showering. The MSBII incorporates several sensors including
behavior characterized by an energy expenditure ≤1.5 metabolic
a 3-axis accelerometer, gyrometer, optical heart-rate sensor,
equivalents (MET), while in a sitting, reclining or lying posture
galvanic skin response sensor, ambient light sensor, ultraviolet
(Tremblay et al., 2017). Evidence exists that elevated levels of
light exposure, and skin temperature sensor. The MSB2 stores the
sedentary behavior in the non-athletic population are associated
data of mean hourly energy expenditure online.
with various adverse health outcomes, such as cardiovascular
disease, diabetes, and all-cause mortality (Chau et al., 2013; de
Rezende et al., 2014). Preliminary Analysis
Within the athletic population it is accepted that active when Beforehand we validated the measurement of energy expenditure
compared to passive (i.e., inactive) recovery (after high-intensity of the multi-sensory MSBII with the energy expenditure from
efforts) (Riganas et al., 2015) is likely to impact overall recovery indirect calorimetry (Metamax 3B, Cortex, Leipzig, Germany)
and sport performance (Laursen and Jenkins, 2002; Buchheit in nine physical education students. Depending on their level
et al., 2009). In elite rowers e.g., active compared to passive of performance they sat, stood, walked at 3, 4, 5 km·h−1 or
recovery provides higher rate of lactate removal compared to jogged at 7.2, 9.0, 10.8, and 12.6 km·h−1 for 3-min. During each
passive recovery (Riganas et al., 2015) and the active recovery 3-min activity the energy expenditure was measured with the
with a more rapid regulation of homeostasis (although not MSBII and a previously validated (Medbo et al., 2002) breath-by-
fully understood) may regulate growth and transcription factors breath metabolic cart (MetaMax 3B, Cortex Biophysik, Leipzig,
(Coffey and Hawley, 2007). In this context, sedentary off-training Germany). In accordance with the manufacturer’s instructions
behavior may negatively affect recovery and in a long-term both, the gas and flow sensor were calibrated prior to all testing.
adaptation to exercise and health. Over the activity range from 1 to 10 MET (Figure 1), i.e.,
In summary, analysis of sedentariness in the elite athletic sitting, standing, walking, and jogging the Pearson correlation
population is rare and only assessed in a team sport setting and coefficient (r) calculation revealed a significant and nearly perfect
not among elite endurance athletes. Potential identification of correlation between the energy expenditure calculated from the
sedentariness could (i) lead to a change in the view of off-training multi-sensory MSBII and the energy expenditure from indirect
procedures (e.g., active recovery) and (ii) could stimulate health calorimetry (r = 0.92; r2 = 0.84, p < 0.001).
advice in light of reducing the risk of sedentary-induced all-cause The correlation coefficients of this preliminary testing are
negative health effects due to accustomed in-career sedentary even higher than previously published correlation coefficients
behavior. Therefore, this pilot study aimed to analyse the off- when comparing the energy expenditure assessed by multi-sensor
training PA profile in national elite German U23 rowers during devices and indirect calorimetry in healthy adults [r ranging from
31 days of their preparation period. Based on a previous analysis 0.56 (Fruin and Rankin, 2004) to 0.85 (Dwyer et al., 2009)].
in football (Weiler et al., 2015) we hypothesized that elite rowers To classify energy expenditure in established PA classifications
display a considerable sedentary off-training behavior. (Ainsworth et al., 2011; Sedentary Behaviour Research Network,
2012) we normalized the energy expenditure by individual body
METHODS mass and categorized received mean METs/hour as sedentary
activity (<1.5 MET), light (1.5–3 MET), moderate (3–6 MET),
Participants and vigorous intense PA (>6 MET). Non-wear time was
Eleven German U23 rowers, competing at national or identified by checking heart rate data, i.e., if no valid heart rate
international level took part in this investigation (peak oxygen was present for an hour it was deemed that the device must
uptake: 66 ± 5 mL·min−1 ·kg−1 , 20 ± 2 years, body mass: 88.4 have been removed in that hour. According to the manufacturer,
± 9.7 kg, height: 189 ± 7 cm). The inclusion criteria were: (i) the MSBII automatically tracks the duration of sleep integrating
age 18–30 years; (ii) male; (iii) squad member of either regional biometric data of heart rate and motion or when the athlete
or national level with seamless periods of rowing before study personally activates the sleeping mode. The hours spent in each
initiation. Exclusion criteria were: (i) medically unfit to perform PA category were calculated for every valid day of data recorded,
the study according to previous recommendations (Steinacker where a valid day consists of at least 480 min of wear time during
et al., 2002). All participants gave their written informed waking hours of the non-training period in correspondence with

Frontiers in Physiology | www.frontiersin.org 89 September 2017 | Volume 8 | Article 655

Sperlich et al. Off-Training Analysis in Elite Rowers

TABLE 1 | Summary of daily activity of 11 rowers during their preparation period.

Mean ± SD 95% CI P; d

Valid days Weekday 11.55 ± 4.25 8.69–14.40 <0.001*

Weekend 4.81 ± 1.53 3.79–5.85 2.11
Mean wear time per day [h] Weekday 22.31 ± 1.14 21.54–23.08 0.67
Weekend 22.05 ± 1.51 21.04–23.07 0.19
Sedentary time (<1.5 MET) Weekday 11.63 ± 1.25 10.80–12.48 0.06
in waking hours [h]
Weekend 12.49 ± 1.10 11.77–13.21 0.73
Light PA (1.5–3 MET) [h] Weekday 1.27 ± 1.15 0.49–2.05 0.04
Weekend 0.67 ± 0.43 0.37–0.96 0.69
Moderate PA (3–6 MET) [h] Weekday 0.76 ± 0.37 0.51–1.00 0.13
Weekend 0.59 ± 0.37 0.31–0.87 0.45
FIGURE 1 | Correlation analysis of energy expenditure measured by MSBII
Vigorous PA (>6 MET) [h] Weekday 0.51 ± 0.44 0.20–0.83 0.97
and indirect calorimetry (Cortex Metamax 3B) of nine subjects sitting, standing,
Weekday 0.53 ± 0.32 0.25–0.80 0.05
walking and jogging.
Sleep time [h] Weekday 8.18 ± 1.24 7.35–9.01 0.80
Weekend 8.07 ± 1.34 7.17–8.97 0.08

(Atkin et al., 2012). Data classified as time in bed and invalid days PA, Physical activity; d, Cohen’s d effect sizes calculated from the Mean ± SD between
(<480 min of wear time) were excluded from the analyses. weekdays and weekend; *Indicates differences between weekdays and weekend for
P < 0.05.

Statistical Analysis
The data to calculate the MET values were processed using the
Python data analysis toolkit “pandas” (0.18.0) and the scientific
computing library “SciPy” (0.17.0) available for the Python
programming language (3.5.1). Further analysis was conducted
using the Statistica software package for Windows R (version 7.1,
StatSoft Inc., Tulsa, OK, USA). That is, a student’s paired t-test
was employed to calculate the differences between weekdays and
weekend activities [i.e., sedentary time (<1.5 MET); light PA
(1.5–3 MET), moderate PA (3–6 MET); vigorous PA (>6 MET)].
An alpha of p < 0.05 was considered as significant. The effect
size, Cohen’s d, (Cohen, 1988) was calculated for all variables,
with the thresholds for small, moderate, and large effects set at
0.20, 0.50, and 0.80, respectively (Cohen, 1988). Medium or large
effects sizes were considered as tendencies if comparisons based
on p-values were insignificant.

FIGURE 2 | Fraction of mean sedentary time [hours], light, moderate, and

RESULTS vigorous PA as well as sleep of total wear time during off-training periods.

A total of 11 rowers provided valid data (i.e., >480 min/day)

for 11.6 week days and 4.8 weekend days during the 31-day
observation period. DISCUSSION
All mean data for sedentary time, light, moderate and
vigorous PA as well as sleep are summarized in Table 1 and In the present study, we aimed to analyze the off-training PA
the corresponding fraction of total wear time during off-training of national elite U23 rowers during their preparation period. To
periods are illustrated in Figure 2. the best of our knowledge this is the first investigation among
The average sedentary time was 11.63 ± 1.25 h per day during endurance athletes.
the week and 12.49 ± 1.10 h per day on the weekend, with The main findings of this investigation were that national elite
moderate effect sizes indicating sedentary time to be higher on U23 rowers when compared to non-athletic population studies
the weekend (p = 0.06; d = 0.73). The average time per day in (Schuna et al., 2013; Owen et al., 2014) display a larger proportion
light, moderate and vigorous PA during the weekdays was 1.27 of time sedentary (<1.5 MET), a lower proportion of light PA,
± 1.15, 0.76 ± 0.37, 0.51 ± 0.44 h per day, and 0.67 ± 0.43, 0.59 but at the same time display a greater amount of moderate
± 0.37, 0.53 ± 0.32 h per weekend day. Light activity was higher to vigorous PA (>3 MET) in addition to their often vigorous
during weekdays compared to weekend (p = 0.04; d = 0.69). training activity. In their secondary analyses of the NHANES

Frontiers in Physiology | www.frontiersin.org 90 September 2017 | Volume 8 | Article 655

Sperlich et al. Off-Training Analysis in Elite Rowers

from 2005 to 2006, Schuna and co-workers present a mean Likewise, young elite futsal players perceive more benefit from
sedentary time of 478.9 (2.6) min/day, 200.0 (1.5) min/day in low immediate postgame (water) exercises compared to dry exercises
PA, 141.3 (1.8) min/day in light PA, 87.8 (1.2) in lifestyle PA, and and seated rest, which is thought to improve their attitude toward
22.8 (0.7) min/day in moderate-to-vigorous intensity PA (Schuna playing (Tessitore et al., 2008). In contrast, results indicate that
et al., 2013). passive and active (i.e., running 5 miles on a flat course on two
The rowers in the present study spent >11.5 h sedentary i.e., consecutive days, at an intensity of 65–75% of maximum heart
expending a mean metabolic equivalent of <1.5 METs per hour rate) recovery result in similar mean 5-km performance (Bosak
which corresponds to sitting, lying and passive transportation etc. et al., 2008). Equally, a single 30-min session of aqua cycling
The present data is in line with a previous investigation (Weiler was not able to attenuate the effects on muscular performance,
et al., 2015) analyzing professional footballers during an English markers of muscle damage, or delayed onset of muscle soreness
league season and demonstrating significant sedentary behavior (DOMS) compared with passive rest (Wahl et al., 2017).
among elite footballers. In the latter study, the footballers spent Finally, muscle activation induces blood flow (Sperlich et al.,
approximately 8 ± 1 h of waking time sedentary. In the present 2013), thereby delivering oxygen and substrates to the muscle
study, the rowers were about 3.5 h more sedentary (hours per and also supports the clearances of metabolites. So, from this
day spend at <1.5 METs) during the weekdays and 4.5 h more perspective, any form of (light) muscle activity during off-
sedentary during the weekend. One reason for the calculated training should support circulatory induced recovery.
sedentariness of our rowers may be attributable to the algorithm Based on our experience, active recovery is employed
(hourly average of activity) of the MSBII neglecting short immediately or with time-delay after exercise and for a certain
interruptions of sedentary time with activities of more than (short) period of time. Since an extremely high variability of
1.5 MET. “best” recovery scheme exists between different athletes (Bishop
However, it is important to note that the sedentariness in our et al., 2008) it is astonishing, that no study so far (at least to
rowers was higher during the weekend compared to weekdays, the best of our knowledge) has investigated the influence of
which has also been confirmed as pattern in other non-athletic different (long-term) off-training PA profiles in athletes. We
populations, such as students (Clemente et al., 2016). Since the acknowledge the fact that certain “sedentary behavior” maybe
rowers were not professional athletes they might not have had necessary for elite athletes to properly recover, however the
enough time (due to work, education, etc.) during the week to impact of prolonged sedentary behavior during off-training and
perform longer and/or (very) intense sessions. Longer session its impact on athletic recovery, performance or injury risk is
(and maybe more intense sessions) would lead to fatigue resulting unknown. From this perspective, future investigation may aim
in less off-training activity. to answer the question whether the manipulation of off-training
However, the rowers in the present study spent clearly more PA may be beneficial or harmful for recovery processes and
time (2 min vs. 30 min) at vigorous activity (>6 MET) when long-term performance development in elite athletes.
compared to elite footballers (Weiler et al., 2015). We can
only speculate to why rowers display more vigorous activity Health Risk of Sedentariness in Athletes?
during their off-training but maybe this mirrors, at least in Although it is well-known that sedentary behavior is related to
part, the typical behavior of rowers preferring more vigorous all-cause mortality (Chau et al., 2013; de Rezende et al., 2014) elite
and exhausting exercise. However, we cannot exclude that some athletes may not be increasingly threatened by this risk (Ekelund
rowers added additional non-scheduled exercise into their free et al., 2016). However, Olympic athletes are not immune toward
time e.g., a soccer game. cardio-vascular disorders and might be exposed to unexpected
high-risk of cardiovascular abnormalities during sport activity
Active vs. Sedentary Recovery (Pelliccia et al., 2017). Additionally, there is some evidence
To improve recovery, various responses of different modalities indicating that elite endurance athletes, when retired, change
have been investigated including macronutrient supplementation their body composition more than aerobic characteristics with
(McLellan et al., 2014), massage techniques (Poppendieck age (Mujika, 2012). From this perspective, the sedentary behavior
et al., 2016), cooling (Poppendieck et al., 2013), self-myofascial of active athletes may not directly be harmful to their health
release (Beardsley and Skarabot, 2015), neuromuscular electrical but, especially after retiring from their sporting career, these
stimulation (Babault et al., 2011), active vs. passive recovery individuals may be at high risk of sedentary-induced all-cause
(Laursen and Jenkins, 2002; Buchheit et al., 2009; Riganas et al., mortality due to accustomed in-career sedentary behavior.
2015) (and many more), all of which are performed rather There is some evidence that interrupting sitting time every
temporarily (minutes to maybe 1 h) and employed promptly 20–30 min by standing up or walking helps to counteract cardio-
after exercise. Short-term active compared to passive recovery metabolic disease (Dunstan et al., 2012) and bodies, such as
in rowers is known to provide a higher rate of lactate removal the American College of Sports Medicine address the issue of
compared to passive recovery (Riganas et al., 2015) and active reducing sedentary behavior (Kravitz and Vella, 2016) repeatedly.
recovery with a more rapid regulation of homeostasis (although The athletic population may not feel addressed, because of their
not fully understood) may regulate growth and transcription high training related PA. In all cases, athletes should be informed
factors (Coffey and Hawley, 2007). Similarly, lactic acid clearance about their current off-training PA profile and the long-term risk
measured 20 min after repeated supramaximal leg exercise (i.e., associated with sedentary behavior. In this context commercially
Wingate tests) is significantly greater with active compared to available wearable sensors (Duking et al., 2016), as long as they
passive recovery and massage in cyclists (Martin et al., 1998). fulfill scientific quality criteria (Sperlich and Holmberg, 2017),

Frontiers in Physiology | www.frontiersin.org 91 September 2017 | Volume 8 | Article 655

Sperlich et al. Off-Training Analysis in Elite Rowers

and do not danger personal data security (Austen, 2015), may as possible to be obtained by continuous 24-h monitoring of
be useful in providing feedback (Duking et al., 2017) of daily PA various PA and also estimate sleep, and various environmental
patterns. conditions. As long as scientific quality is ensured (Duking
et al., 2016; Sperlich and Holmberg, 2017) and personal data
Methodological Considerations secured, such technology can potentially provide a 24-h feedback
Some methodological considerations need acknowledgment: (Duking et al., 2017) to the athlete and supporting staff about
First, we only observed a short period within the season of PA during off-training. Individual feedback to PA may assist
competitive rowers, i.e., 31 days. Although, this observation to counteract exaggerated sedentariness and could stimulate
period is significantly longer compared to other studies health advice in light of reducing the risk of sedentary-induced
investigating PA patterns (Schuna et al., 2013) we cannot judge negative health outcomes due to accustomed in-career sedentary
whether the PA profile during off- and competition season would behavior.
be different. Secondly, since our rowers were among the best
athletes in Germany we cannot estimate whether the result is
CONCLUSION
also true for recreational, female, youth or older rowers. Thirdly,
the data analysis of the MSBII does not allow to record PA Based on our data we conclude that well-trained rowers when
densely, i.e., data every second or minute within a 24-h cycle. compared to other populations display a larger proportion of
Consequently, we could not assess the quantity of possible time sedentary (<1.5 MET) but at the same time display a
micro bouts of PA, which might have been leveled off through greater amount of time in moderate to vigorous PA (>3 MET).
sedentary behavior for the rest of the hour. Also, the position of Future investigation may aim to answer the question whether the
the wrist-worn device could have an error in the calculation of manipulation of off-training PA may be beneficial or harmful for
energy expenditure. Although we instructed all rowers to wear recovery processes and long-term performance development and
the MSBII always on the same arm we cannot be sure if this was health in elite athletes.
the case all the time.
Also, from a methodological point of view, the number
of rowers in the present pilot study was relatively small and AUTHOR CONTRIBUTIONS
more participants would have allowed greater statistical power.
However, the 11 rowers were among the best of their age All designed and approved the methods, analyzed data, and
group in Germany and increasing the sample size would have assisted in manuscript writing. BS, MB, BWS, KW, and GT
meant to integrate “weaker” rowers thereby confounding the performed data collection.
interpretation of the data for the “elite” rowing population.
As this study was designed as pilot study, further research is FUNDING
warranted and the present results should be viewed carefully until
the data is confirmed in other populations. This publication was funded by the German Research
Foundation (DFG) and the University of Wuerzburg in the
Practical Consideration funding programme Open Access Publishing. The project was
As mentioned previously (Sperlich and Holmberg, 2017), supported by the German Federal Institute of Sports Sciences
wearable technology allows to collect as much information (BISp, ZMVI4-070707/16).

REFERENCES Buchheit, M., Cormie, P., Abbiss, C. R., Ahmaidi, S., Nosaka, K. K., and
Laursen, P. B. (2009). Muscle deoxygenation during repeated sprint running:
Ainsworth, B. E., Haskell, W. L., Herrmann, S. D., Meckes, N., Bassett, D. R. effect of active vs. passive recovery. Int. J. Sports Med. 30, 418–425.
Jr., Tudor-Locke, C., et al. (2011). 2011 Compendium of physical activities: a doi: 10.1055/s-0028-1105933
second update of codes and MET values. Med. Sci. Sports Exerc. 43, 1575–1581. Chau, J. Y., Grunseit, A. C., Chey, T., Stamatakis, E., Brown, W. J., Matthews, C.
doi: 10.1249/MSS.0b013e31821ece12 E., et al. (2013). Daily sitting time and all-cause mortality: a meta-analysis. PLoS
Atkin, A. J., Gorely, T., Clemes, S. A., Yates, T., Edwardson, C., Brage, S., et al. ONE 8:e80000. doi: 10.1371/journal.pone.0080000
(2012). Methods of measurement in epidemiology: sedentary behaviour. Int. J. Clemente, F. M., Nikolaidis, P. T., Martins, F. M., and Mendes, R. S. (2016).
Epidemiol. 41, 1460–1471. doi: 10.1093/ije/dys118 Physical activity patterns in university students: do they follow the public health
Austen, K. (2015). What could derail the wearables revolution? Nature 525, 22–24. guidelines? PLoS ONE 11:e0152516. doi: 10.1371/journal.pone.0152516
doi: 10.1038/525022a Coffey, V. G., and Hawley, J. A. (2007). The molecular bases of training adaptation.
Babault, N., Cometti, C., Maffiuletti, N. A., and Deley, G. (2011). Does electrical Sports Med. 37, 737–763. doi: 10.2165/00007256-200737090-00001
stimulation enhance post-exercise performance recovery? Eur. J. Appl. Physiol. Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Hillsdale,
111, 2501–2507. doi: 10.1007/s00421-011-2117-7 NJ: L. Erlbaum Associates.
Beardsley, C., and Skarabot, J. (2015). Effects of self-myofascial release: a systematic de Rezende, L. F., Rodrigues Lopes, M., Rey-Lopez, J. P., Matsudo, V. K., and
review. J. Bodyw. Mov. Ther. 19, 747–758. doi: 10.1016/j.jbmt.2015.08.007 Luiz Odo, C. (2014). Sedentary behavior and health outcomes: an overview
Bishop, P. A., Jones, E., and Woods, A. K. (2008). Recovery from training: of systematic reviews. PLoS ONE 9:e105620. doi: 10.1371/journal.pone.
a brief review: brief review. J. Strength Cond. Res. 22, 1015–1024. 0105620
doi: 10.1519/JSC.0b013e31816eb518 Duking, P., Holmberg, H. C., and Sperlich, B. (2017). Instant biofeedback provided
Bosak, A., Bishop, P., Green, M., and Iosia, M. (2008). Active versus passive by wearable sensor technology can help to optimize exercise and prevent injury
recovery in the 72 hours after a 5-km race. Sport J. 11, 1. and overuse. Front. Physiol. 8:167. doi: 10.3389/fphys.2017.00167

Frontiers in Physiology | www.frontiersin.org 92 September 2017 | Volume 8 | Article 655

Sperlich et al. Off-Training Analysis in Elite Rowers

Duking, P., Hotho, A., Holmberg, H. C., Fuss, F. K., and Sperlich, B. (2016). Poppendieck, W., Wegmann, M., Ferrauti, A., Kellmann, M., Pfeiffer, M.,
Comparison of non-invasive individual monitoring of the training and health and Meyer, T. (2016). Massage and performance recovery: a meta-
of athletes with commercially available wearable technologies. Front. Physiol. analytical review. Sports Med. 46, 183–204. doi: 10.1007/s40279-015-
7:71. doi: 10.3389/fphys.2016.00071 0420-x
Dunstan, D. W., Kingwell, B. A., Larsen, R., Healy, G. N., Cerin, E., Hamilton, M. Riganas, C. S., Papadopoulou, Z., Psichas, N., Skoufas, D., Gissis, I., Sampanis,
T., et al. (2012). Breaking up prolonged sitting reduces postprandial glucose and M., et al. (2015). The rate of lactate removal after maximal exercise: the
insulin responses. Diabetes Care 35, 976–983. doi: 10.2337/dc11-1931 effect of intensity during active recovery. J. Sports Med. Phys. Fitness 55,
Dwyer, T. J., Alison, J. A., McKeough, Z. J., Elkins, M. R., and Bye, P. T. (2009). 1058–1063.
Evaluation of the sensewear activity monitor during exercise in cystic fibrosis Schuna, J. M. Jr., Johnson, W. D., and Tudor-Locke, C. (2013). Adult
and in health. Respir. Med. 103, 1511–1517. doi: 10.1016/j.rmed.2009.04.013 self-reported and objectively monitored physical activity and sedentary
Ekelund, U., Steene-Johannessen, J., Brown, W. J., Fagerland, M. W., Owen, N., behavior: NHANES 2005-2006. Int. J. Behav. Nutr. Phys. Act. 10:126.
Powell, K. E., et al. (2016). Does physical activity attenuate, or even eliminate, doi: 10.1186/1479-5868-10-126
the detrimental association of sitting time with mortality? A harmonised meta- Sedentary Behaviour Research Network (2012). Letter to the editor: standardized
analysis of data from more than 1 million men and women. Lancet 388, use of the terms “sedentary”and “sedentary behaviours”. Appl. Physiol. Nutr.
1302–1310. doi: 10.1016/S0140-6736(16)30370-1 Metab. 37, 540–542. doi: 10.1139/h2012-024
Fiskerstrand, A., and Seiler, K. S. (2004). Training and performance characteristics Sperlich, B., Born, D. P., Kaskinoro, K., Kalliokoski, K. K., and Laaksonen,
among Norwegian international rowers 1970-2001. Scand. J. Med. Sci. Sports M. S. (2013). Squeezing the muscle: compression clothing and muscle
14, 303–310. doi: 10.1046/j.1600-0838.2003.370.x metabolism during recovery from high intensity exercise. PLoS ONE 8:e60923.
Fruin, M. L., and Rankin, J. W. (2004). Validity of a multi-sensor armband in doi: 10.1371/journal.pone.0060923
estimating rest and exercise energy expenditure. Med. Sci. Sports Exerc. 36, Sperlich, B., and Holmberg, H. (2017). The responses of elite athletes to
1063–1069. doi: 10.1249/01.MSS.0000128144.91337.38 exercise: an all-day, 24-hour integrative view is required! Front. Physiol. 8:564.
Kravitz, L., and Vella, C. A. (2016). Reducing Sedentary Behaviors: Sit Less doi: 10.3389/fphys.2017.00564
and Move More Available online at: https://www.acsm.org/docs/default- Sperlich, B., and Holmberg, H. C. (2017). Wearable, yes, but able...?: it is time
source/brochures/reducing-sedentary-behaviors-sit-less-and-move-more.pdf for evidence-based marketing claims! [Letter]. Br. J. Sports Med. 51, 1240.
(Accessed June 02, 2017). doi: 10.1136/bjsports-2016-097295
Laursen, P. B., and Jenkins, D. G. (2002). The scientific basis for high- Steinacker, J., Liu, Y., and Reißnecker, S. (2002). Termination criteria in ergometry
intensity interval training: optimising training programmes and maximising [Article in German language]. Dtsch. Z. Sportmed. 53, 228–229.
performance in highly trained endurance athletes. Sports Med. 32, 53–73. Stoggl, T. L., and Sperlich, B. (2015). The training intensity distribution
doi: 10.2165/00007256-200232010-00003 among well-trained and elite endurance athletes. Front. Physiol. 6:295.
Martin, N. A., Zoeller, R. F., Robertson, R. J., and Lephart, S. M. (1998). The doi: 10.3389/fphys.2015.00295
comparative effects of sports massage, active recovery, and rest in promoting Tessitore, A., Meeusen, R., Pagano, R., Benvenuti, C., Tiberi, M., and
blood lactate clearance after supramaximal leg exercise. J. Athl. Train. 33:30. Capranica, L. (2008). Effectiveness of active versus passive recovery
McLellan, T. M., Pasiakos, S. M., and Lieberman, H. R. (2014). Effects of strategies after futsal games. J. Strength Cond. Res. 22, 1402–1412.
protein in combination with carbohydrate supplements on acute or repeat doi: 10.1519/JSC.0b013e31817396ac
endurance exercise performance: a systematic review. Sports Med. 44, 535–550. Tremblay, M. S., Aubert, S., Barnes, J. D., Saunders, T. J., Carson, V., Latimer-
doi: 10.1007/s40279-013-0133-y Cheung, A. E., et al. (2017). Sedentary Behavior Research Network (SBRN)-
Medbo, J. I., Mamen, A., Welde, B., von Heimburg, E., and Stokke, R. terminology consensus project process and outcome. Int. J. Behav. Nutr. Phys.
(2002). Examination of the metamax, I. and II oxygen analysers during Act. 14:75. doi: 10.1186/s12966-017-0525-8
exercise studies in the laboratory. Scand. J. Clin. Lab. Invest. 62, 585–598. Wahl, P., Sanno, M., Ellenberg, K., Frick, H., Bohm, E., Haiduck, B., et al.
doi: 10.1080/003655102764654321 (2017). Aqua cycling does not affect recovery of performance, damage
Mujika, I. (2012). The cycling physiology of Miguel Indurain 14 years after markers, and sensation of pain. J. Strength Cond. Res. 31, 162–170.
retirement. Int. J. Sports Physiol. Perform. 7, 397–400. doi: 10.1123/ijspp.7.4.397 doi: 10.1519/JSC.0000000000001462
Nikolaidis, P. T. (2012). Overweight and obesity in male adolescent soccer players. Weiler, R., Aggio, D., Hamer, M., Taylor, T., and Kumar, B. (2015).
Minerva Pediatr. 64, 615–622. Sedentary behaviour among elite professional footballers: health and
Nikolaidis, P. T. (2013). Body mass index and body fat percentage are associated performance implications. BMJ Open Sport Exerc. Med. 1:e000023.
with decreased physical fitness in adolescent and adult female volleyball players. doi: 10.1136/bmjsem-2015-000023
J. Res. Med. Sci. 18, 22–26.
Owen, N., Salmon, J., Koohsari, M. J., Turrell, G., and Giles-Corti, B. (2014). Conflict of Interest Statement: The authors declare that the research was
Sedentary behaviour and health: mapping environmental and social contexts conducted in the absence of any commercial or financial relationships that could
to underpin chronic disease prevention. Br. J. Sports Med. 48, 174–177. be construed as a potential conflict of interest.
doi: 10.1136/bjsports-2013-093107
Pelliccia, A., Adami, P. E., Quattrini, F., Squeo, M. R., Caselli, S., Verdile, L., Copyright © 2017 Sperlich, Becker, Hotho, Wallmann-Sperlich, Sareban, Winkert,
et al. (2017). Are Olympic athletes free from cardiovascular diseases? Systematic Steinacker and Treff. This is an open-access article distributed under the terms
investigation in 2352 participants from Athens 2004 to Sochi 2014. Br. J. Sports of the Creative Commons Attribution License (CC BY). The use, distribution or
Med. 51, 238–243. doi: 10.1136/bjsports-2016-096961 reproduction in other forums is permitted, provided the original author(s) or licensor
Poppendieck, W., Faude, O., Wegmann, M., and Meyer, T. (2013). Cooling and are credited and that the original publication in this journal is cited, in accordance
performance recovery of trained athletes: a meta-analytical review. Int. J. Sports with accepted academic practice. No use, distribution or reproduction is permitted
Physiol. Perform. 8, 227–242. doi: 10.1123/ijspp.8.3.227 which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 93 September 2017 | Volume 8 | Article 655

 OPINION
published: 14 August 2017
doi: 10.3389/fphys.2017.00564

The Responses of Elite Athletes to

Exercise: An All-Day, 24-h Integrative
View Is Required!
Billy Sperlich 1* and Hans-Christer Holmberg 2, 3, 4
1
Integrative and Experimental Exercise Science, Institute for Sport Sciences, University of Würzburg, Würzburg, Germany,
2
Swedish Winter Sports Research Centre, Mid Sweden University, Östersund, Sweden, 3 School of Sport Sciences, UiT The
Arctic University of Norway, Tromsø, Norway, 4 School of Kinesiology, University of British Columbia, Vancouver, BC, Canada

Keywords: wearable sensors, training intensity distribution, monitoring, physiological, athletes, biofeedback

The current research topics in Frontiers of Physiology include “Training intensity, volume and
recovery distribution among elite and recreational endurance athletes” (Frontiers in Physiology,
2016) and “Wearable Sensor Technology for Monitoring Training Load and Health in the Athletic
Population” (Frontiers in Physiology, 2017). As editors of both of these topics, we would like to
share some thoughts concerning (a) how they are fundamentally linked and (b) why we believe it is
essential to have an all-day, 24-h integrative view to understand elite athletes’ responses to exercise.
Athletes who train frequently each week schedule their training and off-training for days
(i.e., microcycles, for example, tapering periods, blocks of training) to as long as months (i.e.,
macrocycles, for example, periods of preparation with different focuses or training camps) to
ensure progressive adaptation and prevent fatigue, boredom, and injury. From this perspective,
a fundamental goal is to distribute exercise and off-training effectively over a certain period of time
(for example, one or several seasons) to achieve optimal adaptation.
Here, we highlight the importance of an all-day, 24-h integrative perspective on training,
Edited by:
James (Jim) David Cotter,
emphasizing the fact that conditions outside training significantly modulate adaptation, thereby
University of Otago, New Zealand complicating analysis of the distribution of training intensity.
Elite athletes invest a significant amount of time per year in their training, which in some sports
Reviewed by:
Thierry Busso, amounts to approximately 17% of their waking time (Fiskerstrand and Seiler, 2004). This means
University of Saint-Etienne, France that the remaining 83% is spent on activities such as recovery, including for example, massage,
*Correspondence:
physiotherapy, medical treatments, eating, as well as activities of daily living (including sitting,
Billy Sperlich lying, working, studying, active, and passive transportation) and social engagements (for example,
[email protected] media, sponsor, and family activities). All of these activities modulate psycho-biological responses
to training.
Specialty section: The various approaches for improving recovery include massage (Poppendieck et al., 2016),
This article was submitted to cooling (Poppendieck et al., 2013), stretching and self-myofascial release (Beardsley and Skarabot,
Exercise Physiology, 2015), neuromuscular electrical stimulation (Babault et al., 2011), compression attire (Born et al.,
a section of the journal
2013), active recovery (Laursen and Jenkins, 2002; Buchheit et al., 2009; Riganas et al., 2015), and
Frontiers in Physiology
many more, and most of these modalities are performed for relatively short periods of time (from
Received: 30 June 2017 minutes to perhaps 1–2 h or longer) and usually soon after training. Most of these are designed to
Accepted: 19 July 2017
improve the delivery of oxygen and substrates to muscles and the clearance of metabolites, thereby
Published: 14 August 2017
attenuating or delaying the onset of muscle soreness and rapidly restoring homeostasis, through
Citation: regulation of growth and transcription factors (Coffey and Hawley, 2007). Clearly, recovery must
Sperlich B and Holmberg H-C (2017)
be taken into account when evaluating the different responses of elite athletes to exercise.
The Responses of Elite Athletes to
Exercise: An All-Day, 24-h Integrative
Since moderate-to-light activity (for example, walking or cycling) enhances muscle blood flow,
View Is Required! it is surprising that we can find no studies on the influence of physical activity off-training on
Front. Physiol. 8:564. the biological and psychological outcomes of exercise, especially since the athletic population is
doi: 10.3389/fphys.2017.00564 alarmingly sedentary when not training (Weiler et al., 2015).

Frontiers in Physiology | www.frontiersin.org 94 August 2017 | Volume 8 | Article 564

Sperlich and Holmberg 24-h Integrative View

In addition, other factors such as sleep (Nedelec et al., Thus, it appears virtually impossible to take all of these
2015; Gupta et al., 2017) and nutrition (Thomas et al., factors into consideration when studying a homogenous group
2016) are both influenced by the stress of training and, of elite athletes, not even in a controlled laboratory setting.
in turn, modulate the response to training in a significant However, both retro- and prospective analyses on the responses
fashion. In addition, drug abuse by and/or frequent and adaptation to training should provide as much information
medication of athletes may result in epigenic changes and about such modulators as possible. In this context, we feel
consequently influence physiological adaptation (Kanherkar that a combination of wearable technology and smartphone-
et al., 2014). It seems imperative that such factors also be based applications should prove invaluable, since this is the only
taken into account when comparing different approaches to technology that currently allows as much information as possible
training. to be obtained by continuous 24-h monitoring of, in addition
The busy schedules of (elite) athletes involve a relatively high to the internal and external training loads themselves, sleep,
level of psycho-biological stress, due for example, to frequent traveling, various environmental conditions and psycho-social
traveling for short and long distances, often across time zones status. As long as scientific quality is maintained (Duking et al.,
(Kölling et al., 2016; Fowler et al., 2017), which detracts from 2016; Sperlich and Holmberg, 2017) and personal data protected,
preparedness for subsequent training and competition. More such technology can potentially provide 24-h feedback (Duking
understanding is required here as well. et al., 2017) to the athlete and supporting staff concerning the
Moreover, environmental factors, such as exposure to an various psycho-biological responses to training. In this regards,
elevated (Sperlich et al., 2017) or lowered level of oxygen future findings on “Wearable Sensor Technology for Monitoring
(Girard et al., 2017), variations in temperature (Lorenzo et al., Training Load and Health in the Athletic Population” (Frontiers
2010; Kruger et al., 2015), and atmospheric stressors such as in Physiology, 2017) will hopefully help provide innovative
ozone, particulate matter (Giles and Koehle, 2014), and ultra- approaches to investigating the “Training intensity, volume and
violet radiation, exert an impact on various tissues of the recovery distribution among elite and recreational endurance
human body and thereby potentially modulate responses to athletes.”
training. Accordingly, such factors should also be considered
when judging the responses of elite athletes to exercise. AUTHOR CONTRIBUTIONS
In addition, psycho-social stress resulting from, for example,
media exposure, financial and family concerns, fans, and/or one’s All authors listed have made a substantial, direct and intellectual
own expectations may well influence responses to training. contribution to the work, and approved it for publication.

REFERENCES Girard, O., Brocherie, F., and Millet, G. P. (2017). Effects of altitude/hypoxia on
single- and multiple-sprint performance: a comprehensive review. Sports Med.
Babault, N., Cometti, C., Maffiuletti, N. A., and Deley, G. (2011). Does electrical doi: 10.1007/s40279-017-0733-z. [Epub ahead of print].
stimulation enhance post-exercise performance recovery? Eur. J. Appl. Physiol. Gupta, L., Morgan, K., and Gilchrist, S. (2017). Does elite sport degrade sleep
111, 2501–2507. doi: 10.1007/s00421-011-2117-7 quality? A systematic review. Sports Med. 47, 1317–1333. doi: 10.1007/s40279-
Beardsley, C., and Skarabot, J. (2015). Effects of self-myofascial release: a systematic 016-0650-6
review. J. Bodyw. Mov. Ther. 19, 747–758. doi: 10.1016/j.jbmt.2015.08.007 Kanherkar, R. R., Bhatia-Dey, N., and Csoka, A. B. (2014). Epigenetics across the
Born, D. P., Sperlich, B., and Holmberg, H. C. (2013). Bringing light into the dark: human lifespan. Front. Cell Dev. Biol. 2:49. doi: 10.3389/fcell.2014.00049
effects of compression clothing on performance and recovery. Int. J. Sports Kölling, S., Treff, G., Winkert, K., Ferrauti, A., Meyer, T., Pfeiffer, M., et al. (2016).
Physiol. Perform. 8, 4–18. doi: 10.1123/ijspp.8.1.4 The effect of westward travel across five time zones on sleep and subjective
Buchheit, M., Cormie, P., Abbiss, C. R., Ahmaidi, S., Nosaka, K. K., and jet-lag ratings in athletes before and during the 2015’s World Rowing Junior
Laursen, P. B. (2009). Muscle deoxygenation during repeated sprint running: Championships. J. Sports Sci. doi: 10.1080/02640414.2016.1265141. [Epub
effect of active vs. passive recovery. Int. J. Sports Med. 30, 418–425. ahead of print].
doi: 10.1055/s-0028-1105933 Kruger, M., de Marees, M., Dittmar, K. H., Sperlich, B., and Mester, J.
Coffey, V. G., and Hawley, J. A. (2007). The molecular bases of training adaptation. (2015). Whole-body cryotherapy’s enhancement of acute recovery of running
Sports Med. 37, 737–763. doi: 10.2165/00007256-200737090-00001 performance in well-trained athletes. Int. J. Sports Physiol. Perform. 10,
Duking, P., Holmberg, H. C., and Sperlich, B. (2017). Instant biofeedback provided 605–612. doi: 10.1123/ijspp.2014-0392
by wearable sensor technology can help to optimize exercise and prevent injury Laursen, P. B., and Jenkins, D. G. (2002). The scientific basis for high-
and overuse. Front. Physiol. 8:167. doi: 10.3389/fphys.2017.00167 intensity interval training: optimising training programmes and maximising
Duking, P., Hotho, A., Holmberg, H. C., Fuss, F. K., and Sperlich, B. (2016). performance in highly trained endurance athletes. Sports Med. 32, 53–73.
Comparison of non-invasive individual monitoring of the training and health doi: 10.2165/00007256-200232010-00003
of athletes with commercially available wearable technologies. Front. Physiol. Lorenzo, S., Halliwill, J. R., Sawka, M. N., and Minson, C. T. (2010). Heat
7:71. doi: 10.3389/fphys.2016.00071 acclimation improves exercise performance. J. Appl. Physiol. (1985) 109,
Fiskerstrand, A., and Seiler, K. S. (2004). Training and performance characteristics 1140–1147. doi: 10.1152/japplphysiol.00495.2010
among Norwegian international rowers 1970-2001. Scand. J. Med. Sci. Sports Nedelec, M., Halson, S., Abaidia, A. E., Ahmaidi, S., and Dupont, G. (2015). Stress,
14, 303–310. doi: 10.1046/j.1600-0838.2003.370.x sleep and recovery in elite soccer: a critical review of the literature. Sports Med.
Fowler, P. M., McCall, A., Jones, M., and Duffield, R. (2017). Effects of 45, 1387–1400. doi: 10.1007/s40279-015-0358-z
long-haul transmeridian travel on player preparedness: case study of a Frontiers in Physiology (2016). Training Intensity, Volume and Recovery
national team at the 2014 FIFA World Cup. J. Sci. Med. Sport 20, 322–327. Distribution among Elite and Recreational Endurance Athletes. Available
doi: 10.1016/j.jsams.2016.08.021 online at: http://journal.frontiersin.org/researchtopic/4444/training-
Giles, L. V., and Koehle, M. S. (2014). The health effects of exercising in air intensity-volume-and-recovery-distribution-among-elite-and-recreational-
pollution. Sports Med. 44, 223–249. doi: 10.1007/s40279-013-0108-z endurance-athletes

Frontiers in Physiology | www.frontiersin.org 95 August 2017 | Volume 8 | Article 564

Sperlich and Holmberg 24-h Integrative View

Frontiers in Physiology (2017). Wearable Sensor Technology for Monitoring Thomas, D. T., Erdman, K. A., and Burke, L. M. (2016). American College of
Training Load and Health in the Athletic Population [Online]. Available Sports Medicine Joint Position Statement. Nutrition and Athletic Performance.
online at: http://journal.frontiersin.org/researchtopic/5516/wearable-sensor- Med. Sci. Sports Exerc. 48, 543–568. doi: 10.1249/MSS.00000000000
technology-for-monitoring-training-load-and-health-in-the-athletic- 00852
population [accessed]. Weiler, R., Aggio, D., Hamer, M., Taylor, T., and Kumar, B. (2015).
Poppendieck, W., Faude, O., Wegmann, M., and Meyer, T. (2013). Cooling and Sedentary behaviour among elite professional footballers: health and
performance recovery of trained athletes: a meta-analytical review. Int. J. Sports performance implications. BMJ Open Sport Exerc. Med. 1:e000023.
Physiol. Perform. 8, 227–242. doi: 10.1136/bmjsem-2015-000023
Poppendieck, W., Wegmann, M., Ferrauti, A., Kellmann, M., Pfeiffer, M., and
Meyer, T. (2016). Massage and performance recovery: a meta-analytical review. Conflict of Interest Statement: The authors declare that the research was
Sports Med. 46, 183–204. doi: 10.1007/s40279-015-0420-x conducted in the absence of any commercial or financial relationships that could
Riganas, C. S., Papadopoulou, Z., Psichas, N., Skoufas, D., Gissis, I., Sampanis, M., be construed as a potential conflict of interest.
et al. (2015). The rate of lactate removal after maximal exercise: the effect of
intensity during active recovery. J. Sports Med. Phys. Fitness 55, 1058–1063. Copyright © 2017 Sperlich and Holmberg. This is an open-access article distributed
Sperlich, B., and Holmberg, H.-C. (2017). Wearable, yes, but able...?: it is under the terms of the Creative Commons Attribution License (CC BY). The
time for evidence-based marketing claims! Br. J. Sports Med. 51:1240. use, distribution or reproduction in other forums is permitted, provided the
doi: 10.1136/bjsports-2016-097295 original author(s) or licensor are credited and that the original publication in
Sperlich, B., Zinner, C., Hauser, A., Holmberg, H. C., and Wegrzyk, J. (2017). The this journal is cited, in accordance with accepted academic practice. No use,
impact of hyperoxia on human performance and recovery. Sports Med. 47, distribution or reproduction is permitted which does not comply with these
429–438. doi: 10.1007/s40279-016-0590-1 terms.

Frontiers in Physiology | www.frontiersin.org 96 August 2017 | Volume 8 | Article 564

 ORIGINAL RESEARCH
published: 02 August 2017
doi: 10.3389/fphys.2017.00562

High Intensity Interval Training Leads

to Greater Improvements in Acute
Heart Rate Recovery and Anaerobic
Power as High Volume Low Intensity
Training
Thomas L. Stöggl 1* and Glenn Björklund 2, 3
1
Department of Sport Science and Kinesiology, University of Salzburg, Salzburg, Austria, 2 Department of Health Sciences,
Sports Tech Research Centre, Mid Sweden University, Östersund, Sweden, 3 The Swedish Sports Confederation, Stockholm,
Sweden

The purpose of the current study was to explore if training regimes utilizing diverse training
intensity distributions result in different responses on neuromuscular status, anaerobic
capacity/power and acute heart rate recovery (HRR) in well-trained endurance athletes.
Methods: Thirty-six male (n = 33) and female (n = 3) runners, cyclists, triathletes
Edited by: and cross-country skiers [peak oxygen uptake: (VO2peak ): 61.9 ± 8.0 mL·kg−1 ·min−1 ]
Luca Paolo Ardigò,
University of Verona, Italy
were randomly assigned to one of three groups (blocked high intensity interval
Reviewed by:
training HIIT; polarized training POL; high volume low intensity oriented control group
Alessandro Moura Zagatto, CG/HVLIT applying no HIIT). A maximal anaerobic running/cycling test (MART/MACT)
Universidade Estadual Paulista Júlio was performed prior to and following a 9-week training period.
Mesquita Filho, Brazil
Beat Knechtle, Results: Only the HIIT group achieved improvements in peak power/velocity (+6.4%, P
University of Zurich, Switzerland
< 0.001) and peak lactate (P = 0.001) during the MART/MACT, while, unexpectedly, in
Juliano Dal Pupo,
Universidade Federal de Santa none of the groups the performance at the established lactate concentrations (4, 6, 10
Catarina, Brazil mmol·L−1 ) was changed (P > 0.05). Acute HRR was improved in HIIT (11.2%, P = 0.002)
*Correspondence: and POL (7.9%, P = 0.023) with no change in the HVLIT oriented control group.
Thomas L. Stöggl
[email protected] Conclusion: Only a training regime that includes a significant amount of HIIT improves
the neuromuscular status, anaerobic power and the acute HRR in well-trained endurance
Specialty section:
This article was submitted to
athletes. A training regime that followed more a low and moderate intensity oriented
Exercise Physiology, model (CG/HVLIT) had no effect on any performance or HRR outcomes.
a section of the journal
Frontiers in Physiology Keywords: lactate threshold, peak power, maximal anaerobic running test, mart, MACT, training intensity
distribution
Received: 10 May 2017
Accepted: 18 July 2017
Published: 02 August 2017
INTRODUCTION
Citation:
Stöggl TL and Björklund G (2017) Endurance athletes use different training strategies to improve their performance. Although the
High Intensity Interval Training Leads
bulk of training sessions typically are made up of longer and at slower paced session’s (Tonnessen
to Greater Improvements in Acute
Heart Rate Recovery and Anaerobic
et al., 2014), intervals and higher exercise intensity sessions are a necessity for high performance.
Power as High Volume Low Intensity To implement diverse types of exercise intensities and durations, athletes use various ways of
Training. Front. Physiol. 8:562. periodization in their training. The type of periodization depends on the sport and the length of the
doi: 10.3389/fphys.2017.00562 competition season which decides the duration of the training period. The most common ways for

Frontiers in Physiology | www.frontiersin.org 97 August 2017 | Volume 8 | Article 562

Stöggl and Björklund HIIT Improves Heart Rate Recovery

periodization is (i) high volume low-intensity training (HVLIT), components of endurance performance were shown previously
lactate threshold training (THR), low-volume high-intensity (Stöggl and Sperlich, 2014), while the effects of these training
(interval) training (HIIT) and polarized training, a concept concepts on anaerobic power and HRR are lacking.
consisting of mixing training between low and high intensity The aim of the study was to evaluate the effects of different
(POL) or a gradual decrease in training volume from HVLIT to training concepts (POL vs. HIIT vs. HVLIT) with respect
THR and HIIT in a pyramidal fashion (Stöggl and Sperlich, 2015) to anaerobic power, cardiovascular and metabolic response
Athletes use various tests for evaluation of the training process using key measurements during the MART and HRR. We
to accurately target specific elements that are important for hypothesized that athletes who use training concepts involving
performance. Depending on the sport, different quality’s such as high intensity elements, i.e., HIIT and POL, would display
cardiovascular, muscular, or metabolic variables are key elements superior improvements compared with athletes that use no HIIT
for performance to a various degree. The most common way for (i.e., HVLIT).
endurance athletes to track and overview their daily training is
the use of heart rate (HR) monitoring (Achten and Jeukendrup, MATERIALS AND METHODS
2003). One of the most significant cardiovascular enhancements
is an increased stroke volume which is associated with a lower Participants
exercise HR for a given submaximal work (Blomqvist and Thirty-six competitive endurance athletes (three females and 33
Saltin, 1983). Furthermore, HR recovery (HRR) after cessation males) who participated in either cross-country skiing, cycling,
of exercise has been put forward as a useful indicator to track triathlon, middle- or long-distance running volunteered to take
cardiovascular advances for athletes of various levels (Daanen part in this study (mean ± SD: age: 31 ± 6 yrs, body mass:
et al., 2012). As an example, HRR was tracked in well-trained 74.6 ± 8.9 kg, height: 180 ± 7 cm) were recruited from regional
cyclists throughout a training period of 4 weeks constituting 8 cycling, running, triathlon, athletic, and cross-country skiing
× 4 min HIIT two times per week (Lamberts et al., 2009). It clubs. All participants were well-trained athletes [61.9 ± 8.0
was concluded that well-trained athletes, who responded well mL·kg−1 ·min−1 (range: 54–75 mL·kg−1 ·min−1 )], accustomed
to this type of training, demonstrated a faster HRR after the to a training frequency of more than five sessions per week
interval session and after a 40-km time trial (HR drop during (totally 10–20 h·wk−1 ), participated frequently in endurance
the 60 s post-exercise) that further was related to an enhanced competitions for the last 8–20 years and were healthy throughout
endurance performance. To note, that no other training- or the intervention period. Participants were members or former
control-group was included in this study. Additionally, HRR members of the Austrian cross-country skiing national team (n =
relating to both the drop and the time to reach a certain beat per 8), runners and triathletes (n = 10) or cyclists (n = 13) of regional
minute, responds differently to various forms of interval training sport teams during or since the year before the current study.
regimes (Buchheit et al., 2008). This indicates that HRR could Retrospective analysis of the 6 months training prior to the study
provide beneficial feedback for cardiovascular adjustments, not revealed that none of the participants had regularly engaged
only during exercise, but also in between exercise bouts. HIIT. Instead all had used a HVLIT training protocol with a
Another frequent used tool for determination of physiological maximum of two THR training sessions per week.
response is determination of blood lactate concentration. Most Based on the participants’ baseline VO2max and training mode
often blood lactate is related to an increase in exercise intensity (running or cycling), all athletes were parallelized into three
to identify various lactate thresholds (Beneke et al., 2011). groups: HIIT, POL, and control group (CG; HVLIT oriented with
During variable intensity exercise blood lactate recovery has 1–2 THR sessions per week). At baseline, the three groups were
been suggested as a good indicator for performance both for not statistically different with regard to age, height, body mass, or
cycling and cross-country skiing (Björklund et al., 2007, 2011). VO2max . During an initial visit, study details, and participation
These latter studies were conducted on well-trained athletes. requirements were explained, and all participants gave written
Also, different types of training seem to stimulate the lactate informed consent. The study and protocol received approval
removal abilities as middle distance runners surpass sprint from the University of Salzburg Austria Ethics Committee and
runners regarding lactate recovery in between high intensity was conducted in accordance with the Declaration of Helsinki.
bouts (Bret et al., 2003).
Performance tests with intermittent character are a common Design
and valid instrument to explain performance progress that do not The intervention lasted 9 weeks plus 2 days of pre- and post-
relate to aerobic characteristics. For evaluation of the athletes’ testing. All athletes who were mainly engaged in cycling training
neuromuscular status and anaerobic power maximal anaerobic during the intervention period trained with their own bike and
treadmill tests are a useful assessment, i.e., MART (Paavolainen completed all tests on a bicycle ergometer (Ergoline, Ergoselect
et al., 1999b; Nummela et al., 2006). The test is of intermittent 100P; Bitz, Germany) using their own cycling shoes and pedal
character with an increase in exercise intensity for consecutive system. Other athletes ran during the study and completed their
bouts and is terminated at volitional fatigue. This test relates to pre- and post-testing on a motorized treadmill (HP Cosmos,
the individuals anaerobic input to the exercise performance and Saturn, Traunstein, Germany). All participants were instructed
could therefore provide an estimate which areas the athletes lack not to change their diet throughout the training period and
and need to improve for enhanced performance. To note here, to maintain strength training, if it was part of their training
that the effects of different endurance training concepts on key program. Participants’ nutritional intake was not standardized

Frontiers in Physiology | www.frontiersin.org 98 August 2017 | Volume 8 | Article 562

Stöggl and Björklund HIIT Improves Heart Rate Recovery

or controlled during the study, but for the 3 h prior to all on the protocol of Rusko et al. (1993) in running and Tossavainen
testing in which food intake was not permitted. The training et al. (1996) in cycling.
intensity was controlled by HR based on the baseline incremental On the first test day all athletes completed a VO2max ramp
test: (i) low intensity training (LIT, HR at blood lactate value protocol to determine maximal oxygen uptake (VO2max ) and
<2 mmol·L−1 ); (ii) moderate intensity training (MIT, HR maximal HR (HRmax ). First, the workload for running was set
corresponding to a blood lactate of 3–5 mmol·L−1 ); (iii) high at 8 km·h−1 (inclination: 5%) on the treadmill, and for cycling
intensity interval training (HIIT, >90% HRmax )] (e.g., Seiler, at 200 W with a cadence of >80 rpm for 10 min. The workload
2010; Stöggl and Sperlich, 2015). The HR was measured during was then increased every 30 s by 0.5 km·h−1 (inclination:
each training session and athletes documented training mode, 10%) on the treadmill or 15 W on the cycle ergometer until
exercise duration and intensity in a diary. As a control and exhaustion. VO2 was measured with an open circuit breath-by-
for detailed analysis, HR for all training sessions was stored breath spirograph (nSpire, Zan 600 USB, Oberthulba, Germany),
digitally and analyzed retrospectively. For the quantification of which was calibrated prior to each test using high precision gas
the training intensity distribution within the 9-weeks of training (15.8% O2 , 5% O2 in N; Praxair, Düsseldorf, Germany) and a 1 L
the session goal approach according to Seiler and Kjerland (2006) syringe (nSpire, Oberthulba, Germany). All respiratory data were
was applied. averaged every 30 s.
On the second day athletes performed the MART/MACT.
HIIT Intervention The protocol included stages of 25 s (running) or 30 s (cycling;
The HIIT included two interval blocks of 16 days with one including 3–5 s acceleration time) with 100 s breaks in between.
adaptation week prior to and one recovery week after each block. For the running protocol treadmill speed was increased with
The adaptation week included two 60 min HIIT sessions, three 1.4 km·h−1 increments starting at 14.7 km·h−1 on a grade of
90 min LIT sessions, one 120 min LIT session and 1 day of 7%. For the cycling protocol the test started at 360 W with
recovery. The condensed 16 day interval block included 12 HIIT increments of 40 W. Maximal performance (Vmax ) in the MART
sessions within 15 days, integrating four blocks of three HIIT was calculated by linear interpolation using the formula: Vmax
sessions for 3 consecutive days followed by 1 day of recovery. The = Vf + ((t/25) 1.4 km·h−1 ), where Vf was the velocity of
recovery week contained four LIT sessions of 90 min and 3 days the last completed workload (km·h−1 ), t the duration of the
without any training. All of the HIIT sessions included a 20 min last workload (s) and 1.4 m·s−1 the velocity difference (1V)
warm-up at 75% of HRmax , 4 × 4 min at 90–95% of HRmax with between the last two workloads. For the MACT, the formula
3 min active recovery and a 15 min cool-down at 75% HRmax for maximal power output (Pmax ) was: Pmax = Pf + ((t/30)
based on the protocol proposed earlier (Helgerud et al., 2007). · 40 W), with Pf as the power output of the last completed
The LIT sessions lasted 90–150 min depending on the training stage. A 20 µl blood sample from the right earlobe was collected
mode (running vs. cycling) at an intensity resulting blood lactate within the 60 s of each 100 s rest period, and in the first, third,
of <2 mmol·L−1 . fifth and seventh minutes after the end of the last stage into a
capillary tube (Eppendorf AG, Hamburg, Germany). All samples
POL Intervention were analyzed amperometric-enzymatically (Biosen 5140, EKF-
The POL included three blocks, each lasting 3 weeks: 2 weeks diagnostic GmbH, Magdeburg, Germany) in duplicate, and the
of high volume and intensity training followed by 1 week of mean of the two measures was used for statistical analysis. The
recovery. The high volume and intensity week included six lactate sensor was calibrated before each test using a lactate
sessions with two 60 min HIIT sessions, two 150–240 min long standard sample of 12 mmol·L−1 . Results within a range of ±0.1
duration LIT sessions (duration according to training mode: mmol·L−1 were accepted. Velocity/power output at 4, 6, and 10
cycling, running or roller skiing), which included six to eight mmol·L−1 of blood lactate were calculated. HR recovery (HRR)
maximal sprints of 5 s separated by at least 20 min, and two was calculated as the mean value of all delta changes of each stages
90 min LIT sessions. The recovery week included one 60 min peak HR (highest value at the end or in the first seconds after the
HIIT session, one 120–180 min LIT session and one 90 min LIT end of the stage) and minimal HR (minimum value during the
session. 100 s break; Figure 1).

Control Group (CG/HVLIT) Statistical Analyses

The CG continued their HVLIT dominated training regime with All data exhibited a Gaussian distribution verified by the
a maximum of two THR sessions per week with no HIIT sessions. Shapiro–Wilk’s test and, accordingly, the values are presented
The control group also had three blocks each lasting 3 weeks with as means ± SD. Two-way 2 × 3 repeated-measures ANOVA
2 weeks of high-volume training followed by 1 week of recovery. (2 times: pre-post, 3 groups) to test for main effects of time
(pre- and post-intervention), group (the three training groups)
Pre- and Post-testing and the interaction effect between both factors was applied.
All participants were asked to report well-hydrated and to refrain When a significant main effect over time and/or interaction
from consuming alcohol and caffeine for at least 24-h, as well as effect was observed, paired t-tests within each group were
from engaging in strenuous exercise at least 48-h prior to testing. conducted. Based on the different units of peak velocity/power
The pre- and post-tests included a VO2max ramp protocol and in the MART/MACT percent changes between pre- to post-
a maximal anaerobic running/cycling test (MART/MACT) based values were calculated, and a one-way ANOVA between groups

Frontiers in Physiology | www.frontiersin.org 99 August 2017 | Volume 8 | Article 562

Stöggl and Björklund HIIT Improves Heart Rate Recovery

FIGURE 1 | Illustration of the heart rate-time curve and the heart rate recovery calculation within the MART/MACT of one subject.

was performed using Tukey’s post-hoc analysis. Furthermore, time and interaction effect time × group (both P = 0.001) with
within group changes for these variables were calculated using HIIT demonstrating the greatest increase (6.4 ± 3.4%, P < 0.001)
Wilcoxon tests. In addition, the values obtained were evaluated with no significant change in POL (0.2 ± 5.9%, P = 0.63) and
further by calculating the effect size (p η2 ). The magnitude of the CG/HVLIT (4.7 ± 5.5%, P = 0.087).
difference was classified as trivial (<0.01), small (0.01 ≤ to < For HRR there was a main effect of time (P < 0.001) and
0.06), moderate (0.06 ≤ to < 0.14) or large (≥0.14). An alpha interaction effect time × group (P = 0.011) with HIIT (38.7
value of <0.05 was considered significant. The Statistical Package ± 10.7 to 49.9 ± 14.1 bpm, 11.2%, P = 0.002) and POL (48.9
for the Social Sciences (Version 24.0; SPSS Inc., Chicago, IL, ± 15.9 to 56.8 ± 22.0 bpm, 7.9%, P = 0.023) demonstrating
USA) and Office Excel 2010 (Microsoft Corporation, Redmond, greater increases compared with unchanged levels of 0.1% in
WA, USA) were used for statistical analysis. CG/HVLIT (49.3 ± 7.5 to 49.4 ± 9.3 bpm, P > 0.05) (Figure 2).
All significant main and interaction effects demonstrated large
effect sizes (>0.14).
RESULTS LApeak demonstrated a time × group interaction effect (P =
0.027) with a 7.3% (P = 0.001) increase in HIIT and non-
Thirty-one participants completed the 9-week training protocol,
significantly changed values of −6.6% in POL and +1.3% in
fulfilling more than 95% of the training program and staying
CG/HVLIT (both, P > 0.05).
within the given HR zones. Seven subjects (2 in HIIT and 4
No changes from pre to post and no differences between
in CG) withdrew from the study due to illness (n = 2) or
training groups were detected with respect to HRpeak and velocity
were excluded due to changes in competition schedule (n = 2).
The total training hours, number of training sessions and their /power at 4, 6, and 10 mmol·L−1 blood lactate (all P > 0.05).
percent distribution within LIT, MIT, and HIIT are presented in
Table 1. POL and CG/HVLIT had higher training volume (P < DISCUSSION
0.001) and number of trainings session (P = 0.041) compared
with HIIT. The training intensity distributions with respect to The major findings of the study were that (i) only the HIIT
LIT, MIT, and HIIT were 68/6/26% for POL, 43/0/57% for HIIT group improved their peak velocity or power output in the
and 64/35/1% for CG/HVLIT. HIIT demonstrated the lowest MART/MACT, (ii) HRR was faster in the HIIT and POL groups
number of LIT sessions and CG/HVLIT the highest number of compared with no change in the CG/HVLIT group, (iii) while no
MIT sessions with no difference between the two other groups. training intervention improved the velocity or power output at
HIIT sessions were greatest in HIIT followed by POL and finally the established lactate concentrations during the MART/MACT.
CG/HVLIT.
Percent changes in variables from pre- to post-training and Anaerobic Power
between the training concepts during the MART/MACT are One of the major findings of this training study were the
presented in Table 2. For P/Vpeak there was a main effect of enhanced P/Vpeak for the HIIT group, demonstrating the

Frontiers in Physiology | www.frontiersin.org 100 August 2017 | Volume 8 | Article 562

Stöggl and Björklund HIIT Improves Heart Rate Recovery

TABLE 1 | Volume and intensity training distribution within the 9-weeks training TABLE 2 | Per cent changes in velocity (V) and power (P) and at various lactate
intervention (excluding strength training). thresholds as well as peak velocity and power.

POL HIIT CG/HVLIT P-value POL HIIT CG/HVLIT P-value, Effect

size p η2
Total hours 104 ± 21 66 ± 1* 93 ± 13 <0.001
Number of sessions 54 ± 7 47 ± 1§ 54 ± 8 =0.041 V/P 4 (%) −1.6 ± 13.1 4.1 ± 9.6 3.2 ± 13.04 NS, 0.03

Number of LIT training sessions 37 ± 9 20 ± 1* 36 ± 15 =0.004 V/P 6 (%) 3.3 ± 13.8 1.8 ± 6.5 1.1 ± 8.3 NS, 0.03

Number of MIT training sessions 3±4 0±0 18 ± 9* <0.001 V/P 10 (%) 2.8 ± 9.6 0.1 ± 5.5 2.7 ± 7.5 NS, 0.03

Number of HIIT training sessions 14 ± 3* 27 ± 1* 0 ± 1* <0.001 V/Ppeak (%) 0.2 ± 5.9† 6.4 ± 3.4*** 4.7 ± 5.5 =0.033, 0.22

Percent LIT training sessions 68 ± 12% 43 ± 1%* 64 ± 20% =0.002 LApeak (%) −6.6 ± 13.3† 7.3 ± 4.7*** 1.3 ± 12.3 =0.030, 0.22

Percent MIT training sessions 6 ± 7% 0 ± 0% 35 ± 21%* <0.001 HRpeak (%) −0.8 ± 3.9 −0.5 ± 2.8 0.0 ± 3.1 NS, 0.01

Percent HIIT training sessions 26 ± 7%* 57 ± 1%* 1 ± 1%* <0.001 HRR (%) 7.9 ± 9.7* 11.2 ± 7.7** 0.1 ± 5.6‡ =0.011, 0.28

The values presented are means ± SD. P-values were obtained by one-way ANOVA (3 The values presented are means ± SD. P-values were obtained by one-way ANOVA (three
training groups). POL, polarized training group; HIIT, High intensity interval training group; training groups) calculated over the per cent differences between pre- to post-training
CG /HVLIT, control group with mainly high volume low intensity training; LIT, low intensity (representing the interaction effect time × group). POL, polarized training group; HIIT,
training; MIT, moderate intensity training; HIIT, high intensity interval training. *Different High intensity interval training group; CG/HVLIT, control group with focus on high volume
from all other groups. § Different from training group “CG/HVLIT.” training; V/P4 mmol·L-1 , velocity or power at 4 mmol·L-1 blood lactate; V/P6 mmol·L-1 ,
velocity or power at 6 mmol·L-1 blood lactate; V/P10 mmol·L-1 , velocity or power at 10
mmol·L-1 blood lactate; V/Ppeak , peak velocity or power in the MART/MACT; LApeak , peak
greatest increase (+6.4%), with no notable change in POL lactate during the test and within the first 7 min after end of the last completed stage;
and CG/HVLIT. Earlier data using cross-sectional comparisons HRpeak , peak heart rate value during the MART/MACT; HRR, mean heart rate recovery;
*p < 0.05, **p < 0.01, ***p < 0.001 significant difference within groups from pre- to post-
between different types of athletes in running (Nummela et al., training. † Significant different from HIIT group. ‡ Significant different to both other groups.
1996) or cross-country skiing (Stöggl and Müller, 2009) showed NS, not significant.
that P/Vmax of the MART was determined by metabolic variables
as peak lactate and power output or velocity at 10 mmol·L−1 .
The importance of these variables and especially the velocity at 10
mmol·L−1 have further been strengthened by the same research
group (Nummela et al., 2007). While this cross-sectional data
showed important features for anaerobic power reflected in the
MART, there was no quantification of the athletes training that
preceded the test. In the current study, the HIIT group was
the exclusive training modality that had a positive effect on the
MART/MACT performance. Indeed, the peak lactate and P/Vpeak
was increased in the HIIT group while it remained unchanged
in the other two groups (POL and CG/HVLIT). Certainly, a
greater glycolytic activity, which involves formation of lactate, is
favorable to produce ATP at a higher rate and likewise should add
to the overall performance in the MART/MACT. Therefore, the
HIIT intervention seems to have substantial impact on both the
FIGURE 2 | Per cent changes of mean heart rate recovery within the three
metabolic and neuromuscular components of maximal anaerobic training groups between pre- and post-intervention testing. *p < 0.05, **p <
performance. †
0.01 significant difference within groups from pre- to post-training. p < 0.05,
The relationship between the MART and the maximal ††
p < 0.01 significant different to CG/HVLIT group.
anaerobic oxygen deficit (MAOD) as well as the energy
contribution during the MART was already investigated by
Zagatto et al. (2011). Although, the relationship between The neuromuscular capacity to produce force was shown to
the MART and MAOD was poor, the quantification of the be related to both the MART and 5-k running performance
energetic contribution demonstrated that the aerobic input (Nummela et al., 2006). Furthermore, close relationships
covers the greatest amount of energy production during the between running performance from distances of 400–5,000 m
full test, i.e., including both the work and rest periods (65%), with performance in the MART and short-duration sprint
while the anaerobic glycolytic energy system contributed with performance over 20–30 m were found in numerous studies (e.g.,
approximately only 5%. When analyzing only the effort periods Rusko et al., 1993; Nummela et al., 1996, 2006, 2007; Paavolainen
(25 s) the anaerobic contribution corresponded to ∼74% with et al., 1999a,c). In this context, neuromuscular adaptations
the main energy system being the a-lactic (63%) and not the using HIIT in ice hockey players, improved the general muscle
glycolytic lactic system (11%). Moreover, it has been shown that activation as demonstrated in increased force and rate of force
the aerobic contribution increases already at the second repeated development in an isolated plantar flexion (Kinnunen et al.,
bout in sprint exercise (Bogdanis et al., 1996). Therefore, the 2017). Shortcomings in the study of Kinnunen et al. (2017) were
relation to peak lactate concentration and performance in the the missing transfer of the results to increased sport specific
MART/MACT within the current study seems conflicting. performance (e.g., faster skating times on ice). Compared to the

Frontiers in Physiology | www.frontiersin.org 101 August 2017 | Volume 8 | Article 562

Stöggl and Björklund HIIT Improves Heart Rate Recovery

current study, in the study by Kinnunen et al. (2017) a shorter regarding training hours, it has been shown that severe increase
training period (2.5 vs. 9 weeks) and shorter HIIT bouts (30 s in training hours per week markedly increases HRR along with a
vs. 4 min) were applied. The question remains which training concomitant loss of performance of a single time-trial (Thomson
period durations using different training regimes (e.g., HIIT et al., 2016). Notably in our study, the increase in training hours
protocols) are necessary for these improvements. While these was accompanied with an increase in percent of high intensity
shorter HIIT bouts relate more to team sports as ice hockey, there exercise which makes it difficult to pin point if it is hours or high
seem to be also positive effects for sprint triathlon performance intensity exercise that sole alone explain the outcome.
where subjects used a mix of short and long bouts (Garcia- Another aspect for the more pronounced HRR in the groups
Pinillos et al., 2017). In summary, different HIIT regimes have including high intensity sessions (HIIT and POL) might be the
been proven to have a significant impact on trained endurance intermittent character of HIIT itself. Possibly, the repeated steady
athletes for endurance performance (Laursen et al., 2002). In all, change of high and low intensities within the sessions might be
the improved MART/MACT performance in the current study an appropriate stimulus to enhance the ability of the autonomic
potentially translates to an improved performance in the athlete’s nervous system to acutely adapt toward changing intensities.
specific sport. Future studies would be of a necessity to cover this area to explore
the exact mechanisms.
Heart Rate Recovery (HRR) In all, the training groups that included HIIT (POL and HIIT)
The two training groups including high intensity sessions, HIIT both showed an increased HRR. While an enhanced HRR has
and POL, displayed a superior HRR during the MART/MACT been interpreted to be part of a functional overreaching with
compared with CG/HVLIT. In detail, HIIT demonstrated the a decreased performance, the short tapering period (days) in
greatest per cent change in HRR (11.2%) followed by POL (7.9%) the current study has quickly affected the performance in a
and no change in CG/HVLIT (0.9%). Conventionally, HRR is positive direction. Therefore, the HIIT group likely had sufficient
measured within the first 60 s after termination of a test or time for recovery in between training sessions to show both
training (Lamberts et al., 2009). In this study, the HRR represents improvements in HRR concomitant with performance.
the mean of the acute HRR in between several stages during
the MART/MACT. This is the first time that data during the
specific MART/MACT has been shown. HRR was found to be Velocity or Power Output at Absolute
different between trained and untrained healthy individuals and Lactate Concentrations
that improvements in HRR occur with an increase in training In the current study, neither the velocity nor the power output
status (Daanen et al., 2012). Also this improved HRR has been at any of the established lactate concentrations (4, 6, and
used as a reliable test that relates to various performances as 10 mmol·L−1 ) showed any improvements. More specific, it is
longer time-trials as well as peak power output (Lamberts et al., interesting that the CG/HVLIT as the only training group
2011). However, there are conflicting results showing a decreased that targeted training at the defined lactate concentration
performance for intermittent high intensity exercise in athletes (e.g., approximately two sessions of THR/week) lacked any
that show an enhanced HRR (Le Meur et al., 2016). Furthermore, development in velocity or power output. Interestingly, the
it was suggested that an improved HRR could be biased by a lack of improvement is somewhat unexpected as the especially
decreased maximal HR. None of the groups in the current study enhanced performance in running is explained by a right shift
displayed such a pattern as all of them maintained their maximal for lactate threshold in relation to velocity (vLT; Billat et al.,
HR. Even though, it is well-established that the stroke volume 2002). Notably, this change was apparent already after a 4-week
increases as a result of endurance training (Blomqvist and Saltin, intervention.
1983) the reason for the improved HRR is related rather to the Another study using HIIT and more traditional training as
nervous system. A delayed parasympathetic reactivation has been long slow distance (LSD) displayed increased power outputs for
proposed to be part of the HRR post-exercise (Buchheit et al., both training modalities at 2 and 4 mmol·L−1 but with superior
2007). Interestingly, the HRR seems to be acutely impaired by development in the HIIT group (Ni Cheilleachair et al., 2017).
a high anaerobic contribution. However, the question is if this Their study resembled the adaptation period of the current study
is trainable and might be different if the athlete is accustomed as they used an 8-week training intervention. The use of rowers as
to more anaerobic work. Our results indicate that the athletes subjects could likewise be compared to the cross-country skiers
exposed to HIIT seemed to handle the anaerobic stress better in the current study as both sports use whole body work for
than the HVLIT dominated groups indicated by their superior propulsion. In support of these data a study performed on cyclists
HRR. Differences in training load can impact HRR (Borresen comparing HIIT block periodization with a more traditional
and Lambert, 2007) as demonstrated by an attenuated HRR training regime, i.e., mostly low intensity sessions with a few
following greater training load, defined according to the TRIMP HIIT sessions (Ronnestad et al., 2014), showed that only the
method. In the current study, the HIIT had a markedly lower HIIT block periodization increased power output at 2 mmol·L−1 .
training load when compared to all other modalities when The training performed was rather similar to the current study
counting training hours (<70 vs. ∼100 h). Nevertheless, even using HIIT session of target HR at 88–100% of HRmax . The
though the training load was not calculated according to the approximately accumulated time at this exercise intensity was
TRIMP method, the HIIT clearly showed a lower training load 30 min for each occasion in their study. It might be that the
based on duration and frequency. Contradictory to our results, stimulus that was used in the current study was too short in

Frontiers in Physiology | www.frontiersin.org 102 August 2017 | Volume 8 | Article 562

Stöggl and Björklund HIIT Improves Heart Rate Recovery

duration as it contained 16 min in total per session. However, we CONCLUSION

used a longer training period that spanned more than twice the
weeks (4 vs. 9 weeks). Possibly this could indicate that the amount In this study of elite athletes performing HIIT, POL, or mainly
of time HIIT is performed per session is more important than the HVLIT over a period of 9 weeks, only the HIIT group
accumulated time over a total training period for improvements achieved significant improvements (6.4%) in peak performance
in velocity or power output at lactate concentrations between 4 during the MART/MACT, while, unexpectedly, in no group
and 10 mmol·L−1 . the performance at the established lactate concentrations (4, 6,
Another aspect that should be mentioned here are the 10 mmol·L−1 ) was changed. Acute HRR was improved only
differences in the test protocols when comparing the in the HIIT (11.2%) and POL (7.9%) group with no change
MART/MACT (25 and 30 s stages with 100 s rest) with a in the HVLIT oriented control group. Therefore, it might be
standard incremental protocol (e.g., 3–5 min stages with concluded that only a training regime that includes a significant
20–30 s rest for blood sampling). In this context it is worth amount of HIIT improves the neuromuscular characteristics,
noting that in the study of Stöggl and Sperlich (2014) the anaerobic power, and the acute HRR in well-trained endurance
peak velocity/power at 4 mmol·L−1 increased in both POL athletes. A training regime that followed more a HVLIT
(+8.1%, P < 0.01) and HIIT (+5.6%, P < 0.05). Therefore, oriented model had no effect on any performance outcomes.
performance changes at lactate thresholds cannot be directly Practically, if HIIT is incorporated during pre-race preparation,
transferred among different test protocols. This might also be i.e., tapering HRR could provide a useful tool for monitoring
attributed toward the different energy system contributions adaptions related to anaerobic power and physiological response.
between the MART/MACT vs. a standard incremental test These findings shed new light into the cardiovascular, central
protocol. nervous and anaerobic adaptations in response to training
regimes with different training intensity distributions and
Limitations, Perspectives, and Practical should be of special interest in sports with high intensity
Applications intermittent character (e.g., game sports like soccer, ice hockey,
One limitation of the current study can be seen in the and handball) with a substantial anaerobic energy contribution.
mix between test modalities applied across the participants Also, the results might be of interest for endurance athletes
by using running or cycling tests specific to their preferred competing in sports using a masstart that involves repetitive
training exercise (e.g., cyclist vs. runner). However, because high intensity elements that are decisive for the race outcome
it is not an easy task to recruit large numbers of well- (e.g., fast accelerations during the start, sprint attacks, and finish
trained to elite athletes for such an experiment, various spurt).
types of endurance athletes were included. Furthermore,
although the MART was shown to be associated with AUTHOR CONTRIBUTIONS
neuromuscular factors/characteristics (Paavolainen et al., 1994,
1999a,c; Nummela et al., 2006) no specific parameters about Conception and design of the experiments: TS, Performance
effects on neuromuscular components were measured in the of the experiments: TS, and GB. Data analysis: TS, and GB.
current study. Therefore, only indirect conclusions from MART Preparation of the manuscript: TS, and GB. Both authors read
performance changes toward neuromuscular components can be and approved the final manuscript.
drawn.
Future research about long-term effects of different ACKNOWLEDGMENTS
training intensity distributions in well-trained athletes on
aerobic and anaerobic key components of performance is The authors would like to thank the athletes, coaches,
warranted. Still the measurement of HRR in between HIIT and research assistants involved in this study for their
bouts, or when using the MART for diagnostics as in the participation, enthusiasm, and cooperation. Special thanks to
current study, could be a practical tool to track physiological Julia Stöggl, Andreas Hochwimmer, and Thomas Damisch for
adaptations. In addition, the transfer of these enhanced their great assistance in recruitment, care, and control of
capacities toward real competition situations has still to be the athletes during the training intervention and pre- and
proven. post-testing.

REFERENCES after severe exhausting intensity runs. Eur. J. Appl. Physiol. 87, 496–505.
doi: 10.1007/s00421-002-0648-7
Achten, J., and Jeukendrup, A. E. (2003). Heart rate monitoring: applications Björklund, G., Laaksonen, M. S., and Holmberg, H. C. (2011). Blood lactate
and limitations. Sports Med. 33, 517–538. doi: 10.2165/00007256-200333070- recovery and respiratory responses during diagonal skiing of variable
00004 intensity. Eur. J. Sport Sci. 11, 317–326. doi: 10.1080/17461391.2010.
Beneke, R., Leithauser, R. M., and Ochentel, O. (2011). Blood lactate diagnostics 521580
in exercise testing and training. Int. J. Sports Physiol. Perform. 6, 8–24. Björklund, G., Pettersson, S., and Schagatay, E. (2007). Performance
doi: 10.1123/ijspp.6.1.8 predicting factors in prolonged exhausting exercise of varying
Billat, V. L., Mille-Hamard, L., Demarle, A., and Koralsztein, J. P. (2002). intensity. Eur. J. Appl. Physiol. 99, 423–429. doi: 10.1007/s00421-006-
Effect of training in humans on off- and on-transient oxygen uptake kinetics 0352-0

Frontiers in Physiology | www.frontiersin.org 103 August 2017 | Volume 8 | Article 562

Stöggl and Björklund HIIT Improves Heart Rate Recovery

Blomqvist, C. G., and Saltin, B. (1983). Cardiovascular adaptations to physical Paavolainen, L. M., Nummela, A. T., and Rusko, H. K. (1999a).
training. Annu. Rev. Physiol. 45, 169–189. doi: 10.1146/annurev.ph.45.030183. Neuromuscular characteristics and muscle power as determinants
001125 of 5-km running performance. Med. Sci. Sport Exer. 31, 124–130.
Bogdanis, G. C., Nevill, M. E., Boobis, L. H., and Lakomy, H. K. (1996). doi: 10.1097/00005768-199901000-00020
Contribution of phosphocreatine and aerobic metabolism to energy supply Paavolainen, L., Hakkinen, K., Hamalainen, I., Nummela, A., and Rusko,
during repeated sprint exercise. J. Appl. Physiol. 80, 876–884. H. (1999b). Explosive-strength training improves 5-km running time by
Borresen, J., and Lambert, M. I. (2007). Changes in heart rate recovery in improving running economy and muscle power. J. Appl. Physiol. 86,
response to acute changes in training load. Eur. J. Appl. Physiol. 101, 503–511. 1527–1533.
doi: 10.1007/s00421-007-0516-6 Paavolainen, L., Hakkinen, K., Nummela, A., and Rusko, H. (1994).
Bret, C., Messonnier, L., Nouck Nouck, J. M., Freund, H., Dufour, A. B., and Neuromuscular characteristics and fatigue in endurance and sprint athletes
Lacour, J. R. (2003). Differences in lactate exchange and removal abilities in during a new anaerobic power test. Eur. J. Appl. Physiol. Occup. Physiol. 69,
athletes specialised in different track running events (100 to 1500 m). Int. J. 119–126. doi: 10.1007/BF00609403
Sports Med. 24, 108–113. doi: 10.1055/s-2003-38201 Paavolainen, L., Nummela, A., Rusko, H., and Hakkinen, K. (1999c).
Buchheit, M., Laursen, P. B., and Ahmaidi, S. (2007). Parasympathetic reactivation Neuromuscular characteristics and fatigue during 10 km running. Int. J.
after repeated sprint exercise. Am. J. Physiol. Heart Circ. Physiol. 293, H133– Sports Med. 20, 516–521. doi: 10.1055/s-1999-8837
H141. doi: 10.1152/ajpheart.00062.2007 Ronnestad, B. R., Hansen, J., and Ellefsen, S. (2014). Block periodization of
Buchheit, M., Millet, G. P., Parisy, A., Pourchez, S., Laursen, P. B., and Ahmaidi, high-intensity aerobic intervals provides superior training effects in trained
S. (2008). Supramaximal training and postexercise parasympathetic cyclists. Scand. J. Med. Sci. Sports 24, 34–42. doi: 10.1111/j.1600-0838.2012.0
reactivation in adolescents. Med. Sci. Sports Exerc. 40, 362–371. 1485.x
doi: 10.1249/mss.0b013e31815aa2ee Rusko, H., Nummela, A., and Mero, A. (1993). A new method for the evaluation
Daanen, H. A., Lamberts, R. P., Kallen, V. L., Jin, A., and Van Meeteren, N. of anaerobic running power in athletes. Eur. J. Appl. Physiol. Occup. Physiol. 66,
L. (2012). A systematic review on heart-rate recovery to monitor changes 97–101. doi: 10.1007/BF01427048
in training status in athletes. Int. J. Sports Physiol. Perform. 7, 251–260. Seiler, K. S. (2010). What is best practice for training intensity and duration
doi: 10.1123/ijspp.7.3.251 distribution in endurance athletes? Int. J. Sport Physiol. 5, 276–291.
Garcia-Pinillos, F., Camara-Perez, J. C., Soto-Hermoso, V. M., and Latorre- doi: 10.1123/ijspp.5.3.276
Roman, P. A. (2017). A high intensity interval training (HIIT)-based running Seiler, K. S., and Kjerland, G. O. (2006). Quantifying training intensity distribution
plan improves athletic performance by improving muscle power. J. Strength in elite endurance athletes: is there evidence for an “optimal” distribution?
Cond. Res. 31, 146–153. doi: 10.1519/JSC.0000000000001473 Scand. J. Med. Sci. Sports 16, 49–56. doi: 10.1111/j.1600-0838.2004.00418.x
Helgerud, J., Hoydal, K., Wang, E., Karlsen, T., Berg, P., Bjerkaas, Stöggl, T. L., and Müller, E. (2009). Kinematic determinants and physiological
M., et al. (2007). Aerobic high-intensity intervals improve VO2max response of cross-country skiing at maximal speed. Med. Sci. Sports Exerc. 41,
more than moderate training. Med. Sci. Sports Exerc. 39, 665–671. 1476–1487. doi: 10.1249/MSS.0b013e31819b0516
doi: 10.1249/mss.0b013e3180304570 Stöggl, T. L., and Sperlich, B. (2015). The training intensity distribution
Kinnunen, J. V., Piitulainen, H., and Piirainen, J. M. (2017). Neuromuscular among well-trained and elite endurance athletes. Front. Physiol. 6:295.
adaptations to short-term high-intensity interval training in female ice hockey doi: 10.3389/fphys.2015.00295
players. J. Strength Cond. Res. doi: 10.1519/JSC.0000000000001881. [Epub Stöggl, T., and Sperlich, B. (2014). Polarized training has greater impact on key
ahead of print]. endurance variables than threshold, high intensity, or high volume training.
Lamberts, R. P., Swart, J., Noakes, T. D., and Lambert, M. I. (2009). Changes in Front. Physiol. 5:33. doi: 10.3389/fphys.2014.00033
heart rate recovery after high-intensity training in well-trained cyclists. Eur. J. Thomson, R. L., Bellenger, C. R., Howe, P. R., Karavirta, L., and Buckley, J. D.
Appl. Physiol. 105, 705–713. doi: 10.1007/s00421-008-0952-y (2016). Improved heart rate recovery despite reduced exercise performance
Lamberts, R. P., Swart, J., Noakes, T. D., and Lambert, M. I. (2011). A novel following heavy training: a within-subject analysis. J. Sci. Med. Sport 19,
submaximal cycle test to monitor fatigue and predict cycling performance. Br. 255–259. doi: 10.1016/j.jsams.2015.02.010
J. Sports Med. 45, 797–804. doi: 10.1136/bjsm.2009.061325 Tonnessen, E., Sylta, O., Haugen, T. A., Hem, E., Svendsen, I. S., and Seiler,
Laursen, P. B., Shing, C. M., Peake, J. M., Coombes, J. S., and Jenkins, D. G. (2002). S. (2014). The road to gold: training and peaking characteristics in the
Interval training program optimization in highly trained endurance cyclists. year prior to a gold medal endurance performance. PLoS ONE 9:e101796.
Med. Sci. Sports Exerc. 34, 1801–1807. doi: 10.1097/00005768-200211000-00017 doi: 10.1371/journal.pone.0101796
Le Meur, Y., Buchheit, M., Aubry, A., Coutts, A. J., and Hausswirth, Tossavainen, M., Nummela, A., Paavolainen, L., Mero, A., and Rusko, H. (1996).
C. (2016). Assessing overreaching with HRR: what is the minimal Comparison of two maximal anaerobic cycling tests. Int. J. Sports Med.
exercise intensity required? Int. J. Sports Physiol. Perform. 12, 569–573. 17(Suppl. 2), S120–S124. doi: 10.1055/s-2007-972912
doi: 10.1123/ijspp.2015-0675 Zagatto, A., Redkva, P., Loures, J., Kalva Filho, C., Franco, V., Kaminagakura, E.,
Ni Cheilleachair, N. J., Harrison, A. J., and Warrington, G. D. (2017). et al. (2011). Anaerobic contribution during maximal anaerobic running test:
HIIT enhances endurance performance and aerobic characteristics more correlation with maximal accumulated oxygen deficit. Scand. J. Med. Sci. Sports
than high-volume training in trained rowers. J. Sports Sci. 35, 1052–1058. 21, e222–e230. doi: 10.1111/j.1600-0838.2010.01258.x
doi: 10.1080/02640414.2016.1209539
Nummela, A. T., Paavolainen, L. M., Sharwood, K. A., Lambert, M. I., Noakes, Conflict of Interest Statement: The authors declare that the research was
T. D., and Rusko, H. K. (2006). Neuromuscular factors determining 5 km conducted in the absence of any commercial or financial relationships that could
running performance and running economy in well-trained athletes. Eur. J. be construed as a potential conflict of interest.
Appl. Physiol. 97, 1–8. doi: 10.1007/s00421-006-0147-3
Nummela, A., Hamalainen, I., and Rusko, H. (2007). Comparison of maximal Copyright © 2017 Stöggl and Björklund. This is an open-access article distributed
anaerobic running tests on a treadmill and track. J. Sports Sci. 25, 87–96. under the terms of the Creative Commons Attribution License (CC BY). The use,
doi: 10.1080/02640410500497717 distribution or reproduction in other forums is permitted, provided the original
Nummela, A., Mero, A., Stray-Gundersen, J., and Rusko, H. (1996). Important author(s) or licensor are credited and that the original publication in this journal
determinants of anaerobic running performance in male athletes and non- is cited, in accordance with accepted academic practice. No use, distribution or
athletes. Int. J. Sports Med. 17(Suppl. 2), S91–S96. doi: 10.1055/s-2007-972907 reproduction is permitted which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 104 August 2017 | Volume 8 | Article 562

 ORIGINAL RESEARCH
published: 02 August 2017
doi: 10.3389/fphys.2017.00515

Eleven-Week Preparation Involving

Polarized Intensity Distribution Is Not
Superior to Pyramidal Distribution in
National Elite Rowers
Gunnar Treff 1*, Kay Winkert 1 , Mahdi Sareban 1, 2 , Jürgen M. Steinacker 1 , Martin Becker 3
and Billy Sperlich 4
1
Division of Sports and Rehabilitation Medicine, Ulm University Hospital, Ulm, Germany, 2 Institute of Sports Medicine,
Prevention and Rehabilitation, Paracelsus Medical University, Salzburg, Austria, 3 Data Mining and Information Retrieval
Group, Julius-Maximilian University, Würzburg, Germany, 4 Integrative and Experimental Training Science, Institute of Sport
Science, University of Würzburg, Würzburg, Germany

Polarized (POL) training intensity distribution (TID) emphasizes high-volume low-intensity

exercise in zone (Z)1 (<first lactate threshold) with a greater proportion of high-intensity Z3
(>second lactate threshold) compared to Z2 (between first and second lactate threshold).
In highly trained rowers there is a lack of prospective controlled evidence whether POL
Edited by:
is superior to pyramidal (PYR; i.e., greater volume in Z1 vs. Z2 vs. Z3) TID. The aim of
James (Jim) David Cotter,
University of Otago, New Zealand the study was to compare the effect of POL vs. PYR TID in rowers during an 11-wk
Reviewed by: preparation period. Fourteen national elite male rowers participated (age: 20 ± 2 years,
Brett Tiaki Smith, maximal oxygen uptake (V̇O2max ): 66 ± 5 mL/min/kg). The sample was split into PYR
University of Waikato, New Zealand
Anthony Rice, and POL by varying the percentage spent in Z2 and Z3 while Z1 was clamped to ∼93%
Australian Institute of Sport, Australia and matched for total and rowing volume. Actual TIDs were based on time within heart
*Correspondence: rate zones (Z1 and Z2) and duration of Z3-intervals. The main outcome variables were
Gunnar Treff
average power in 2,000 m ergometer-test (P2,000 m ), power associated with 4 mmol/L
[email protected]
[blood lactate] (P4[BLa] ), and V̇O2max . To quantify the level of polarization, we calculated
Specialty section: a Polarization-Index as log (%Z1 × %Z3 / %Z2). PYR and POL did not significantly differ
This article was submitted to
regarding rowing or total volume, but POL had a higher percentage of Z3 intensities
Exercise Physiology,
a section of the journal (6 ± 3 vs. 2 ± 1%; p < 0.005) while Z2 was lower (1 ± 1 vs. 3 ± 2%; p < 0.05) and
Frontiers in Physiology Z1 was similar (94 ± 3 vs. 93 ± 2%, p = 0.37). Consequently, Polarization-Index was
Received: 27 February 2017 significantly higher in POL (3.0 ± 0.7 vs. 1.9 ± 0.4 a.u.; p < 0.01). P2,000 m did not
Accepted: 05 July 2017
Published: 02 August 2017
significantly change with PYR (1.5 ± 1.7%, p = 0.06) nor POL (1.5 ± 2.6%, p = 0.26).
Citation:
V̇O2max did not change (1.7 ± 5.6%, p = 0.52 or 0.6 ± 2.6, p = 0.67) and a small
Treff G, Winkert K, Sareban M, increase in P4[BLa] was observed in PYR only (1.9 ± 4.8%, p = 0.37 or −0.5 ± 4.1%,
Steinacker JM, Becker M and
p = 0.77). Changes from pre to post were not significantly different between groups in
Sperlich B (2017) Eleven-Week
Preparation Involving Polarized any performance measure. POL did not prove to be superior to PYR, possibly due to the
Intensity Distribution Is Not Superior to high and very similar percentage of Z1 in this study.
Pyramidal Distribution in National Elite
Rowers. Front. Physiol. 8:515. Keywords: rowing, training intensity distribution, elite athletes, interval training, high intensity, high volume,
doi: 10.3389/fphys.2017.00515 training zones

Frontiers in Physiology | www.frontiersin.org 105 August 2017 | Volume 8 | Article 515

Treff et al. Polarized Training in Rowers

INTRODUCTION Z2 and Z3 (Steinacker et al., 1998). Single case studies reported

a TID of 85% in Z1 (Nybo et al., 2014) in an elite lightweight
Olympic rowers aim to cover the distance of 2,000 m faster rower from Denmark and ∼81% in Z1 in a double Olympic
than their opponents, with world best times varying from 5:18 champion from Norway (Seiler and Tønnessen, 2009), with both
to 7:08 min. For this purpose, rowers generate an average reports not specifying the percentages spent in Z2 and Z3. Most
power of ∼590 W (Voliantis and Secher, 2009) and exhibit analyses so far report a pyramidal TID, i.e., decreasing amount
an outstanding maximal oxygen uptake (V̇O2max ; Secher, 1993; of training spent in Z1, Z2 and Z3. German junior rowers e.g.,
Ingham et al., 2002) with extreme acidosis and metabolic stress exhibited a TID of ∼95-3-2 (i.e., percentage in Z1, Z2, Z3) during
(Nielsen, 1999). the last 9 weeks before the first competition (Guellich et al.,
To prepare for this kind of physical exertion and performance, 2009) and nine successful Olympic rowers from New Zealand
elite rowers perform most of their training time with high- featured a TID of 77-17-6 (Plews et al., 2014). A successful French
volume and (relatively) low-intensity exercise (Nybo et al., rower employed a TID of 45% Z1 and 55% Z2 (Lacour et al.,
2014), often referred to as “basic endurance training,” which is 2009), notably a TID emphasizing “threshold” intensity. Only
quite demanding in rowing as indicated by an oxygen uptake one investigation so far reported a polarized (POL) TID of 93-
exceeding 3.5 L/min (Secher, 1993) and forces >500 N per 2-5 in a Belgian elite sculler (Bourgois et al., 2013). POL is
stroke (Roth et al., 1993). The high metabolic and muscular characterized by a relatively high amount of volume performed
demands limit high-volume low-intensity rowing sessions to 90– in Z1 and Z3, with less volume in Z2. Taking into account that
100 min to avoid impaired stroke technique. With increasing boat rowing is a high-intensity sport and being aware of several reports
velocity the energy expenditure increases 2.2- to 2.4-fold (Secher, from other endurance disciplines like e.g. running, cycling or
1993) ultimately forcing the athlete to row as efficiently as cross-country skiing (Stöggl and Sperlich, 2015), the long-term
possible during each session. In addition to the rowing sessions, stimulus of POL may improve endurance performance with
most rowers implement 2–3 strength-training sessions per week potentially less autonomic and hormonal stress and boredom,
into their training schedules, thereby increasing the exhaustive which is supported by experiments in club rowers who especially
character of a rower’s preparation for competition. emphasized Z3-training (Driller et al., 2009; Ní Chéilleachair
Over the last three decades the total training volume in elite et al., 2016).
rowing elevated by 66%, augmenting to ∼23 h/wk in Norwegian Several observational studies of national or world-class
rowers during the 90’s (Fiskerstrand and Seiler, 2004) and is athletes from various sport disciplines like running (Billat
estimated to peak at ∼29 h/wk nowadays (Nielsen, 2009/2017). et al., 2001) or cross-country skiing (Seiler and Kjerland, 2006;
Based on temporal limits and high workload of today’s elite Sandbakk et al., 2011; Tønnessen et al., 2014) successfully applied
rowers we may assume that total training volume is near its a POL TID. Only one controlled study in 18 club rowers
functional maximum. As the medal winners’ 2,000 m boat speed following a 28-day detraining period reported a similar increase
still increases by ∼0.12% per year (Kleshnev and Nolte, 2011), of ergometer performance with POL (72-0-28) compared to
optimization of training schedules, especially by altering the a control group exaggerating low-intensity rowing (98-2-0;
intensity distribution (TID) might be a worthwhile resource to Ingham et al., 2008).
enhance performance. Integrating the findings of rowing studies as well as findings
Within the literature a three-zone intensity model is applied from other endurance sports (Neal et al., 2013; Stöggl and
to quantify the TID. The model is based on the following Sperlich, 2014) strong evidence exists that POL may be applied in
physiological benchmarks: Zone 1 (Z1) is defined as low-intensity high performance rowing, but this notion is drawn on two serious
exercise with low levels of blood lactate below first lactate limitations: Firstly, performance benefits of POL have been
or ventilatory threshold. Zone 2 (Z2) refers to elevated and concluded based on retrospective observations, but prospective
accumulated blood lactate concentration also called “lactate randomized-controlled data on sub-elite or elite level rowers do
threshold training” identified as an intensity between the first not exist. Secondly, POL has been compared to static TIDs which
and second lactate or ventilatory threshold. Finally, Zone 3 do not change over weeks or months. From a methodological
(Z3) refers to high-intensity exercise above the second lactate point of view, experiments involving static TIDs are convenient
or ventilatory threshold (Kindermann et al., 1979; Seiler and for scientists to compare differences between groups, but a static
Kjerland, 2006). Common additional parameters to demarcate TID does not mirror real-training scenarios in high performance
the three intensity zones are based on blood lactate levels [BLa] of sports, in which TIDs are shaped “dynamically” with increasing
<2 mmol/L (Z1), 2–4 mmol/L (Z2), and >4 mmol/L (Z3) as well percentages of Z2 and Z3 before competitions. Notably, this is
as certain percentages of maximal heart rate or maximal oxygen recommended by the current scientific literature (Bangsbo et al.,
uptake. Noteworthy, national sport governing bodies and rowing 2010; Tønnessen et al., 2014).
federations often apply five-zone models to differentiate training Altogether, numerous successful TIDs exist in rowing (i.e.,
intensities more detailed (Seiler, 2010). POL and PYR), data from various disciplines and rowing
So far, few studies have described the TID in rowing, especially are conflicting, and no prospective randomized-controlled
on an elite level. In 1998 Steinacker et al. reported German, investigation exists comparing POL to a dynamic TID in
Danish, Dutch, and Norwegian elite rowers spending 90–96% elite rowers. Therefore, we aimed to compare 11 weeks of a
of their total training volume in Z1 but it remains uncertain to competition-preparation period involving POL to a dynamic
the reader how the remaining percentage was exactly split into PYR distribution in national elite rowers.

Frontiers in Physiology | www.frontiersin.org 106 August 2017 | Volume 8 | Article 515

Treff et al. Polarized Training in Rowers

MATERIALS AND METHODS mode, duration, distance, and intensity as well as information
on days off and illness or injury. Entries of Z1 and Z2 sessions
Design were based on time spent in corresponding HR-zones. HR was
The present prospective study was conducted during the final measured by the athletes’ own HR-monitors and/or with the
11 weeks of the preparation period, i.e., calendar week 3– smartphone based rowing in Motion–App (In Motion Software
14. Immediately after the study, the national qualification & Sports Technology, Hanau, Germany) that was connected to a
regatta to apply for the national team was scheduled. Fourteen chest belt with Bluetooth data transmission (H7, Polar Electro,
national elite rowers participated in the study. Twelve (86%) Oy, Finland). To avoid underestimation of Z3 sessions due to
rowed for Germany on international regattas in 2016, three the delayed HR-response at high intensities, Z3 sessions were not
rowers were lightweights. Table 1 summarizes the rowers’ documented by time in corresponding HR-zone, but by the total
anthropometric data. All rowers provided written informed duration of the performed Z3-interval, as long as the maximal
consent to participate. The experimental protocol was approved HR of the interval reached the individually defined Z3 HR-zone.
by the ethical review board of the University of Ulm. Otherwise the interval was rated as Z2.
The rowers trained in two different training facilities (A;B) The diary logs were checked by the coaches and crosschecked
within Germany. Athletes could not be randomly assigned to by the research team for plausibility. After completion of the
PYR or POL, because several rowers trained in crew boats. study, all data were exported (.csv files) and subsequently
Moreover, it was not possible to separate existing training squads analyzed using the Python data analysis toolkit “pandas”
for organizational reasons. To overcome this limitation, we (version 0.18.0, PyData Development Team) and the Scientific
allocated two groups in each of the two facilities to either PYR or Computing Library “Scipy” (version 0.17.0, SciPy developers).
POL. In each facility, one training group followed the traditional Notably, a basic framework of adequate training intensities
rowing schedule emphasizing high-volume low-intensity exercise was provided by rowing stroke frequency and pace prescribed by
and a pyramidal TID (PYR). The other group targeted a polarized the coaches, which is a common practice in rowing (Plews et al.,
TID model (POL). In facility A, one athlete of each group was 2014).
excluded from the study due to illness or injury not related to
the intervention. The 11-wk duration included pre- and post-test Training Modes
procedures to evaluate the changes in rowing performance and Training differentiated four modes, namely (i) Rowing: involving
physiological variables. boat and ergometer rowing, (ii) Endurance: other endurance
training like running, cycling, swimming, etc., (iii) Strength:
Training Intervention resistance training, machine-based or weight lifting, and (iv)
Training Intensity Zones Other: stretching, stability training, etc.
A three-zone training model was applied to quantify TID (Foster
et al., 2001; Seiler and Kjerland, 2006; Seiler, 2010). The following Training Intensity Distribution
three intensity zones were established based on a 5 × 4-min The overall training of both groups included all four training
ergometer step test as described in detail below (Section Power modes (Rowing, Endurance, Strength, and Other). Based on the
Output at 2 and 4 mmol/L Blood Lactate): Z1 was defined as the coaches’ experience with their athletes, both groups targeted 16–
intensity between 65% of maximal heart rate and the first lactate 18 h total training volume per week and ∼120 km of rowing
threshold or lactate-equivalent as described by Kindermann et al. per week. The general training schedule provided 2–3 sessions of
(1979). Z2 was defined as the intensity between first lactate strength training and 6–8 rowing sessions per week. The primary
threshold and the second or individual lactate threshold as distinction between the two groups was the prescribed intensity
described by Dickhuth et al. (1991). Z3 was defined as an intensity distribution, with PYR including two to three Z2-sessions (e.g. 1
above the second lactate threshold. × 4 or 2 × 3 or 3 × 2 km) with not more than one session in
All intensities were related to the corresponding heart rates Z3. In contrast, POL included 2-3 sessions of Z3 training (e.g. 2
(HR) during the ergometer step test to allow for objective entries × 2 km, 10 × 250 m; 6 × 1 km) while avoiding Z2 as much as
into the mandatory online training diary of the German Rowing possible.
Federation. The diary included information about the training To ensure high compliance, the general training prescription
was discussed with the coaches of both groups before initiating
the study. However, since preparation period is fundamental
for competition success, the coaches were permitted to adapt
TABLE 1 | Participants’ anthropometric characteristics.
the schedule depending on the athletes’ particular needs, health
Variable PYR POL p dCohen status and environmental conditions.

Standing height (cm) 193 ± 2 185 ± 7 0.029 −1.55

Pre- and Post-measurements
Body mass (kg) 93 ± 3 85 ± 11 0.138 −0.99
During the first visit all rowers were medically examined by a
Age (years) 19 ± 1 21 ± 2 0.062 0.26
licensed sports physician. The medical examination also included
V̇O2max (mL/min/kg) 64 ± 3 68 ± 7 0.171 0.74 an electrocardiography at rest (CardioPart 12 Blue, Amedtec,
PYR (n = 7), pyramidal training intensity distribution; POL (n = 7), polarized training Aue, Germany), an echocardiogram (Philips CX50, Phillips
intensity distribution; V̇O2max , maximal oxygen uptake. Medical Systems, Andover, MA, USA), and blood analysis to

Frontiers in Physiology | www.frontiersin.org 107 August 2017 | Volume 8 | Article 515

Treff et al. Polarized Training in Rowers

exclude iron deficiency and anemia. All rowers were declared free of measurement on Concept 2 rowing ergometers for P2,000 m
from cardiovascular disease and eligible to perform the exercise amounts to 1.3% (95%CI 0.9-2-9; Soper and Hume, 2004). We
protocol and the study. used this value to estimate the smallest worthwhile change in
Afterwards, a series of rowing ergometer tests was conducted P2,000 m (Smith and Hopkins, 2011).
on two days (Figure 1) employing a Concept 2 Type D ergometer
(Concept 2, Morrisville, USA) for all tests. The ergometer was Measurements of Maximal Oxygen Uptake
modified with a load cell for force measurement and a rotary V̇O2max was measured with a ramp test protocol that enables a
transducer to calculate the power output [Institut für Forschung linear increase in power and objective test termination (Winkert
und Entwicklung von Sportgeräten (FES), Berlin, Germany]. et al., 2016). Briefly, target power of the rowing ergometer was
initially set to 160 W and increased 30 W/min (lightweight
Power Output at 2 and 4 mmol/L Blood Lactate rowers) or 35 W/min (open weight class). The test automatically
After the physical examination, all rowers performed a 5 × 4- terminated in case the rowers failed to increase power within a
min incremental step test with 50 W increments per stage. The 7-W range of five strokes. Gas exchange and ventilation were
workloads ranged from 150 to 350 W in the lightweight and measured using a metabolic analyzer with a dynamic mixing
from 200 to 400 W in the open weight class rowers. During a chamber (Metamax 3x, Cortex Biophysics, Leipzig, Germany).
30-s break between each stage, 20 µL of capillary blood were The technical error of measurement of this device amounts to
sampled from the hyperemic earlobe and the level of blood lactate 0.03–0.21 L/min (95% confidence interval; Larsson et al., 2004).
was immediately analyzed amperometric-enzymatically (C-Line, The system was calibrated prior to each test using ambient
EKF, Barleben, Germany). A specialized software calculated air and the manufacturers’ calibration gas (16% O2 , 5% CO2 ).
power output at 2 and 4 mmol/L [blood lactate] (P2[BLa] and A precision 3-L syringe (Hans Rudolph, Shawnee, USA) was
P4[BLa] ) using a polynomic fitting of the power and lactate data employed to calibrate the flow sensor.
(Winlactat, Mesics, Münster, Germany). P4[BLa] is an accepted V̇O2max was defined as the highest V̇O2 with increasing
measure of rowing performance with standard errors of the workload and averaged over a 30-s interval. V̇O2max was
estimate of 2,000 m ergometer performance amounting to 1.4– considered when V̇O2 failed to increase with progressive work
3.3% (Smith and Hopkins, 2012). We used 3.3% as the lower limit rate (leveling off) or at least a plateau of V̇O2 was present.
to identify worthwhile changes, since the performance level of A plateau was defined as an increase in V̇O2 < 150 mL/min,
our rowers was similar to those of previous reports (Nevill et al., which is the most frequent definition in literature (Midgley et al.,
2011). 2007). A leveling-off or plateau V̇O2 was found in all cases. In
addition, respiratory exchange ratio at exertion was always >1.1
Two Thousand Meters Ergometer Test with [BLa] ≥ 8 mmol/L.
All rowers performed an all-out 2,000 m ergometer test to
evaluate maximal rowing ergometer performance, by covering Polarization-Index
the virtual distance of 2,000 m as fast as possible. The average To quantify the individual level of periodization, we calculated
power (P2,000 m ) was recorded from the Concept 2 monitor a Polarization-Index based on the percentage, time, or distance
afterwards. This test is employed worldwide in elite rowing to trained in each intensity zone. The Polarization-Index was
determine changes of maximal performance (Hahn et al., 2000; calculated as follows:
Mäestu et al., 2005; Smith and Hopkins, 2012). The standard
error of the estimate of 2,000 m single-scull performance has
been calculated to be 2.6% (Jürimäe et al., 2000). The error Polarization−Index (a.u.) = log (Z1/Z2 × Z3) (1)

FIGURE 1 | Study and test design. P[BLa]4 , Power at blood lactate concentration of 4 mmol/L; P2,000 m , Average power in 2,000 m rowing ergometer test; V̇O2max ,
Maximal oxygen uptake; PYR, Pyramidal training intensity distribution; POL, Polarized training intensity distribution; CW, Calendar week.

Frontiers in Physiology | www.frontiersin.org 108 August 2017 | Volume 8 | Article 515

Treff et al. Polarized Training in Rowers

If Zone 2 = 0, following formula avoided zero in the applied a Fishers’ exact test to calculate if distributions between
denominator: 1P2,000 m and Group (PYR; POL), or Polarization-Index (≤2;
>2) were different.
Polarization−Index (a.u.) = log (Z1/0.1 × (Z3 − 0.1)) (2)

If Z3 = 0, Polarization-Index was defined as zero. RESULTS

In case of a Polarization-Index > 2.0 a.u., the TID was defined Training
to be polarized, indicating an increasing level of polarization with
Specific Training (Boat & Rowing Ergometer)
higher values. If Polarization-Index was ≤2 a.u. the TID was
The rowing volume and sum of rowing sessions are summarized
defined as being not polarized.
in Table 2. Rowing training was not significantly different
In the context of this study, all TIDs with a Polarization-
between groups regarding absolute volume, duration or
Index ≤ 2.0 a.u. were pyramidal or indifferent, but from a
frequency. Small effect sizes indicated a slightly higher distance
theoretical perspective, a Polarization-Index ≤ 2.0 can indicate
covered by PYR, while the number of training sessions was
at least five different TIDs, namely (i) a pyramidal TID (80-15-
slightly higher in POL (Table 2).
5; Polarization-Index = 1.4 a.u.), (ii) an inverse pyramidal TID
(20-30-50; Polarization-Index = 1.5 a.u.), (iii) a Lactate Threshold
Training Mode
TID (60-38-2; Polarization-Index = 0.5 a.u.), an indifferent TID
The distribution of rowing (boat and ergometer) and strength
(90-5-5; Polarization-Index = 2.0 a.u.), or a Long-Slow-Distance
training were not significantly different between groups, but
TID (100-0-0; Polarization-Index = 0.0 a.u.).
there was a tendency for higher volume of Endurance and
A Polarization-Index > 2.0 a.u. does not indicate a polarized
Other training in POL with moderate to large effects. Notably,
distribution, if Z1 is smaller than Z3 (e.g., 40-0-60; PI = 4.4 a.u.).
the number of Endurance and Other sessions were significantly
This kind of TID would probably be classified as High Intensity
higher in POL compared to PYR, underlined by large effect sizes
Training or HIT, since polarized distributions necessitate Z1
(Table 3).
volume to be highest.
Both groups did not differ regarding the days of illness
Even if we assume that a given Polarization-Index reflects
(p = 0.81) amounting to a median of 3 days (min-max: 1–7) in
the same degree of polarization, we cannot expect the same
PYR and 4 days (1-11) in POL.
physiological response from different TIDs emerging in the
same Polarization-Index. For example, a Polarization-Index of
2.3 a.u. can result out of 90-3-7 and 50-10-40. Nevertheless, Intensity Distribution
the Polarization-Index seems useful to quantify the level of Figure 2 shows that percentage of training in Z1 was similar
polarization of TIDs with Z1-percentages between 75 and 95%, between PYR and POL (94 ± 3% and 93 ± 2%; p = 0.37,
which are frequently used in high performance sports (Seiler, dCohen = −0.33), but Z2 was significantly higher in PYR (3 ± 2%
2010). and 1 ± 1%; p < 0.01; dCohen = −1.27) and Z3 was significantly
higher in POL (2 ± 1% and 6 ± 3%; p < 01; dCohen = 1.79). This
Statistical Analysis emerged into a significantly higher Polarization-Index in POL
All statistical procedures were calculated using the statistical (1.9 ± 0.4 and 3.0 ± 0.7; p < 0.01; dCohen = 1.93).
package SPSS 21. Average data are expressed as arithmetic The longitudinal differences in TID of both groups during
mean ± standard deviation, unless otherwise stated. To calculate the 11-wk intervention expressed by means of the Polarization-
differences between groups, a t-test was employed after testing for Index are displayed in Figure 3. The Polarization-Index was
normal distribution using a Shapiro–Wilkinson-Test. A paired significantly higher in POL in calendar week 4, 5, 6, 7, 8, and 10
t-test was applied to calculate differences between pre and post- (p = 0.00 to 0.02). No significant differences in the Polarization-
test. An unpaired t-test analyzed significant differences between Index were found in the other weeks, although with a tendency to
training groups. Cohens d (dCohen ) was calculated to estimate be greater in POL (p = 0.19 to 0.36) except of weeks 11 and 12. To
effect sizes (Cohen, 1988), defined as follows: trivial: 0–|0.2|, note, the Polarization-Index clearly increased in PYR toward the
small: |0.2|–|0.6|, moderate |0.6|–|1.2|, large: |1.2|–|2.0|, very end of the study period, starting in calendar week 11 (3.0 a.u.),
large: |2.0|–|4.0|, and infinite: > |4.0| (Hopkins, 2003). thereby indicating a considerable increase of Z3-intensities.
A correlation analysis using Pearson’s coefficient identified
possible effects of volume spent in the specific training modes,
number of days without training, training volume in Z1-Z3,
TABLE 2 | Mean characteristics of the total rowing volume and sum of rowing
Polarization-Index and changes in endurance variables (i.e., sessions.
P2,000 m , P2[BLa] , P4[BLa] , V̇O2max ). Effect sizes of correlation were
defined as follows: trivial: 0.0, small: 0.1–0.3, moderate: 0.3–0.5, Variable PYR POL p dCohen

high: 0.5–0.7, very high: 0.7–0.9, nearly perfect: 0.9, and perfect
Rowing distance (km) 1334 ± 67 1255 ± 264 0.466 −0.41
1.0 (Hopkins, 2003).
Rowing duration (min) 5953 ± 315 5919 ± 1216 0.945 −0.04
We dichotomized the outcome of the main variable P2,000 m
Sessions (n) 80 ± 4 84 ± 13 0.414 0.42
into 1P2,000 m (≤1.3%; >1.3%) to distinguish between changes
smaller or higher than the smallest worthwhile change and PYR, pyramidal training intensity distribution; POL, polarized training intensity distribution.

Frontiers in Physiology | www.frontiersin.org 109 August 2017 | Volume 8 | Article 515

Treff et al. Polarized Training in Rowers

TABLE 3 | Distribution of training modes during observation period.

Training mode Duration (min/wk) Percent (%/wk) Sessions (1/wk)

PYR POL p dCohen PYR POL p dCohen PYR POL p dCohen

Rowing 541 ± 28 537 ± 110 0.93 −0.05 58 ± 5 54 ± 9 0.35 −0.55 7.2 ± 0.3 7.7 ± 1.2 0.41 0.57
Strength 178 ± 32 149 ± 72 0.36 −0.52 19 ± 3 15 ± 7 0.18 −0.74 1.6 ± 0.4 1.3 ± 0.7 0.39 −0.53
Endurance 144 ± 30 202 ± 68 0.07 1.10 15 ± 3 21 ± 6 0.08 1.27 1.7 ± 0.4 3.0 ± 0.8 0.00 1.37
Other 68 ± 28 100 ± 29 0.06 1.12 7±3 10 ± 3 0.07 1.00 2.2 ± 0.6 3.5 ± 1.2 0.03 1.37
Total 931 ± 50 990 ± 100 0.19 0.75 12.7 ± 0.8 15.5 ± 2.3 0.002 1.63

PYR, pyramidal training intensity distribution; POL, polarized training intensity distribution; Rowing, boat & rowing ergometer; Strength, resistance training; Endurance, other modes of
endurance training than rowing (e.g., spinning, cycling, running); Other, all other kinds of training not mentioned before (e.g., stretching, stabilization training, Yoga).

††
FIGURE 2 | Intensity distribution in boat & ergometer rowing. ** and :
FIGURE 3 | Polarization-Index of a pyramidal (PYR) and a polarized rowing
Percentage rowed in Z2 and Z3 differed between pyramidal (PYR) and
training group (POL) during 11 weeks of intervention and pre-/post-testing.
polarized training intensity distribution (POL) (p < 0.01).
Polarization-Index = log (%Zone 1 × %Zone 3 / %Zone 2). See text for details.
Broken line indicates Polarization-Index of 2 a.u., which is defined being the
cut off between polarized (> 2) and not polarized (≤2 a.u.) training intensity
distribution. The high percentage of Zone 3 in calendar weeks 3 and 13 was
Performance due to the pre-/post-testing.

In the complete sample, P2,000 m increased significantly (p = 0.03)

from 443 ± 30 W to 449 ± 26 W, corresponding to an
improvement in 2,000 m time of ∼2 s (pre: 370.1 ± 8.7 s to post: Training and Performance: Correlation
368.2 ± 7.2 s, representing a small effect (dCohen = −0.24). Analysis
Average changes between PYR and POL or within PYR and POL Correlation analysis indicated, that changes in P2,000 m became
from pre to post-test were trivial or small and not significant smaller with greater volume of Other training like e.g., stretching
(Table 4). (r = −0.61; p = 0.03).
V̇O2max , P2[BLa] , P4[BLa] did not significantly change from pre Changes in performance variables (P2[BLa] , P4[BLa] , P2,000 m ,
to post or between groups. A small increase of average P2[BLa] V̇O2max , and duration in 2,000 m ergometer test) were not
and P4[BLa] was detected in PYR only, where three rowers (43%) significantly correlated to total days without training (highest
improved P4[BLa] by more than 3.3% in contrast to POL with absolute r = 0.11 with p = 0.71), neither health related (highest
only one rower (14%) who improved above this threshold. The absolute r = 0.42; p = 0.18) nor related to scheduled days off
unbalanced improvement in P2[BLa] and P4[BLa] between groups (highest absolute r = 0.26; p = 0.41).
is also reflected by small effect sizes in PYR and trivial effect sizes A high correlation was found between changes in V̇O2max
in POL (Table 4). and absolute weekly volume (r = 0.58; p = 0.05) or percentage
On an individual level five out of six (83%) rowers with POL (r = 0.59; p = 0.04) spent in Z2. However, average changes in
improved P2,000 m above the estimated error of measurement V̇O2max were small, amounting to 0.1 L/min which is within
of 1.3% (95%CI 0.9-2-9; Soper and Hume, 2004). In PYR, the technical error of measurement of 0.03–0.21 L/min (95%
four out of seven (57%) rowers improved P2,000 m above 1.3% confidence interval) for the device used in this study (Larsson
(Figure 4). et al., 2004).

Frontiers in Physiology | www.frontiersin.org 110 August 2017 | Volume 8 | Article 515

Treff et al. Polarized Training in Rowers

TABLE 4 | Changes in performance after 11 weeks of training in national elite rowers.

Variable PYR POL PYR vs. POL

Pre Post 1% n p dCohen Pre Post 1% n p dCohen p dCohen

V̇O2max 64 ± 3 64 ± 2 1.7 ± 5.6 6 0.522 0.00 68 ± 7 68 ± 7 0.6 ± 2.8 6 0.686 0.00 0.712 0.22
(mL/min/kg)
Duration 368.8 ± 7.6 367.0 ± 6.4 −0.5 ± 0.6 7 0.060 −0.26 372.0 ± 10 369.8 ± 8.4 −0.5 ± 0.9 6 0.221 −0.24 0.962 0.03
2,000 m test (s)
P2,000m (W) 447 ± 27 454 ± 24 1.5 ± 1.7 7 0.057 0.27 438 ± 36 444 ± 30 1.5 ± 2.6 6 0.258 0.00 0.916 0.06
P2[BLa] (W) 291 ± 26 298 ± 19 3.0 ± 5.6 7 0.232 0.31 297 ± 16 297 ± 27 0.2 ± 5.9 7 0.897 0.00 0.441 0.42
P4[BLa] (W) 336 ± 25 341 ± 20 1.9 ± 4.8 7 0.369 0.22 337 ± 17 336 ± 24 −0.5 ± 4.1 7 0.770 −0.05 0.369 0.50

PYR, Pyramidal training intensity distribution; POL, polarized training intensity distribution; V̇O2max , maximal oxygen uptake; P2,000m , average power in 2,000 m rowing ergometer test;
P2[BLa] and P4[BLa] , Power output with [blood lactate] 2 and 4 mmol/L. PYR vs. POL was calculated from the absolute difference between pre- and post-test of each group.

FIGURE 4 | Individual change of average power in 2,000 m rowing-ergometer FIGURE 5 | Individual change (%) of average power in 2,000 m rowing
test (P2,000 m ) after 11-weeks between rowers following a polarized (POL) or ergometer test (P2,000 m ) in 14 highly trained competitive U23 rowers.
pyramidal (PYR) training intensity distribution. The dashed line corresponds to Horizontal dashed line indicates Polarization-Index of 2, which is defined being
estimated error of measurement for P2,000 m on Concept 2 rowing ergometers the cut off for polarized training intensity distribution (see text for
of 1.3% according to Soper and Hume (2004). Short lines indicate median of details).Vertical dashed line indicates 1.3%, being the estimated error of
each group. measurement for P2,000 m according to Soper and Hume (2004).

Negative correlation coefficients indicated, that the higher the Figure 5 indicates that—irrespective of the group allocation—
absolute volume of Z3-training, the smaller the increase in P2[BLa] six out of seven (86%) rowers with a Polarization-Index >2
(r = −0.56; p = 0.02) and P4[BLa] (r = −0.53; p = 0.05). Similar increased P2,000 m more than 1.3%. In contrast, only three out
results were obtained for percentage of Z3 and P2[BLa] (r = 0.63; of six rowers (50%) with a Polarization-Index ≤2 improved
p = 0.02) and P4[BLa] (r = −0.59; p = 0.03). In line with the their P2,000 m above the smallest worthwhile change. However,
previous result, smaller changes in P2[BLa] (r = −0.58; p = 0.03) distribution between the dichotomized variable Polarization-
and P4[BLa] (r = −0.64; p = 0.01) were correlated with higher Index (≤2.0; >2.0) vs. 1P2,000 m (≤1.3%; >1.3%) was not
Polarization-Index. significantly different (Fishers exact test: p = 0.27).

Distribution of Worthwhile Change in

DISCUSSION
2,000 m Ergometer Test, Group, and
Polarization-Index We compared two TIDs (POL vs. real-life PYR) in national elite
Even though more rowers in the POL-group increased P2,000 m rowers with increasing percentage of Z3-training during the last
(Figure 4, Table 4) the distribution between groups regarding the 3 weeks of a 11-wk training period. Notably, percentage of Z1 was
dichotomized variable change ≥1.3% vs. change <1.3% was not clamped to ∼93% in both groups. In summary, POL did not show
different (Fisher’s exact: p = 0.56), basically due to an outlier with to be significantly superior to PYR regarding any physiological
1P2,000 m of −3.3% (Figure 4). determinant of rowing performance, but small effects indicated

Frontiers in Physiology | www.frontiersin.org 111 August 2017 | Volume 8 | Article 515

Treff et al. Polarized Training in Rowers

improved P2[BLa] and P4[BLa] in PYR only. However, irrespective volume of elite rowers amounts to ∼120 km/week. In our study
of the groups, we observed a higher percentage of worthwhile rowing training averaged 114 km/wk (POL) to 121 km/wk (PYR)
improvements regarding P2,000 m , which is the key variable in which appears to be relatively high since most of our rowers were
rowing, when the Polarization-Index was ≥2.1 a.u. U23-rowers, who generally train less then world-class rowers of
On a group level both TID models allowed for improvements higher age, but reasonably more than juniors who row ∼97.1 ±
in P2,000 m , which is the most reliable and accepted surrogate 19.5 km/wk (Guellich et al., 2009). We therefore assume, that
measure of rowing performance (Hahn et al., 2000; Mäestu et al., training volume in our study was per-se high enough to allow for
2005). Also Ingham et al. who compared POL to a low-intensity changes in performance, independent of alterations in TID.
group, did not find POL to be superior in rowers, but in contrast Volume of Endurance and Other training was moderately
to our results, the maximal oxygen uptake increased in both of higher in POL. The number of the according sessions was
their groups probably attributable to the much lower training significantly higher in POL with even large effect sizes (Table 3).
status of those club rowers who were 13% slower in the 2000 m While the high volume of non-specific endurance training was
ergometer test compared to our athletes (Ingham et al., 2008). without negative effects, Other training (e.g., stretching) was
The athletes in our study were highly trained rowers including obviously not effective above a certain threshold, as indicated by
several medalists of Junior and U23 world championships, and it the negative correlation with P2,000 m .
is well-known that significant and worthwhile improvements in
maximal oxygen uptake are not easily achieved in this group of Duration of the Intervention
athletes. The intervention period of 11 weeks is generally long enough to
Data from other endurance athletes (e.g., cyclists and allow for physiological adaptations and is comparable to studies
runners) suggested, that POL might be superior regarding key in elite (Stöggl and Sperlich, 2014) or sub-elite athletes (Neal
endurance variables including V̇O2max , P2[BLa] , P4[BLa] , or time- et al., 2013).
trial performance (Neal et al., 2013; Stöggl and Sperlich, 2014)
and confirmed by several uncontrolled training studies using Training Intensity Distribution
observational data (Billat et al., 2001; Seiler and Kjerland, 2006; In our study the percentage of Z2 and Z3 was significantly
Sandbakk et al., 2011; Tønnessen et al., 2014) and also studies different between PYR and POL with large effect sizes, indicating
emphasizing high-intensity in rowers (Ní Chéilleachair et al., a relevant difference between groups in intensities near and
2016). above lactate threshold. However, the accumulated percentage of
We will therefore briefly discuss possible reasons why training in Z2 and Z3 did not exceed 7% in any group, which
there were no clear differences in important physiological is very similar to classical rowing data (Steinacker et al., 1998;
determinants between POL and PYR, including illness, training Guellich et al., 2009) but appears to be relatively low compared
volume, mode, and TID. to the majority of current data in rowers varying between 7%
(Bourgois et al., 2013), 15% (Nybo et al., 2014), 19% (Seiler
Illness and Tønnessen, 2009), and 23% (Plews et al., 2014), and as well
Due to the typical rough weather conditions between January as studies investigating POL involving other disciplines [20%
and March in central Europe, the pre-competition period of the (Neal et al., 2013) and 32% (Stöggl and Sperlich, 2014)]. Thus,
rowers in our study was frequently disturbed by minor illnesses we assume that the equally low percentage spent in Z2 and Z3
like colds and upper respiratory tract infections. Several athletes was not a sufficient stimulus to improve e.g., oxygen uptake,
in our group experienced minor illnesses with the cancellations since high-intensity exercise is more effective in inducing central
of single and multiple training sessions. However, based on the adaptations, as reported in highly trained cyclists (Laursen et al.,
training diaries and our statistical analysis we did not detect any 2002). In addition, our study clearly indicates that a polarized
significant correlation between days of illness and performance TID is not superior as such, but necessitates an optimal and
outcome or any differences between groups. As the athletes may probably higher sum of Z2 and Z3 intensities than realized by our
not have always reported re-scheduling of training due to minor POL-group. Obviously, POL requires an optimal and probably
illness (e.g., substituting cycle ergometer for rowing) it may be greater overall proportion of Z2- and Z3-intensitiy in contrast to
possible that mild infectious diseases have affected the outcome of the TID accomplished by our POL-group.
the study. However, it is noteworthy that frequent re-scheduling As recommend by the scientific literature (Bangsbo et al.,
is a real-life circumstance within the training process thereby 2010; Tønnessen et al., 2014) our study involved an increase
constantly altering the prescribed TID. in intensity over time especially in PYR to taper for the first
national trials. Since especially PYR increased Z3 toward the end
Training Volume of the study, which is a real-training procedure, and since greater
Data concerning the precise training volumes of high amounts of Z3 are reported to introduce rapid adaptions (Driller
performance rowers are scarce. The average weekly volumes et al., 2009; Ní Chéilleachair et al., 2016), the great amount of Z3
of groups and single cases vary from 102 km/wk (Seiler and in PYR during the last 2 weeks of the study period suggests, that
Tønnessen, 2009), 111.9 ± 43.7 km/wk (Tran et al., 2013), pronounced and short periods of polarized training after several
119 km/wk (Lacour et al., 2009), 124 km/wk (Mikulic, 2011), weeks of Z1 and Z2 training are a sufficient stimulus to improve
127 km/wk (Bourgois et al., 2013) to 135 km/wk (Nybo et al., performance. We assume, this effect of real-life TID very likely
2014). Based on the aforementioned reports, the average training contributed to the lack of differences between POL and PYR.

Frontiers in Physiology | www.frontiersin.org 112 August 2017 | Volume 8 | Article 515

Treff et al. Polarized Training in Rowers

Due to the aforementioned reasons including illness, fatigue minor illnesses and environmental factors caused elimination
or environmental conditions, some rowers in POL and PYR or altering of sessions with higher intensity, which might not
showed a greater “polarization” than others, as expressed by have been reported in the diaries. In addition, the athletes in
the Polarization-Index. Interestingly, we found more frequent our study rowed in different boat types, ranging from single
improvements of P2,000 m in those rowers who trained more sculls to crew boats like four and quadruple sculls. Rowing
polarized, irrespective of the group they were allocated to. in crew boats hinders the strict individual adherence to a
Judging from the plot in Figure 5, we observed (with the prescribed TID, because the individual rower has to adapt to
exception of one outlier) that higher levels of polarization led a given pace that emerges from the skills and physiological
to an increase of more than 1.3% (i.e., the estimated error capacity of the crew, which also contributed to the scattering of
of measurement) in P2,000 m , especially if Polarization-Index individual TIDs and physiological changes, furthermore partly
was >2.3 a.u. The notion, that variation including polarization explaining the differences to studies in e.g., cyclists. Another
needs to be consequently implemented to offer considerable limitation of our study is that the calculation of the TIDs is
advantages is plausible, because other studies reported increases not based on HR-logfiles, but on the rowers’ diary entries, based
of performance and/or V̇O2max in rowers after high intensity- on their HR-measurements, time, and distance trained in each
interventions between four (Driller et al., 2009) and eight weeks zone. The lack of HR-logfiles for direct analysis hindered us to
(Ní Chéilleachair et al., 2016). Since the average changes in distinguish between training intensities within a given intensity
P2,000 m were not related to any physiological variable, other zone. Nevertheless, we cross-checked the entries with the coaches
factors including efficiency and pacing may have accounted for and associated rowers to overcome this limitation. An additional
the changes. measurement after 8 weeks of the study would have theoretically
P4[BLa] is an established and valuable parameter to assess allowed to quantify the impact of the high-intensity training in
performance in rowing (Ingham et al., 2002; Smith and Hopkins, PYR during the last three weeks of the study (Figure 3). However,
2012) and a relevant fitness marker for many high performance the interruption of the training process due to several days of
coaches in rowing (Altenburg et al., 2012). Surprisingly, we did obligatory tapering before the measurements and two additional
not find increases of P2[BLa] and P4[BLa] in POL, but small and days of testing would have been an unrealistic training scenario,
more frequent improvements of P4[BLa] in PYR. This is in line thereby causing another limitation and moreover not acceptable
with others, also reporting minor improvements in P4[BLa] with by the coaches and athletes. Finally, athletes were inevitably
POL compared to a control group emphasizing low-intensity aware of the group (POL or PYR) they were allocated to. We
(Ingham et al., 2008). therefore cannot exclude expectancy effects.
According to our data, the higher percentage of Z3 (or the
lower percentage of Z2) in POL contributes to the unaltered CONCLUSION
or even lowered P4[BLa] values in POL, as indicated by the
high negative correlation between P4[BLa] and Polarization-Index We conclude that on a group level, POL is not superior to a
or percentage of Z3. The notion, that Z2 training is essential dynamic, real-life PYR distribution when percentage of Z1 is
for improvements in rowing performance is in line with other clamped to ∼93%. However, it seems POL can have ergogenic
reports on successful elite rowers, whose training schedule always effects, i.e., improved P2,000 m , if applied consistently and with a
incorporated higher percentage of Z2-training, as indicated by Polarization-Index ≥2.3 a.u. Taking previous data of elite non-
data from New Zealand (Plews et al., 2014), Norway (Seiler and rowing athletes into account, we assume that higher percentage
Tønnessen, 2009), and Denmark (Nybo et al., 2014). In addition, of Z3 is necessary to achieve potential ergogenic superiority of
coaches of the POL-group reported (but did not quantify) that POL compared to PYR.
during the intervention period, the speed in Z1 was partially
lower than before the study, to allow for recovery from fatigue ETHICS STATEMENT
inducing Z3 sessions. Since energy expenditure increases with
boat speed by 2.2- to 2.4-fold power (Secher, 1993), we assume This study was carried out in accordance with the
that the sessions at the lower end of the T1-range with low recommendations of ethical review board of the University
metabolic cost and muscular force were insufficient to further of Ulm with written informed consent from all subjects. All
stimulate adaptions, thereby explaining the lack of improvement subjects gave written informed consent in accordance with the
in POL. Differences of TID within Z1 are probably relevant to Declaration of Helsinki. The protocol was approved by ethical
induce further adaptations, however, since we did not collect the review board of the University of Ulm.
necessary data, this notion warrants further investigation.
AUTHOR CONTRIBUTIONS
Limitations
Coaches and athletes volunteered to participate in the study GT: Planned and designed the study, conducted measurements,
and were fully informed about the two training regimes, but analyzed the data, prepared the manuscript. KW: Planned and
the scattering of the TIDs within the POL group indicates that designed the study, conducted measurements, analyzed the data,
some rowers did not entirely follow the training program in the edited the manuscript. MS: Conducted measurements, edited the
same consequence. This limitation of our study was attributable manuscript. JS: Designed the study, edited the manuscript. MB:
to concerns by coaches and athletes to adopt a new TID Analyzed the data. BS: Designed the study, analyzed the data,
with the possibility for non-functional overreaching. Further, edited the manuscript.

Frontiers in Physiology | www.frontiersin.org 113 August 2017 | Volume 8 | Article 515

Treff et al. Polarized Training in Rowers

FUNDING ACKNOWLEDGMENTS
The project was supported by the German Federal Institute of The authors thank all rowers and their coaches for their
Sports Sciences (BISp, ZMVI4-070707/16). enthusiasm and the time they dedicated to the study.

REFERENCES Laursen, P. B., Shing, C. M., Peake, J. M., Coombes, J. F., and Jenkins, D. G. (2002).
Interval training program optimization in highly trained endurance cyclists.
Altenburg, D., Mattes, K., and Steinacker, J. M. (2012). Manual for Rowing Training Med. Sci. Sports Exerc. 34, 1801–1807. doi: 10.1097/00005768-200211000-
- Technique, Performance and Planning, 2nd Edn. Wiebelsheim: Limpert. 00017
Bangsbo, J., Elbe, J. M., Andersen, M., and Poulsen, A. L. (2010). Mäestu, J., Jürimäe, J., and Jürimäe, T. (2005). Monitoring of
International Consensus Conference “Performance in Top Sports performance and training in rowing. Sports Med. 35, 597–617.
Involving Intense Exercise”. Scand. J. Med. Sci. Sports 20(Suppl. 2), 2–4. doi: 10.2165/00007256-200535070-00005
doi: 10.1111/j.1600-0838.2010.01183.x Midgley, A. W., McNaughton, L. R., Polman, R., and Marchant, D. (2007).
Billat, V. L., Demarle, A., Slawinski, J., Paiva, M., and Koralsztein, J. P. (2001). Criteria for determination of maximal oxygen uptake: a brief critique
Physical and training characteristics of top-class marathon runners. Med. Sci. and recommendations for future research. Sports Med. 37, 1019–1028.
Sports Exerc. 33, 2089–2097. doi: 10.1097/00005768-200112000-00018 doi: 10.2165/00007256-200737120-00002
Bourgois, J., Steyaert, A., and Boone, J. (2013). Physiological and anthropometric Mikulic, P. (2011). Maturation to elite status: a six-year physiological case study
progression in An International Oarsman: a 15 year case study. Int. J. Sports of a world champion rowing crew. Eur. J. Appl. Physiol. 111, 2363–2368.
Physiol. Perform. 9, 723–726. doi: 10.1123/ijspp.2013-0267 doi: 10.1007/s00421-011-1870-y
Cohen, J. (1988). Statistical Power Analysis for the Behavioural Sciences. Hillside. Neal, C. M., Hunter, A. M., Brennan, L., O’Sullivan, A., Hamilton, D. L.,
NJ: Lawrence Earlbaum Associates. De Vito, L., et al. (2013). Six weeks of a polarized training-intensity
Dickhuth, H. H., Huonker, M., Münzel, T., Drexler, H., Berg, A., and Keul, distribution leads to greater physiological and performance adaptations
J. (1991). “Individual Anaerobic Threshold for Evaluation of Competitive than a threshold model in trained cyclists. J. Appl. Physiol. 114, 461–471.
Athletes and Patients with Left Ventricular Dysfunction,” in Advances in doi: 10.1152/japplphysiol.00652.2012
Ergometry, eds N. Bachl, T. E. Graham, H. Löllgen (Berlin, Heidelberg: Nevill, A. M., Allen, S. V., and Ingham, S. A. (2011). Modelling the
Springer), 173–179. determinants of 2000 m rowing ergometer performance: a proportional,
Driller, M. W., Fell, J. W., Gregory, J. R., Shing, C. M., and Williams, A. D. (2009). curvilinear allometric approach. Scand. J. Med. Sci. Sports 21, 73–78.
The effects of high-intensity interval training in well-trained rowers. Int. J. doi: 10.1111/j.1600-0838.2009.01025.x
Sports Physiol. Perform. 4, 110–121. doi: 10.1123/ijspp.4.1.110 Ní Chéilleachair, N. J. N., Harrison, A. J., and Warrington, G. D. (2016).
Fiskerstrand, A., and Seiler, K. S. (2004). Training and performance characteristics HIIT enhances endurance performance and aerobic characteristics more
among Norwegian International Rowers 1970-2001. Scand. J. Med. Sci. Sports than high-volume training in trained rowers. J. Sports Sci. 35, 1052–1058.
14, 303–310. doi: 10.1046/j.1600-0838.2003.370.x doi: 10.1080/02640414.2016.1209539
Foster, C., Florhaug, J., Franklin, J., Gottschall, L. A., Hrovatin, L., Parker, S., et al. Nielsen, H. B. (1999). PH after competitive rowing: the lower physiological range?
(2001). A new approach to monitoring exercise training. J. Strength Cond. Res. Acta Physiol. Scand. 165, 113–114. doi: 10.1046/j.1365-201x.1999.00485.x
15, 109–115. doi: 10.1519/00124278-200102000-00019 Nielsen, T. (ed.) (2009/2017). Training Program for Clubs and Individuals - World
Guellich, A., Seiler, S., and Emrich, E. (2009). Training methods and intensity Rowing. Training Program for Clubs and Individuals - World Rowing.
distribution of young world-class rowers. Int. J. Sports Physiol. Perform. 4, Nybo, L., Schmidt, J. F., Fritzdorf, S., and Nordsborg, N. B. (2014). Physiological
448–460. doi: 10.1123/ijspp.4.4.448 characteristics of an aging olympic athlete. Med. Sci. Sports Exerc. 46,
Hahn, A., Bourdon, P., and Tanner, R. (2000). “Protocols for the physiological 2132–2138. doi: 10.1249/MSS.0000000000000331
assessment of rowers” in Physiological Tests for Elite Athletes 1st Edn., ed C. Plews, D. J., Laursen, P. B., Kilding, A. E., and Buchheit, M. (2014). Heart-
J. Gore, (Champaign, IL: Human Kinetics), 311–326. rate variability and training-intensity distribution in elite rowers. Int. J. Sports
Hopkins, W. G. (2003). A Scale of Magnitudes for Effect Statistics. A New View of Physiol. Perform. 9, 1026–1032. doi: 10.1123/ijspp.2013-0497
Statistics. Available online at: http://www.sportsci.org/resource/stats/effectmag. Roth, W., Schwanitz, P., Pas, P., and Bauer, P. (1993). Force-time
html (Accessed April 24, 2017). characteristics of the rowing stroke and corresponding physiological muscle
Ingham, S. A., Carter, H., Whyte, G. P., and Doust, J. H. (2008). Physiological and adaptations. Int. J. Sports Med. 14(Suppl. 1), 32–34. doi: 10.1055/s-2007-
performance effects of low- versus mixed-intensity rowing training. Med. Sci. 1021221
Sports Exerc. 40, 579–584. doi: 10.1249/MSS.0b013e31815ecc6a Sandbakk, A., Holmberg, H.-C., Leirdal, S., and Ettema, G. (2011). The
Ingham, S. A., Whyte, G. P., Jones, K., and Nevill, A. M. (2002). Determinants of Physiology of World-class Sprint Skiers. Scand. J. Med. Sci. Sports 21, e9–e16.
2000 m rowing ergometer performance in elite rowers. Eur. J. Appl. Physiol. 88, doi: 10.1111/j.1600-0838.2010.01117.x
243–246. doi: 10.1007/s00421-002-0699-9 Secher, N. H. (1993). Physiological and biomechanical aspects
Jürimäe, J., Mäestu, J., Jürimäe, T., and Pihl, E. (2000). Prediction of rowing of rowing. implications for training. Sports Med. 15, 24–42.
performance on single sculls from metabolic and anthropometric variables. J. doi: 10.2165/00007256-199315010-00004
Hum. Movement Stud. 38, 123–136. Seiler, K. S., and Kjerland, G. A. (2006). Quantifying training intensity distribution
Kindermann, W., Simon, G., and Keul, G. (1979). The significance of the aerobic- in elite endurance athletes: is there evidence for an “optimal” distribution?
anaerobic transition for the determination of work load intensities during Scand. J. Med. Sci. Sports 16, 49–56. doi: 10.1111/j.1600-0838.2004.00418.x
endurance training. Eur. J. Appl. Physiol. 42, 25–34. doi: 10.1007/BF00421101 Seiler, S. (2010). What is best practice for training intensity and duration
Kleshnev, V., and Nolte, V. (2011). “Learning from racing” in Rowing Faster distribution in endurance athletes? Int. J. Sports Physiol. Perform. 5, 276–291.
(Serious Training for Serious Rowers) 2nd Edn, ed V. Nolte (New York, NY: doi: 10.1123/ijspp.5.3.276
Human Kinetics Inc.), 251–265. Seiler, S., and Tønnessen, E. (2009). Intervals, thresholds, and long slow distance:
Lacour, J. R., Messonnier, L., and Bourdin, M. (2009). Physiological correlates the role of intensity and duration in endurance training. Sportscience 13, 32–53.
of performance. case study of a world-class rower. Eur. J. Appl. Physiol. 106, Available online at: http://www.sportsci.org/2009/ss.htm
407–413. doi: 10.1007/s00421-009-1028-3 Smith, T. B., and Hopkins, W. G. (2011). Variability and predictability
Larsson, P. U., Wadell, K. M. E., Jakobsson, E. J. I., Burlin, L. U., of finals times of elite rowers. Med. Sci. Sports Exerc. 43, 2155–2160.
and Henriksson-Larsén, K. B. (2004). Validation of the metamax II doi: 10.1249/MSS.0b013e31821d3f8e
portable metabolic measurement system. Int. J. Sports Med. 25, 115–123. Smith, T., B., and Hopkins, W. G. (2012). Measures of rowing performance. Sports
doi: 10.1055/s-2004-819953 Med. 42, 343–358. doi: 10.2165/11597230-000000000-00000

Frontiers in Physiology | www.frontiersin.org 114 August 2017 | Volume 8 | Article 515

Treff et al. Polarized Training in Rowers

Soper, C., and Hume, P. A. (2004). Rowing: reliability of power output during Physiol. Funct. Imaging 29, 241–244. doi: 10.1111/j.1475-097X.2009.
rowing changes with ergometer type and race distance. Sports Biomech. 3, 00867.x
237–248. doi: 10.1080/14763140408522843 Winkert, K., Pieper, L., Steinacker, J. M., and Treff, G. (2016). “High
Steinacker, J. M., Lormes, W., Lehmann, M., and Altenburg, D. (1998). Training test-retest reliability of test-duration and V̇O2max with a software
of rowers before world championships. Med. Sci. Sports Exerc. 30, 1158–1163. supported ramp-test for rowing ergometry [Abstract],” in 21st
doi: 10.1097/00005768-199807000-00022 Annual Congress of the European College of Sport Science, Vienna.
Stöggl, T. L., and Sperlich, B. (2015). The training intensity distribution Available online at: http://ecss-congress.eu/2016/16/index.php/
among well-trained and elite endurance athletes. Front. Physiol. 6:295. programme/scientific-programme/mini-oral-sessions (Accessed October
doi: 10.3389/fphys.2015.00295 25, 2016).
Stöggl, T., and Sperlich, B. (2014). Polarized training has greater impact on key
endurance variables than threshold, high intensity, or high volume training. Conflict of Interest Statement: The authors declare that the research was
Front. Physiol. 5:33. doi: 10.3389/fphys.2014.00033 conducted in the absence of any commercial or financial relationships that could
Tønnessen, E., Sylta, A., Haugen, T. A., Hem, E., Svendsen, I. S., and Seiler, be construed as a potential conflict of interest.
S. (2014). The road to gold: training and peaking characteristics in the
year prior to a gold medal endurance performance. PLoS ONE 9:e101796. Copyright © 2017 Treff, Winkert, Sareban, Steinacker, Becker and Sperlich. This
doi: 10.1371/journal.pone.0101796 is an open-access article distributed under the terms of the Creative Commons
Tran, J., Rice, J., Main, L., and Gastin, P. (2013). Training characteristics of elite Attribution License (CC BY). The use, distribution or reproduction in other forums
australian rowers: training volume, loading patterns, and training variety. J. Sci. is permitted, provided the original author(s) or licensor are credited and that the
Med. Sport 16:e94. doi: 10.1016/j.jsams.2013.10.226 original publication in this journal is cited, in accordance with accepted academic
Voliantis, S., and Secher, N. H. (2009). Rowing, the ultimate challenge practice. No use, distribution or reproduction is permitted which does not comply
to the human body implications for physiological variables. Clin. with these terms.

Frontiers in Physiology | www.frontiersin.org 115 August 2017 | Volume 8 | Article 515

 ORIGINAL RESEARCH
published: 11 July 2017
doi: 10.3389/fphys.2017.00485

The Mucosal Immune Function Is Not

Compromised during a Period of
High-Intensity Interval Training. Is It
Time to Reconsider an Old
Assumption?
Dennis-Peter Born *, Christoph Zinner and Billy Sperlich

Integrative and Experimental Exercise Science, Institute for Sport Sciences, University of Wuerzburg, Wuerzburg, Germany

Purpose: The aim of the study was to evaluate the mucosal immune function and
circadian variation of salivary cortisol, Immunoglobin-A (sIgA) secretion rate and mood
during a period of high-intensity interval training (HIIT) compared to long-slow distance
training (LSD).
Methods: Recreational male runners (n = 28) completed nine sessions of either
HIIT or LSD within 3 weeks. The HIIT involved 4 × 4 min of running at 90–95%
of maximum heart rate interspersed with 3 min of active recovery while the LSD
Edited by: comprised of continuous running at 70–75% of maximum heart rate for 60–80 min.
Robert Aughey,
The psycho-immunological stress-response was investigated with a full daily profile of
Victoria University, Australia
salivary cortisol and immunoglobin-A (sIgA) secretion rate along with the mood state on
Reviewed by:
Andrew McKune, a baseline day, the first and last day of training and at follow-up 4 days after the last day
University of Canberra, Australia of training. Before and after the training period, each athlete’s running performance and
John Hough,
University of Bedfordshire,
peak oxygen uptake (V· O2peak ) was determined with an incremental exercise test.
United Kingdom Results: The HIIT resulted in a longer time-to-exhaustion (P = 0.02) and increased
*Correspondence: V· O2peak compared to LSD (P = 0.01). The circadian variation of sIgA secretion rate
Dennis-Peter Born
[email protected] showed highest values in the morning immediately after waking up followed by a
decrease throughout the day in both groups (P < 0.05). With HIIT, the wake-up response
Specialty section: of sIgA secretion rate was higher on the last day of training (P < 0.01) as well as the
This article was submitted to
Exercise Physiology, area under the curve (AUCG ) higher on the first and last day of training and follow-up
a section of the journal compared to the LSD (P = 0.01). Also the AUCG for the sIgA secretion rate correlated
Frontiers in Physiology
with the increase in V· O2peak and running performance. The AUCG for cortisol remained
Received: 06 March 2017
unaffected on the first and last day of training but increased on the follow-up day with
Accepted: 23 June 2017
Published: 11 July 2017 both, HIIT and LSD (P < 0.01).
Citation: Conclusion: The increased sIgA secretion rate with the HIIT indicates no compromised
Born DP, Zinner C and Sperlich B
(2017) The Mucosal Immune Function
mucosal immune function compared to LSD and shows the functional adaptation of the
Is Not Compromised during a Period mucosal immune system in response to the increased stress and training load of nine
of High-Intensity Interval Training. Is It
sessions of HIIT.
Time to Reconsider an Old
Assumption? Front. Physiol. 8:485. Keywords: circadian rhythm, cortisol, diurnal profile, endurance, high-volume training, immunoglobin-A,
doi: 10.3389/fphys.2017.00485 periodization

Frontiers in Physiology | www.frontiersin.org 116 July 2017 | Volume 8 | Article 485

Born et al. Immune Function and High-Intensity Training

INTRODUCTION during periods of intensified training discussed (Gleeson, 2007;

He et al., 2010). Especially high-intensity exercise substantially
The exercise duration and intensity are key components to increases the levels of cortisol (Allgrove et al., 2008). Therefore,
design training programs, stimulate adaptation and maximize the question arises whether the repeated exposure to HIIT
performance of the modern day endurance athlete (Hydren and over a prolonged period of time would compromise the
Cohen, 2015). Especially short intervals of 2–8 min at 90–95% mucosal immune function despite the tempting ergogenic effects
of maximal heart rate (HRmax ) interspersed with periods of including improved V· O2peak .
incomplete recovery known as high-intensity interval training Earlier studies however mostly investigated the mucosal
(HIIT) are applied to maximize central and peripheral adaptation immune function with single pre- and post-training saliva
(Buchheit and Laursen, 2013a,b; Ronnestad et al., 2014, 2016; samples (He et al., 2010; Gill et al., 2014; Ihalainen et al.,
Stoggl and Sperlich, 2014, 2015; Sylta et al., 2016). When 2016). Due to the dramatic decrease of cortisol and sIgA
performing three training sessions per week for 8 weeks the HIIT from early morning throughout the day (Hucklebridge et al.,
(e.g., 4 × 4 min) significantly increased the peak oxygen uptake 1998; Rohleder et al., 2007), a full daily profile is warranted
(V· O2peak ) while no improvements occurred with the long-slow to investigate the circadian variation of these biomarkers of
distance (LSD) running at 70% of HRmax for 45 min (Helgerud interest. Therefore, the aim of the study was to evaluate the
et al., 2007). The LSD sessions in the latter study however were mucosal immune function and circadian variation of salivary
fairly short while longer training sessions including low intensity cortisol, sIgA secretion rate and mood during a period of nine
have a small effect on V· O2peak (Stoggl and Sperlich, 2014). sessions of HIIT performed within 3 weeks compared to LSD.
Still the superior benefits of HIIT to improve V· O2peak The hypothesis was that the HIIT would increase levels of
have been shown multiple times in recreational and elite salivary cortisol, reduce sIgA secretion rate and impair mood
athletes among various endurance sports including running, thereby showing a more pronounced psycho-immunological
cycling, rowing, and cross-country skiing (Helgerud et al., 2007; stress-response and compromised mucosal immune function
Ronnestad et al., 2014, 2016; Ni Cheilleachair et al., 2017). compared to the LSD.
Despite the promising effects of HIIT however, the question
arises among athletes, coaches and scientists whether the training
METHODS
load of several HIIT sessions within a short period of time
might compromise the mucosal immune function. Especially Subject Characteristics
the chronic exposure to high training loads has generally been For the present investigation, 28 recreational endurance runners
assumed to increase the incidence of upper-respiratory tract were assigned into two groups performing either HIIT (n =
infection (URTI; Trochimiak and Hubner-Wozniak, 2012). 16, age: 25 ± 4 years, body mass: 76 ± 5 kg, body height:
A practical method to assess the stress-response from a 179 ± 6 cm) or LSD (n = 12, age: 25 ± 3 years, body mass:
psycho-immunological perspective during periods of intensified 77 ± 11 kg, body height: 182 ± 5 cm). After being informed
training is the combined assessment of the circadian variation about the potential risks and benefits of the study involved,
of biomarkers in saliva and mood state (Papacosta et al., all runners gave their written consent to participate. The study
2013; Born et al., 2016). While the concentrations of enzymes, was approved by the Ethical Committee of the University of
hormones, and anti-bacterial compounds are far lower in saliva Wuppertal and performed in accordance with the Declaration of
than in blood samples the relative changes in response to Helsinki.
exercise are highly correlated to the blood serum (Cadore et al.,
2008; VanBruggen et al., 2011; Tanner et al., 2014). The non- Study Design
invasive collection of saliva allows a greater sampling rate and Each athlete completed nine sessions of either HIIT or LSD
in an athletic population the entire circadian variation can be within a period of 3 weeks with at least 1 day between the sessions
investigated with minimal interference in the daily training and (Stoggl and Sperlich, 2014) in addition to their routine aerobic
recovery routines (Gatti and De Palo, 2011; Papacosta and Nassis, training (4.0 ± 2.0 vs. 3.8 ± 1.6 h/week with the HIIT and
2011). LSD group, respectively) as performed previously (Ronnestad
Salivary immunoglobin-A (sIgA) and cortisol are et al., 2014; Faiss et al., 2015). Before (Pre-) and after (Post-)
biomarkers of particular interest when investigating the the training period all participants performed an incremental
psycho-immunological stress-response during periods of test to exhaustion for the determination of running performance,
intensified training. In ultra-marathon runners the extreme i.e., time-to-exhaustion (TTE), as well as variables related to the
competition load acutely decreased the sIgA secretion rate and cardio-respiratory and metabolic capacity. For each individual
increased the levels of cortisol immediately after the exercise Pre- and Post- was scheduled on the same time of the day
(Gill et al., 2014). Interestingly, low sIgA correlated with a high and performed with the same pair of running shoes within the
susceptibility of URTI and number of sick days during a period same ambient air condition (20 ± 1◦ C and 36 ± 4% relative
of polarized endurance training including both, continuous and humidity).
interval training session (Ihalainen et al., 2016). Also an inverse Each HIIT session was initiated with a 10-min warm-up
correlation between levels of cortisol and sIgA secretion rate has of moderate intensity running at 70% of maximum heart
been shown (Hucklebridge et al., 1998) and a potential decisive rate (HRmax ) including short bouts (30–45 s) with a higher
role of cortisol on the exercise-induced immune suppression running intensity to prepare the cardio-respiratory and metabolic

Frontiers in Physiology | www.frontiersin.org 117 July 2017 | Volume 8 | Article 485

Born et al. Immune Function and High-Intensity Training

system for the upcoming intervals. After the warm-up, the Psycho-Immunological Stress-Response
participants performed 4 × 4-min intervals with an exercise In order to assess the acute exercise induced stress-response as
intensity corresponding to the individuals 90–95% of HRmax well as the circadian variation of the markers of interest, on each
interspersed with 3 min of active recovery corresponding to 70% day of measurement (baseline, T1, T9, and follow-up) five saliva
of HRmax (Helgerud et al., 2007). The HR data were recorded samples were taken: (1) immediately after waking up, (2) 30 min
and each runner reached in at least 94% of all intervals (i.e., after waking up, (3) immediately before training, (4) immediately
34 of 36 possible intervals during the training period) the after training, and (5) before going to bed as described previously
targeted exercise intensity of >90% of HRmax to be included (Rohleder et al., 2007; Born et al., 2016).
in the statistical analysis. Due to the delayed HR response To collect the saliva samples all athletes received the following
at the onset of exercise the athletes were instructed to reach instructions: (1) not to eat and drink (other than plain water)
the targeted HR zone (90–95% of HRmax ) within the first 60– or brushing teeth for 30 min beforehand to avoid blood
90 s of each interval as recommended previously (Helgerud contamination of the saliva sample, (2) rinse the mouth with
et al., 2007). The LSD was performed continuously at 70– water and swallow any remaining fluid, (3) start a stopwatch and
75% of HRmax . The duration for each LSD session was 60, 70, passively collect saliva in the mouth while resting in a seated
and 80 min for the first, second and third week of training, position with the head tilted slightly forward for exactly 2 min, (4)
respectively. spit the accumulated saliva through a sterile polypropylene straw
In order to investigate the psycho-immunological stress- into a polypropylene tube (Sali-Cap Tubes, IBL International,
response to nine sessions of either HIIT or LSD, saliva samples Hamburg, Germany), (5) store the saliva tube in the freezer at
along with questionnaires were taken on a baseline day before −18◦ C and bring the sample to the lab the next morning for
the start of the study, the first (T1) and last (T9) day of training. the analysis of cortisol and sIgA, as recommended previously
The follow-up measurement was performed on the day of the (Granger et al., 2007). For the first sample of the day, the
post-test 4 days after the last day of training (Figure 1). participants were instructed to collect the saliva immediately after
waking up when the alarm went off while sitting at the edge of the
Data Collection bed. For the next 30 min the participants were allowed to walk
Cardio-Respiratory and Metabolic Response around, take a shower and prepare the breakfast. They however
The incremental running test was performed on a treadmill were not allowed to eat, drink or continue sleeping. The exact
(H/P Cosmos, Mercury, Nussdorf-Traunstein, Germany) and time of the day for each saliva sample was reported in a specific
initiated with a running velocity of 2.4 m/s. Subsequently, the protocol.
running velocity was increased by 0.4 m/s in 5-min intervals until A scaling on the transparent polypropylene tube showed the
voluntary exhaustion. The incremental test was performed at 1% volume of saliva collected. In case the collected saliva was <1
inclination to simulate the missing air resistance and drag forces mL within the first 2 min the athletes continued collecting
of outdoor running (Gore, 2000). Maximal effort was considered saliva for another 2 min. The additional collection time was
when the runners met three of the following four criteria: (1) reported in the protocol for the subsequent determination of
V· O2 showed a leveling-off defined as an increase of V· O2 of <2.1 saliva flow rate. Especially high-intensity exercise reduces the
mL/kg/min (Taylor et al., 1955), (2) respiratory exchange ratio flow rate of saliva by enhancing sympathetic and/or attenuating
> 1.05, (3) HR ≥ 90% of the age-predicted HR, (4) ratings of the parasympathetic activation (Papacosta and Nassis, 2011).
perceived exertion ≥18 on Borg’s 6–20 scale (Borg, 1970). Therefore, the saliva flow rate was employed to calculate the
During the incremental running test, the participants were secretion rate of sIgA from the absolute concentrations for any
equipped with an open-circuit breath-by-breath gas analyzer further statistical analysis (Allgrove et al., 2008; Papacosta and
(MetaMax3B_R2, Cortex Biophysik GmbH, Leipzig, Germany) Nassis, 2011).
breathing through a turbine flowmeter which was attached to Along with the saliva sample, the athletes rated their current
a proper fitting face mask covering the mouth and nose (7,450 mood on a questionnaire adapted by Wilhelm and Schoebi
Series V2 TM Mask, Hans Rudolph Inc., Shawnee, USA). The (2007). The items included levels of stress, anxiety, annoyance,
HR was collected time aligned with the V· O2 data using a happiness, exhaustion, and energy ranked on a Likert scale
chest belt (H7, Polar Electro Oy, Kempele, Finland). Before each from (1) not at all to (5) very much. After decoding, the
test the oxygen (O2 ) and carbon dioxide (CO2 ) sensors of the mood was assessed by the sum of all items while a high
gas analyzer were 2-point calibrated to ambient air (20.93% score indicated vitality and well-being and a low score a
O2 and 0.03% CO2 ) and calibration gas containing 15% O2 suppressed mood state as described previously (Born et al.,
and 5% CO2 (UN 1950 Aerosols, Cortex Biophysik GmbH, 2016). The athletes were asked to report any unusual and
Leipzig, Germany) to anticipate the expiratory gas compound. stressful events including signs of URTI immediately prior to
The turbine’s flow volume was calibrated using a 3-L syringe the saliva sampling since a sympathetic stress-response would
(M9474-C, Medikro Oy, Kuopio, Finland). The levels of blood dramatically affect the concentration of cortisol, saliva flow rate
lactate concentration were determined in the capillary blood and mood. In only one occasion a participant reported to be
sampled from the left ear lobe (LactatePro2, LT-1730, Arkray, upset from such an event (i.e., a private matter that cased him
Kyoto, Japan) and used for the subsequent linear extrapolation stress but was not related to the study) and the sample from
of the running velocities at 2 and 4 mmol/L blood lactate this particular point of measurement was excluded from the
concentration. analysis.

Frontiers in Physiology | www.frontiersin.org 118 July 2017 | Volume 8 | Article 485

Born et al. Immune Function and High-Intensity Training

FIGURE 1 | The study design with the time-points for the saliva sampling and questionnaire on the baseline day, the first (T1) and last (T9) day of training and the
follow-up for both groups.

All saliva samples were centrifuged at 2,000 g for 10 for the psycho-immunological stress-response on each day of
min to separate the firm mucus at the bottom of the tube. measurement. Pearson’s product moment correlation coefficient
The supernatant aqueous fraction was used to analyze the was used to identify potential variables that were related to
concentration of cortisol and sIgA using commercially available the change in V· O2peak and TTE. All data were recorded and
enzyme-immunoassay kits (DRG Instruments, Marburg, prepared using Excel 2010 (Microsoft Corp., Redmond, USA)
Germany). The standard ranges for the determination of cortisol and analyzed subsequently with Statistical 10.0 (StatSoft Inc.,
and sIgA were 2–80 ng/mL and 6.9–400 µg/mL with a sensitivity Tulsa, USA).
of 0.5 ng/mL and 0.5 µg/mL at the 95% confidence limit,
respectively. The intra-assay coefficient of variation (CV) for
RESULTS
cortisol and sIgA were 5.2 and 3.6% with an inter-assay CV of 5.7
and 5.5%, respectively. In order to cope with the inter-individual Physiological Assessment
variation in the concentration of cortisol and sIgA secretion rate, The performance data as well as cardio-respiratory and metabolic
the values from the baseline day were used to normalize the response to nine sessions of HIIT and LSD are presented in
data. Table 1. While HIIT and LSD increased the TTE from Pre- to
Post- (P < 0.01), the interaction effect revealed that the HIIT
Statistical Analysis resulted in a longer TTE at Post- compared to the LSD (P =
The data are presented as mean values ± standard deviations 0.02). The V· O2peak increased with the HIIT only (interaction
(SD), normal distribution was confirmed with Shapiro-Wilk’s test effect, P = 0.01) and the levels of blood lactate concentration
and an alpha-level <0.05 considered as statistically significant. lessened with the LSD only (interaction effect, P = 0.01). The
A 2-way analysis of variance (ANOVA) with repeated measure running velocities at 2 and 4 mmol/L blood lactate concentration
using Fisher’s post-hoc test was performed to detect significant increased with both, HIIT and LSD (P < 0.01 and P < 0.01,
differences between the training intensity (HIIT vs. LSD) and respectively).
time-points of measurement. Additionally, effect size (partial
eta2 ) and statistical power were calculated for each variable. As Psycho-Immunological Stress-Response
suggested previously (Pruessner et al., 2003), the area under Table 2 illustrates the detailed analysis of the AUCG for the
the curve with respect to ground level (AUCG ) was determined psycho-immunological stress-response during the first and last

Frontiers in Physiology | www.frontiersin.org 119 July 2017 | Volume 8 | Article 485

Born et al. Immune Function and High-Intensity Training

TABLE 1 | The performance data as well as cardio-respiratory and metabolic response to nine sessions of HIIT compared to LSD (mean ± SD).

HIIT LSD F-value P-value Partial eta2 Test power

Time-to-exhaustion (s) Pre- 1887 ± 290 1764 ± 221 a) F (1, 26) = 4 n.s.
Post- 2129 ± 298*+ 1861 ± 220+ b) F (1, 26) = 33 P < 0.01 0.56 0.99
c) F (1, 26) = 6 P = 0.02 0.19 0.65

Peak oxygen uptake (mL/kg/min) Pre- 49 ± 4.5 52.4 ± 4.8 a) F (1, 26) = 1 n.s.
Post- 51.5 ± 3.9+ 51.8 ± 5.3 b) F (1, 26) = 0 n.s.
c) F (1, 26) = 8 P = 0.01 0.24 0.79

Maximum heart rate (beats/min) Pre- 196 ± 8 194 ± 8 a) F (1, 26) = 0 n.s.
Post- 194 ± 8 195 ± 8 b) F (1, 26) = 3 n.s.
c) F (1, 26) = 4 n.s.

Maximum blood lactate concentration (mmol/L) Pre- 8 ± 2.2 9.2 ± 2.1 a) F (1, 26) = 0 n.s.
Post- 8.7 ± 2.2 8.2 ± 1.7+ b) F (1, 26) = 0 n.s.
c) F (1, 26) = 9 P = 0.01 0.25 0.81

Velocity (m/s) at 2 mmol/L blood lactate concentration Pre- 2.6 ± 0.5 2.4 ± 0.4 a) F (1, 26) = 3 n.s.
Post- 2.8 ± 0.4+ 2.6 ± 0.3+ b) F (1, 26) = 13 P < 0.01 0.33 0.93
c) F (1, 26) = 0 n.s.

Velocity (m/s) at 4 mmol/L blood lactate concentration Pre- 3.3 ± 0.4* 3 ± 0.3 a) F (1, 26) = 8 P = 0.01 0.25 0.80
Post- 3.5 ± 0.3*+ 3.1 ± 0.3+ b) F (1, 26) = 20 P < 0.01 0.44 0.99
c) F (1, 26) = 2 n.s.

Significant differences were identified with a 2-way ANOVA: training intensity (HIIT vs. LSD) x time (Pre- vs. Post-). HIIT, High-intensity training; LSD, Long-slow distance training.
a) Main effect: training intensity (HIIT vs. LSD).
b) Main effect: time (Pre- vs. Post-).
c) Interaction effect: training intensity x time.
*Significant difference compared to LSD. + significant difference compared to Pre-. n.s., not significant.

day of training and during the follow-up. A main effect for the in V· O2peak was related to the sIgA secretion rate on day T1 (r =
time was evident as the AUCG for the levels of cortisol increased 0.39, P = 0.04).
during the follow-up in both groups (P < 0.01). The post-hoc
analysis revealed that, the levels of cortisol were increased from
before to after exercise (P < 0.05) and increased compared to DISCUSSION
the corresponding values at T1 (P < 0.05) with both groups
on the day of follow-up. The cortisol values normalized by the The main findings of the present study were that, the HIIT
end of the day however, showing lower values before going to results in a longer TTE and increased V· O2peak compared to the
bed compared to immediately after waking up on all days of LSD. The ergogenic effects of HIIT were accompanied with an
measurement (i.e., T1, T9, and follow-up) with the HIIT (P < increased sIgA secretion rate evident as a larger AUCG on the
0.01) and LSD (P < 0.01). first and last day of training as well as follow-up. The levels of
A main effect for the training intensity was evident as the cortisol were unaffected by the training intensity (HIIT vs. LSD)
AUCG for sIgA secretion rate was higher with the HIIT on T1, but increased over time on the day of the follow-up with both,
T9 and follow-up (P = 0.01). The post-hoc analysis showed the HIIT and LSD. Mood remained unaffected with both groups
highest sIgA secretion rate in the morning immediately after during the entire training period.
waking up with decreasing values throughout the day in both, The results of the present study are in line with previous
the HIIT (P < 0.05) and LSD (P < 0.05). The wake-up response findings showing the benefits of HIIT to improve important
by the end of the training period at T9 was higher with the HIIT variables related to the endurance performance, i.e., TTE, and
compared to the corresponding value at T1 (P = 0.01) as well V· O2peak (Helgerud et al., 2007; Buchheit and Laursen, 2013a,b;
as compared to the LSD (P < 0.01; Figure 2). Mood remained Ronnestad et al., 2014, 2016; Stoggl and Sperlich, 2014, 2015;
unaffected with respect to training intensity and time. Sylta et al., 2016). The research focus however was, to investigate
the circadian variation of biomarkers in saliva and mood during
Correlation Analysis such a period of HIIT. The question was whether the exposure
Person correlation analysis detected that, the increased TTE from to nine sessions of HIIT compared to LSD within a 3-week time
pre- to post- correlated with the AUCG of sIgA secretion rate on frame would compromise the mucosal immune function besides
the day of follow-up (r = 0.45, P = 0.02). As well, the increase the promising effects on TTE and V· O2peak .

Frontiers in Physiology | www.frontiersin.org 120 July 2017 | Volume 8 | Article 485

Born et al. Immune Function and High-Intensity Training

TABLE 2 | The AUCG for the psycho-immunological stress-response to nine sessions of HIIT compared to LSD on the first (T1) and last (T9) day of training as well as the
follow-up (mean ± SD).

HIIT LSD F-value P-value Partial eta2 Test power

Saliva flow rate (mL/min) T1 1476 ± 198 1346 ± 445 a) F (1, 26) = 1 n.s.
T9 1461 ± 316 1332 ± 393 b) F (1, 26) = 7 P < 0.01 0.21 0.91
Follow-up 1225 ± 233+# 1157 ± 426 c) F (1, 26) = 0 n.s.

Levels of cortisol (ng/mL) T1 1809 ± 384 1554 ± 426 a) F (1, 26) = 1 n.s.
T9 1597 ± 565 1394 ± 257 b) F (1, 26) = 11 P < 0.01 0.3 0.99
Follow-up 1941 ± 354# 2104 ± 855+# c) F (1, 26) = 2 n.s.

Salivary immunoglobin A secretion rate (µg/min) T1 1917 ± 1178* 1010 ± 506 a) F (1, 26) = 8 P = 0.01 0.23 0.77
T9 2123 ± 1394* 1142 ± 628 b) F (1, 26) = 1 n.s.
Follow-up 1968 ± 1164* 962 ± 427 c) F (1, 26) = 0 n.s.

Mood (a.u.) T1 1363 ± 204 1413 ± 150 a) F (1, 26) = 0 n.s.

T9 1391 ± 152 1444 ± 180 b) F (1, 26) = 2 n.s.
Follow-up 1465 ± 164 1420 ± 92 c) F (1, 26) = 3 n.s.

Significant differences were identified with a 2-way ANOVA with repeated measure: training intensity (HIIT vs. LSD) x time (T1 vs. T9 vs. Follow-up).
a) Main effect: training intensity (HIIT vs. LSD).
b) Main effect: time (T1 vs. T9 vs. Follow-up).
c) Interaction effect: training intensity x time.
*Significant difference compared to LSD.
+ Significant difference compared to T1.
# Significant difference compared to T9.
n.s., Not significant.

In contrast to the initial hypothesis and the general From a mechanistic perspective, the sIgA is one of the
assumption that chronic exposure to high training loads most abundant antimicrobial proteins in the saliva (Papacosta
compromises the mucosal immune function evident as and Nassis, 2011) and synthesized locally in the submucosa
a decreased sIgA secretion rate (Tiollier et al., 2005; (Allgrove et al., 2008). The activation of the sympathetic nervous
Trochimiak and Hubner-Wozniak, 2012) the HIIT in the system and hypothalamic-pituitary-adrenal-axis promotes the
present investigation actually increased the sIgA secretion transepithelial transport of sIgA to the mucosal surface
rate throughout the entire training period. An earlier review (Goodrich and McGee, 1998). Especially the sIgA secretion rate
concluded that extreme efforts, such as HIIT, would increase increased acutely immediately after a bout of high-intensity
the infection risk of the upper-respiratory tract while moderate exercise (Allgrove et al., 2008). The concentration of salivary
intensity exercise, such as LSD, would improve the mucosal cortisol however did not respond until 1 h after exercise
immune function (Trochimiak and Hubner-Wozniak, 2012). (Allgrove et al., 2008) explaining why the cortisol response
A “J”-shaped relationship was generally accepted between the remained unaffected in the present study taking the saliva
immune function and the training load, while both, too low samples immediately after the HIIT and LSD. While the sIgA
as well as too high training loads, would impair the immune secretion rate responded immediately to the increased stress of
function (Trochimiak and Hubner-Wozniak, 2012). The latter HIIT, sIgA could therefore be a marker of stress being more
assumption however presumes that the mucosal immune system sensitive than salivary cortisol.
needs to be stressed in some way to adapt and improve its The question remains, whether the increased sIgA secretion
capacity to neutralize and defend viral pathogens in a similar way rate when waking up on T9 with the HIIT shows the stress
as muscles becomes stronger when exposed to regular training response from the previous training session, which was still
stress and adequate overload. evident during recovery, or if the increased sIgA secretion rate
The question arises, how much training and overload is indicates a chronically enhanced mucosal immune function. In
necessary to improve the mucosal immune function. In the both cases, with the repeated exposure to high training loads
present study, the stimulus of nine sessions of HIIT followed by (nine sessions of HIIT) the enhanced secretion of sIgA must be
at least 1 day of recovery was adequate to increase the number of matched by an increased synthesis in the submucosal plasma cells
antimicrobial proteins in the saliva as indicated by an elevated (Goodrich and McGee, 1998). Otherwise, over the time course of
sIgA secretion rate. In contrast, the LSD did not provide a nine HIIT sessions the store of IgA in the submucosa available for
sufficient stimulus to adapt and augment the mucosal immune transport across the epithelium would become depleted (Proctor
function evident as a sIgA secretion rate that was unaltered over et al., 2003). The lack of any peak elevation of sIgA with the
the time course of the training period. LSD indicates that the more intense training stimulus with the

Frontiers in Physiology | www.frontiersin.org 121 July 2017 | Volume 8 | Article 485

Born et al. Immune Function and High-Intensity Training

FIGURE 2 | The circadian variation of the (A) levels of cortisol and (B) salivary immunoglobin A (sIgA) secretion rate in response to the high-intensity interval training
(HIIT) and the long-slow distance training (LSD) normalized to the concentrations obtained from the baseline day before the start of the training period. Samples were
taken on the first (T1) and ninth (T9) day of training as well as the follow-up (1) immediately after waking up, (2) 30 min after waking up, (3) before training, (4) after
training and (5) before going to bed. For the sake of clarity, the standard error is illustrated for the corresponding mean values. Significant differences are indicated as
follows: * between groups, + in comparison to immediately after waking up of the same day, # in comparison to the corresponding value on day T1, † in comparison
to before exercise.

HIIT must has induced functional adaptation that elevated the this matter based on the current literature and concluded
sIgA synthesis and augmented the mucosal immune function in “that international athletes performing high-volume training suffer
addition to the ergogenic effects of a longer TTE and increased fewer, not greater, URTI episodes than lower-level performers”.
V· O2peak . The authors also concluded that the immune function is actually
Recently, particular interest has been drawn to the correlation improved with the regular but intermittent exposure to various
of reduced sIgA secretion rate and increased risk of URTI. During forms of stress (Walsh and Oliver, 2016). Linked with the results
polarized endurance training including both, continuous and of the present study it might be time to consider the mucosal
interval training, the runners who showed greater basal sIgA immune function as a highly adaptable system that, at least in
concentrations suffered less from URTI. Interestingly, the sIgA a 3-week time period, responds well to the stress of HIIT.
concentration before the start of the study predicted the number In the present investigation, salivary cortisol remained
of sick-days (r = −0.76, P < 0.01) during the following 12- unaffected immediately after exercise on the first and last day of
week training period (Ihalainen et al., 2016). The aforementioned training with both training groups. During the day of the follow-
findings indicate that the mucosal immune function might up however the incremental exercise test to exhaustion increased
be predisposed by either genetic factors or affected by the the levels of cortisol immediately after exercise with both, the
individual’s training history. The findings of the present study HIIT and LSD. Investigations evidenced that the participation in
support the latter by showing an increased sIgA secretion rate in official competitions induced greater levels of cortisol compared
response to the more severe training stress of the HIIT compared to training matches or race simulations indicating that the
to the moderate intensity exercise with the LSD. psychological stress itself rather than the actual physical load
Therefore, the generally accepted “J”-shaped relationship affects the response in cortisol (Rohleder et al., 2007; Moreira
between training load and the risk of URTI needs to be et al., 2012a,b, 2013). When comparing two volleyball matches
questioned. In a recent review Walsh and Oliver (2016) discussed that were played against the same opponent, the more important

Frontiers in Physiology | www.frontiersin.org 122 July 2017 | Volume 8 | Article 485

Born et al. Immune Function and High-Intensity Training

final championship match induced a greater cortisol response increased V· O2peak , were even accompanied with an increased
compared to the regular season match (Moreira et al., 2013). sIgA secretion rate indicating that the mucosal immune system
Similarly, in elite basketball players the cortisol values were still adapted over the time course of the training period by increasing
elevated largely above baseline during the competition phase the number of antimicrobial proteins and improving the capacity
even with a physical load that was almost half as much as during to neutralize and defend viral pathogens. The training stimulus
the preceding training phase (He et al., 2010). In the present of the LSD on the other hand was insufficient to improve
study, the incremental exercise test that was performed on the day the mucosal immune function or V· O2peak . Based on our
of the follow-up created a semi-competitive situation since each data we cannot generally accept the assumption that high
runner attempted to compel themselves mentally and physically training loads necessarily compromises the mucosal immune
as hard as possible in order to profit from the past training period function. Connecting the data of the present study with
and outrun their training colleagues. The arousal, anxiety, and previous findings (Born et al., 2016; Walsh and Oliver, 2016),
pressure to perform well during this type of performance test it might be time to consider the mucosal immune function
seemed to induce a substantial cortisol response which was not as a highly adaptable system that responds well to the stress
evident at any time point during the training phase neither with and load of training, in particular nine sessions of HIIT within
the HIIT nor LSD. 3 weeks.
Recent studies showed that the compromised mucosal The HIIT had no effect on the levels of cortisol and mood.
immune function, i.e., reduced sIgA secretion rate, induced by Therefore, the psychological stress, i.e., the arousal, anxiety,
the stress of competition was accompanied with high levels of mental stress, and pressure to perform well during competition
cortisol (He et al., 2010; Moreira et al., 2013). Therefore, an (Rohleder et al., 2007; Moreira et al., 2012a,b, 2013), rather than
antagonistic activity of increased cortisol values that inhibit the the actual physical load of exercise might be responsible for an
sIgA response has been discussed (He et al., 2010). In the present impaired mucosal immune function. Future studies should apply
study, especially the HIIT required the runners to perform each the circadian variation of sIgA secretion rate, cortisol and mood
training session close to their physical limit at 90–95% of HRmax . to further distinguish between the psychological and physical
The mood data however show, that our athletes did not feel stressors and how both could impact the mucosal immune
pressured or psychologically stressed due to the high training function during periods with an intensified training load and
loads of HIIT shown by a mood state that was not different to the competition.
LSD. Also the levels of cortisol on the first and last day of training
were fairly low with the HIIT and not different to LSD. Assuming AUTHOR CONTRIBUTIONS
an antagonistic activity of cortisol and sIgA (He et al., 2010;
Moreira et al., 2013), the low levels of cortisol on the first and Conception of the experimental design, data collection, analysis,
last day of training could explain why our athletes did not show interpretation, preparing and critically revising the manuscript:
any reduced sIgA secretion rate with the HIIT but adapted to the DB, CZ, and BS. All authors read and approved the final version
training stress and improved their mucosal immune function. of the manuscript.

CONCLUSION ACKNOWLEDGMENTS
In contrast to the hypothesis, we could not investigate any We would like to thank all the athletes for their enthusiastic
signs of a compromised mucosal immune function with the contribution to the study. This publication was supported by the
HIIT compared to LSD. The ergogenic effects of HIIT, i.e., Open Access Publication Fund of the University of Wuerzburg.

REFERENCES Buchheit, M., and Laursen, P. B. (2013b). High-intensity interval training,

solutions to the programming puzzle: Part I: cardiopulmonary emphasis. Sports
Allgrove, J. E., Gomes, E., Hough, J., and Gleeson, M. (2008). Effects of Med. 43, 313–338. doi: 10.1007/s40279-013-0029-x
exercise intensity on salivary antimicrobial proteins and markers of stress Cadore, E., Lhullier, F., Brentano, M., Silva, E., Ambrosini, M., Spinelli,
in active men. J. Sports Sci. 26, 653–661. doi: 10.1080/026404107017 R., et al. (2008). Correlations between serum and salivary hormonal
16790 concentrations in response to resistance exercise. J. Sports Sci. 26, 1067–1072.
Borg, G. (1970). Perceived exertion as an indicator of somatic stress. Scand. J. doi: 10.1080/02640410801919526
Rehabil. Med. 2, 92–98. Faiss, R., Willis, S., Born, D., Sperlich, B., Vesin, J., Holmberg, H.,
Born, D. P., Faiss, R., Willis, S. J., Strahler, J., Millet, G. P., Holmberg, H. et al. (2015). Double-poling repeated sprint training in hypoxia by
C., et al. (2016). Circadian variation of salivary immunoglobin A, alpha- competitive cross-country skiers. Med. Sci. Sports Exerc. 47, 809–817.
amylase activity and mood in response to repeated double-poling sprints doi: 10.1249/MSS.0000000000000464
in hypoxia. Eur. J. Appl. Physiol. 116, 1–10. doi: 10.1007/s00421-015- Gatti, R., and De Palo, E. F. (2011). An update: salivary hormones
3236-3 and physical exercise. Scand. J. Med. Sci. Sports 21, 157–169.
Buchheit, M., and Laursen, P. B. (2013a). High-intensity interval training, doi: 10.1111/j.1600-0838.2010.01252.x
solutions to the programming puzzle. Part II: anaerobic energy, Gill, S. K., Teixeira, A. M., Rosado, F., Hankey, J., Wright, A., Marczak, S., et al.
neuromuscular load and practical applications. Sports Med. 43, 927–954. (2014). The impact of a 24-h ultra-marathon on salivary antimicrobial protein
doi: 10.1007/s40279-013-0066-5 responses. Int. J. Sports Med. 35, 966–971. doi: 10.1055/s-0033-1358479

Frontiers in Physiology | www.frontiersin.org 123 July 2017 | Volume 8 | Article 485

Born et al. Immune Function and High-Intensity Training

Gleeson, M. (2007). Immune function in sport and exercise. J. Appl. Physiol. 103, represent measures of total hormone concentration versus time-dependent
693–699. doi: 10.1152/japplphysiol.00008.2007 change. Psychoneuroendocrinology 28, 916–931. doi: 10.1016/S0306-4530(02)
Goodrich, M. E., and McGee, D. W. (1998). Regulation of mucosal B cell 00108-7
immunoglobulin secretion by intestinal epithelial cell-derived cytokines. Rohleder, N., Beulen, S. E., Chen, E., Wolf, J. M., and Kirschbaum, C. (2007).
Cytokine 10, 948–955. doi: 10.1006/cyto.1998.0385 Stress on the dance floor: the cortisol stress response to social-evaluative
Gore, C. J. (2000). Physiological Tests for Elite Athletes. Leeds: Human Kinetics. threat in competitive ballroom dancers. Pers. Soc. Psychol. Bull. 33, 69–84.
Granger, D. A., Kivlighan, K. T., Fortunato, C., Harmon, A. G., Hibel, L. doi: 10.1177/0146167206293986
C., Schwartz, E. B., et al. (2007). Integration of salivary biomarkers Ronnestad, B. R., Hansen, J., and Ellefsen, S. (2014). Block periodization of high-
into developmental and behaviorally-oriented research: problems intensity aerobic intervals provides superior training effects in trained cyclists.
and solutions for collecting specimens. Physiol. Behav. 92, 583–590. Scand. J. Med. Sci. Sports 24, 34–42. doi: 10.1111/j.1600-0838.2012.01485.x
doi: 10.1016/j.physbeh.2007.05.004 Ronnestad, B. R., Hansen, J., Thyli, V., Bakken, T. A., and Sandbakk, O.
He, C. S., Tsai, M. L., Ko, M. H., Chang, C. K., and Fang, S. H. (2010). (2016). 5-week block periodization increases aerobic power in elite cross-
Relationships among salivary immunoglobulin A, lactoferrin and cortisol in country skiers. Scand. J. Med. Sci. Sports 26, 140–146. doi: 10.1111/sms.
basketball players during a basketball season. Eur. J. Appl. Physiol. 110, 989–995. 12418
doi: 10.1007/s00421-010-1574-8 Stoggl, T. L., and Sperlich, B. (2015). The training intensity distribution
Helgerud, J., Hoydal, K., Wang, E., Karlsen, T., Berg, P., Bjerkaas, among well-trained and elite endurance athletes. Front. Physiol. 6:295.
M., et al. (2007). Aerobic high-intensity intervals improve VO2max doi: 10.3389/fphys.2015.00295
more than moderate training. Med. Sci. Sports Exerc. 39, 665–671. Stoggl, T., and Sperlich, B. (2014). Polarized training has greater impact on key
doi: 10.1249/mss.0b013e3180304570 endurance variables than threshold, high intensity, or high volume training.
Hucklebridge, F., Clow, A., and Evans, P. (1998). The relationship between Front. Physiol. 5:33. doi: 10.3389/fphys.2014.00033
salivary secretory immunoglobulin A and cortisol: neuroendocrine response Sylta, O., Tonnessen, E., Hammarstrom, D., Danielsen, J., Skovereng, K.,
to awakening and the diurnal cycle. Int. J. Psychophysiol. 31, 69–76. Ravn, T., et al. (2016). The effect of different high-intensity periodization
doi: 10.1016/S0167-8760(98)00042-7 models on endurance adaptations. Med. Sci. Sports Exerc. 48, 2165–2174.
Hydren, J. R., and Cohen, B. S. (2015). Current scientific evidence for a polarized doi: 10.1249/MSS.0000000000001007
cardiovascular endurance training model. J. Strength Cond. Res. 29, 3523–3530. Tanner, A. V., Nielsen, B. V., and Allgrove, J. (2014). Salivary and plasma cortisol
doi: 10.1519/JSC.0000000000001197 and testosterone responses to interval and tempo runs and a bodyweight-
Ihalainen, J. K., Schumann, M., Hakkinen, K., and Mero, A. A. (2016). only circuit session in endurance-trained men. J. Sports Sci. 32, 680–689.
Mucosal immunity and upper respiratory tract symptoms in recreational doi: 10.1080/02640414.2013.850594
endurance runners. Appl. Physiol. Nutr. Metab. 41, 96–102. doi: 10.1139/apnm- Taylor, H. L., Buskirk, E., and Henschel, A. (1955). Maximal oxygen intake as an
2015-0242 objective measure of cardio-respiratory performance. J. Appl. Physiol. 8, 73–80.
Moreira, A., Crewther, B., Freitas, C. G., Arruda, A. F., Costa, E. C., and Aoki, M. Tiollier, E., Gomez-Merino, D., Burnat, P., Jouanin, J. C., Bourrilhon, C., Filaire, E.,
S. (2012a). Session RPE and salivary immune-endocrine responses to simulated et al. (2005). Intense training: mucosal immunity and incidence of respiratory
and official basketball matches in elite young male athletes. J. Sports Med. Phys. infections. Eur. J. Appl. Physiol. 93, 421–428. doi: 10.1007/s00421-004-1231-1
Fitness 52, 682–687. Trochimiak, T., and Hubner-Wozniak, E. (2012). Effect of exercise on
Moreira, A., Franchini, E., de Freitas, C. G., Schultz de Arruda, A. F., de Moura, the level of immunoglobulin a in saliva. Biol. Sport 29, 255–261.
N. R., Costa, E. C., et al. (2012b). Salivary cortisol and immunoglobulin A doi: 10.5604/20831862.1019662
responses to simulated and official Jiu-Jitsu matches. J. Strength Cond. Res. 26, VanBruggen, M. D., Hackney, A. C., McMurray, R. G., and Ondrak, K. S. (2011).
2185–2191. doi: 10.1519/JSC.0b013e31823b8702 The relationship between serum and salivary cortisol levels in response to
Moreira, A., Freitas, C. G., Nakamura, F. Y., Drago, G., Drago, M., and Aoki, M. S. different intensities of exercise. Int. J. Sports Physiol. Perform. 6, 396–407.
(2013). Effect of match importance on salivary cortisol and immunoglobulin A doi: 10.1123/ijspp.6.3.396
responses in elite young volleyball players. J. Strength Cond. Res. 27, 202–207. Walsh, N. P., and Oliver, S. J. (2016). Exercise, immune function and respiratory
doi: 10.1519/JSC.0b013e31825183d9 infection: an update on the influence of training and environmental stress.
Ni Cheilleachair, N. J., Harrison, A. J., and Warrington, G. D. (2017). Immunol. Cell Biol. 94, 132–139. doi: 10.1038/icb.2015.99
HIIT enhances endurance performance and aerobic characteristics more Wilhelm, P., and Schoebi, D. (2007). Assessing mood in daily life - Structural
than high-volume training in trained rowers. J. Sports Sci. 35, 1052–1058. validity, sensitivity to change, and reliability of a short-scale to measure
doi: 10.1080/02640414.2016.1209539 three basic dimensions of mood. Eur. J. Psychol. Assess. 23, 258–267.
Papacosta, E., Gleeson, M., and Nassis, G. P. (2013). Salivary hormones, IgA, and doi: 10.1027/1015-5759.23.4.258
performance during intense training and tapering in judo athletes. J. Strength
Cond. Res. 27, 2569–2580. doi: 10.1519/JSC.0b013e31827fd85c Conflict of Interest Statement: The authors declare that the research was
Papacosta, E., and Nassis, G. P. (2011). Saliva as a tool for monitoring steroid, conducted in the absence of any commercial or financial relationships that could
peptide and immune markers in sport and exercise science. J. Sci. Med. Sport be construed as a potential conflict of interest.
14, 424–434. doi: 10.1016/j.jsams.2011.03.004
Proctor, G. B., Garrett, J. R., Carpenter, G. H., and Ebersole, L. E. (2003). Salivary Copyright © 2017 Born, Zinner and Sperlich. This is an open-access article
secretion of immunoglobulin A by submandibular glands in response to distributed under the terms of the Creative Commons Attribution License (CC BY).
autonomimetic infusions in anaesthetised rats. J. Neuroimmunol. 136, 17–24. The use, distribution or reproduction in other forums is permitted, provided the
doi: 10.1016/S0165-5728(02)00466-6 original author(s) or licensor are credited and that the original publication in this
Pruessner, J. C., Kirschbaum, C., Meinlschmid, G., and Hellhammer, D. journal is cited, in accordance with accepted academic practice. No use, distribution
H. (2003). Two formulas for computation of the area under the curve or reproduction is permitted which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 124 July 2017 | Volume 8 | Article 485

 ORIGINAL RESEARCH
published: 13 June 2017
doi: 10.3389/fphys.2017.00413

No Additional Benefits of Block- Over

Evenly-Distributed High-Intensity
Interval Training within a Polarized
Microcycle
Kerry McGawley 1*, Elisabeth Juudas 1 , Zuzanna Kazior 1, 2 , Kristoffer Ström 1, 3 ,
Eva Blomstrand 2 , Ola Hansson 3 and Hans-Christer Holmberg 1
1
Department of Health Sciences, Swedish Winter Sports Research Centre, Mid Sweden University, Östersund, Sweden,
2
Åstrand Laboratory, Swedish School of Sport and Health Sciences, Stockholm, Sweden, 3 Diabetes and Endocrinology,
Department of Clinical Sciences, Lund University Diabetes Centre, Lund University, Malmö, Sweden

Introduction: The current study aimed to investigate the responses to block- versus
evenly-distributed high-intensity interval training (HIT) within a polarized microcycle.
Methods: Twenty well-trained junior cross-country skiers (10 males, age 17.6 ± 1.5
and 10 females, age 17.3 ± 1.5) completed two, 3-week periods of training (EVEN and
BLOCK) in a randomized, crossover-design study. In EVEN, 3 HIT sessions (5 × 4-min
Edited by:
Thomas Leonhard Stöggl, of diagonal-stride roller-skiing) were completed at a maximal sustainable intensity each
University of Salzburg, Austria week while low-intensity training (LIT) was distributed evenly around the HIT. In BLOCK,
Reviewed by: the same 9 HIT sessions were completed in the second week while only LIT was
Karim Chamari,
Aspetar, Qatar
completed in the first and third weeks. Heart rate (HR), session ratings of perceived
Patrick Wahl, exertion (sRPE), and perceived recovery (pREC) were recorded for all HIT and LIT
German Sport University Cologne,
sessions, while distance covered was recorded for each HIT interval. The recovery-stress
Germany
questionnaire for athletes (RESTQ-Sport) was completed weekly. Before and after EVEN
*Correspondence:
Kerry McGawley and BLOCK, resting saliva and muscle samples were collected and an incremental test
[email protected] and 600-m time-trial (TT) were completed.

Specialty section: Results: Pre- to post-testing revealed no significant differences between EVEN and
This article was submitted to BLOCK for changes in resting salivary cortisol, testosterone, or IgA, or for changes in
Exercise Physiology,
a section of the journal
muscle capillary density, fiber area, fiber composition, enzyme activity (CS, HAD, and
Frontiers in Physiology PFK) or the protein content of VEGF or PGC-1α. Neither were any differences observed
Received: 01 November 2016 in the changes in skiing economy, V̇O2 max or 600-m time-trial performance between
Accepted: 30 May 2017 interventions. These findings were coupled with no significant differences between
Published: 13 June 2017
EVEN and BLOCK for distance covered during HIT, summated HR zone scores, total
Citation:
McGawley K, Juudas E, Kazior Z, sRPE training load, overall pREC or overall recovery-stress state. However, 600-m TT
Ström K, Blomstrand E, Hansson O performance improved from pre- to post-training, irrespective of intervention (P = 0.003),
and Holmberg H-C (2017) No
and a number of hormonal and muscle biopsy markers were also significantly altered
Additional Benefits of Block- Over
Evenly-Distributed High-Intensity post-training (P < 0.05).
Interval Training within a Polarized
Microcycle. Front. Physiol. 8:413.
Discussion: The current study shows that well-trained junior cross-country skiers are
doi: 10.3389/fphys.2017.00413 able to complete 9 HIT sessions within 1 week without compromising total work done

Frontiers in Physiology | www.frontiersin.org 125 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

and without experiencing greater stress or reduced recovery over a 3-week polarized
microcycle. However, the findings do not support block-distributed HIT as a superior
method to a more even distribution of HIT in terms of enhancing physiological or
performance adaptions.
Keywords: cross-country skiing, endurance, junior athletes, muscle, periodization, recovery, stress, training load

INTRODUCTION training (i.e., a series of concentrated HIT sessions followed by

LIT), but not after a more traditional (even) distribution of LIT
Various theories pertaining to training periodization (i.e., the and HIT (Rønnestad et al., 2014b, 2016).
division of an athlete’s seasonal training into smaller units and Despite these positive findings associated with blocking HIT,
periods of time) have been discussed in the scientific literature prescribing short periods of intensified training in this manner
(see Issurin, 2010 for a comprehensive review). However, these has also been shown to result in the development of overreaching
theories have been criticized for lacking an evidence base (Kiely, (OR) symptoms. For example, Halson et al. (2002) showed that
2012), which seems to be due, at least in part, to the logistical 2 weeks of intensified HIT in the middle of a 6-week training
challenges associated with comparing different forms of training period led to OR among trained cyclists, reporting reductions
within athlete groups. In one study that reported the effects in MPO, maximal heart rate (HRmax ), V̇O2 max and cycling
of two different training periodization models used over two performance, as well as increases in global mood disturbance.
seasons with elite kayakers (García-Pallarés et al., 2010), rigorous In addition, Jürimäe et al. (2004) observed reductions in
scientific control was limited due to the applied and longitudinal performance capacity, resting testosterone levels and recovery, as
nature of the study. The outcomes of group studies are further well as increases in stress levels, following 6 days of intensified
complicated by the variability of individual responses to long- training with trained rowers. Although block HIT studies
term training programs (Mann et al., 2014). On the other hand, have not typically monitored subjective markers of well-being,
case studies of effective organizational strategies over seasonal these findings highlight the importance of understanding the
periods (e.g., Støren et al., 2012) provide limited support for multi-dimensional responses to intense training interventions,
general use. especially given the potential for short-term OR to develop into
Experimental studies designed to investigate the effects the more chronic overtraining syndrome (Meeusen et al., 2013).
of different training organization models on physiological At present there is a lack of information regarding the
and performance adaptations are typically limited to short physiological, psychological and performance-based responses of
intervention periods. This results in a need for potent junior athletes performing block HIT, which would be relevant
training stimuli, particularly where athlete populations are to coaches working in a variety of endurance sports. Therefore,
concerned. One such method used to develop aerobic power and the present investigation was designed to compare the effects
performance has been to concentrate a large number of high- of two polarized training interventions in well-trained junior
intensity interval training (HIT) sessions (e.g., 5–15 sessions) into male and female cross-country skiers. Nine HIT sessions were
a short period of time (e.g., 6–14 days) (Stølen et al., 2005; Breil concentrated in the middle of a 3-week training period in the
et al., 2010; Wahl et al., 2013, 2014; Rønnestad et al., 2014b, 2016). experimental intervention (BLOCK), while the same 9 HIT
This strategy has been referred to as “block training” and is based sessions were distributed evenly over the 3-week period in the
on the overload principle, with a super-compensation in selected control intervention (EVEN). Supplementary LIT and strength
fitness components thought to occur after a period of focused training were also matched across the two interventions, with
loading followed by a short recovery period (Issurin, 2010). only the organization of sessions differing. It was hypothesized
While HIT is considered necessary to elicit physiological and that BLOCK would lead to a greater relative increase in V̇O2 max
performance gains among endurance-trained athletes (Laursen and 600-m time-trial (TT) performance compared with EVEN,
et al., 2002; Iaia et al., 2009; Buchheit and Laursen, 2013a,b; despite covering less distance during the HIT sessions and
Gunnarsson et al., 2013), low-intensity training (LIT) remains attaining lower heart rates (HRs) during BLOCK compared with
a fundamental component of endurance programs (Sandbakk EVEN. Higher perceived exertion and stress, as well as reduced
and Holmberg, 2014). Combining these two forms of contrasting perceived recovery, were also expected during BLOCK compared
training stimuli, while performing relatively little moderate- with EVEN.
intensity training (MIT), is referred to as polarized training
(Seiler and Kjerland, 2006; Laursen, 2010). Some studies MATERIALS AND METHODS
have suggested that more polarized training distributions are
beneficial to endurance performance (Esteve-Lanao et al., 2007; Participants
Neal et al., 2012; Stöggl and Sperlich, 2014). Moreover, two Twenty well-trained cross-country skiers (10 males: age 17.6
recent studies have demonstrated improvements in V̇O2 max, ± 1.5 years, body mass 72.3 ± 4.8 kg, V̇O2 max 67.1 ±
maximal power output (MPO) and power output at blood lactate 2.6 mL·kg−1 ·min−1 ; 10 females: age 17.3 ± 1.5 years, body
concentrations of 2–4 mmol·L−1 among cyclists and cross- mass 61.1 ± 7.5 kg, V̇O2 max 54.2 ± 4.0 mL·kg−1 ·min−1 ) were
country skiers following a period of block-distributed polarized recruited from two specialist Swedish ski schools. All participants

Frontiers in Physiology | www.frontiersin.org 126 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

had at least 6 years of experience racing in cross-country concentrations using commercial ELISA kits (Salimetrics LLC,
skiing and competed at a national level, while eight were Pennsylvania, USA), as described by Beaven et al. (2008).
also members of national junior development teams. Average
endurance training volume was typically 500–750 h per year, or RESTQ-Sport
9–14 h per week, and an additional 60–80 h of functional strength The RESTQ-Sport was completed on arrival at the laboratory
training was completed annually. Weekly training frequency was after providing a saliva sample. The questionnaire consists of
periodized, with 4–5 endurance sessions (6–8 h) completed per 12 general scales and 7 additional sport-specific scales, with 4
week during low volume periods and up to 12 endurance sessions questions per scale, and assesses the balance between perceived
(25 h) completed per week during high volume periods. After recovery and stress (Kellmann and Kallus, 2001). The total stress
being informed of the aims and possible risks of the study the score corresponds to the sum of the scores of all of the stress
participants provided written informed consent to take part and subscales (7 general plus 3 sport-specific), while the total recovery
informed parental consent was obtained for those aged under score represents the sum of the scores of all of the recovery
18 years. The study was pre-approved by the Regional Ethical subscales (5 general plus 4 sport-specific). A general indicator
Review Board, Umeå University, Umeå, Sweden. of recovery-stress was calculated as the total recovery minus the
total stress score (Kellmann and Kallus, 2001).
Study Overview
The study was conducted at the end of the cross-country ski Sub-maximal and Maximal Incremental Tests
racing season, from April to June. A crossover design was used, Following the measurement of height and body mass (Seca 764,
whereby one group of athletes (EVEN-BLOCK; 6 males and Hamburg, Germany), sub-maximal and maximal incremental
7 females) completed 3 weeks of EVEN followed by 3 weeks tests were carried out on a motor-driven treadmill (Rodby RL
of BLOCK, separated by a 4-day break, and the other group 3000, Rodby, Vänge, Sweden) according to procedures previously
(BLOCK-EVEN; 4 males and 3 females) completed 3 weeks of described by McGawley and Holmberg (2014). Briefly, the
BLOCK followed by 3 weeks of EVEN, also separated by a 4-day diagonal roller-skiing technique and the same pair of pre-warmed
break (Table 1). As such, all 20 athletes performed both training roller-skis (Pro-Ski Classic C2, Sterners, Dala-Järna, Sweden)
interventions. The number of athletes in each group was not were used for both tests. The sub-maximal test was fixed at a
equal for the logistical reason that all HIT sessions needed to be 7◦ gradient and included a 4-min warm-up followed by four, 4-
completed on the same hill. Laboratory-based testing was carried min continuous stages. Speeds differed for individuals depending
out before and after the two interventions. on age, sex, and skiing ability, with the warm-up and first stage
completed at 5.2–7.0 km/h and increases of either 0.8 or 1.0
Pre- and Post-testing km/h per stage to final speeds of 7.6–10.0 km/h. At the end of
All participants attended the laboratory once during each test the sub-maximal test there was a 1-min break before participants
period to provide a resting saliva sample and to complete the commenced the maximal test. Depending on age, sex, and ability,
76-question recovery-stress questionnaire for athletes (RESTQ- the starting speed for the maximal test was 10, 11, or 12 km/h and
Sport), as well as sub-maximal and maximal incremental tests the initial gradient was 3◦ or 4◦ . The gradient was then increased
and a 600-m TT (Figure 1). Participants were familiar with the by 1◦ every minute, up to a maximum of 9◦ , after which speed
sub-maximal and maximal treadmill protocols, having completed was increased by 0.4 km/h every minute. The test was terminated
them routinely as part of their seasonal testing. The specific 600- when participants were unable to continue. Respiratory variables
m TT protocol was new to all participants, so a familiarization were measured using a mixed expired air procedure with an
was included in the test battery. Participants arrived at the ergospirometry system (AMIS 2001 model C, Innovision A/S,
laboratory in a rested state having consumed a standardized Odense, Denmark) equipped with a flow meter. The gas analysers
breakfast. The sub-maximal and maximal incremental tests, as were calibrated with a high-precision mixture of 16.0% O2 and
well as a familiarization to the 600-m TT, were completed in the 4.0% CO2 (Air Liquide, Kungsängen, Sweden) and the flow meter
morning prior to a standardized lunch, while the 600-m TT was was calibrated at three rates with a 3-L air syringe (Hans Rudolph,
completed after lunch. In addition to the main test day, muscle Kansas City, USA). The V̇O2 values were calculated from 10-s
biopsies were taken on a separate day during each test period epochs and skiing economy is expressed as the 1-min steady-
from a sub-group of 11 athletes aged 18 years and over (6 males: state value measured during the final minute of the second sub-
age 18.7 ± 0.8 years; 5 females: age 18.6 ± 0.9 years). maximal stage, while V̇O2 max is expressed as the highest 30-s
average recorded over any three consecutive 10-s samples (i.e., a
Saliva Samples sliding average).
Resting saliva samples were collected by passive drool (Beaven
et al., 2008) on arrival at the laboratory, between 08:00–08:30 600-m Time-Trial
during each test period. The participants consumed only water A familiarization to the 600-m TT, which was limited to 400 m
in the 1 h prior to collection (Sperlich et al., 2012). Samples were in order to minimize any impact of fatigue, was carried out on
collected in sterile tubes and stored at −20◦ C until analysis. After the same morning as the sub-maximal and maximal incremental
thawing and centrifuging at 2,000 rpm for 10 min, the samples tests, after at least 1 h of rest. The 600-m TT was then completed
were analyzed in duplicate and average values were used to in the afternoon, following a standardized lunch and at least 2 h
determine cortisol, testosterone and immunoglobulin A (IgA) of rest, again according to the methods described by McGawley

Frontiers in Physiology | www.frontiersin.org 127 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

TABLE 1 | An overview of the 53-day crossover study design incorporating two, 3-week training interventions (EVEN and BLOCK) flanked by pre- and post-testing.

Day

Group 1–3 4–24 25–28 29–49 50–53

EVEN-BLOCK (n = 13) Test period 1 3-week EVEN Test period 2 3-week BLOCK Test period 3
BLOCK-EVEN (n = 7) 3-week BLOCK 3-week EVEN

FIGURE 1 | A schematic of the laboratory tests performed before and after the training interventions. RESTQ-Sport: 76-question recovery-stress questionnaire for
athletes; Sub-max + Max: sub-maximal and maximal incremental tests; TT famil: 400-m familiarization to the 600-m time-trial (TT).

and Holmberg (2014). Following a 15-min warm-up the test to evaluate capillary density, fiber composition and fiber areas, as
protocol began with 100 m at a fixed speed (8.8 km·h−1 for the described by Kazior et al. (2016). Maximal enzyme activities of
females and 10.8 km·h−1 for the males) to avoid over-pacing, citrate synthase (CS), 3-hydroxyacyl CoA dehydrogenase (HAD)
followed by a self-paced maximal effort for the remaining 500 and phosphofructokinase (PFK) were carried out according to
m. The treadmill gradient was fixed at 7◦ and the diagonal- the procedures described by Opie and Newsholme (1967), Essén
stride technique was used throughout. The same motor-driven et al. (1975), Alp et al. (1976), respectively. The protein content
treadmill as that described in Section Sub-maximal and Maximal of VEGF and PGC-1α were measured using western blots.
Incremental Tests was used, fitted with lasers that automatically Briefly, 20 mg of tissue was homogenized in 250 µl of RIPA
increased or decreased the speed if the athlete moved to the buffer (Sigma) using glass/teflon homogenization. Following
front or rear of the belt, respectively, maintaining a constant centrifugation at 13,000 g, total protein concentration of the
speed otherwise (Swarén et al., 2013). Expired air was collected supernatant was estimated using the PierceTM BCA Protein
throughout using the procedures described above. Assay kit (Thermo Fisher Scientific). Protein (25 µg) was
separated by SDS-PAGE (NuPAGE R Bis-Tris Precast Gels, 4–
Muscle Biopsies 12%), transferred to nitrocellulose membranes and detection was
Resting muscle biopsies were taken from the vastus lateralis 1– made using SuperSignalWestPico Chemiluminescent Substrate
3 days before the main test day during each of the three test (Thermo Fisher Scientific). Antibodies used were Anti-PGC-
periods (see Table 1). After 10 min of supine rest the skin above 1α Mouse mAb (4C1.3) (ST1202) and Anti-VEGF (Ab-3)
the middle portion of the vastus lateralis was anesthetized with Mouse mAb (14–124) (GF25) (Merck Millipore). Results are
2% lidocaine (B. Braun Medical, Danderyd, Sweden) and biopsies presented as ratios of VEGF or PGC-1α expression to a loading
were taken using the needle technique with suction enhancement control, beta Actin (ab8227) (Abcam), ensuring equal loading
(Bergström, 1962; Hennessey et al., 1997). The tissue obtained on the gel (Ruas et al., 2012; Andrzejewski et al., 2015).
was rapidly cleaned from blood and fat and divided into three Also, gel-to-gel variation was adjusted for using an internal
parts. One part was mounted in an embedding medium (Tissue standard.
Tek R O.C.T. Compound) for subsequent histochemical analyses
and frozen immediately in isopentane that was cooled to its
freezing point in liquid nitrogen. The other two parts were Training
immediately frozen in liquid nitrogen for subsequent enzyme EVEN and BLOCK Interventions
and protein-content analyses. The samples were stored at −80◦ C The two, 3-week training interventions were developed in close
until analyzed. For histochemical analysis, serial 10-µm cross- cooperation with the coaches of the participating athletes. The
sections were cut in a cryostat at −20◦ C. Following preincubation EVEN intervention replicated a typical 3-week polarized training
at pH 4.3, 4.6, and 10.3, the sections were stained for myofibrillar cycle, while BLOCK involved 1 week of LIT only, both before
ATPase at pH 9.4 and the muscle fibers were classified as type I, and after an intensified week of HIT only. The EVEN and
IIA, IIB, or IIC (Brooke and Kaiser, 1969). To visualize capillaries, BLOCK interventions were workload matched and included 7
the cross-sections were stained by the amylase-PAS procedure and 9 LIT sessions, respectively (matched for total training time),
(Andersen, 1975). Computer image analysis (Leica QWin Runner as well as 9 HIT sessions and 6 functional strength sessions
V 3.5.1, Leica Microsystems, Bromma, Sweden) was performed (Table 2).

Frontiers in Physiology | www.frontiersin.org 128 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

TABLE 2 | The number and distribution of low-intensity training (LIT), approach to quantifying training load used a modification of the
high-intensity interval training (HIT) and functional strength (STR) training sessions 0–10 category ratio rating scale (CR-10) originally presented by
during three weeks of evenly-distributed (EVEN) and block (BLOCK) training.
Borg (1982). As described by Foster et al. (2001), within 30 min
EVEN BLOCK after every training session participants responded to the simple
question “How was your workout?” using a 10-point scale, with 0
Week 1 Week 2 Week 3 Week 1 Week 2 Week 3
and 10 corresponding to “rest” and “maximal,” respectively. The
LIT 2 3 2 4 0 5
text on the scale was presented to the athletes in both English and
HIT 3 3 3 0 9 0
Swedish. The session rating of perceived exertion (sRPE) training
STR 2 2 2 3 0 3
load was calculated by multiplying the 0–10 rating by the total
session duration (in min) and expressed in arbitrary units (Foster
et al., 2001). Total sRPE training load scores were calculated for
each individual by summing the sRPE training loads for all LIT
Durations of the LIT sessions ranged from 58 ± 4 to 127 and HIT sessions during EVEN and BLOCK.
± 11 min (∼60-, 90-, 105-, or 120-min sessions) and were
completed at ∼60–80% of HRmax as either skiing (∼60% of all Recovery and Stress Measures
sessions, either roller- or on-snow skiing, depending on weather Following the warm-up prior to all LIT and HIT training sessions
conditions), running (∼35% of all sessions) or cycling (∼5% the participants reported their perceived recovery (pREC) on a
of all sessions). The distribution of different LIT activities was scale from 0 to 10, with 0 and 10 corresponding to “very poorly
similar between the two interventions. The HIT sessions were recovered/extremely tired” and “very well recovered/highly
standardized at 75 min and consisted of a warm-up, 5 × 4-min energetic,” respectively (Laurent et al., 2011). The RESTQ-Sport
intervals separated by 6 min of active recovery and a warm- was completed weekly on a rest day prior to starting each training
down. The intervals were completed using the diagonal-stride week at a standardized time of day.
cross-country skiing technique on the same pair of roller-skis
throughout the study for each individual. The aim was to cover Data Analysis
as much total distance as possible, as evenly as possible, over the The Statistical Package for the Social Sciences (SPSS, Version
five intervals within each session. Distance covered was measured 22) was used to carry out statistical analyses. Interval and ratio
to the nearest meter for all athletes during each interval for data are expressed as mean ± standard deviation, while ordinal
every HIT session. All intervals were completed on the same data (sRPE, pREC, and RESTQ-Sport) are expressed as median
uphill asphalt slope (∼12%/7◦ ) and the active recovery involved [range]. Paired t-tests were used to compare responses to HIT
a downhill jog back to the start. In the case of a problem with the versus LIT and EVEN vs. BLOCK, while unpaired t-tests were
roller-skis participants completed the session by running on the used to compare responses between males and females. Two-
same uphill slope with poles (distance covered was not analyzed way ANOVAs with repeated measures were used to identify the
for running intervals). All HIT sessions were performed in intervention (EVEN vs. BLOCK), time (pre- to post-training)
groups of 6–8 athletes supervised by at least two researchers and and interaction effects. The magnitude of the training effect for
one coach. A standardized 20-min warm-up, including a 2-min EVEN vs. BLOCK was also assessed using effect size (ES), where
uphill interval on the training hill, was performed before each differences of <0.2, <0.6, <1.2, and <2.0 are interpreted as
HIT session and a 15-min cool-down was performed afterwards. trivial, small, moderate and large, respectively (Hopkins et al.,
The 9 HIT sessions in week 2 of BLOCK were organized such 2009). For the interval and ratio data, two-way ANOVAs with
that 2 sessions were completed on days 1, 2, and 5 (separated post-hoc Tukey tests were used to identify interaction effects and
by at least 5 h of rest and a meal), 1 session was completed differences between the interventions (EVEN vs. BLOCK) and
on days 3, 6, and 7, leaving day 4 as a rest day. The strength training weeks and HR zones. For the ordinal data, Friedman
sessions completed throughout the study involved functional and tests were used to compare weeks within each intervention
complex exercises and were supervised as part of the athletes’ and Wilcoxon signed rank tests were used to compare pairwise
regular training program. responses to HIT versus LIT and EVEN versus BLOCK, while
Mann-Whitney tests were used to compare responses between
Training Loads males and females. The level of statistical significance was set at
Heart rate was monitored for all LIT and HIT training sessions P < 0.05.
(Polar RS800CX, Polar Electro Oy, Kempele, Finland) and data
were subsequently analyzed for the determination of average HR RESULTS
(HRav ), HRmax , session duration and time spent in each of the
HR zones. The summated HR zone (sHRZ) method was used to Changes from Pre- to Post-training
quantify the HR-based training load for each session using the Pre- to post-training effects of EVEN and BLOCK are
following five HR zones: 1 = 50–60% of HRmax , 2 = 60–70% of displayed in Table 3. There were no interaction effects between
HRmax , 3 = 70–80% of HRmax , 4 = 80–90% of HRmax and 5 = 90– intervention and time for any of the performance, saliva or
100% of HRmax (Edwards, 1993; Foster et al., 2001). Cumulated muscle biopsy variables, as demonstrated by the change (1) data
time spent in each zone (in min) was multiplied by the zone value (P > 0.05). Neither were there any significant intervention
(i.e., 1–5) to obtain an overall sHRZ training load. The second (group) effects (P > 0.05). However, there were significant

Frontiers in Physiology | www.frontiersin.org 129 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

TABLE 3 | Mean ± SD pre- to post-training and change (1) data following three weeks of evenly-distributed (EVEN) or block (BLOCK) training.

EVEN BLOCK 1

Pre Post 1 Pre Post 1 P-value ES

n = 20

Skiing economy (O2 L·min−1 ) 2.83 ± 0.57 2.78 ± 0.57* −0.06 ± 0.10 2.77 ± 0.58 2.79 ± 0.58 0.02 ± 0.16 0.117 0.79
VO2 max(mL·kg−1 · min−1 ) 60.3 ± 7.2 61.2 ± 7.9 0.9 ± 2.6 61.4 ± 8.1 60.6 ± 8.2 −0.8 ± 2.4 0.071 0.67
600-m time-trial (s)T 187 ± 23 183 ± 25* −3 ± 5 185 ± 23 184 ± 22 −1 ± 6 0.280 0.44
Resting cortisol (µg/dL) 0.45 ± 0.21 0.54 ± 0.30 0.09 ± 0.21 0.53 ± 0.32 0.48 ± 0.19 −0.05 ± 0.30 0.208 0.59
Resting testosterone (pg/mL)T 108 ± 75 96 ± 45 −12 ± 51 108 ± 53 86 ± 39 −20 ± 41 0.592 0.15
Resting testosterone:cortisolT 264 ± 203 216 ± 157 −48 ± 94 257 ± 190 218 ± 170 −39 ± 139 0.817 0.18
Resting IgA (µg/mL) 36 ± 34 76 ± 83 40 ± 96 72 ± 83 45 ± 43 −26 ± 69 0.081 0.68

n = 11

Capillary density (per mm2 ) 377 ± 31 379 ± 44 3 ± 34 385 ± 42 365 ± 36 −20 ± 39 0.253 0.66
Mean fiber area (µm2 )T 4594 ± 761 4661 ± 764 68 ± 522 4596 ± 776 4968 ± 1000 372 ± 655 0.379 0.11
Type I (%) 66.2 ± 7.5 69.1 ± 6.7 2.9 ± 6.5 67.5 ± 6.5 67.2 ± 8.3 −0.3 ± 7.4 0.203 0.50
Type IIA (%) 25.2 ± 6.4 22.7 ± 5.6 −2.5 ± 4.8 24.1 ± 5.0 21.9 ± 7.9 −2.2 ± 8.3 0.904 0.07
Type IIB (%) 7.2 ± 4.1 7.2 ± 3.8 0.0 ± 3.2 7.5 ± 4.4 9.4 ± 5.9 1.9 ± 3.9 0.177 0.60
Type IIC (%) 1.4 ± 1.9 1.0 ± 2.4 −0.4 ± 3.4 1.0 ± 2.4 1.5 ± 2.6 0.5 ± 3.9 0.680 0.27
CS activity (µmol/min/g) 23.2 ± 2.9 22.5 ± 2.9 −0.7 ± 2.5 23.6 ± 2.7 22.5 ± 1.7 −1.1 ± 2.1 0.743 0.15
HAD activity (µmol/min/g)T 7.4 ± 1.0 7.4 ± 1.1 0.0 ± 1.1 7.6 ± 1.1 6.9 ± 0.9* −0.7 ± 1.0 0.317 0.60
PFK activity (µmol/min/g)T 21.7 ± 2.9 20.7 ± 2.1 −1.0 ± 2.3 20.8 ± 2.1 20.1 ± 2.4 −0.7 ± 1.9 0.753 0.14
VEGF protein content (AU) 20.0 ± 9.3 15.4 ± 10.7 −4.7 ± 15.2 16.5 ± 10.9 19.9 ± 15.5 3.5 ± 12.4 0.235 0.53
PGC-1α protein content (AU)T 0.07 ± 0.02 0.09 ± 0.02* 0.02 ± 0.02 0.08 ± 0.01 0.11 ± 0.04* 0.03 ± 0.04 0.862 0.11

CS, Citrate synthase; HAD, 3-hydroxyacyl CoA dehydrogenase; PFK, phosphofructokinase; AU, arbitrary units. T Significant time (pre- to post-training) effect, irrespective of intervention
(P < 0.05); *Significantly different from pre-training (P < 0.05).

time (i.e., pre- to post-training) effects, irrespective of group, for Performance during HIT
600-m TT performance, resting testosterone concentration and Average distances covered during each of the 5 × 4-min intervals
testosterone:cortisol ratio, mean muscle fiber area, HAD and PFK were similar during EVEN and BLOCK and are presented in
activity and PGC-1α protein content. Table 6 for all participants, as well as the males and females.

Adherence to Training Heart Rate Zones

In total the athletes completed 97 ± 6% (range: 90–100%) of The % of total time spent in each of the five HR zones during
the HIT sessions during EVEN and all (i.e., 100 ± 0%) of the HIT and LIT in EVEN and BLOCK for both the males and
HIT sessions during BLOCK and 91 and 99% of these sessions, females is displayed in Figure 2. Overall there were no significant
respectively, were performed using diagonal roller-skiing. A total differences between interventions in the proportion of total
of 95 ± 4% (range: 90–100%) and 98 ± 4% (range: 90–100%) training time spent in each of the HR zones (∼16, 37, 27,
of the LIT sessions were completed during EVEN and BLOCK, 10, and 10% of total time in zones 1–5, respectively, for both
respectively. interventions). Furthermore, the total sHRZ scores did not differ
between EVEN and BLOCK (3,739 ± 440 and 3,684 ± 449,
Responses to Training: LIT vs. HIT respectively; P = 0.329). However, the females demonstrated
A description of the HR and sRPE responses during LIT and a significantly higher total sHRZ score during EVEN compared
HIT for EVEN and BLOCK combined (i.e., independent of with the males (4,012 ± 392 and 3,466 ± 298, respectively; P =
intervention type) is presented in Table 4 for all participants, and 0.003) and a non-significant tendency for the same difference
for males and females separately. during BLOCK (3,869 ± 520 and 3,498 ± 281, respectively; P =
0.067).
Responses to Training: Even vs. Block
Training Time for LIT Perceived Exertion and Recovery
The LIT durations for the three separate weeks differed during The median [range] sRPE and pREC scores during each of the
EVEN and BLOCK (Table 5), while total time spent performing separate training weeks for EVEN and BLOCK are displayed
LIT did not differ between the two interventions (742 ± 33 and in Table 7. The total sRPE training loads did not significantly
754 ± 28 min for EVEN and BLOCK, respectively; P = 0.218). differ between EVEN and BLOCK (7,751 [5,758–9,462] and 8,127

Frontiers in Physiology | www.frontiersin.org 130 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

TABLE 4 | Mean ± SD average heart rate (HRav ), maximal heart rate (HRmax ), % of the total training time spent in zones 1–5 (Z1–Z5) and summated heart rate zone
(sHRZ) scores and median [range] session rating of perceived exertion (sRPE) scores during low-intensity training (LIT) and high-intensity interval training (HIT).

All participants Males Females

HRav (beats·min−1 ) LIT 133 ± 9 130 ± 7 136 ± 10

HIT 153 ± 8** 152 ± 7** 155 ± 9**

HRmax (beats·min−1 ) LIT 159 ± 8 156 ± 3 161 ± 10

HIT 191 ± 6** 191 ± 6** 192 ± 7**

% of total training time Z1 LIT 21 ± 15 25 ± 15 16 ± 15

spent in each zone
HIT 11 ± 5** 11 ± 6* 10 ± 5
Z2 LIT 48 ± 13 56 ± 13 40 ± 7††
HIT 26 ± 8** 30 ± 6** 21 ± 8†
Z3 LIT 29 ± 18 18 ± 10 40 ± 18††
HIT 25 ± 7 22 ± 8 27 ± 6*
Z4 LIT 2±3 1±1 4 ± 3††
HIT 19 ± 4** 19 ± 4** 19 ± 5**
Z5 LIT 0±0 0±0 0±0
HIT 20 ± 8** 17 ± 6** 23 ± 8**

Total sHRZ score LIT 3,196 ± 509 2,896 ± 280 3,496 ± 517††
HIT 4,227 ± 377** 4,069 ± 303** 4,385 ± 390**

Total sRPE score LIT 5,440 [3,735–6,983] 4,230 [3,735–6,489] 6,186 [3,989–6,983]†
HIT 10,463 [9,225–11,592]** 10,200 [9,225–11,550]* 10,725 [9,488–11,592]*

Significantly different from LIT: *P < 0.05, **P < 0.001.

† ††
Significantly different from the males: P < 0.05, P < 0.01.

males (6 [3–10] vs. 4 [1–8], respectively; P < 0.001) and the

TABLE 5 | Mean ± SD weekly durations (min) for low-intensity training (LIT) during
three weeks of evenly-distributed (EVEN) and block (BLOCK) training. females (6 [3–8] vs. 4 [1–8], respectively; P < 0.001).

Week 1 Week 2 Week 3 Recovery-Stress State

The overall recovery-stress state (as measured by the RESTQ-
EVEN 231 ± 29 272 ± 5 239 ± 11 Sport) did not significantly differ between EVEN and BLOCK
BLOCK 319 ± 9* 0 ± 0* 435 ± 26* (P = 0.510), but was significantly lower (indicating a less
Significantly different from EVEN: *P < 0.001.
recovered/more stressed state) after week two for BLOCK
compared with EVEN (14 [1–28] vs. 18 [5–35], respectively;
P = 0.033). In addition, the females had a significantly lower
TABLE 6 | Mean ± SD average distance covered (m) during the 4-min
high-intensity intervals during three weeks of evenly-distributed (EVEN) and block
recovery-stress state compared with the males during BLOCK
(BLOCK) training. after week two (11 [1–20] vs. 21 [13–28], P < 0.001) and three
(15 [0–23] vs. 22 [11–30], P = 0.014).
All participants Males Females

EVEN 740 ± 71 795 ± 38 675 ± 35* DISCUSSION

BLOCK 736 ± 75 792 ± 46 671 ± 43*
The present investigation has shown that a 3-week polarized
Significantly different from the males: *P < 0.001. training intervention incorporating a block distribution of HIT
is well-tolerated by a group of male and female junior cross-
[6,338–10,085], respectively; P = 0.286). However, the females country skiers. In contrast to the hypothesis, distance covered
demonstrated a significantly higher total sRPE training load for during HIT was not lower during BLOCK compared with EVEN.
EVEN compared with the males (8,076 [6,140–9,462] and 7,530 In addition, despite weekly differences, overall total sHRZ scores,
[5,758–8,155], respectively; P = 0.031) and a non-significant time spent in each of the HR zones, perceived exertion scores and
tendency for the same difference during BLOCK (8,423 [7,337– training loads, perceived recovery and the overall recovery-stress
8,948] and 7,740 [6,338–10,085], respectively; P = 0.190). state were not different following BLOCK compared with EVEN.
The pREC prior to each of the HIT sessions was significantly In terms of pre- to post-training, no differences were observed
improved for EVEN compared with BLOCK for all participants between the changes in any of the performance or physiological
(6 [3–10] vs. 4 [1–8], respectively; P < 0.001), as well as for the measures following the two interventions (see Table 3). However,

Frontiers in Physiology | www.frontiersin.org 131 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

FIGURE 2 | Percentage of the total time spent by the male and female skiers in the five different heart rate zones during the HIT (A) and LIT (B) sessions for the
evenly-distributed (EVEN) and block (BLOCK) training interventions. Significantly different between the males and females: *P < 0.05.

there were significant time effects for 600-m TT performance and positive improvements in V̇O2 max, MPO, sub-maximal power
a number of hormonal and muscle biopsy markers, irrespective output and TT performance among trained athletes following
of intervention. Based on these findings the current study does 13–15 HIT sessions completed within 10–14 days (Stølen et al.,
not support block-distributed HIT as a superior method to 2005; Breil et al., 2010; Wahl et al., 2013). While these studies
evenly-distributed HIT in terms of enhancing physiological or highlight the potential benefits of blocking HIT, they did not
performance adaptions. compare different types of training organization. Therefore, the
Block training is based on the theory that a period of highly observed effects may simply be due to the training stimulus,
concentrated, specialized loading will generate an increase in per se, rather than the specific distribution of HIT sessions.
the training stimulus such that, following a period of recovery, More recently, Rønnestad and colleagues have completed a
work capacity and performance will increase due to a super- series of studies comparing block- with more evenly-distributed
compensation (Issurin, 2008, 2010). At the same time, aerobic- HIT interventions (Rønnestad et al., 2014a,b, 2016). Greater
based HIT has been shown to maximize the time spent exercising improvements in V̇O2 max were reported for trained cyclists
close to V̇O2 max, which is considered the most effective stimulus following block- (5 HIT sessions in week 1, 1 HIT session in
for developing the oxygen transport and utilization systems weeks 2–4) compared with evenly- (2 HIT sessions per week)
(Billat, 2001; Midgley et al., 2006). A number of studies have distributed HIT (Rønnestad et al., 2014a,b). Among competitive
combined the concepts of block training and HIT, showing cross-country skiers and biathletes, by contrast, improvements

Frontiers in Physiology | www.frontiersin.org 132 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

TABLE 7 | Median [range] session rating of perceived exertion (sRPE) and perceived recovery (pREC) scores during the individual evenly-distributed (EVEN) and block
(BLOCK) training weeks.

sRPE pREC

All Males Females All Males Females

EVEN Week 1 7 [1–10]** 7 [1–9]** 7 [2–10]** 6 [3–9] 6 [4–9] 6 [3–8]

Week 2 6 [1–9]** 4 [1–9]** 7 [2–9]** 6 [3–10]** 7 [3–10]** 6 [3–8]**
Week 3 6 [2–10]** 5 [2–10]** 6 [3–10]** 5 [3–8]**† 6 [3–8]† 5 [3–8]**†
All weeks 6 [1–10] 5 [1–10] 6 [2–10] 6 [3–10] 6 [3–10] 6 [3–8]

BLOCK Week 1 3 [2–6]†† 3 [2–5]†† 3 [2–6]†† 6 [3–9]†† 7 [5–9]†† 6 [3–9]††

Week 2 8 [2–10] 8 [2–10] 8 [4–10] 4 [1–8] 4 [1–8] 4 [1–8]
Week 3 4 [2–7]†† 3 [2–6]†† 4 [2–7]†† 7 [3–9]†† 7 [3–9]†† 7 [4–8]††
All weeks 6 [2–10] 5 [2–10] 7 [2–10] 5 [1–9] 6 [1–9] 5 [1–9]

sRPE, assessed using a modified CR-10 scale with 0 and 10 corresponding to rest and maximal, respectively; pREC, assessed on a scale with 0 and 10 corresponding to very poorly
recovered/extremely tired and very well recovered/highly energetic, respectively.
Significantly different from the corresponding BLOCK week: **P < 0.001.
† ††
Significantly different from week 2: P < 0.01, P < 0.001.

in V̇O2 max were not greater following block- versus evenly- period may be critical in detecting beneficial effects of block
distributed HIT, but the block training group improved MPO training.
and sub-maximal power output to a greater extent than the even As well as investigating a range of pre- to post-training
training group (Rønnestad et al., 2016). The current study aimed markers to assess the efficacy of BLOCK compared with EVEN,
to investigate the potential mechanisms for the superior effects of a large focus of the current study was directed toward examining
block- compared with evenly-distributed HIT by systematically the responses during training, in order to explain any potential
monitoring the daily responses to training (through HR, sRPE, differences between the two interventions. It was expected that
recovery and performance measures), as well as by examining less total distance would be covered during the HIT sessions
the peripheral adaptations in the muscle through pre- and post- in BLOCK compared with EVEN, due to the reduced recovery
intervention muscle biopsies. between sessions and subsequent accumulation of fatigue. For
Despite a more intense HIT stimulus applied in the current instance, power output produced by endurance-trained athletes
study compared with Rønnestad et al. (2014a,b); Rønnestad has been observed to be lower during a second session of HIT
et al. (2016), no significant differences were observed between performed on the same day compared to on a separate day
BLOCK and EVEN for any of the variables measured pre- to (Yeo et al., 2008). Unexpectedly, however, the average distance
post-training (i.e., skiing economy, V̇O2max, TT performance, covered during the 5 × 4-min intervals was similar between the
resting salivary markers or muscle biopsy markers). This could two interventions in the present study (∼740 m per interval).
be due to a number of reasons relating to the study design. In contrast to Yeo et al. (2008), who allowed only 2 h of rest
Firstly, the current study was conducted at the end of the and water consumption between sessions, participants in the
cross-country season, whereas all other published block training current study rested for 4–5 h and ate a meal between any two
studies have been completed during pre-season (Breil et al., HIT sessions on the same day. Therefore, longer recovery and
2010; Wahl et al., 2013; Rønnestad et al., 2014a,b, 2016). energy replacement may help to maintain performance when
Since Losnegard et al. (2013) have shown that cross-country completing two HIT sessions on the same day. Alternatively,
skiers perform more HIT and less LIT toward the end of the the whole-body nature of cross-country skiing exercise may
competitive season, it is possible that the timing and resulting lead to reduced local fatigue and allow training intensities to
training status of the athletes in the present study affected the be maintained during a second training session within a day.
efficacy of the BLOCK intervention. Another factor could be The relatively long recovery duration of 6 min between each
the lack of any HIT sessions in the final training week during interval, which resulted from the logistical requirement for
BLOCK. Anecdotally, athletes in the current study reported athletes to jog back down the hill after each interval, may
feelings of lethargy as a result of only having performed LIT also have enabled the maintenance of work done over the five
in the week prior to laboratory testing. This was not the repetitions. In fact, as little as 2 min of recovery between 4-
case following EVEN, whereby three HIT sessions had been min HIT bouts has been shown to be sufficient in maintaining
performed in the week prior to testing. Bosquet et al. (2007) performance in a set of repeated intervals, although a higher
refer to maintenance of training intensity during an optimal average oxygen consumption was possible during intervals
taper and in support of this, Rønnestad et al. (2014a,b, 2016) with a 2- vs. 4-min recovery period (Seiler and Hetlelid,
prescribed at least 1 HIT session per week during recovery 2005).
following their block intervention. Therefore, the maintenance Similar to distance covered, it was also expected that the
of some HIT sessions in the weeks following the overload athletes in the current study would attain lower HRs during

Frontiers in Physiology | www.frontiersin.org 133 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

BLOCK compared with EVEN, due to the more concentrated relative intensities and more time spent in higher HR zones may
training load and reduced recovery between HIT sessions. This have resulted in a higher perception of effort among the females
response has previously been demonstrated for competitive compared with the males.
cyclists during maximal exercise following a period of intensified Previous studies investigating periods of intensified training
training predominantly consisting of HIT (Jeukendrup et al., among endurance athletes have shown short-term reductions in
1992). However, no differences were identified in total sHRZ recovery and well-being, as well as increases in mood disturbance
scores or the time spent in each of the HR zones during EVEN and stress levels (Jeukendrup et al., 1992; Halson et al., 2002;
compared with BLOCK. With no differences observed for the Jürimäe et al., 2004; Coutts et al., 2007). Since an excess of stress
group as a whole, further analyses were conducted to compare can result in long-term performance decline that is manifested as
differences in the responses between the males and females. overtraining, or non-functional OR (Meeusen et al., 2013), there
Interestingly, the females demonstrated a higher sHRZ score than was a potential risk for the young athletes in the current study
the males during EVEN, with a non-significant tendency for the performing so many HIT sessions within 1 week. Therefore,
same difference during BLOCK. This appears attributable to the the RESTQ-Sport, which has been identified as a useful tool
fact that the females spent more time in zones 3 and 4 during for monitoring perceived stress and recovery among athletes
the LIT sessions, while the males spent more time in zones 1 or (Saw et al., 2016), was administered weekly (in contrast to the
2. While only speculative, it is possible that some of the females session-based pREC scale). Despite a significant difference during
worked relatively harder during the LIT sessions in order to “keep week two, the overall recovery-stress state was not different
up” with other members of the training group. While the athletes following BLOCK compared with EVEN. A rapid restoration
typically trained with others of a similar standard, a group session of the recovery-stress state is consistent with previous findings
may have put pressure on the weaker members (often females) that have shown global mood state to recover to baseline after
to work at a higher relative intensity than the stronger members 4–6 days of easy training (Halson et al., 2002). Therefore, it
(often males). Furthermore, any common undulating training seems that non-functional OR may be avoided by limiting the
routes, where specific techniques (and therefore velocities, to a duration of the intense training period and allowing sufficient
certain extent) are employed on given inclines, would also likely recovery afterwards. An interesting and unexpected finding in the
lead to higher relative intensities among the females due to lower current study was that the females demonstrated lower recovery-
maximal aerobic capacities. In addition to differences during the stress states compared with the males, with the largest differences
LIT sessions, the males also spent significantly more time in zone observed after week two of BLOCK. Thus, this study provides
2 during the BLOCK HIT sessions, with the females tending to novel data to suggest that female athletes are more vulnerable
spend more time in the higher HR zones. This indicates a more than males to the stressors associated with block-distributed HIT
rapid HR recovery among the males between intervals. Overall within a polarized microcycle, perhaps due to higher internal
these findings highlight the need for coaches to carefully monitor workloads during training sessions.
the internal loads (i.e., HR responses) of individuals within a The current study is the first to have comprehensively
training group, especially in mixed-sex groups, to ensure that compared the responses during, and effects of, two polarized
specified training targets are achieved. training models differing only in the distribution of training
Subjective measures have been reported to be more sensitive sessions. Due to the high adherence rates (90–100% of sessions
and consistent than objective measures when monitoring completed by all individuals during HIT and LIT), the results
changes in athlete well-being in response to training (Saw et al., may be considered a true representation of the prescribed
2016), hence the use of sRPE, pREC, and RESTQ-Sport in the interventions. Findings have shown distinct demands on the
current study. An analysis of the separate weeks highlighted athletes during the three separate weeks of EVEN and BLOCK,
clear distinctions in the differing demands during BLOCK and demonstrated by the significant weekly differences in time spent
EVEN, with sRPE scores significantly higher and pREC scores performing LIT and HIT, perceived exertion and recovery scores
significantly lower during week two of BLOCK compared with and recovery-stress states. Despite this, the overall responses
weeks one and three, as well as compared with week two of during the two interventions were typically similar in terms
EVEN. In addition, pREC was improved prior to the HIT sessions of performance and subjective measures (i.e., distance covered
during EVEN compared to BLOCK. This indicates an improved during HIT, session ratings of perceived exertion, perceived
readiness to train when HIT sessions are spread out over 3 recovery and recovery-stress states). Moreover, changes pre-
weeks rather than being condensed into 1 week. The extreme to post-training did not differ between EVEN and BLOCK.
training load prescribed in week two of BLOCK was the basis Some limitations of the present study may be related to the
for hypothesizing that perceived exertion would have been higher experimental design, specifically the lack of any HIT sessions
and perceived recovery would have been lower after BLOCK following the intensified training week, the relatively short
compared with EVEN. However, results showed no differences duration between the intensified training week and follow-up
in sRPE scores, sRPE training loads or average pREC scores after laboratory tests, the short time period (3 weeks) over which
the two, 3-week interventions. Interestingly, and consistent with the interventions were prescribed and/or the relatively long
the sHRZ data, the females demonstrated a significantly higher recovery duration (6 min) between the 4-min HIT intervals.
sRPE training load during EVEN compared with the males and Nevertheless, a novel aspect of the study is the comparison
a tendency for the same difference during BLOCK. This may be between males and females, which revealed some real practical
for a similar reason to that previously proposed; that is, higher issues for coaches whereby the females typically demonstrated

Frontiers in Physiology | www.frontiersin.org 134 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

higher HR responses and sRPE scores, as well as higher stress the Declaration of Helsinki. The protocol was approved by the
scores and lower recovery-stress states, compared to the males. Regional Ethical Review Board, Umeå University, Umeå Sweden.
In light of these specific differences, future research may be
directed toward investigating how higher internal training loads, AUTHOR CONTRIBUTIONS
perceived exertion and subjective recovery-stress states in females
influence long-term training adaptations and potential OR or KM, EJ, and HH made substantial contributions to the
overtraining. In conclusion, the current study has shown that conception and design of the work while KM, EJ, ZK, KS, EB, OH,
well-trained junior cross-country skiers are able to complete 9 and HH all made substantial contributions to the acquisition,
HIT sessions within 1 week without compromising total work analysis and interpretation of data for the work. All authors (KM,
done or experiencing greater stress or reduced recovery in EJ, ZK, KS, EB, OH, and HH) were involved in the drafting and
comparison to completing 3 HIT sessions per week over 3 weeks. critical revision of the work, as well as the final approval of the
However, a short training intervention using block-distributed version to be published, and agree to be accountable for all aspects
HIT is not supported as being superior to evenly-distributed HIT of the work.
when applied to well-trained, junior cross-country skiers.
ACKNOWLEDGMENTS
ETHICS STATEMENT
The authors would like to thank Hannah Wishart for her
This study was carried out in accordance with the involvement in the data collection process, Professor Bengt
recommendations of the Regional Ethical Review Board Saltin for his contribution to the muscle sampling procedures,
(Umeå Sweden) with written informed consent from all subjects Bengt Stattin for his unfaltering enthusiasm and cooperation, the
(and informed parental consent for those aged under 18 years). coaches who assisted throughout the project and the athletes for
All subjects gave written informed consent in accordance with their dedication to the experiment.

REFERENCES neuromuscular load and practical applications. Sports Med. 43, 927–954.
doi: 10.1007/s40279-013-0066-5
Alp, P. R., Newsholme, E. A., and Zammit, V. A. (1976). Activities of citrate Coutts, A. J., Slattery, K. M., and Wallace, L. K. (2007). Practical tests for
synthase and NAD+-linked and NADP+-linked isocitrate dehydrogenase in monitoring performance, fatigue and recovery in triathletes. J. Sci. Med. Sport
muscle from vertebrates and invertebrates. Biochem. J. 154, 689–700. 10, 372–381. doi: 10.1016/j.jsams.2007.02.007
Andersen, P. (1975). Capillary density in skeletal muscle of man. Acta Physiol. Edwards, S. (1993). The Heart Rate Monitor Book. Sacramento, CA: Fleet Feet
Scand. 95, 203–205. doi: 10.1111/j.1748-1716.1975.tb10043.x Press.
Andrzejewski, W., Kassolik, K., Kobierzycki, C., Grzegrzolka, J., Ratajczak- Essén, B., Jansson, E., Henriksson, J., Taylor, A. W., and Saltin, B. (1975). Metabolic
Wielgomas, K., Jablonska, K., et al. (2015). Increased skeletal muscle expression characteristics of fibre types in human skeletal muscle. Acta Physiol. Scand. 95,
of VEGF induced by massage and exercise. Folia Histochem. Cytobiol. 53, 153–165. doi: 10.1111/j.1748-1716.1975.tb10038.x
145–151. doi: 10.5603/FHC.a2015.0013 Esteve-Lanao, J., Foster, C., Seiler, S., and Lucia, A. (2007). Impact of training
Beaven, C. M., Gill, N. D., and Cook, C. J. (2008). Salivary testosterone and intensity distribution on performance in endurance athletes. J. Strength Cond.
cortisol responses in professional rugby players after four resistance exercise Res. 21, 943–949. doi: 10.1519/R-19725.1
protocols. J. Strength Cond. Res. 22, 426–432. doi: 10.1519/JSC.0b013e31816 Foster, C., Florhaug, J. A., Franklin, J., Gottschall, L., Hrovatin, L. A., Parker, S.,
35843 et al. (2001). A new approach to monitoring exercise training. J. Strength Cond.
Bergström, J. (1962). Muscle electrolytes in man determined by neutron activation Res. 15, 109–115.
analysis on needle biopsy specimens. Scand. J. Clin. Lab. Invest. 14(Suppl.), 68. García-Pallarés, J., García-Fernández, M., Sánchez-Medina, L., and Izquierdo,
Billat, L. V. (2001). Interval training for performance: a scientific and M. (2010). Performance changes in world-class kayakers following two
empirical practice. Special recommendations for middle- and long- different training periodization models. Eur. J. Appl. Physiol. 110, 99–107.
distance running. Part I: aerobic interval training. Sports Med. 31, 13–31. doi: 10.1007/s00421-010-1484-9
doi: 10.2165/00007256-200131010-00002 Gunnarsson, T. P., Christensen, P. M., Thomassen, M., Nielsen, L. R., and
Borg, G. A. (1982). Psychophysical bases of perceived exertion. Med. Sci. Sports Bangsbo, J. (2013). Effect of intensified training on muscle ion kinetics,
Exerc. 14, 377–381. fatigue development, and repeated short-term performance in endurance-
Bosquet, L., Montpetit, J., Arvisais, D., and Mujika, I. (2007). Effects of tapering trained cyclists. Am. J. Physiol. Regul. Integr. Comp. Physiol. 305, R811–R821.
on performance: a meta-analysis. Med. Sci. Sports Exerc. 39, 1358–1365. doi: 10.1152/ajpregu.00467.2012
doi: 10.1249/mss.0b013e31806010e0 Halson, S. L., Bridge, M. W., Meeusen, R., Busschaert, B., Gleeson, M., Jones, D.
Breil, F., Weber, S., Koller, S., Hoppeler, H., and Vogt, M. (2010). Block A., et al. (2002). Time course of performance changes and fatigue markers
training periodization in alpine skiing: effects of 11-day HIT on during intensified training in trained cyclists. J. Appl. Physiol. 93, 947–956.
VO2 max and performance. Eur. J. Appl. Physiol. 109, 1077–1086. doi: 10.1152/japplphysiol.01164.2001
doi: 10.1007/s00421-010-1455-1 Hennessey, J. V., Chromiak, J. A., Della Ventura, S., Guertin, J., and MacLean,
Brooke, M. H., and Kaiser, K. K. (1969). Some comments on the histochemical D. B. (1997). Increase in percutaneous muscle biopsy yield with a suction-
characterization of muscle adenosine triphosphatase. J. Histochem. Cytochem. enhancement technique. J. Appl. Physiol. 82, 1739–1742.
17, 431–432. Hopkins, W. G., Marshall, S. W., Batterham, A. M., and Hanin, J. (2009).
Buchheit, M., and Laursen, P. B. (2013a). High-intensity interval training, Progressive statistics for studies in sports medicine and exercise science. Med.
solutions to the programming puzzle: Part I: cardiopulmonary emphasis. Sports Sci. Sports Exerc. 41, 3–13. doi: 10.1249/MSS.0b013e31818cb278
Med. 43, 313–338. doi: 10.1007/s40279-013-0029-x Iaia, F. M., Hellsten, Y., Nielsen, J. J., Fernstrom, M., Sahlin, K., and
Buchheit, M., and Laursen, P. B. (2013b). High-intensity interval training, Bangsbo, J. (2009). Four weeks of speed endurance training reduces
solutions to the programming puzzle: Part II: anaerobic energy, energy expenditure during exercise and maintains muscle oxidative capacity

Frontiers in Physiology | www.frontiersin.org 135 June 2017 | Volume 8 | Article 413

McGawley et al. Block Distributed High-Intensity Interval Training

despite a reduction in training volume. J. Appl. Physiol. 106, 73–80. Rønnestad, B. R., Ellefsen, S., Nygaard, H., Zacharoff, E. E., Vikmoen, O., Hansen,
doi: 10.1152/japplphysiol.90676.2008 J., et al. (2014a). Effects of 12 weeks of block periodization on performance
Issurin, V. (2008). Block periodization versus traditional training theory: a review. and performance indices in well-trained cyclists. Scand. J. Med. Sci. Sports 24,
J. Sports Med. Phys. Fitness 48, 65–75. 327–335. doi: 10.1111/sms.12016
Issurin, V. B. (2010). New horizons for the methodology and Rønnestad, B. R., Hansen, J., and Ellefsen, S. (2014b). Block periodization of high-
physiology of training periodization. Sports Med. 40, 189–206. intensity aerobic intervals provides superior training effects in trained cyclists.
doi: 10.2165/11319770-000000000-00000 Scand. J. Med. Sci. Sports 24, 34–42. doi: 10.1111/j.1600-0838.2012.01485.x
Jeukendrup, A. E., Hesselink, M. K., Snyder, A. C., Kuipers, H., and Rønnestad, B. R., Hansen, J., Thyli, V., Bakken, T. A., and Sandbakk, O. (2016). 5-
Keizer, H. A. (1992). Physiological changes in male competitive cyclists week block periodization increases aerobic power in elite cross-country skiers.
after two weeks of intensified training. Int. J. Sports Med. 13, 534–541. Scand. J. Med. Sci. Sports 26, 140–146. doi: 10.1111/sms.12418
doi: 10.1055/s-2007-1021312 Ruas, J. L., White, J. P., Rao, R. R., Kleiner, S., Brannan, K. T., Harrison, B. C.,
Jürimäe, J., Mäestu, J., Purge, P., and Jürimäe, T. (2004). Changes in stress et al. (2012). A PGC-1α isoform induced by resistance training regulates skeletal
and recovery after heavy training in rowers. J. Sci. Med. Sport 7, 335–339. muscle hypertrophy. Cell 151, 1319–1331. doi: 10.1016/j.cell.2012.10.050
doi: 10.1016/s1440-2440(04)80028-8 Sandbakk, Ø., and Holmberg, H. C. (2014). A reappraisal of success factors
Kazior, Z., Willis, S. J., Moberg, M., Apro, W., Calbet, J. A., Holmberg, H. C., et al. for Olympic cross-country skiing. Int. J. Sports Physiol. Perform. 9, 117–121.
(2016). Endurance exercise enhances the effect of strength training on muscle doi: 10.1123/ijspp.2013-0373
fiber size and protein expression of Akt and mTOR. PLoS ONE 11:e0149082. Saw, A. E., Main, L. C., and Gastin, P. B. (2016). Monitoring the athlete
doi: 10.1371/journal.pone.0149082 training response: subjective self-reported measures trump commonly used
Kellmann, M., and Kallus, K. W. (2001). Recovery-Stress Questionnaire for Athletes: objective measures: a systematic review. Br. J. Sports Med. 50, 281–291.
User Manual. Champaign, IL: Human Kinetics. doi: 10.1136/bjsports-2015-094758
Kiely, J. (2012). Periodization paradigms in the 21st century: evidence- Seiler, K. S., and Kjerland, G. Ø. (2006). Quantifying training intensity distribution
led or tradition-driven? Int. J. Sports Physiol. Perform. 7, 242–250. in elite endurance athletes: is there evidence for an “optimal” distribution?
doi: 10.1123/ijspp.7.3.242 Scand. J. Med. Sci. Sports 16, 49–56. doi: 10.1111/j.1600-0838.2004.00418.x
Laurent, C. M., Green, J. M., Bishop, P. A., Sjökvist, J., Schumacker, R. E., Seiler, S., and Hetlelid, K. J. (2005). The impact of rest duration on work
Richardson, M. T., et al. (2011). A practical approach to monitoring recovery: intensity and RPE during interval training. Med. Sci. Sports Exerc. 37,
development of a perceived recovery status scale. J. Strength Cond. Res. 25, 1601–1607. doi: 10.1249/01.mss.0000177560.18014.d8
620–628. doi: 10.1519/JSC.0b013e3181c69ec6 Sperlich, B., Achtzehn, S., Buhr, M., Zinner, C., Zelle, S., and Holmberg, H.
Laursen, P. B. (2010). Training for intense exercise performance: high- C. (2012). Salivary cortisol, heart rate, and blood lactate responses during
intensity or high-volume training? Scand. J. Med. Sci. Sports 20, 1–10. elite downhill mountain bike racing. Int. J. Sports Physiol. Perform. 7, 47–52.
doi: 10.1111/j.1600-0838.2010.01184.x doi: 10.1123/ijspp.7.1.47
Laursen, P. B., Shing, C. M., Peake, J. M., Coombes, J. S., and Jenkins, Stöggl, T., and Sperlich, B. (2014). Polarized training has greater impact on key
D. G. (2002). Interval training program optimization in highly endurance variables than threshold, high intensity, or high volume training.
trained endurance cyclists. Med. Sci. Sports Exerc. 34, 1801–1807. Front. Physiol. 5:33. doi: 10.3389/fphys.2014.00033
doi: 10.1249/01.MSS.0000036691.95035.7D Stølen, T., Chamari, K., Castagna, C., and Wisløff, U. (2005). Physiology of soccer:
Losnegard, T., Myklebust, H., Spencer, M., and Hallén, J. (2013). an update. Sports Med. 35, 501–536. doi: 10.2165/00007256-200535060-00004
Seasonal variations in VO2 max, O2 -cost, O2 -deficit and performance Støren, Ø., Sanda, S. B., Haave, M., and Helgerud, J. (2012). Improved VO2 max
in elite cross-country skiers. J. Strength Cond. Res. 27, 1780–1790. and time trial performance with more high aerobic intensity interval training
doi: 10.1519/JSC.0b013e31827368f6 and reduced training volume; a case study on an elite national cyclist. J. Strength
Mann, T. N., Lamberts, R. P., and Lambert, M. I. (2014). High responders Cond. Res. 26, 2705–2711. doi: 10.1519/JSC.0b013e318241deec
and low responders: factors associated with individual variation Swarén, M., Supej, M., Eriksson, A., and Holmberg, H. C. (2013). “Treadmill
in response to standardized training. Sports Med. 44, 1113–1124. simulation of Olympic cross-country ski tracks,” in Science and Nordic Skiing
doi: 10.1007/s40279-014-0197-3 II, eds A. Hakkarainen, V. Linnamo, and S. Lindinger (London: E & FN Spon),
McGawley, K., and Holmberg, H. C. (2014). Aerobic and anaerobic contributions 237–242.
to energy production among junior male and female cross-country Wahl, P., Guldner, M., and Mester, J. (2014). Effects and sustainability of a 13-day
skiers during diagonal skiing. Int. J. Sports Physiol. Perform. 9, 32–40. high-intensity shock microcycle in soccer. J. Sports Sci. Med. 13, 259–265.
doi: 10.1123/ijspp.2013-0239 Wahl, P., Zinner, C., Grosskopf, C., Rossmann, R., Bloch, W., and Mester,
Meeusen, R., Duclos, M., Foster, C., Fry, A., Gleeson, M., Nieman, D., et al. J. (2013). Passive recovery is superior to active recovery during a
(2013). Prevention, diagnosis, and treatment of the overtraining syndrome: high intensity shock microcycle. J. Strength Cond. Res. 27, 1384–1393.
joint consensus statement of the European College of Sport Science and the doi: 10.1519/JSC.0b013e3182653cfa
American College of Sports Medicine. Med. Sci. Sports Exerc. 45, 186–205. Yeo, W. K., Paton, C. D., Garnham, A. P., Burke, L. M., Carey, A. L., and Hawley, J.
doi: 10.1249/MSS.0b013e318279a10a A. (2008). Skeletal muscle adaptation and performance responses to once a day
Midgley, A. W., McNaughton, L. R., and Wilkinson, M. (2006). Is there versus twice every second day endurance training regimens. J. Appl. Physiol.
an optimal training intensity for enhancing the maximal oxygen uptake 105, 1462–1470. doi: 10.1152/japplphysiol.90882.2008
of distance runners? Empirical research findings, current opinions,
physiological rationale and practical recommendations. Sports Med. 36, Conflict of Interest Statement: The authors declare that the research was
117–132. doi: 10.2165/00007256-200636020-00003 conducted in the absence of any commercial or financial relationships that could
Neal, C. M., Hunter, A. M., Brennan, L., O’Sullivan, A., Hamilton, D. L., be construed as a potential conflict of interest.
De Vito, G., et al. (2012). Six weeks of a polarised training intensity
distribution leads to greater physiological and performance adaptations Copyright © 2017 McGawley, Juudas, Kazior, Ström, Blomstrand, Hansson and
than a threshold model in trained cyclists. J. Appl. Physiol. 114, 461–471. Holmberg. This is an open-access article distributed under the terms of the Creative
doi: 10.1152/japplphysiol.00652.2012 Commons Attribution License (CC BY). The use, distribution or reproduction in
Opie, L. H., and Newsholme, E. A. (1967). The activities of fructose other forums is permitted, provided the original author(s) or licensor are credited
1,6-diphosphatase, phosphofructokinase and phosphoenolpyruvate and that the original publication in this journal is cited, in accordance with accepted
carboxykinase in white muscle and red muscle. Biochem. J. 103, academic practice. No use, distribution or reproduction is permitted which does not
391–399. comply with these terms.

Frontiers in Physiology | www.frontiersin.org 136 June 2017 | Volume 8 | Article 413

 ORIGINAL RESEARCH
published: 29 May 2017
doi: 10.3389/fphys.2017.00298

Is Marathon Training Harder than the

Ironman Training? An ECO-method
Comparison
Jonathan Esteve-Lanao 1 , Diego Moreno-Pérez 1, 2 , Claudia A. Cardona 1, 3 ,
Eneko Larumbe-Zabala 4 , Iker Muñoz 5 , Sergio Sellés 6 and Roberto Cejuela 1, 6*
1
All In Your Mind TM, Madrid, España, 2 Fundamentos de los Deportes, Universidad Europea de Madrid, Madrid, España,
3
Departamento de Fisioterapia, Universidad del Valle de México, Mérida, Mexico, 4 Texas Tech University Health Sciences
Center, Lubbock, TX, USA, 5 Ciencias de la Salud, Universidad Europea del Atlántico, Santander, España, 6 Departmental
Section of Physical Education and Sports, University of Alicante, Alicante, Spain

Purpose: To compare the absolute and relative training load of the Marathon (42k) and
the Ironman (IM) training in recreational trained athletes.
Methods: Fifteen Marathoners and Fifteen Triathletes participated in the study. Their
performance level was the same relative to the sex’s absolute winner at the race. No
Edited by: differences were presented neither in age, nor in body weight, height, BMI, running
Thomas Leonhard Stöggl,
University of Salzburg, Austria
VO2max max, or endurance training experience (p > 0.05). They all trained systematically
Reviewed by:
for their respective event (IM or 42k). Daily training load was recorded in a training log,
Christoph Zinner, and the last 16 weeks were compared. Before this, gas exchange and lactate metabolic
Universität Würzburg, Germany
tests were conducted in order to set individual training zones. The Objective Load Scale
Jeremiah John Peiffer,
Murdoch University, Australia (ECOs) training load quantification method was applied. Differences between IM and 42k
*Correspondence: athletes’ outcomes were assessed using Student’s test and significance level was set at
Roberto Cejuela p < 0.05.
[email protected]
Results: As expected, Competition Time was significantly different (IM 11 h 45 min ± 1
Specialty section: h 54 min vs. 42k 3 h 6 min ± 28 min, p < 0.001). Similarly, Training Weekly Avg Time (IM
This article was submitted to
12.9 h ± 2.6 vs. 42k 5.2 ± 0.9), and Average Weekly ECOs (IM 834 ± 171 vs. 42k 526
Exercise Physiology,
a section of the journal ± 118) were significantly higher in IM (p < 0.001). However, the Ratio between Training
Frontiers in Physiology Load and Training Time was superior for 42k runners when comparing ECOs (IM 65.8 ±
Received: 19 December 2016 11.8 vs. 42k 99.3 ± 6.8) (p < 0.001). Finally, all ratios between training time or load vs.
Accepted: 24 April 2017
Published: 29 May 2017
Competition Time were superior for 42k (p < 0.001) (Training Time/Race Time: IM 1.1 ±
Citation:
0.3 vs. 42k 1.7 ± 0.5), (ECOs Training Load/Race Time: IM 1.2 ± 0.3 vs. 42k 2.9 ± 1.0).
Esteve-Lanao J, Moreno-Pérez D,
Conclusions: In spite of IM athletes’ superior training time and total or weekly training
Cardona CA, Larumbe-Zabala E,
Muñoz I, Sellés S and Cejuela R (2017) load, when comparing the ratios between training load and training time, and training
Is Marathon Training Harder than the time or training load vs. competition time, the preparation of a 42k showed to be harder.
Ironman Training? An ECO-method
Comparison. Front. Physiol. 8:298. Keywords: training intensity distribution, polarized training, training load quantification, endurance training,
doi: 10.3389/fphys.2017.00298 marathon, ironman

Frontiers in Physiology | www.frontiersin.org 137 May 2017 | Volume 8 | Article 298

Esteve-Lanao et al. Ironman or Marathon: What Training Is Harder?

INTRODUCTION their best at an Ironman distance triathlon (3.8k swim, 180k

cycle, and 42.1k run) or a Marathon (42.195k) race. Training
The interest of recreational athletes in long distance events has and competitive experience in endurance sports was similar
been constantly growing in the last 30 years. As a sample of between subjects (∼7 years). Subjects’ descriptive characteristics
this, the world’s marathon majors circuit reached about 200,000 are shown in Table 1.
runners in 2015 (WMM, 2016), while in triathlon, the Ironman R
corporation events reached about 70,000 (WTC, 2016). The Main Characteristics of Training and
Ironman (IM) and Marathon (42k) distances are the classical Periodization
longest endurance events in their respective sports. These events Both 42k (runners) and IM (triathletes) groups completed a
require large amounts of training (O’Brien et al., 1993; Laursen 16 weeks macrocycle (Figure 1). Global load was designed to
and Rhodes, 2001), so monitoring the training load becomes a alternate every 3 weeks of hard training load with an easy, lower
must in order to prevent over reaching or overtraining (Halson, load week (4, 4 weeks mesocycles). They followed an inverse
2014). periodization model, so that the peak training volumes were
“Training Load” (or training stimulus) implies the prescribed between weeks #10 and #11.
combination of the mode of exercise and the dose of the Hydration and nutritional guidelines were followed during
volume, intensity, and density or frequency (Wenger and Bell, these sessions, based on personal interviews with a sport
1986; Bompa and Haff, 2009). Quantifying the training load nutrition specialist, including the calculation of the sweating
becomes of key importance, since it helps to consider the real rate. Time of day was set early in the morning for running or
demands of a given sport discipline (Bompa and Haff, 2009). cycling sessions in order to minimize potential heart rate drift
There are many studies describing the physiological demands effects (Poole et al., 1994). Swimming sessions for triathletes were
of a long distance event competition (Föhrenbach et al., 1987; conducted twice to three times a week in the evening. Racing
O’Brien et al., 1993; Laursen and Rhodes, 2001), but few studies conditions were similar for both types of events (including
focused on the training load leading to a given performance temperature, humidity, altitude, and profile) and they were held
(Esteve-Lanao et al., 2007; Guellich and Seiler, 2010; Seiler, 2010; in Spain during the spring season.
Neal et al., 2013; Muñoz et al., 2014a,b; Stöggl and Sperlich, No differences were applied to the programs in strength
2014). Bearing in mind that it is difficult to establish the precise training. This training was based on maximal strength
amount of training load that an athlete needs (Seiler and development with moderate loads during the initial 11 weeks.
Tønnessen, 2009), several studies have been focused on training It consisted in progressive workouts starting with resistance
intensity distribution between professional and recreational training machines. Loads progressed from 2 to 4 sets per
athletes including different disciplines (Robinson et al., 1991; muscle group, 25–8 reps, 40–75% of estimated 1RM through
Lucía et al., 2000, 2003; Billat et al., 2001; Esteve-Lanao et al., submaximal testing calculations. Those exercises were replaced
2005; Seiler and Kjerland, 2006; Guellich et al., 2009). Different at week 4 by multi-joint exercises. Loads were gradually increased
methods have been proposed to quantify training load (Borresen in a similar fashion as mentioned before. Resistance training
and Lambert, 2009). One of the few methods (if any) that allow was gradually combined with specific strength methods in every
the comparisons of the training load between different modes of sport (paddles for swimming, hills, or harder gears for cycling,
exercise (i.e., running vs. swimming or cycling) is the Objective light weighted belts for running) between weeks 5 and 13. By
Load Scale (ECO in Spanish) method (Cejuela and Esteve-Lanao, the end of the macrocycle, some basic maintenance resistance
2011). training on machines was conducted with moderate loads
Since there are no data comparing athletes of different sports (60–70% of estimated 1RM), low number, explosive-velocity
who perform at similar level for IM and 42k (i.e., trained
recreational athletes), and starting from the simplistic athletes’
question of “how hard it is” to be ready for these challenges,
TABLE 1 | Samplee characteristics.
the goals of our study were: (1) To observe the differences
between IM and 42k training load and (2) To observe the Ironman Marathon t28 p
correlation between the competition time and the training (n = 15) (n = 15)
intensity distribution at each group of endurance athletes.
Age (years) 39.1 (7.5) 38.2 (7.6) 0.34 0.737
Weight (kg) 70.9 (8.3) 69.7 (8.1) 0.42 0.675
METHODS Height (cm) 174.5 (7.3) 175.2 (5) −0.32 0.751
BMI (kg/m2 ) 23.3 (2.4) 22.7 (2.3) 0.71 0.482
Participants VO2max (ml/min/kg) 56.1 (6.1) 57.5 (6.8) −0.61 0.550
Thirty recreational level athletes (15 long distance triathletes and Training experience (year) 7.1 (2.1) 7.3 (2.3) −0.41 0.672
15 marathon runners) participated in the study. Both groups Relative performance (%) 140.4 (24.5) 143.7 (22.2) −0.38 0.705
had 13 male and 2 females. They all volunteered and gave Competition time (min) 704.6 (113.6) 186.5 (27.9) 17.15 <0.001
written informed consent to participate in the study, which had Same sex’s absolute winner 503.7 (33.7) 130.1 (4.9) 42.49 <0.001
been approved by the Universidad Europea Ethical Advisory time (min)
Committee. They all lived and trained in Spain, with the same Relative performance was calculated dividing competition time by absolute winner’s time,
coach (J. E-L). Their main goal for the season was to perform considering the sex of each participant.

Frontiers in Physiology | www.frontiersin.org 138 May 2017 | Volume 8 | Article 298

Esteve-Lanao et al. Ironman or Marathon: What Training Is Harder?

FIGURE 1 | Experimental design. Planned training load is described with arbitrary units, with the purpose of showing general trend of hard/easy weeks.

reps. Two sessions a week were programed (for most of the Running tests were conducted on a treadmill (Technogym
weeks with exception of the taper phase and the 4th weeks Run Race 1400 HC, Gambettola, Italy), with a starting velocity
during each mesocycle). About 1.75 (0.20) sessions/week were of 8 km·h−1 , increased by 0.3 km·h−1 every 30 s until volitional
reported. However, the training load for strength training was exhaustion (Esteve-Lanao et al., 2007; Muñoz et al., 2014b).
not quantified for this study. For both cycling and running tests, standardized criteria were
No speed training or any other workouts beyond VO2max zone used for VO2max achievement, AnT and AeT determination
were prescribed. and HR recordings, following previously described procedures
(Muñoz et al., 2014a,b).

Baseline Physiological Testing and Training Training Intensity Distribution and Training
Zones Settings Load Quantification
Before starting the training program, all athletes participated Based on the classical 3-phase model of Skinner and McLellan
in short-distance events macrocycles (10k or Half-Marathon for (1980), and for practical purposes in terms of training intensity
runners, Sprint or Olympic distance for triathletes), followed by distribution analysis, three main zones were differentiated: <AeT
a 3–4 weeks transition period. One week before starting the 16- (at or below AeT), BAeT-AnT (between thresholds, precisely
week macrocycle, during that transition period, graded exercise beyond AeT and below AnT) and >AnT (at or beyond AnT).
tests were used to determine training zones. Of note, all data were included, both warm-ups and cool-downs,
Swimming tests were performed as a graded multi-stage test thus not following the so-called “Session Goal Aproach” applied
consisting of 7 repetitions of 200 m with 2 min rests (Pyne in other studies (Seiler, 2010).
and Sweetenham, 2003). Heart rate (HR, beats·min−1 ) and For daily training workouts, these three zones were subdivided
blood lactate (bLA, mMol·L−1 ) samples from the earlobe were into narrower ranges (dividing each zone for being more precise
analyzed with a portable lactate analyzer (Lactate Pro, Arkray Inc, in some workouts, and adding a Maximal Aerobic Power zone for
Amstelveen, NED). Triathletes were familiar with the Rating of some swimming workouts), up to a total of 6 zones from <AeT
Perceived Exertion (RPE) (Borg, 1998), so RPE 0–10 scale was to Maximal Aerobic Power plus 2 “anaerobic” zones. Microsoft R
applied immediately after collecting HR data and LA samples. Excel R Training logs were designed to calculate training load
Threshold criteria were defined as follows: blood lactate 0.5 based on the methodology of ECOs (Cejuela and Esteve-Lanao,
mMol·L−1 increase toward previous stage for Aerobic Threshold 2011) which was specially developed for training quantification
(AeT), >1.0 mMol·L−1 increase for Anaerobic Threshold (AnT), in triathlons, so that it is suitable for comparisons to any single
and 8–9 mMol·L−1 for Maximal Aerobic Velocity (MAV) event sport.
(Beneke, 2003; Billat et al., 2003). This methodology seems the most appropriate when
Cycling and running tests were conducted with a gas exchange comparing different endurance activities, as different exercise
analyzer (VO2000, Medical Graphics, St, Paul, MA, USA). A modes show different degrees of muscle damage, energy cost,
ramp-protocol test was conducted for cycling on an ergometer effort densities, and differences at the ability of maintaining
(Sensormedics, Yorba Linda, CA, USA) starting at 50 W and technique (Cejuela and Esteve-Lanao, 2011).
increasing 5 W every 12 s (Lucía et al., 2000). Two independent The Percentages of time spent at <AeT, BAeT-AnT, and >AnT
observers identified AeT and AnT. The following variables were calculated by the training log dividing the total time spent
were measured: oxygen uptake (VO2 ), pulmonary ventilation during a workout at a given zone by the total exercising time
(VE), ventilatory equivalents for oxygen (VE·VO−1 2 ), and carbon during the workout and multiplied by 100. The training logs were
dioxide (VE·CO−1 2 ), and end-tidal partial pressure of oxygen prepared in order to record every session and to differentiate
(PET O2 ) and carbon dioxide (PET CO2 ). running or triathlon (swim/bike/run) sessions and calculations.

Frontiers in Physiology | www.frontiersin.org 139 May 2017 | Volume 8 | Article 298

Esteve-Lanao et al. Ironman or Marathon: What Training Is Harder?

Briefly, the ECOs were calculated by multiplying the total version 22 (SPSS Inc., Chicago, Illinois, USA). The significance
duration of a training session (in minutes) with a scoring value level was set at 0.05.
between 1 and 50, depending on the heart rate-based training
zone (1–8) and by a factor of 1.0, 0.75, or 0.5 for running,
swimming or biking, respectively. Daily and weekly training RESULTS
loads (ECOs) of each subject were quantified. For example, a 60
min running session at Zone 1 is scored like this: 60 × 1 × 1, As shown in Table 1, IM and 42k athletes did not show any
since Running has a factor of “1,” and Zone 1 has a factor of significant difference in age, weight, height, body mass index, nor
“1.” However, a 2-h cycling session at zone 1 also scores “60,” VO2max . Additionally, their relative performance, compared to
since Cycling has a factor of “0.5” (120 × 1 × 0.5). Another their sex’s absolute winner, was also equivalent (40.4 and 43.7%
example (in the case of an interval training workout), would be over the winner’s time, respectively). Mean performance time was
a 12 × 100 m swimming workout at zone 4 plus other 20 min 11 h 45 min for IM and 3 h 06 min for 42k.
of zone 1 swimming adding the warm up and the cool down. If Training logs were reported weekly and were filled by all
the swimmer would we performing at 1 min 40 s per rep, this the athletes. No relevant injuries, tactical or technical problems
would be a total of 20 min at zone 4. Thus, if the swimmer would appeared along the training program or competitions. Based
we performing 20 (min) × Zone 1 is 20, and multiplied by 0.75 on total ECOs, an average of 98.7(20.2)% accomplishment was
scores 15. When adding the other 20 min interval net time at found.
Zone 4 per 0.75 is 60, so the total session scores 75 ECOs. As shown in Table 2 (or equivalently in Figure 2), IM athletes’
Both runners and triathletes were filling manually personal loads were significantly higher during the greater part of training
training logs with the information recorded in their HR cycle. IM athletes invested significantly more time and ended
monitors, considering net training time from the whole session, up at higher loads, with higher weekly averages. Training peaks
in terms of the amount of time spent per training zone at each in time and loads were also significantly higher for IM athletes
sport (Cejuela and Esteve-Lanao, 2011). (Table 3).
Speed, Power or Pace values corresponding to the training Peak Training Load was located about the same period
zones were increased during the program according to RPE/HR from competition in both groups (between 6th and 8th before
initial training zones, as previously described (Muñoz et al., competition). During the last 4 weeks, the loads were reduced
2014a) and based on the reported validity of these lab references in both groups, but more markedly in IM, so that no differences
during a subsequent period of several months (Lucía et al., were found during that last mesocycle.
2000). RPE was considered appropriate for technical swimming Although no significant differences were found in relative time
exercises and also when HR recording showed any anomalous and load in zone <AeT, IM athletes showed significantly higher
display or abnormality. relative times and loads in zone BAeT-AnT, while 42k runners’
Training Loads were designed to meet a mean (SD) of relative time and load in zone >AnT were significantly higher
∼15,313(2,087) ECOs for the IM and 8,333(699) for 42k. (Table 3).
The reason for this SD was that different programs were Ratios of training load by training time in 42k runners
designed according to performance level differences, plus other showed significantly higher relative training load [t (28) = −9.50,
considerations such as time availability or training experience.
General training intensity distribution was scheduled week by
week, with a global mean of 84/7/9%, respectively, in Zones 1/2/3 TABLE 2 | Training Load per week (ECOs, weeks 1–16).
(IM 78/19/3, 42k 86/2/12).
Inclusion criteria were the following: (1) to complete 85% of Week (#) Ironman (n = 15) Marathon (n = 15) t28 p

total training sessions, (2) to record 95% of total training sessions,

1 677.2 (169.4) 348.6 (162.5) 5.42 <0.001
and (3) to complete and perform continuously, without any
2 834.8 (163.6) 540 (146.7) 5.20 <0.001
relevant health, tactical, or technical problems, the full distance
3 961.1 (276) 452.4 (190.7) 5.87 <0.001
in competition performing as best as possible.
4 826.6 (260.3) 575.5 (143) 3.27 0.003
5 977.1 (296) 535.9 (143) 5.20 <0.001
6 842.7 (251) 608.7 (143) 3.14 0.005
Data Analysis
7 946.7 (404.9) 554.7 (136.8) 3.55 0.002
Training accomplishment was calculated dividing the training
8 863.5 (193.5) 529.6 (162.2) 5.12 <0.001
logs reported data by the prescribed training loads. Relative
performance (%) was calculated dividing competition time by 9 970.7 (310.2) 561.2 (185) 4.39 <0.001

absolute winner’s time, considering the sex of each participant, 10 1044.3 (381.3) 665.4 (218.5) 3.34 0.002

and multiplying the result by 100. Data were summarized 11 1064.9 (400.8) 647.7 (227.5) 3.51 0.002

as mean (standard deviation). Differences between IM and 12 897.3 (248.9) 555.3 (225.2) 3.95 <0.001

42k athletes’ outcomes were assessed using Student’s t-test. 13 710.7 (274.5) 621.7 (214.2) 0.99 0.331
Association between performance and training time and loads 14 807.8 (347.1) 561.3 (229.6) 2.30 0.031
were assessed using Pearson’s product-moment correlation 15 592.1 (267.6) 453.0 (239.5) 1.50 0.145
coefficients. This statistical analysis was performed using SPSS 16 210.5 (160.8) 206.0 (72.8) 0.10 0.923

Frontiers in Physiology | www.frontiersin.org 140 May 2017 | Volume 8 | Article 298

Esteve-Lanao et al. Ironman or Marathon: What Training Is Harder?

between performance and “easy” (<AeT) or “moderate” (BAeT-

AnT) zones time or load, “intense” >AnT zone seems to be linked
to a better 42k performance, but not in IM. Again, the superior
intensity which marathoners perform seems to explain why this
association might be found. For instance, a 3 h marathoner
(like those in our study) will average around 70% of VO2max
during the race (O’Brien et al., 1993), which would be BAeT-
AnT zone in our runners’ assessments. However, only during
the swimming it might be reasonable to exert beyond AeT (and
maybe at some particular moments on the bike) during an IM
distance triathlon (Muñoz et al., 2014a). Thus, this opens new
insights for the general assumption of the benefits of “Polarized
Training Distribution,” so that it might not be “always better”
for all disciplines. In fact, previous studies reported polarized
distribution in top elite marathoners (Billat et al., 2001), but not
in recreational IM distance triathletes (Neal et al., 2011; Muñoz
FIGURE 2 | Comparison of IM (black) and 42k (white) weekly training et al., 2014a).
loads (ECOs) during 16 weeks. *p < 0.05; **p < 0.01; ***p < 0.001. The training intensity distribution found in this study
(considering both groups) is in line with other sports (Billat et al.,
2001; Lucía et al., 2003; Fiskerstrand and Seiler, 2004; Seiler and
Kjerland, 2006; Guellich and Seiler, 2010). One of the more active
p < 0.001]. Additionally, ratios of training time [t (28) = −3.99, p
groups of researchers in this issue is Seiler’s (Seiler and Kjerland,
< 0.001] and load relative to competition time [t (28) = −6.69, p
2006; Seiler and Tønnessen, 2009; Seiler, 2010) who pointed
< 0.001] also favored 42k runners.
out a 75-5-20 or even 80-0-20 distribution as “optimal,” which
When observing correlations with performance time, both
seems to be really difficult to achieve unless discarding the warm-
groups showed significant, inversely related correlations between
ups and cool-downs. Had we done it, we would have probably
total training time and performance (the more training time,
reached superior >AnT zone percentage. However, as previously
the better performance). However, only the 42k group showed
described, since they were recreational athletes (i.e., reduced
large and significant correlation between total training load and
time and training frequency) the training programs increased the
performance.
weekly time spent in easy zones by using relatively long warm-ups
Relative time and loads in zone <AeT showed a negative
and cool-downs.
association to performance only for IM, but not for 42k (where
Load dynamics were not exactly as originally designed. Peak
only time in Zone <AeT was significantly related). Conversely,
load week was approximately the same between groups, although
absolute and relative time in zone >AnT, and relative load in zone
tapering was different. In any case, the tapering technique was
>AnT, showed a negative association to relative performance
progressive, particularly in 42k. As previously discussed in the
only for 42k, but not for IM. Relative time and loads in
literature, progressive, non-linear tapering techniques seem to
zone BAeT-AnT showed a significant association with poor
have a more pronounced positive impact on performance than
performance for both groups (so more training time or load
step-taper strategies (Mujika and Padilla, 2003).
BAeT-AnT, the worst performance). Total training time in Zone
As suspected, IM athletes trained more than 42k runners.
BAeT-AnT did not show significant associations (see Table 4).
Particularly, IM group trained about more than twice in time and
about 1/3 in training load. Recently, it was reported that national
DISCUSSION level triathletes trained an average of 1,256 ECOs per week over
the same period of time as our study (unpublished data from
Both groups presented a significant association between more Saugy et al., 2016). Triathletes in our study averaged 834. Top
<AeT training and better performance. However, the opposite Class Marathoners (according to our calculations using the ECOs
association was found with BAeT-AnT, so that the more you train method), were 1,200 ECOs whilst trained marathoners would
“moderate,” the worse in IM or 42k. This agrees with previous be around 1,000 weekly (Billat et al., 2001). The marathoners
studies conducted with age-group IM triathletes (Muñoz et al., in our study averaged 526 ECOs. These and future data might
2014a), recreational 10k runners (Muñoz et al., 2014b), and help general references toward optimal doses in relation to
trained Cross Country runners (Esteve-Lanao et al., 2005, 2007). performance level, sports and events, although it is known that
To our knowledge, this is the first study to identify these training response and performance outputs are multifactorial,
associations in recreational trained marathon runners. including genetics (Smith, 2003).
Interestingly, >AnT was associated to better performance The original question of this study was about “how hard it
in 42k but not in IM. According to the existing literature is” to train for an IM vs. for a 42k. The approach that has
(O’Brien et al., 1993; Laursen and Rhodes, 2001), we suspect been conducted compares training time and load in relation
that physiological intensity differences between these two events to competition. For instance, classical swimming volumes
might explain this. In contrast with the opposite pattern shown are tremendously high compared to the competition distance

Frontiers in Physiology | www.frontiersin.org 141 May 2017 | Volume 8 | Article 298

Esteve-Lanao et al. Ironman or Marathon: What Training Is Harder?

TABLE 3 | Load description as total values, weekly averages, percentages of load by zone, peak values, and relative loads to training and competition
time.

Ironman (n = 15) Marathon (n = 15) t28 p

TOTALS
Total training time (h) 206.7 (40.8) 84.3 (15.5) 10.86 <0.001
Total training load (ECOs) 13347.1 (2732.5) 8416.6 (1887.8) 5.75 <0.001
WEEKLY AVERAGES
Training weekly avg time (h) 12.9 (2.6) 5.2 (0.9) 10.90 <0.001
Training weekly avg time (min) 775.5 (153.7) 311.4 (56.2) 10.98 <0.001
Average weekly load (ECOs) 834.1 (170.7) 526 (118.1) 5.75 <0.001
TRAINING BY ZONES
% of Time in zone <AeT 67.5 (13.6) 74.6 (3.9) −1.94 0.070
% of Time in zone BAeT-AnT 28.4 (11.8) 15.6 (4.9) 3.87 0.001
% of Time in zone >AnT 4.3 (2.3) 9.7 (3.9) −4.64 <0.001
% of Load in zone <AeT 45.8 (11.9) 48.3 (5) −0.74 0.470
% of Load in zone BAeT-AnT 40.8 (9.8) 26.5 (8.1) 4.37 <0.001
% of Load in zone >AnT 13.7 (5.7) 25.1 (10.4) −3.75 0.001
PEAKS
Training time peak (h) 20.9 (4.8) 8.8 (1.8) 9.17 <0.001
Peak load (ECOs) 1345.5 (355.1) 838.1 (178.9) 4.94 <0.001
Peak load week (#) 7.9 (1.9) 6.5 (2.7) 1.63 0.115
RATIOS BY TRAINING TIME
Training load (ECOs) 65.8 (11.8) 99.3 (6.8) −9.50 <0.001
RATIOS BY COMPETITION TIME
Training time (h) 1.1 (0.3) 1.7 (0.5) −3.99 <0.001
Training load (ECOs) 1.2 (0.3) 2.9 (1) −6.69 <0.001

TABLE 4 | Correlation coefficients (and p-value) between performance Training Load per every minute spent in competition is higher
(min) and training load distribution. for 42k (2.9 vs. 1.2), as well as the training time invested per every
IM 42k
minute in competition (1.7 vs. 1.1, 42k vs. IM). This is the first
study to compare athletes from similar level. Further studies will
TOTAL have to present new comparisons in terms of performance level,
Total training time (h) −0.59 (0.021) −0.80 (<0.001) sports and disciplines.
Total training load (ECOs) −0.04 (0.894) −0.73 (0.002) The main limitation in our study was that it was conducted
ZONE <AeT with athletes who were trained by the same coach, so the
Training time in zone <AeT −0.74 (0.002) −0.82 (<0.001) applicability of the results should be restricted to similar
% of Time in zone <AeT −0.70 (0.004) −0.10 (0.723) conditions. This was done because it allowed a better control
% of Load in zone <AeT −0.60 (0.019) 0.37 (0.175) of many aspects of the program (such as baseline loads,
ZONE BAeT-AnT mesocycles distributions, program length, peak volume location,
Training Time in zone BAeT-AnT 0.47 (0.075) 0.11 (0.700) taper design, strength, and speed training methodology, and
% of Time in zone BAeT-AnT 0.71 (0.003) 0.65 (0.009) training supervision). General training intensity distribution
% of Load in zone BAeT-AnT 0.54 (0.038) 0.79 (<0.001) was scheduled to be ∼84/7/9%, respectively, in Zones 1/2/3.
ZONE >AnT However, it showed a global mean of ∼72/20/8%, respectively,
Training time in zone >AnT 0.33 (0.223) −0.77 (0.001)
in Zones 1/2/3. IM should be ∼78/19/3% but it was ∼68/28/4%,
% of Time in zone >AnT 0.44 (0.099) −0.74 (0.002)
42k should be ∼86/2/12%, but it was ∼74/16/10%. This
% of Load in zone >AnT 0.33 (0.234) −0.81 (<0.001)
training intensity distribution has been classified as “pyramidal”
(Stöggl and Sperlich, 2015). According to Zone 1 and Zone 2
distributions, both showed a standard deviation of ∼11% among
all athletes and both groups. Consequently, in spite of sharing
(Mujika et al., 1995). The findings of our study state that the same coach supervision, a high variability between programs’
42k training is harder in relation to the competition demands. loads, accomplishment and intensity distribution were given.
Training Load per training hour is significantly higher (99.3 The ECOs method has been recently used in a dozen of
ECOs in 42k vs. 65.8 in IM), which is about 1.5 ECOs in 42k peer-reviewed papers with elite and high level athletes (Debevec
training, vs. 1 ECO per min in IM. Moreover, Average Weekly et al., 2015; Hauser et al., 2016; Saugy et al., 2016; Villaño et al.,

Frontiers in Physiology | www.frontiersin.org 142 May 2017 | Volume 8 | Article 298

Esteve-Lanao et al. Ironman or Marathon: What Training Is Harder?

2016). Further studies should focus on training tolerance between Europea de Madrid, with written informed consent from
different athletes’ disciplines but including biomarkers to relate all subjects. All subjects gave written informed consent in
these theoretical training load comparisons and discriminate accordance with the Declaration of Helsinki. The protocol was
between hormonal status, muscle damage, oxidative stress levels approved by the Ethics Committee of the Universidad Europea
or others. de Madrid.

CONCLUSIONS AUTHOR CONTRIBUTIONS

The highest associations between performance and training were JE and RC made the general design. JE trained the athletes.
found with time and % of time in zone <AeT for IM, whilst JE, RC, IM, and EL wrote the manuscript. DM and CC
in 42k it was related to the zone <AeT time or load, plus the participated in the data acquisition conducting the performance
accumulation of any variable related to >AnT. tests, supervising the training workouts, supplying on line
In spite of IM athletes’ superior training time and total or assistance and preparing the data. EL analyzed the data and
weekly training load, and according to the ratios between training wrote a substantial part of the results section. SS reviewed
load and training time, and training time or training load vs. the manuscript and made important contributions to the final
competition time, the preparation of a 42k showed to be harder. edition.

ETHICS STATEMENT ACKNOWLEDGMENTS

This study was carried out in accordance with the We thank Jesica de la Mata for her grammatical assistance
recommendations of the Ethics Committee of the Universidad throughout this manuscript.

REFERENCES Guellich, A., and Seiler, S. (2010). Lactate profile changes in relation to training
characteristics in junior elite cyclists. Int. J. Sports Physiol. Perform. 5, 316–327.
Beneke, R. (2003). Methodological aspects of maximal lactate steady state- doi: 10.1123/ijspp.5.3.316
implications for performance testing. Eur. J. Appl. Physiol. 89, 95–99. Guellich, A., Seiler, S., and Emrich, E. (2009). Training methods and intensity
doi: 10.1007/s00421-002-0783-1 distribution of young world-class rowers. Int. J. Sports Physiol. Perform. 4,
Billat, V. L., Demarle, A., Slawinski, J., Paiva, M., and Koralsztein, J. P. (2001). 448–460. doi: 10.1123/ijspp.4.4.448
Physical and training characteristics of top-class marathon runners. Med. Sci. Halson, S. L. (2014). Monitoring training load to understand fatigue in athletes.
Sports Exerc. 33, 2089–2097. doi: 10.1097/00005768-200112000-00018 Sports Med. 44, 139–147. doi: 10.1007/s40279-014-0253-z
Billat, V., Sirvent, P., Py, G., Koralsztein, J., and Mercier, J. (2003). Hauser, A., Schmitt, L., Troesch, S., Saugy, J. J., Cejuela-Anta, R., Faiss,
The concept of maximal lactate steady state: a bridge between R., et al. (2016). Similar hemoglobin mass response in hypobaric and
biochemistry, physiology and sport science. Sports Med. 33, 407–426. normobaric hypoxia in athletes. Med. Sci. Sports Exerc. 48, 734–741.
doi: 10.2165/00007256-200333060-00003 doi: 10.1249/MSS.0000000000000808
Bompa, T., and Haff, G. (2009). Training Theory and Methodology. Champaign, IL: Laursen, P. B., and Rhodes, E. C. (2001). Factors affecting performance
Human Kinetics. in an ultraendurance triathlon. Sports Med. 31, 195–209.
Borg, G. (1998). Borg’s Perceived Exertion and Pain Scales. Champaign, IL: Human doi: 10.2165/00007256-200131030-00004
Kinetics. Lucía, A., Hoyos, J., Pérez, M., and Chicharro, J. L. (2000). Heart rate and
Borresen, J., and Lambert, M. (2009). The quantification of training load, the performance parameters in elite cyclists: a longitudinal study. Med. Sci. Sports
training response and the effect on performance. Sports Med. 39, 779–795. Exerc. 32, 1777–1782. doi: 10.1097/00005768-200010000-00018
doi: 10.2165/11317780-000000000-00000 Lucía, A., Hoyos, J., Santalla, A., Earnest, C., and Chicharro, J. L. (2003). Tour
Cejuela, R., and Esteve-Lanao, J. (2011). Training load quantification in triathlon. de France versus Vuelta a Espa-a: which is harder? Med. Sci. Sports Exerc. 35,
J. Hum. Sport Exerc. 6, 218–232. doi: 10.4100/jhse.2011.62.03 872–878. doi: 10.1249/01.MSS.0000064999.82036.B4
Debevec, T., Pialoux, V., Saugy, J., Schmitt, L., Cejuela, R., Mury, P., et al. Mujika, I., Chatard, J. C., Busso, T., Geyssant, A., Barale, F., and Lacoste, L. (1995).
(2015). Prooxidant/antioxidant balance in hypoxia: a cross-over study on Effects of training on performance in competitive swimming. Can. J. Appl.
normobaric vs. Hypobaric “live high-train low.” PLoS ONE 10:e0137957. Physiol. 20, 395–406. doi: 10.1139/h95-031
doi: 10.1371/journal.pone.0137957 Mujika, I., and Padilla, S. (2003). Scientific bases for precompetition
Esteve-Lanao, J., Foster, C., Seiler, S., and Lucia, A. (2007). Impact of training tapering strategies. Med. Sci. Sports Exerc. 35, 1182–1187.
intensity distribution on performance in endurance athletes. J. Strength Cond. doi: 10.1249/01.MSS.0000074448.73931.11
Res. 21, 943–949. doi: 10.1519/00124278-200708000-00048 Muñoz, I., Cejuela, R., Seiler, S., Larumbe, E., and Esteve-Lanao, J. (2014a).
Esteve-Lanao, J., San Juan, A. F., Earnest, C. P., Foster, C., and Lucia, Training-intensity distribution during an ironman season: relationship with
A. (2005). How do endurance runners actually train? Relationship competition performance. Int. J. Sports Physiol. Perform. 9, 332–339.
with competition performance. Med. Sci. Sports Exerc. 37, 496–504. doi: 10.1123/ijspp.2012-0352
doi: 10.1249/01.MSS.0000155393.78744.86 Muñoz, I., Seiler, S., Bautista, J., España, J., Larumbe, E., and Esteve-Lanao, J.
Fiskerstrand, A., and Seiler, K. S. (2004). Training and performance characteristics (2014b). Does polarized training improve performance in recreational runners?
among Norwegian International Rowers 1970-2001. Scand. J. Med. Sci. Sports Int. J. Sports Physiol. Perform. 9, 265–272. doi: 10.1123/ijspp.2012-0350
14, 303–310. doi: 10.1046/j.1600-0838.2003.370.x Neal, C. M., Hunter, A. M., Brennan, L., O’Sullivan, A., Hamilton, D. L.,
Föhrenbach, R., Mader, A., and Hollmann, W. (1987). Determination of endurance De Vito, G., et al. (2013). Six weeks of a polarized training-intensity
capacity and prediction of exercise intensities for training and competition distribution leads to greater physiological and performance adaptations
in marathon runners. Int. J. Sports Med. 8, 11–18. doi: 10.1055/s-2008-10 than a threshold model in trained cyclists. J. Appl. Physiol. 114, 461–471.
25633 doi: 10.1152/japplphysiol.00652.2012

Frontiers in Physiology | www.frontiersin.org 143 May 2017 | Volume 8 | Article 298

Esteve-Lanao et al. Ironman or Marathon: What Training Is Harder?

Neal, C. M., Hunter, A. M., and Galloway, S. D. R. (2011). A 6-month analysis Smith, D. (2003). A framework for understanding the training
of training-intensity distribution and physiological adaptation in Ironman process leading to elite performance. Sports Med. 33, 1103–1126.
triathletes. J. Sports Sci. 29, 1515–1523. doi: 10.1080/02640414.2011.596217 doi: 10.2165/00007256-200333150-00003
O’Brien, M. J., Viguie, C. A., Mazzeo, R. S., and Brooks, G. A. (1993). Carbohydrate Stöggl, T., and Sperlich, B. (2014). Polarized training has greater impact on key
dependence during marathon running. Med. Sci. Sports Exerc. 25, endurance variables than threshold, high intensity, or high volume training.
1009–1017. Front. Physiol. 5:33. doi: 10.3389/fphys.2014.00033
Poole, D. C., Barstow, T. J., Gaesser, G. A., Willis, W. T., and Whipp, B. J. (1994). Stöggl, T., and Sperlich, B. (2015). The training intensity distribution
VO2 slow component: physiological and functional significance. Med. Sci. among well-trained and elite endurance athletes. Front. Physiol. 27:295.
Sports Exerc. 26, 1354–1358. doi: 10.3389/fphys.2015.00295
Pyne, S., and Sweetenham, B. (2003). “Chapter 16: Program evaluation,” in Villaño, D., Vilaplana, C., Medina, S., Algaba-Chueca, F., Cejuela-Anta, R.,
Championship Swim Training, eds B. Sweetenham and J. Atkinson (Champaign Martínez-Sanz, J., et al. (2016). Relationship between the ingestion of a
IL: Human Kinetics), 279–288. polyphenol-rich drink, hepcidin hormone, and long-term training. Molecules
Robinson, D. M., Robinson, S. M., Hume, P. A., and Hopkins, W. G. (1991). 21:1333. doi: 10.3390/molecules21101333
Training intensity of elite male distance runners. Med. Sci. Sports Exerc. 23, Wenger, H. A., and Bell, G. J. (1986). The interactions of intensity, frequency and
1078–1082. doi: 10.1249/00005768-199109000-00013 duration of exercise training in altering cardiorespiratory fitness. Sports Med. 3,
Saugy, J. J., Schmitt, L., Hauser, A., Constantin, G., Cejuela, R., Faiss, R., et al. 346–356. doi: 10.2165/00007256-198603050-00004
(2016). Same performance changes after live high-train low in normobaric vs. WMM (2016). Facts and Figures. Chicago, IL: Abbott World Marathon Majors.
hypobaric hypoxia. Front. Physiol. 7:138. doi: 10.3389/fphys.2016.00138 WTC (2016). World Triathlon Corporation. Tampa, FL.
Seiler, K. S., and Kjerland, G. Ø. (2006). Quantifying training intensity distribution
in elite endurance athletes: is there evidence for an “optimal” distribution? Conflict of Interest Statement: The authors declare that the research was
Scand. J. Med. Sci. Sports 16, 49–56. doi: 10.1111/j.1600-0838.2004.00418.x conducted in the absence of any commercial or financial relationships that could
Seiler, S. (2010). What is best practice for training intensity and duration be construed as a potential conflict of interest.
distribution in endurance athletes? Int. J. Sports Physiol. Perform. 5, 276–291.
doi: 10.1123/ijspp.5.3.276 Copyright © 2017 Esteve-Lanao, Moreno-Pérez, Cardona, Larumbe-Zabala, Muñoz,
Seiler, S., and Tønnessen, E. (2009). Intervals, thresholds, and long slow distance : Sellés and Cejuela. This is an open-access article distributed under the terms
the role of intensity and duration in endurance training. Training 13, 32–53. of the Creative Commons Attribution License (CC BY). The use, distribution or
Available online at: http://sportsci.org/2009/ss.pdf reproduction in other forums is permitted, provided the original author(s) or licensor
Skinner, J. S., and McLellan, T. H. (1980). The transition from are credited and that the original publication in this journal is cited, in accordance
aerobic to anaerobic metabolism. Res. Q. Exerc. Sport 51, 234–248. with accepted academic practice. No use, distribution or reproduction is permitted
doi: 10.1080/02701367.1980.10609285 which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 144 May 2017 | Volume 8 | Article 298

 METHODS
published: 24 May 2017
doi: 10.3389/fphys.2017.00337

Intensity- and Duration-Based

Options to Regulate Endurance
Training
Peter Hofmann * and Gerhard Tschakert

Exercise Physiology, Training and Training Therapy Research Group, Institute of Sports Science, University of Graz, Graz,
Austria

The regulation of endurance training is usually based on the prescription of exercise

intensity. Exercise duration, another important variable of training load, is rarely prescribed
by individual measures and mostly set from experience. As the specific exercise
duration for any intensity plays a substantial role regarding the different kind of cellular
stressors, degree, and kind of fatigue as well as training effects, concepts integrating
the prescription of both intensity and duration within one model are needed. An
according recent approach was the critical power concept which seems to have a
physiological basis; however, the mathematical approach of this concept does not
allow applying the three zones/two threshold model of metabolism and its different
physiological consequences. Here we show the combination of exercise intensity and
duration prescription on an individual basis applying the power/speed to distance/time
relationship. The concept is based on both the differentiation of intensities by two
Edited by:
Thomas Leonhard Stöggl, lactate or gas exchange variables derived turn points, and on the relationship between
University of Salzburg, Austria power (or velocity) and duration (or distance). The turn points define three zones of
Reviewed by: intensities with distinct acute metabolic, hormonal, and cardio-respiratory responses for
Beat Knechtle,
University of Zurich, Switzerland endurance exercise. A maximal duration exists for any single power or velocity such as
Patrick Wahl, described in the power-duration relationship. Using percentages of the maximal duration
German Sport University Cologne,
allows regulating fatigue, recovery time, and adaptation for any single endurance training
Germany
session. Four domains of duration with respect to induced fatigue can be derived from
*Correspondence:
Peter Hofmann maximal duration obtained by the power-duration curve. For any micro-cycle, target
[email protected] intensities and durations may be chosen on an individual basis. The model described
here is the first conceptual framework of integrating physiologically defined intensities
Specialty section:
This article was submitted to and fatigue related durations to optimize high-performance exercise training.
Exercise Physiology,
a section of the journal Keywords: exercise prescription, intensity, duration, endurance exercise, athletes
Frontiers in Physiology

Received: 28 October 2016

Accepted: 08 May 2017
INTRODUCTION
Published: 24 May 2017
The regulation of endurance training is usually based on the prescription of individual exercise
Citation:
intensity zones/domains, (Meyer et al., 2005; Pescatello, 2014, p. 168) whereas, in contrast, exercise
Hofmann P and Tschakert G (2017)
Intensity- and Duration-Based Options
duration is rarely prescribed by individual measures and mostly set from personal experience
to Regulate Endurance Training. or “usual” settings. Tremblay et al. (2005) critically mentioned that little research tempting to
Front. Physiol. 8:337. isolate the effect of exercise duration has been done but they suggested a duration threshold
doi: 10.3389/fphys.2017.00337 for hormonal responses especially for low intensity exercise. This is in line with earlier results

Frontiers in Physiology | www.frontiersin.org 145 May 2017 | Volume 8 | Article 337

Hofmann and Tschakert Intensity and Duration Framework of Training

by Viru et al. (1996) proposing that in exercise performed PRESCRIPTION OF INTENSITY

below a certain threshold intensity, hormonal responses will only
occur after a certain long duration. As the hormonal changes The prescription of exercise intensity for endurance-type exercise
trigger acute and chronic adaptation it is suggested that not is usually based on exercise markers from maximal and/or
only intensity but also the duration for any specific intensity sub-maximal incremental exercise tests (Meyer et al., 2005).
is crucial to induce training effects or to avoid overload (Viru, Guidelines recommend using percentages of maximal oxygen
1995, p. 1–20). Recently, it was shown by Skovgaard et al. (2016) uptake (VO2max ), maximal heart rate (HRmax ), or maximal
that muscle PGC-1α mRNA, identified as a key regulator of power output (Pmax ) for setting exercise intensity (Pescatello,
mitochondrial biogenesis and oxidative genes, did not change 2014, p. 168). However, threshold or turn point concepts
significantly after 60 min of endurance exercise in their study. are suggested to be the gold standard for exercise intensity
This short duration and a high fitness level did not allow to prescription in practice (Meyer et al., 2005) although still
sufficiently challenge muscle PGC-1α mRNA for the relative low critically discussed (Mann et al., 2013). Actually, most authors
exercise intensity (60% of VO2max ) applied. However, including agree to set training intensities by a three phase and two threshold
high intensity speed endurance exercise provided a stimulus model (Meyer et al., 2005; Hofmann and Tschakert, 2010)
for muscle mitochondrial biogenesis, substrate regulation, and indicated by a first lactate (LT1 /LTP1 ) or ventilatory (VT1 ) and a
angiogenesis. second lactate (LT2 /LTP2 ) or ventilatory (VT2 ) threshold or turn
Consequently, concerning training effects (Platonov, 1999; point which has been successfully integrated into the practice
Noakes, 2000; Abbiss and Laursen, 2005), concepts integrating (Seiler and Kjerland, 2006; Seiler, 2010; Algrøy et al., 2011; Muñoz
the prescription of both intensity and duration within one model et al., 2014a,b; Tønnessen et al., 2014, 2015). Figure 1 shows an
are needed with respect to the main aims in endurance training example of the time course of selected variables and the according
which are to increase maximal oxygen uptake, the maximal turn points LTP1 /VT1 and LTP2 /VT2 for a trained cyclist.
sustainable speed, or power (threshold speed), and to increase Several variables enable to discern three distinct phases
economy and time to exhaustion (Lundby and Robach, 2015). of metabolism and cardio-respiratory responses which allow
Several authors prescribed the distribution of various intensity setting defined intensities for continuous or interval-type exercise
domains for endurance training (Esteve-Lanao et al., 2005; (Hofmann and Tschakert, 2010; Tschakert and Hofmann, 2013).
Seiler and Kjerland, 2006; Seiler, 2010; Stöggl and Sperlich, According to the lactate shuttle theory (Brooks, 1986, 2009) the
2015), but explicit prescriptions of an optimal duration for first lactate turn point (LTP1 ) is defined as the first increase in
each individual intensity domain are still missing. Pettitt (2016) blood lactate concentration (La) accompanied by a first change
recently combined exercise intensity and duration by introducing of increase in ventilation (VT1 ) and distinct changes in other
a critical velocity similar to the critical power concept (Vanhatalo ventilatory variables. The second lactate turn point (LTP2 ) is
et al., 2011; Poole et al., 2016). The CP model itself is not based on defined as the second abrupt increase in La accompanied by
physiological measures, although it seems to have a physiological a sharp increase in ventilation (VT2 ) and distinct changes in
basis which was shown to be related to the maximal lactate steady other ventilatory variables (Figures 1, 2). It has to be mentioned
state intensity (Jones et al., 2008, 2010). This concept however, that the chosen incremental test protocol influences the accuracy
does not include a differentiation of all intensity domains of any threshold determination and the validity to prescribe
(Meyer et al., 2005; Hofmann and Tschakert, 2010) which are constant load or intermittent-type exercise. A careful choice of
known to trigger specific acute physiological responses, which the protocol is accordingly substantial. A detailed discussion of
are suggested to be crucial for a successful training adaptation this problem, however, is not within the scope of this article but
(Hoppeler, 2016). Dekerle et al. (2003) as well as Pringle and discussed elsewhere (McLellan, 1985; Amann et al., 2004).
Jones (2002) showed that the critical power calculated from
a given range of exhaustion time did not correspond to the Continuous Exercise
maximal lactate steady state (mLaSS) similar to Brickley et al. The first and the second turn points are sub-maximal markers
(2002) indicating the need to combine both aspects into one from incremental exercise which can be used to prescribe defined
model recently shown by Burnley and Jones (2016). exercise workloads with distinct and defined metabolic, cardio-
It is well-prescribed that competitive endurance athletes using respiratory, and hormonal responses as shown recently by our
the polarization model train up to 13 training sessions per week working group for constant load and matched intermittent-
with an intensity distribution of about 80% of total training type exercise (Tschakert and Hofmann, 2013; Moser et al., 2015;
volume performed at low intensity and about 20% high-intensity Tschakert et al., 2015). During exercise below LTP1 , no increase
work such as interval training (Esteve-Lanao et al., 2005; Seiler in La above baseline level was detected for constant load exercise,
and Kjerland, 2006; Seiler, 2010; Stöggl and Sperlich, 2015). and it was shown recently that this intensity can be sustained for a
From this point of view, a focus on optimization of both the low very long duration of up to 24 h in trained ultra-distance athletes
intensity-high volume and the high intensity-low volume parts (Pokan et al., 2014). Increasing the workload above LTP1 leads
of the training as well as concepts and models to prescribe both to an increase in La above baseline, but after several minutes a
intensity and duration including physiologically relevant zones La steady state is built up. The mLaSS is reached at LTP2 power
are required. Aim of the paper is to give a theoretical framework output, but this intensity is clearly limited in time (Dittrich et al.,
prescribing both intensity and duration for endurance 2014) but independent of exercise mode (Fontana et al., 2009),
training. training status, and temperature (Périard et al., 2012). Although

Frontiers in Physiology | www.frontiersin.org 146 May 2017 | Volume 8 | Article 337

Hofmann and Tschakert Intensity and Duration Framework of Training

FIGURE 1 | Time course of heart rate (HR), oxygen uptake (VO2 ), and lactate (La) as well as the first (LTP1 ) and the second (LTP2 ) lactate turn point
during an incremental cycle ergometer exercise test in a well-trained cyclist (A). Time course of ventilation (VE), oxygen uptake (VO2 ), carbon dioxide output
(VCO2 ), and lactate (La) as well as the first (VT1 ) and the second (VT2 ) ventilatory turn point during incremental cycle ergometer exercise in a well-trained cyclist (B).

mLaSS intensity can be determined rather precisely, time to maximal duration, grade of fatigue and the subsequent recovery
exhaustion at the mLaSS still can vary distinctively between time which has been described recently by Burnley and Jones
athletes. Faude et al. (2017) showed a low reliability of time-to- (2016). These authors suggested distinct fatigue mechanisms for
exhaustion and blood lactate concentration at mLaSS indicating each intensity domain. A paucity of fatigue-related mechanistic
that a precise individual prescription of exercise still remains studies was shown for the moderate and high-intensity domains
challenging especially with respect to duration. but lee attention has focused on the low intensity part so far.
Workloads above LTP2 /VT2 lead to a constant increase in La As can be seen in Figure 3, an intensity slightly above LTP1
until the individual level of tolerance has been reached. This is already increases La which indicates that the critical lactate
also reflected in the response of adrenaline and noradrenaline clearance rate for the working muscle has been exceeded, and
(Moser et al., 2015). Figure 3 shows schematically the time course therefore, different hormonal and cardio-respiratory responses
of La for the three different exercise intensity zones. Percentages are suggested for this intensity level (Moser et al., 2015). In very
of maximal heart rate (%HRmax ), of maximal oxygen uptake prolonged exercise with blood lactate remaining at resting level
(%VO2max ), or % HR reserve (%HRR) as well as % oxygen throughout exercise it was prescribed that only after several
uptake reserve (VO2 R) are not able to discriminate these phases hours fatigue occurs and increases with time until the limit of
correctly on an individual basis (Hofmann et al., 2001; Meyer tolerance (Burnley and Jones, 2016). It has to be mentioned
et al., 2005; Scharhag-Rosenberger et al., 2010). that energy stores play a substantial role regarding the maximal
Usually, the first and/or the second turn points duration until the point of fatigue (Johnson et al., 2004).
(corresponding to the mLaSS) are applied to prescribe exercise
intensity limits for prolonged endurance exercise training Interval Exercise
(Esteve-Lanao et al., 2005; Muñoz et al., 2014a; Tønnessen et al., The prescription of intermittent exercise is somehow more
2014) whereas the first turn point is rarely investigated (Mann complicated compared to continuous exercise as the number of
et al., 2013). It is obvious that LTP2 /VT2 clearly discern between variables is higher. In addition to the workload intensity for the
sustainable, metabolically balanced or non-sustainable, not intervals (Ppeak ), the total duration (ttotal ) (number of intervals),
metabolically balanced workloads, whereas during exercise near the duration of the single workloads (tpeak ) as well as recovery
the first turn point a difference in acute responses can only be intensity (Prec ) and duration (trec ) and the corresponding mean
detected after a very long duration of exercise (Tremblay et al., load (Pmean ) have to be considered (Buchheit and Laursen,
2005). Mostly, these small differences in intensity slightly below 2013a,b; Tschakert and Hofmann, 2013). Similar to constant load
or above LTP1 /VT1 are not detected and recognized by athletes exercise (CLE), the mean intensity and total duration are main
although it might be important with respect to the particular markers of the overall workload, but Pmean is influenced by the

Frontiers in Physiology | www.frontiersin.org 147 May 2017 | Volume 8 | Article 337

Hofmann and Tschakert Intensity and Duration Framework of Training

FIGURE 3 | Time course of blood lactate concentration (La) below and

above the first (LTP1 ) and the second (LTP2 ) lactate turn points.
Intensities below LTP1 give a lactate steady state at resting level because the
critical lactate clearance rate within the working muscle is not exceeded (zone
I). For intensities above LTP2 , some La produced within the muscle is shuttled
to the system but La clearance is high enough to establish an equilibrium, the
so called systemic La steady state (LaSS) (zone II). Just below LTP2 the
FIGURE 2 | Time course of blood lactate concentration (La) and first maximal lactate clearance rate, the so-called maximal systemic LaSS is
(LTP1 ) and second (LTP2 ) lactate turn points. LTP1 is the first increase approached. Above LTP2 , no equilibrium can be established and La increases
above baseline. Lactate production within the muscle (PM ) is equal to the with time with early termination of exercise due to non-sustainable acidosis
elimination rate within the working muscle (EM ) and, consequently, no La has (zone III).
to be shuttled to the system (S). La within the system stays at resting level
because the critical lactate clearance of the working muscle is not exceeded.
This is the first zone which is metabolically balanced at muscle level (M) (EM =
PM ). Above LTP1 , blood La increases with increasing workloads. La we could show that even with a high-intensity running speed,
production within the muscle (PM ) exceeds the muscular rate of La elimination short 10 s intervals, 20 s passive recovery but a very low mean load
(EM ) and La has to be shuttled to the system (S). Within the system, La from
below LTP1 , lactate levels were only slightly higher than resting
the muscle (PS ) can be eliminated (by resting muscles, heart, brain,...) (ES +
EM = PM ) and La builds up an equilibrium (PS = ES ) at an elevated level, a so
level, and 30 min of exercise was clearly below the maximal
called lactate steady state on a systemic level. The critical La clearance rate of duration which was, however, not obtained in this study (Wallner
the system is not exceeded which gives a metabolically balanced situation et al., 2014).
(zone II). If the rate of La production within the muscle exceeds the maximal Similar to constant load exercise (Chidnok et al., 2012; Soares-
elimination rate of the system and the muscle (PM > EM + ES ), La increases
Caldeira et al., 2012), the problem arises how to prescribe the
exponentially without a metabolically balanced situation (zone III). The
transition from zone II to zone III is indicated by the LTP2 . optimal total duration (number of intervals) for intermittent
exercise to identify optimal adaptation effects for any specific
micro-cycle of a training period (Platonov, 1999; Lyakh et al.,
2014). As long as there is a metabolically balanced situation
aforementioned variables with respect to the degree and the kind (aerobic interval training), we may treat this problem similar
of fatigue and recovery (Burnley and Jones, 2016). Nonetheless, to constant load exercise. In case of increasing La (anaerobic
also for intermittent exercise, intensities (Ppeak , Prec , Pmean ) interval training), the optimal number of intervals may be set
are suggested to be set in relation to sub-maximal (LTP1 /VT1 , similar as it is performed in resistance-type exercise (Richens
LTP2 /VT2 ) and maximal markers (Pmax ) from an incremental and Cleather, 2014). Again there is urgent need to identify
exercise test: Ppeak = Pmax , Prec = %PLTP1 , Pmean = %PLTP2 any markers of optimal duration for both constant load and
(Tschakert and Hofmann, 2013). interval-type exercise. Burnley and Jones (2016) highlighted that
In addition, we could recently show that aerobic high- the power-duration relationship exists not only for constant-
intensity interval exercise (HIIE) with short workload durations power laboratory-based exercise, but also for variable-paced, self-
and Pmean -matched constant load exercise produced similar paced, and intermittent or stochastic exercise, which more closely
acute metabolic, hormonal and cardio-respiratory responses reflects the “real-world” athletic performance.
(Moser et al., 2015; Tschakert et al., 2015). In contrast, HIIE
with long workload durations but the same mean load yielded PRESCRIPTION OF DURATION
significantly higher acute physiological responses compared to
short HIIE and CLE (Tschakert et al., 2015). This indicated that It is a fact that any certain intensity has its own critical time
strictly planning interval-type exercise respecting all variables limit which is dependent on the type of exercise and the kind
allows the regulation and the predictability of the acute of athletes, but may be used as an individual diagnostic tool
physiological responses (Tschakert et al., 2015). In a recent paper to prescribe exercise duration (Vanhatalo et al., 2011; Pettitt,

Frontiers in Physiology | www.frontiersin.org 148 May 2017 | Volume 8 | Article 337

Hofmann and Tschakert Intensity and Duration Framework of Training

durations with selective adaptation (Viru, 1995, p. 251). This

author suggested “very heavy maximal,” “heavy sub-maximal,”
“moderate,” and “low” workloads with respect to duration, but
independent of the chosen intensity (Table 1).
Platonov (1999, p. 51) suggested that only “maximal exercise”
(75–100% of the maximal duration until clear fatigue and loss
of performance) induces distinct adaptation processes. This
maximal exercise needs long recovery of about 48 h and longer
but induces considerable performance increments (Kenttä and
Hassmén, 1998; Issurin, 2009) (Figure 5) indicated by hormonal
responses and signal-cascades yet not fully understood (Russell
et al., 2013; Hoppeler, 2016; Kirby and McCarthy, 2016).
Shortening the duration to 60–75% of the maximal duration until
clear fatigue with the same intensity only leads to a compensated
fatigue (some signs of fatigue which can be compensated without
a loss in performance) which reduces recovery duration to half
of the maximal exercise domain (about 24 h). The moderate
workload is suggested between 20 and 60% of the maximal
duration which does neither induce a compensated nor a clear
fatigue and, therefore, does not increase performance but rather
stabilizes it. Lastly, low workload defined as duration of less than
20% of maximal duration (again with the same intensity) induces
regeneration and maintains exercise performance. Figure 6
shows an example of the recovery of heart rate variability
(HRV) after maximal, sub-maximal, moderate, and low duration
exercise with the same intensity applied (unpublished results).
HRVwas shown to be sensitive for intensity and duration of
exercise (Kaikkonen et al., 2010; Myllymäki et al., 2012).
From this concept (Platonov, 1999, p. 51; Viru, 1995, p.
251), we assume that for any single intensity within the three
different intensity domains (Zone I: below LTP1 /VT1 ; Zone II:
between LTP1 /VT1 and LTP2 /VT2 ; Zone III: above LTP2 /VT2 ),
FIGURE 4 | Speed-duration relationship of endurance-type world we may chose four different duration domains with clearly
records in running from 800 m to marathon race distance (A) and distinct adaptation effects on various physiological processes and
swimming (B) was shown to be linear in a logarithmic scale for male and
female athletes.
exercise performance. To prescribe these duration domains one
needs the maximal duration for at least 2–3 different intensities to
draw the power-duration relationship as can be seen in Figure 7
which shows the same athlete as Figure 1.
2016; Poole et al., 2016). Figure 4A shows the running speed
for all endurance-type world records in continuous running
from 800 m to marathon distance which has been described as Model to Combine the Prescription of Both
a most perfect logarithm relationship for both men and women Intensity and Duration
(Nikolaidis et al., 2017). A similar relationship can be shown Earlier approaches such as the model from Garcin and Billat
for free style swimming (Figure 4B). The relationship between (2001) applied a perceived exertion scale to attest both intensity
speed and distance is linear applying a logarithmic x-axis within and duration, but the optimal duration may not be obtained
this wide range of race distances. It is obvious that no speed- from this model. As can be seen in Figures 7, 8, the power
distance pairs above the linear line are possible (Burnley and output (or speed) to time (or distance) relationship allows
Jones, 2016). This speed-distance or power-duration relationship discerning these specific durations for any intensity of interest
can therefore be applied to detect the maximal speed or power by using four different duration domains according to Platonov
output for any distance or duration but, no less importantly, (in (Viru, 1995), p. 251). As shown in Table 1, zone 1 is
to detect any maximal duration or distance for an arbitrarily defined as low, zone 2 as moderate, zone 3 as sub-maximal,
chosen speed or power on an individual basis. Additionally, and zone 4 as maximal workload each producing different states
independent of the chosen intensity, this concept allows setting of fatigue and, consequently, different effects of adaptation,
a targeted duration (% of maximal duration) for endurance which is in line with recent data from Tremblay et al. (2005).
exercise training with respect to improvement, maintenance or These authors showed a duration threshold for various hormone
recovery purposes. To regulate these distances we apply the responses for a comparable low intensity of 50–55% VO2max
concept of Platonov (1999) who differentiated four domains of whereas a longer duration induced a favorable hormone profile

Frontiers in Physiology | www.frontiersin.org 149 May 2017 | Volume 8 | Article 337

Hofmann and Tschakert Intensity and Duration Framework of Training

TABLE 1 | Definition of specific duration domains for endurance-type exercise (modified from Platonov, 1999).

Workload Phase Duration Targets

Low 1st phase of stable performance 15–20% of maximal duration until clear fatigue Maintaining exercise performance and accelerated recovery
Moderate 2nd phase of stable performance 20–60% of maximal duration until clear fatigue Maintaining exercise performance
Sub-maximal Phase of compensated fatigue 60–75% of maximal duration until clear fatigue Stabilization and moderate increases in performance
Maximal Phase of clear fatigue 75–100% of maximal duration until clear fatigue Distinct increases in performance

FIGURE 5 | Relationship between duration of exercise and fatigue (modified from Platonov, 1999). There is a certain duration for each intensity leading to
distinct grade of fatigue which terminates exercise (4), prolongs recovery times and increases “super-compensation” with increased performance. Reducing the
duration to less than 75% of the maximal duration does not induce distinct but compensated fatigue without a loss in performance (3) and, consequently, a much
shorter recovery time and less if any “super-compensation.” Reducing duration to less than 60% of the maximal duration (2) does not induce any fatigue and,
therefore, does not increase performance, but rather stabilizes the given performance level. A duration less than 20% of the maximal duration is just a functional
stimulation suggested adequate for regeneration.

which was suggested to support the mobilization of fuels for both thresholds. It is, however, important to note that some
recovery and restoration of glycogen stores. In this combined authors also use fixed reference values for lactate such as 2 and
exercise prescription model, the setting of work intensities 4 mmol.l−1 (Seiler and Kjerland, 2006; Guellich et al., 2009;
should also be individualized and physiologically based by Orie et al., 2014) which may overestimate the volume especially
using turn point intensities as discerning markers for distinctly for the low intensity volumes. Our own results showed that
different metabolic, hormonal and cardio-respiratory responses La at LTP1 was found at 1.2–1.6 mmol.l−1 (Hofmann et al.,
(Tschakert and Hofmann, 2013; Moser et al., 2015). 1997, 2001). An individual and accurate intensity prescription
This concept (Table 1 and Figure 8) enables athletes and is crucial even at low power outputs near LTP1 since allowing
coaches to fine-tune training volume and/or intensity to further intensities just 10% above LTP1 definitely shortens the time to
optimize training processes which is of particular relevance when clear fatigue by ∼40% (Figure 8). As a consequence, high-volume
the limits of tolerance are reached. In addition, it allows a training set above LTP1 may get too close to a fatigue state
retrospective analysis of distances covered with given intensities that avoids repeating high volumes on a regular daily basis. In
in the past. addition, for low intensity exercise training, it is usually NOT
intended to reach maximal duration (tmax ) but to apply a certain
DISCUSSION AND CONCLUSIONS percentage of tmax (Table 1) in order to avoid fatigue and to
guarantee the ability to repeat high volumes of training on a
The concept to combine turn point derived intensities and daily basis. However, we like to point out that dependent on the
optimized durations may be specifically interesting with respect aim of a specific training period specific types of micro-cycles
to the new polarized training concept (Seiler and Kjerland, 2006; need to be structured combining exercise type, intensity, and
Seiler, 2010; Muñoz et al., 2014a,b; Tønnessen et al., 2014) duration.
where 80–90% of training volume is set below LTP1 /VT1 and Beside the attractiveness of the concept, several open
up to 22% above LTP2 /VT2 with very low volumes between questions and limits have to be addressed. Firstly, the chosen

Frontiers in Physiology | www.frontiersin.org 150 May 2017 | Volume 8 | Article 337

Hofmann and Tschakert Intensity and Duration Framework of Training

FIGURE 8 | Maximal and optimal (% of maximal) duration for exercise

FIGURE 6 | Relative changes of heart rate variability markers (%HRV) intensities at LTP1 and LTP2 in a well-trained cyclist applying the
during constant load exercise just below the second lactate turn point modified concept of Platonov (1999). For any specific metabolic, hormonal
with different duration of the maximal sustainable distance (100%). It or cardio-respiratory target intensity (<LTP1 /VT1 ; between LTP1 /VT1 and
can be seen that recovery of the HRV is dependent on % maximal duration LTP2 /VT2; > LTP2 /VT2 ), the optimal duration with respect to the four workload
with early recovery at low to moderate distances (20, 40%) according to domains “low” (1), “moderate” (2), “sub-maximal” (3), and “maximal” (4) may
Platonov (1999). be derived from this graph of the maximal intensity-duration relationship.

training status and sex, and individual physiological markers are

needed to guide the training. As could be shown in a pilot test
(Figure 6), heart rate variability might be a potential parameter
to identify these cut-off points for duration as discussed recently
(Kaikkonen et al., 2010; Saboul et al., 2016). Additionally,
ratings of perceived exertion (RPE) scales may be helpful to
identify these reference markers on a daily individual basis
(Garcin and Billat, 2001; Seiler and Sjursen, 2004; Coquart et al.,
2012).
A second important limit is the method to derive the
power-duration or velocity—distance relationship. To obtain
a valid regression line, data points must be obtained from
highly motivated athletes from competitions. However,
maximal performance changes during the training year due
FIGURE 7 | Example of exercise intensity-duration relationship in a to periodization will make it a bit more difficult to obtain
single well-trained cyclist. As described in Figure 1, for any individual optimal distances throughout the year. Additionally, not
athlete, a plot of maximal speed and distance relationships can be drawn
describing the maximal distance for any specific velocity or power output or
all sports allow obtaining these markers under comparable
the maximal velocity or power output for any chosen distance. According to and possibly standardized conditions such as in cycling
the modified concept from Platonov (1999), these individual draws allow to or on-snow cross-country skiing. Semi-specific tests such
prescribe the optimal distance (% of maximal duration) for selected intensities as ergometer or ski-roller tests may help to overcome this
for various endurance-type training situations. To prescribe both intensity and
problem.
duration on an individual basis, this concept needs to be combined to the
usual sub-maximal performance diagnostic markers such as LTP1 / VT1 and
A third limit may be the idea to maximize volumes by
LTP2 /VT2 . reducing intensity below LTP1 /VT1 . Although athletes may
withstand such volumes from an acute metabolic state of
view, some long-term problems such as orthopedic complaints
(arthrosis, stress fractures) as well as disturbances in energy or
percentages of maximal duration are just marginally evidenced. fluid supply may arise from such as concept (Noakes, 2000;
To discern the low (regenerative) zone from a moderate zone Cymet and Sinkov, 2006; Krampla et al., 2008; Weber, 2009;
without fatigue, a zone with compensated fatigue and finally zone Warden et al., 2014).
4 with clear fatigue needs to be taken with caution. Carefully Despite those limits, this concept gives a solid theoretical
conducted studies and retrospective analysis of distances covered framework that allows optimizing both intensity and duration of
at defined intensities are needed to identify the stability or the whole spectrum of endurance training load for the first time.
variability of these percentages for athletes with different age, It may help to improve exercise training for top level performance

Frontiers in Physiology | www.frontiersin.org 151 May 2017 | Volume 8 | Article 337

Hofmann and Tschakert Intensity and Duration Framework of Training

even though it is already close to the limits of tolerance for the recommendations of Declaration of Helsinki. The subject gave
human body. written informed consent in accordance with the Declaration of
Helsinki.
ETHICS STATEMENT
AUTHOR CONTRIBUTIONS
This methodological consideration included a single pilot tests
which were not part of a formal study but a proof of PH: Author wrote the manuscript, draw the figures, and
principle determination of markers from standard performance graphs. GT: Author contributed equally in writing the
diagnostic tests which was performed in accordance with the manuscript.

REFERENCES in moderately trained subjects. Eur. J. Appl. Physiol. 107, 187–192.

doi: 10.1007/s00421-009-1111-9
Abbiss, C. R., and Laursen, P. B. (2005). Models to explain fatigue Garcin, M., and Billat, V. (2001). Perceived exertion scales attest to
during prolonged endurance cycling. Sports Med. 35, 865–898. both intensity and exercise duration. Percept. Mot. Skills 93, 661–671.
doi: 10.2165/00007256-200535100-00004 doi: 10.2466/pms.2001.93.3.661
Algrøy, E. A., Hetlelid, K. J., Seiler, S., and Stray Pedersen, J. I. (2011). Quantifying Guellich, A., Seiler, S., and Emrich, E. (2009). Training methods and intensity
training intensity distribution in a group of Norwegian professional soccer distribution of young world-class rowers. Int. J. Sports. Physiol. Perform. 4,
players. Int. J. Sports Physiol. Perform. 6, 70–81. doi: 10.1123/ijspp.6.1.70 448–460. doi: 10.1123/ijspp.4.4.448
Amann, M., Subudhi, A., and Foster, C. (2004). Influence of testing protocol on Hofmann, P., Pokan, R., von Duvillard, S. P., Seibert, F. J., Zweiker, R., and
ventilatory thresholds and cycling performance. Med. Sci. Sports Exerc. 36, Schmid, P. (1997). Heart rate performance curve during incremental cycle
613–622. doi: 10.1249/01.MSS.0000122076.21804.10 ergometer exercise in healthy young male subjects. Med. Sci. Sports Exerc. 29,
Brickley, G., Doust, J., and Williams, C. A. (2002). Physiological responses during 762–768.
exercise to exhaustion at critical power. Eur. J. Appl. Physiol. 88, 146–151. Hofmann, P., and Tschakert, G. (2010). Special needs to prescribe
doi: 10.1007/s00421-002-0706-1 exercise intensity for scientific studies. Cardiol. Res. Pract. 2011:209302.
Brooks, G. A. (1986). The lactate shuttle during exercise and recovery. Med. Sci. doi: 10.4061/2011/209302
Sports Exerc. 18, 360–368. doi: 10.1249/00005768-198606000-00019 Hofmann, P., Von Duvillard, S. P., Seibert, F. J., Pokan, R., Wonisch, M.,
Brooks, G. A. (2009). Cell-cell and intracellular lactate shuttles. J. Physiol. 587(Pt Lemura, L. M., et al. (2001). %HRmax target heart rate is dependent on
23), 5591–5600. doi: 10.1113/jphysiol.2009.178350 heart rate performance curve deflection. Med. Sci. Sports Exerc. 33, 1726–1731.
Buchheit, M., and Laursen, P. B. (2013a). High-intensity interval training, doi: 10.1097/00005768-200110000-00017
solutions to the programming puzzle: part I: cardiopulmonary emphasis. Sports Hoppeler, H. (2016). Molecular networks in skeletal muscle plasticity. J. Exp. Biol.
Med. 43, 313–338. doi: 10.1007/s40279-013-0029-x 219 (Pt 2), 205–213. doi: 10.1242/jeb.128207
Buchheit, M., and Laursen, P. B. (2013b). High-intensity interval training, Issurin, V. B. (2009). Generalized training effects induced by athletic preparation.
solutions to the programming puzzle. Part II: anaerobic energy, A review. J. Sports Med. Phys. Fitness 49, 333–345.
neuromuscular load and practical applications. Sports Med. 43, 927–954. Johnson, N. A., Stannard, S. R., and Thompson, M. W. (2004). Muscle triglyceride
doi: 10.1007/s40279-013-0066-5 and glycogen in endurance exercise: implications for performance. Sports Med.
Burnley, M., and Jones, A. M. (2016). Power-duration relationship: physiology, 34, 151–164. doi: 10.2165/00007256-200434030-00002
fatigue, and the limits of human performance. Eur. J. Sport Sci. 3, 1–12. Jones, A. M., Vanhatalo, A., Burnley, M., Morton, R. H., and Poole,
doi: 10.1080/17461391.2016.1249524 D. C. (2010). Critical power: implications for determination of
Chidnok, W., Dimenna, F. J., Bailey, S. J., Vanhatalo, A., Morton, R. H., VO2max and exercise tolerance. Med. Sci. Sports Exerc. 42, 1876–1890.
Wilkerson, D. P., et al. (2012). Exercise tolerance in intermittent cycling: doi: 10.1249/MSS.0b013e3181d9cf7f
application of the critical power concept. Med. Sci. Sports Exerc. 44, 966–976. Jones, A. M., Wilkerson, D. P., DiMenna, F., Fulford, J., and Poole, D. C. (2008).
doi: 10.1249/MSS.0b013e31823ea28a Muscle metabolic responses to exercise above and below the “critical power”
Coquart, J. B., Eston, R. G., Noakes, T. D., Tourny-Chollet, C., L’hermette, M., assessed using 31P-MRS. Am. J. Physiol. Regul. Integr. Comp. Physiol. 294,
Lemaître, F., et al. (2012). Estimated time limit: a brief review of a perceptually R585–R593. doi: 10.1152/ajpregu.00731.2007
based scale. Sports Med. 42, 845–855. doi: 10.2165/11635370-000000000-00000 Kaikkonen, P., Hynynen, E., Mann, T., Rusko, H., and Nummela, A. (2010). Can
Cymet, T. C., and Sinkov, V. (2006). Does long-distance running cause HRV be used to evaluate training load in constant load exercises? Eur. J. Appl.
osteoarthritis? J. Am. Osteopath. Assoc. 106, 342–345. Physiol. 108, 435–442. doi: 10.1007/s00421-009-1240-1
Dekerle, J., Baron, B., Dupont, L., Vanvelcenaher, J., and Pelayo, P. (2003). Maximal Kenttä, G., and Hassmén, P. (1998). Overtraining and recovery. A conceptual
lactate steady state, respiratory compensation threshold and critical power. Eur. model. Sports Med. 26, 1–16. doi: 10.2165/00007256-199826010-00001
J Appl. Physiol. 89, 281–288. doi: 10.1007/s00421-002-0786-y Kirby, T. J., and McCarthy, J. J. (2016). MicroRNAs in skeletal muscle
Dittrich, N., de Lucas, R. D., Beneke, R., and Guglielmo, L. G. (2014). biology and exercise adaptation. Free Radic. Biol. Med. 64, 95–105.
Time to exhaustion at continuous and intermittent maximal lactate steady doi: 10.1016/j.freeradbiomed.2013.07.004
state during running exercise. Int. J. Sports Physiol. Perform. 9, 772–776. Krampla, W. W., Newrkla, S. P., Kroener, A. H., and Hruby, W. F.
doi: 10.1123/ijspp.2013-0403 (2008). Changes on magnetic resonance tomography in the knee joints of
Esteve-Lanao, J., San Juan, A. F., Earnest, C. P., Foster, C., and Lucia, marathon runners: a 10-year longitudinal study. Skeletal. Radiol. 37, 619–626.
A. (2005). How do endurance runners actually train? Relationship doi: 10.1007/s00256-008-0485-9
with competition performance. Med. Sci. Sports Exerc. 37, 496–504. Lundby, C., and Robach, P. (2015). Performance Enhancement: what are the
doi: 10.1249/01.MSS.0000155393.78744.86 physiological limits? Physiology 30, 282–292. doi: 10.1152/physiol.00052.2014
Faude, O., Hecksteden, A., Hammes, D., Schumacher, F., Besenius, E., Lyakh, V., Mikołajec, K., Bujas, P., and Litkowycz, R. (2014). Review of
Sperlich, B., et al. (2017). Reliability of time-to-exhaustion and selected platonov’s “sports training periodization. general theory and its practical
psycho-physiological variables during constant-load cycling at the application” - kiev: olympic literature, 2013. J. Hum. Kinet. 30, 259–263.
maximal lactate steady-state. Appl. Physiol. Nutr. Metab. 42, 142–147. doi: 10.2478/hukin-2014-0131
doi: 10.1139/apnm-2016-0375 Mann, T., Lamberts, R. P., and Lambert, M. I. (2013). Methods of prescribing
Fontana, P., Boutellier, U., and Knöpfli-Lenzin, C. (2009). Time to exhaustion relative exercise intensity: physiological and practical considerations. Sports
at maximal lactate steady state is similar for cycling and running Med. 43, 613–625. doi: 10.1007/s40279-013-0045-x

Frontiers in Physiology | www.frontiersin.org 152 May 2017 | Volume 8 | Article 337

Hofmann and Tschakert Intensity and Duration Framework of Training

McLellan, T. M. (1985). Ventilatory and plasma lactate response with different metabolic responses between individuals. J. Sci. Med. Sports 13, 74–79.
exercise protocols: a comparison of methods. Int. J. Sports Med. 6, 30–35. doi: 10.1016/j.jsams.2008.12.626
doi: 10.1055/s-2008-1025809 Seiler, K. S., and Kjerland, G. Ø. (2006). Quantifying training intensity distribution
Meyer, T., Lucía, A., Earnest, C. P., and Kindermann, W. (2005). A conceptual in elite endurance athletes: is there evidence for an “optimal” distribution?
framework for performance diagnosis and training prescription from Scand. J. Med. Sci. Sports 16, 49–56. doi: 10.1111/j.1600-0838.2004.
submaximal gas exchange parameters–theory and application. Int. J. Sports 00418.x
Med. 26(Suppl. 1), S38–S48. doi: 10.1055/s-2004-830514 Seiler, S. (2010). What is best practice for training intensity and duration
Moser, O., Tschakert, G., Mueller, A., Groeschl, W., Pieber, T. R., Obermayer- distribution in endurance athletes? Int. J. Sports Physiol. Perform. 5, 276–291.
Pietsch, B., et al. (2015). Effects of high-intensity interval exercise versus doi: 10.1123/ijspp.5.3.276
moderate continuous exercise on glucose homeostasis and hormone response Seiler, S., and Sjursen, J. E. (2004). Effect of work duration on physiological
in patients with type 1 diabetes mellitus using novel ultra-long-acting insulin. and rating scale of perceived exertion responses during self-paced interval
PLoS ONE 10:e0136489. doi: 10.1371/journal.pone.0136489 training. Scand. J. Med. Sci. Sports 14, 318–325. doi: 10.1046/j.1600-0838.2003.
Muñoz, I., Cejuela, R., Seiler, S., Larumbe, E., and Esteve-Lanao, J. (2014a). 00353.x
Training-intensity distribution during an ironman season: relationship with Skovgaard, C., Brandt, N., Pilegaard, H., and Bangsbo, J. (2016). Combined speed
competition performance. Int. J. Sports Physiol. Perform. 9, 332–339. endurance and endurance exercise amplify the exercise-induced PGC-1α and
doi: 10.1123/ijspp.2012-0352. PDK4 mRNA response in trained human muscle. Physiol. Rep. 4:e12864.
Muñoz, I., Seiler, S., Bautista, J., España, J., Larumbe, E., and Esteve-Lanao, J. doi: 10.14814/phy2.12864
(2014b). Does polarized training improve performance in recreational runners? Soares-Caldeira, L. F., Okuno, N. M., Magalhães Sales, M., Campbell, C. S.,
Int. J. Sports Physiol. Perform. 9, 265–272. doi: 10.1123/ijspp.2012-0350 Simões, H. G., and Nakamura, F. Y. (2012). Similarity in physiological
Myllymäki, T., Rusko, H., Syväoja, H., Juuti, T., Kinnunen, M. L., and and perceived exertion responses to exercise at continuous and intermittent
Kyröläinen, H. (2012). Effects of exercise intensity and duration on nocturnal critical power. Eur. J. Appl. Physiol. 112, 1637–1644. doi: 10.1007/s00421-011-
heart rate variability and sleep quality. Eur J Appl Physiol. 112, 801–809. 2123-9
doi: 10.1007/s00421-011-2034-9 Stöggl, T. L., and Sperlich, B. (2015). The training intensity distribution
Nikolaidis, P. T., Zingg, M. A., and Knechtle, B. (2017). Performance trends in among well-trained and elite endurance athletes. Front. Physiol. 6:295.
age-group runners from 100 m to marathon-The World Championships from doi: 10.3389/fphys.2015.00295
1975 to 2015. Scand. J. Med. Sci. Sports doi: 10.1111/sms.12821. [Epub ahead of Tønnessen, E., Svendsen, I. S., Rønnestad, B. R., Hisdal, J., Haugen, T.
print]. A., and Seiler, S. (2015). The annual training periodization of 8 world
Noakes, T. D. (2000). Physiological models to understand exercise fatigue and the champions in orienteering. Int. J. Sports Physiol. Perform. 10, 29–38.
adaptations that predict or enhance athletic performance. Scand. J. Med. Sci. doi: 10.1123/ijspp.2014-0005
Sports. 10, 123–145. doi: 10.1034/j.1600-0838.2000.010003123.x Tønnessen, E., Sylta, Ø., Haugen, T. A., Hem, E., Svendsen, I. S., and Seiler,
Orie, J., Hofman, N., de Koning, J. J., and Foster, C. (2014). Thirty-eight years of S. (2014). The road to gold: training and peaking characteristics in the
training distribution in Olympic speed skaters. Int. J. Sports Physiol. Perform. 9, year prior to a gold medal endurance performance. PLoS ONE 9:e101796.
93–99. doi: 10.1123/ijspp.2013-0427 doi: 10.1371/journal.pone.0101796
Périard, J. D., Caillaud, C., and Thompson, M. W. (2012). The role Tremblay, M. S., Copeland, J. L., and Van Helder, W. (2005). Influence of exercise
of aerobic fitness and exercise intensity on endurance performance in duration on post-exercise steroid hormone responses in trained males. Eur. J.
uncompensable heat stress conditions. Eur. J. Appl. Physiol. 112, 1989–1999. Appl. Physiol. 94, 505–513. doi: 10.1007/s00421-005-1380-x
doi: 10.1007/s00421-011-2165-z Tschakert, G., and Hofmann, P. (2013). High-intensity intermittent exercise:
Pescatello, L. S. (ed.) (2014). ACSM’s Guidelines for Exercise Testing and methodological and physiological aspects. Int. J. Sports Physiol. Perform. 8,
Prescription, 9th Edn. Baltimore, MD: Wolters Kluwer—Lippincott Williams 600–610. doi: 10.1123/ijspp.8.6.600
& Wilkins. Tschakert, G., Kroepfl, J., Mueller, A., Moser, O., Groeschl, W., and Hofmann,
Pettitt, R. W. (2016). Applying the critical speed concept to racing strategy and P. (2015). How to regulate the acute physiological response to “aerobic”
interval training prescription. Int. J. Sports Physiol. Perform. 11, 842–847. high-intensity interval exercise. J. Sports Sci. Med. 14, 29–36.
doi: 10.1123/ijspp.2016-0001 Vanhatalo, A., Jones, A. M., and Burnley, M. (2011). Application of critical power
Platonov, V. N. (1999). Workload – Fatigue – Performance: The modern in sport. Int. J. Sports Physiol. Perform. 6, 128–136. doi: 10.1123/ijspp.6.1.128
framework of exercise training. (Belastung - Ermüdung - Leistung: Der moderne Viru, A. (1995). Adaptation in Sports Training. Boca Raton, FL: CRC Press.
Trainingsaufbau). Münster: Phillippka Sportverlag (german). Viru, A., Smirnova, T., Karelson, K., Snegovskaya, V., and Viru, M. (1996).
Pokan, R., Ocenasek, H., Hochgatterer, R., Miehl, M., Vonbank, K., Von Determinants and modulators of hormonal responses to exercise. Biol. Sports
Duvillard, S. P., et al. (2014). Myocardial dimensions and hemodynamics 13, 169–187.
during 24-h ultraendurance ergometry. Med. Sci. Sports Exerc. 46, 268–275. Wallner, D., Simi, H., Tschakert, G., and Hofmann, P. (2014). Acute physiological
doi: 10.1249/MSS.0b013e3182a64639 response to aerobic short-interval training in trained runners. Int. J. Sports
Poole, D. C., Burnley, M., Vanhatalo, A., Rossiter, H. B., and Jones, A. M. (2016). Physiol. Perform. 9, 661–666. doi: 10.1123/ijspp.2013-0385
Critical power: an important fatigue threshold in exercise physiology. Med. Sci. Warden, S. J., Davis, I. S., and Fredericson, M. (2014). Management and prevention
Sports Exerc. 48, 2320–2334. doi: 10.1249/MSS.0000000000000939 of bone stress injuries in long-distance runners. J. Orthop. Sports Phys. Ther. 44,
Pringle, J. S., and Jones, A. M. (2002). Maximal lactate steady state, critical 749–765. doi: 10.2519/jospt.2014.5334
power and EMG during cycling. Eur. J. Appl. Physiol. 88, 214–226. Weber, J. M. (2009). The physiology of long-distance migration: extending
doi: 10.1007/s00421-002-0703-4 the limits of endurance metabolism. J. Exp. Biol. 212 (Pt 5), 593–597.
Richens, B., and Cleather, D. J. (2014). The relationship between the number of doi: 10.1242/jeb.015024
repetitions performed at given intensities is different in endurance and strength
trained athletes. Biol. Sport. 31, 157–161. doi: 10.5604/20831862.1099047 Conflict of Interest Statement: The authors declare that the research was
Russell, A. P., Lamon, S., Boon, H., Wada, S., Güller, I., Brown, E. L., et al. (2013). conducted in the absence of any commercial or financial relationships that could
Regulation of miRNAs in human skeletal muscle following acute endurance be construed as a potential conflict of interest.
exercise and short-term endurance training. J. Physiol. 591, 4637–4653.
doi: 10.1113/jphysiol.2013.255695 Copyright © 2017 Hofmann and Tschakert. This is an open-access article distributed
Saboul, D., Balducci, P., Millet, G., Pialoux, V., and Hautier, C. (2016). A pilot study under the terms of the Creative Commons Attribution License (CC BY). The use,
on quantification of training load: the use of HRV in training practice. Eur. J. distribution or reproduction in other forums is permitted, provided the original
Sport Sci. 16, 172–181. doi: 10.1080/17461391.2015.1004373 author(s) or licensor are credited and that the original publication in this journal
Scharhag-Rosenberger, F., Meyer, T., Gässler, N., Faude, O., and Kindermann, is cited, in accordance with accepted academic practice. No use, distribution or
W. (2010). Exercise at given percentages of VO2max: heterogeneous reproduction is permitted which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 153 May 2017 | Volume 8 | Article 337

 ORIGINAL RESEARCH
published: 27 December 2016
doi: 10.3389/fphys.2016.00642

Training Intensity Distribution and

Changes in Performance and
Physiology of a 2nd Place Finisher
Team of the Race across America
Over a 6 Month Preparation Period
Christian Manunzio 1 , Joachim Mester 1, 2 , Walter Kaiser 3 and Patrick Wahl 1, 2, 4*
1
Institute of Training Science and Sport Informatics, German Sport University, Cologne, Germany, 2 The German Research
Centre of Elite Sport, German Sport University Cologne, Cologne, Germany, 3 Private Practice for Oral and Maxillofacial
Surgery, Neuss, Germany, 4 Department of Molecular and Cellular Sport Medicine, Institute of Cardiovascular Research and
Sport Medicine, German Sport University Cologne, Cologne, Germany

Aim: To monitor the training intensity distribution (TID) and the development of
physiological and performance parameters.
Edited by:
Thomas Leonhard Stöggl, Methods: During their preparation period for the RAAM, 4 athletes (plus 1 additional
University of Salzburg, Austria
backup racer) performed 3 testing sessions; one before, one after 3, and one after 6
Reviewed by:
months of training. VO2max , maximal rate of lactate accumulation (dLa/dtmax ), critical
Gianluca Vernillo,
University of Calgary, Canada power, power output at lactate minimum (MLSSP ), peak and mean power output during a
Beat Knechtle, sprint test, heart rate recovery, isometric strength, jumping height, and body composition
University of Zurich, Switzerland
Øyvind Sandbakk,
were determined. All training sessions were recorded with a power meter. The endurance
Norwegian University of Science and TID was analyzed based on the time in zone approach, according to a classical 3-zone
Technology, Norway
model, including all power data of training sessions, and a power specific 3-zone model,
*Correspondence:
where time with power output below 50% of MLSSP was not considered.
Patrick Wahl
[email protected] Results: The TID using the classical 3-zone model reflected a pyramidal TID (zone 1: 63
± 16, zone 2: 28 ± 13 and zone 3: 9 ± 4%). The power specific 3-zone model resulted
Specialty section:
This article was submitted to in a threshold-based TID (zone 1: 48 ± 13, zone 2: 39 ± 10, zone 3: 13 ± 4%). VO2max
Exercise Physiology, increased by 7.1 ± 5.3% (P = 0.06). dLa/dtmax decreased by 16.3 ± 8.1% (P = 0.03).
a section of the journal
Power output at lactate minimum and critical power increased by 10.3 ± 4.1 and 16.8
Frontiers in Physiology
± 6.2% (P = 0.01), respectively. No changes were found for strength parameters and
Received: 24 May 2016
Accepted: 07 December 2016 jumps.
Published: 27 December 2016
Conclusion: The present study underlines that a threshold oriented TID results in only
Citation:
Manunzio C, Mester J, Kaiser W and
moderate increases in physiological parameters. The amount of training below 50%
Wahl P (2016) Training Intensity of MLSSp (∼28% of total training time) is remarkably high. Researchers, trainers, and
Distribution and Changes in
athletes should pay attention to the different ways of interpreting training power data, to
Performance and Physiology of a 2nd
Place Finisher Team of the Race gain realistic insights into the TID and the corresponding improvements in performance
across America Over a 6 Month and physiological parameters.
Preparation Period.
Front. Physiol. 7:642. Keywords: critical power, VO2max , maximal rate of lactate accumulation, MLSS, lactate minimum intensity,
doi: 10.3389/fphys.2016.00642 ultra-endurance performance

Frontiers in Physiology | www.frontiersin.org 154 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

INTRODUCTION high-volume low-intensity training (HVLIT), (2) training at or

near the lactate threshold (THR), (3) low-volume high-intensity
The Race Across America (RAAM) is an annual 4800 km non- interval training (HIT), (4) “pyramidal” (PYR) TID consisting
stop cycling race from the west coast of America to the east coast. of high volume of HVLIT, medium volume of THR, and small
The physiological and environmental challenges encountered by volume of HIT, (5) as well as the combination of HVLIT and HIT,
the athletes participating in this event are numerous: insufficient named “polarized” training (POL).
energy intake (Knechtle et al., 2005; Hulton et al., 2010), sleep There are a number of prospective and intervention studies
deprivation (Hulton et al., 2010; Lahart et al., 2013) and tough in cycling (Neal et al., 2013), running (Esteve-Lanao et al., 2007;
climate conditions like extreme heat in the desert or very Muñoz et al., 2014), different endurance disciplines (Stöggl and
high humidity resulting in a decline in performance due to Sperlich, 2014), as well as retrospective studies in speed skating
dehydration (Bowen et al., 2006; Paulin et al., 2015). Additionally, (Yu et al., 2012), running (Billat et al., 2001), and cross-country
over 30,000 m of altitude difference and the corresponding time skiing (Seiler and Kjerland, 2006), reporting the superiority of a
under hypoxic conditions make the event one of the toughest bike POL TID compared to other TIDs.
races in the world and the unofficial world championship of ultra- However, there are also many retrospective studies, generally
endurance cycling. The race can be performed as a team event, reporting a PYR TID for endurance athletes in running (Esteve-
with 2, 4, or 8 racers as a relay team or solo. Lanao et al., 2005), cycling (Luciá et al., 2000; Schumacher
The performance requirements for a team relay in the RAAM and Mueller, 2002; Zapico et al., 2007), and triathlon (Neal
can be described as an enormous number of individual time et al., 2011). In a recent review, it was summarized that most
trials, interspersed by longer and shorter phases of recovery, retrospective studies on well-trained athletes report on PYR
depending on the team tactics. The physiological load can, TID, although some world-class athletes performed POL TID
therefore, be compared to the bike split in triathlon events, where during certain phases of the season (Stöggl and Sperlich, 2015).
performance (after preload by the swim split) is ideally as high Sandbakk et al. (2016), Tønnessen et al. (2014), and Guellich et al.
as possible without impeding the following run split. Under ideal (2009) reported a general focus on HVLIT TID in the preparation
conditions, each team member in a 4 person relay team would period, which became more polarized in the competition period
have to cover 1200 km on the bike as fast as possible. of world/national class cross-country skiers, biathletes, and
Concerning the physiological prerequisites, Laursen et al. young world-class rowers.
(1999) described a RAAM relay team with 4 racers, finishing Most studies reporting TID used the heart rate (HR) or the
in fourth place with a mean maximal oxygen uptake (VO2max ) rate of perceived exertion (RPE) to determine training time in
of 71.5 mL·min−1 ·kg−1 ,while a solo racer winning the RAAM different intensity zones. To the best of our knowledge, there
was described by Ice et al. (1988) with a VO2max of 79.6 are no studies using power output for the determination of TID.
mL·min−1 ·kg−1 . These values correspond roughly to the The ability to accurately quantify the mechanical work during
range of VO2max values for professional cyclists from 69.7 training makes cycling unique in allowing such insights into the
mL·min−1 ·kg−1 to 84.8 mL·min−1 ·kg−1 summarized by Padilla demands of sporting preparation. Power output as a marker of
et al. (1999). However, Faria et al. (2005) showed in their the external load is zero in phases like downhill sections or during
review of cycling physiology, that VO2max alone is neither a coasting, but the internal load represented by the HR might still
good predictor of endurance performance nor a valid distinction appear higher (Halson, 2014), resulting in a “smoothing effect.”
between elite and amateur cyclists. Lactate threshold (LT) might Especially very short high-intensity accelerations (e.g., initial
be more suitable as a predictor of endurance performance since pedal strokes or short intervals), as well as abrupt reduction of
it was the single variable that correlated best with time trial external load, will not be reflected by the HR. Additionally, Seiler
performance (r = 0.86, p = 0.01) in the study of Støren et al. and Kjerland (2006) demonstrated very big differences in TID
(2013). This might be associated with the findings of Jacobs between the total time-in-zone approach and the session-goal
et al. (2011), who demonstrated that endurance performance method. Many factors may influence the relationship between
among highly trained cyclists correlated with skeletal muscle workload and HR (Buchheit, 2014), and day-to-day variation in
oxidative capacity, accounting for 47% of the variation in 26 km HR was shown to be approximately 6.5%, which might influence
time trial performance. However, no previous study did report the analysis of HR data and consequently the time spent in each
comprehensively on the development of aerobic [e.g., VO2max , zone (Bagger et al., 2003). However, Davies and Knibbs (1971)
maximal lactate steady state (MLSS)] and anaerobic parameters found, that an intensity of at least 50% of VO2max is necessary
[e.g., maximal rate of lactate accumulation (dLa/dtmax ), maximal to result in performance enhancements. This raises the question
voluntary contraction (MVC)] in combination with performance how TID (time in each zone in a 3 zone-model) based on power
measures [peak power output (PPO), sprint PPO, sprint mean data might change: 1. if all power data of training are included in
power output (MPO), critical power (CP)] over the course of a the analysis and are assigned to one of the 3 zones, or 2. if power
preparation period for an ultra-endurance cycling race. data below a certain threshold are excluded from the analysis.
On the topic of training to optimize performance for such We accompanied 6 months of preparation time of a 4 person
extreme endurance challenges, different information can be RAAM team (plus 1 additional backup racer). They later finished
found in scientific literature. Stöggl and Sperlich (2015) proposed the actual race in second place. All athletes were amateurs, and
different concepts of training intensity and volume distribution their training time was restricted by social and environmental
(TID) in the physical conditioning of endurance athletes: (1) conditions. The first aim of this study was to add a more

Frontiers in Physiology | www.frontiersin.org 155 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

detailed physiological and performance profile for long-distance Srl, Italy) system, which calculated the jump height. Subjects
cyclists over the course of their preparation period, including performed three SJ and three CMJ and the maximal jump height
not only aerobic (VO2max , MLSS, CP, PPO), but also anaerobic was taken for later analysis (Brown and Weir, 2001).
(dLa/dtmax , sprint PPO, MPO, MVC) measurements. The second
aim was to analyze which of the physiological variables (VO2max , Strength-Tests
MLSS, dLa/dtmax ) might correlate with performance (CP, sprint Maximal voluntary isometric strength (MVC) was tested with
MPO), including the values of all three testing sessions, in both legs on a leg press (LP), a leg extension (LE), and a
order to see which of the laboratory parameters might be useful leg curl (LC) machine (Edition-Line, gym80, Gelsenkirchen,
to predict performance. The third aim was to examine the Germany), which were equipped with the digital measurement
composition of the training load based on power data with technique Digimax (mechaTronic; Hamm, Germany) to make
regards to TID. More specifically, changes in the TID are measurements of force-time and velocity-time variables (5 kN
compared in two conditions: when all power data are included, strength sensor type KM1506, distance sensor type S501D,
or when power data below a certain threshold are eliminated. megaTron; Munich, Germany) with the included software
IsoTest and DynamicTest 2.0, as described previously (Wahl
METHODS et al., 2016). The sensors were installed in line with the steel
band of the machines that lifts the weight plates. Maximum force
Subjects relative to body weight was calculated for statistical analysis and
Four male experienced cyclists and triathletes (more than 15 data presentation. Subjects performed three trials of each test,
years of regular endurance training) and one backup athlete and the maximal value was taken for later analysis.
(45 ± 6.5 years; 182.2 ± 8.9 cm; 79.5 ± 6.6 kg) in preparation for
the RAAM volunteered to participate in this study over 6 months. Cycling-Tests
The preparation period started after the offseason. During that All tests were performed on an SRM Ergometer (Schoberer
offseason athletes had a mean training time of ∼8 h per week, Radmesstechnik, GmbH, Germany, Jülich), with seat and
consisting of cycling, running and swimming. The study protocol handlebar height kept identical for each subject throughout all
was approved by the University’s ethics review board and is in tests.
accordance with the declaration of Helsinki. Each subject gave it’s
written informed consent and was informed about possible risks
Sprint-Test
of participation.
After an initial warm-up at 2 W·kg−1 for 10 min, followed by
5 min of passive rest, a 15 s all-out sprint test was performed in
Design
an isokinetic mode set to a cadence of 120 rpm. The subjects
Each athlete underwent three single day testing sessions: The first
performed the test in a sitting position on the ergometer and
at the beginning of the preparation period, the second after 3
were verbally encouraged to achieve maximal power output
months of training (phase 1) and the third at the end of the 6
throughout the test. Afterward, peak power (sprint PPO) and
months training period (phase 2), 2 weeks before the race. Each
mean power (sprint MPO) were determined. Capillary samples
diagnostic session was completed in the same manner at the
from the earlobe were collected before and in minute intervals
same time of day. Prior to performance testing, the subject’s body
(1′ –10′ ) after the test to determine the “maximal rate of lactate
mass and lean body mass were measured using a four-electrode
accumulation (dLa/dtmax )” according to Heck et al. (2003) and
bioimpedance body scale (BC 418 MA, Tanita Corp., Tokyo,
Hauser et al. (2014a):
Japan). Performance testing consisted of jump and strength tests,
followed by endurance tests on a cycle ergometer. Subjects were
instructed to arrive in the laboratory in a rested, 2 h postprandial dLa/dtmax (mmol·L−1 ·s−1 ) = ([La]max − [La]rest ) (texerc − talac )−1
and fully hydrated state. They were ordered to avert strenuous
exercise for at least 24 h before each test. The order of the tests was where [La]max (mmol·L−1 ) = maximal lactate concentration after
kept identical for each individual in the following order. Between the exercise; [La]rest (mmol · L−1 ) = lactate concentration before
the different tests, adequate resting time was ensured. exercise; texerc (seconds) = duration of exercise; talac (seconds)
= period at the beginning of exercise for which (fictitiously) no
Squat Jump (SJ) and Counter Movement lactate formation is assumed. The talac for each subject was set as
Jump Test (CMJ) the time to sprint PPO (seconds).
For the SJ, the subjects were instructed to place the hands on
the hips and to lower the hip into a squat position with a knee VO2max -Test
angle of 90◦ . Out of this position they had to jump with both legs After the sprint test and additional 10 min of recovery, the
for maximal height. For the CMJ, the subjects were instructed to athletes performed a maximal incremental exercise test (initial
place the hands on the hips and to lower the hip dynamically load 100 W + 20 W 1 min−1 ) until exhaustion to determine
down to a self-selected level before jumping with both legs for the VO2max and the corresponding workload (PPO). Heart
maximal height. Hands remained on the hips for the entire rate (Polar, Kempele, Finland), VO2 and carbon dioxide
movement in both tests to eliminate any influence of the arm output (VCO2 ) (Cortex Metalyzer II, Leipzig, Germany) were
swing. Flight time was measured using the Optojump (Microgate continuously measured during the test.

Frontiers in Physiology | www.frontiersin.org 156 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

Afterward, the VO2max and dLa/dtmax were used to calculate TABLE 1 | Changes of power based training zones over the course of the
the maximal lactate steady state (MLSSc ) and the lactate turn preparation period for RAAM.

point 1 (LT1) according to Hauser et al. (2014a). Test Nr. 1 2 3

Lactate-Minimum-Test (LMT) Zone 1 0/136* ± 24– 0/141* ± 17– 0/146* ± 25–

A modified lactate minimum test (Knöpfli-Lenzin and Boutellier, 190 ± 42 197 ± 37 215 ± 42
2011) was performed exactly 7 min after finishing the VO2max Zone 2 191 ± 51– 198 ± 41– 216 ± 51–
test. After these 7 min of passive recovery, RPE, lactate and heart 272 ± 48 282 ± 34 292 ± 50

rate were measured. The drop in heart rate within this time was Zone 3 >273 ± 48 >283 ± 34 >293 ± 50

used to calculate a heart rate recovery index (HRR). Subjects *lower limit of zone 1 for the power specific 3-zone model.
started a second incremental test, beginning with a workload
50 W below the MLSSc . The workload was increased by 10 w
every 90 s, until complete exhaustion. Heart rate (Polar, Kempele, relative power output, average power output, average power
Finland) was continuously measured during the test, RPE and output in percent of MLSSP , average power output in percent of
lactate samples were obtained out of the earlobe at the end of each LT1, and average cadence).
step. Power output at the step which elicited lactate minimum
(LM) was considered as maximal lactate steady state power Statistical Analysis
(MLSSP ), according to Knöpfli-Lenzin and Boutellier (2011). For all statistical analysis of the data Statistica (Version 7.1,
StatSoft Inc., USA) software package for Windows R was used.
Short Power-Profile Descriptive statistics of the data are presented as means ±
Besides the laboratory measurements, the athletes underwent 3 standard deviation (± SD). Data were tested via skewness and
mean maximal power (MMP) tests in the field over the duration kurtosis test for normal distribution. Indices smaller than 2 were
of 2, 5, and 10 min. The MMP tests were each done on different considered to be normal distributed (Vincent, 2005). ANOVA
days around the laboratory measurements. The results were used repeated-measures with Bonferroni post-hoc test was used to
to calculate critical power (CP) according to Monod and Scherrer compare the three testing sessions and the percentage of training
(1965). time in each zone for each of the 6 months for both TID
models. Statistical differences were considered to be significant
Training Data for p < 0.05. The relationship between different parameters
Training mainly consisted of endurance training on the bike. was investigated with Pearson’s correlation coefficient. For these
Additionally, athletes carried out individual core stability correlation analyses, all three time points were included. Cohen’s
training, which was not recorded. Endurance training data of the effect size (d) was calculated for the comparison of all tests with
subjects were captured using wireless SRM Cranks (Schoberer each other. The thresholds for small, moderate, and large effects
Radmesstechnik, GmbH, Germany, Jülich) and the Powercontrol were defined as 0.20, 0.50, and 0.80, respectively (Cohen, 1988;
8 head unit (Schoberer Radmesstechnik, GmbH, Germany, Wahl et al., 2014).
Jülich), saving power data in 1-s intervals for later analysis.
The aggregation of the raw data was done in SRM Win 6.42.18
RESULTS
(Schoberer Radmesstechnik, GmbH, Germany, Jülich), further
analysis was performed in Microsoft Excel 2010 (Microsoft The decrease in body mass during the total training period nearly
Corporation, Redmond, USA) and Statistica 7.1 (StatSoft Inc., reached statistical significance (−2.6 ± 2.3%, P = 0.07, d = 0.29).
Tulsa, USA). Body fat was significantly reduced by −28.7 ± 19.5% (P = 0.01,
Power data were retrospectively analyzed and were used to d = 1.91) as shown in Figures 1A,B.
determine the percentage of training time spent in each of three
training zones for each individual training session. Strength-Tests
The classical 3-zone model was defined as follows: zone 1 There was no significant change in relative leg curl MVC (P =
(below the calculated first rise of lactate/LT1, Hauser et al., 1.0, d = 0.06), relative leg extension MVC (P = 1.0, d = 0.46) and
2014a), zone 2 (between LT1 and MLSSP ), and zone 3 (above leg press MVC (P = 1.0, d = 0.06). Jumping height in the CMJ
MLSSP ). The average training time in each zone for all sessions (P = 1.0, d = 0.15) and in the SJ (P = 1.0, d = 0.16) changed not
of each subject was then determined according to Seiler and significantly (Table 2).
Kjerland (2006). Additionally, a power specific 3-zone model was
defined by deleting all time with power output below 50% of Sprint-Test
MLSSP , in order to reduce the impact of coasting phases and The decreases in the relative PPO and MPO in the sprint nearly
similar events on the TID: zone 1 (between 50% of MLSSP and reached statistical significance [PPO: −3.2 ± 2.7%, P = 0.06,
LT1, Hauser et al., 2014a), zone 2 (between LT1 and MLSSP ) d = 0.25 (Figure 3A) and MPO: −3.6 ± 4.1%, P = 0.06, d = 0.45
and zone 3 (above MLSSP ). Training zones were adjusted after (Figure 3B)].
each testing session according to the changes in LT1 and MLSSP dLa/dtmax decreased significantly by −16.3 ± 8.1% (P = 0.03,
(Table 1). Individual race data were retrospectively analyzed for d = 0.73), relative dLa/dtmax kg−1 decreased significantly by
each of the 4 racers (time, distance, altitude difference, average −16.5 ± 6.7% (P = 0.02, d = 0.67) as shown in Figure 3C. A

Frontiers in Physiology | www.frontiersin.org 157 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

FIGURE 1 | Bodyweight (A) body fat (B). *Significant difference; (*) values nearly reached statistical significance (P = 0.07).

TABLE 2 | Maximal isometric strength (ISO MAX) and jump performance. Relative MLSSP showed a high positive correlation with relative
VO2max (r = 0.92, P < 0.001) and a negative correlation with
Test Nr. 1 2 3
absolute dLa/dtmax (r = −0.78, P < 0.001).
ISO MAX Leg Press [n·kg−1 ] 52.0 ± 12.1 43.9 ± 9.9 52.8 ± 17.0
Leg Extension [n·kg−1 ] 28.5 ± 2.9 28.0 ± 3.0 29.84 ± 3.2 Short Power-Profile
Leg Curl [n·kg−1 ] 16.3 ± 1.0 16.5 ± 0.7 16.3 ± 1.2 Relative MMP tested in the field over a duration of 2 min
increased significantly by 18.7 ± 13.4% (P = 0.01, d = 1.81),
Jumps Squat Jump [cm] 31.68 ± 5.97 32.62 ± 7.58 32.02 ± 5.76
5 min increased significantly by 16.9 ± 11.6% (P = 0.002 d
CMJ [cm] 32.68 ± 6.43 34.04 ± 8.47 33.00 ± 6.89
= 1.56), and 10 min increased significantly by 17.1 ± 6.3% (P
= 0.001, d = 1.44) during the training phase. The calculated
relative CP increased significantly by 16.8 ± 6.2% (P = 0.007,
high correlation was found between relative MPO in the sprint d = 1.20) (Figures 4A–D). Relative CP showed a high positive
and relative dLa/dtmax (r = 0.85, P < 0.001). correlation with relative MLSSP (r = 0.86, P < 0.001), relative
MLSSC (r = 0.88, P < 0.001), and relative VO2max (r = 0.85, P <
VO2max -Test 0.001). Relative CP and dLa/dtmax showed a negative correlation
The increase in relative VO2max in the time course of the 6- of r = −0.61 (P < 0.02).
month preparation phase nearly reached statistical significance
(7.1 ± 5.3%, P = 0.06, d = 0.69) (Figure 2A). Relative peak power
output in the ramp test significantly increased by 9.5 ± 7.1% Training Data
(P = 0.02, d = 0.71) (Figure 2B). The drop in heart rate during 7 Total mean training time was 366 ± 41 h leading to a pyramidal
min of passive recovery showed no significant changes (−14.4 ± TID (zone 1: 63 ± 16%, zone 2: 28 ± 13%, zone 3: 9 ±
17.2%, P = 0.13, d = 0.74) (Figure 3D), as well as peak heart rate 5%) (Figure 5 top) and approximately ∼15.3 h of training per
in the ramp test (174 ± 10 to 173 ± 9 bpm). week using the classical 3-zone model. Total mean training time
without power output below 50% MLSSP (power specific 3-zone
Lactate-Minimum-Test (LMT) model) was 261 ± 47 h and reflected a THR TID (zone 1: 48 ±
The relative MLSSP increased significantly by 10.3 ± 4.1% 13%, zone 2: 39 ± 10%, zone 3: 13 ± 4%) (Figure 5 bottom).
(P = 0.01, d = 0.61) similar to the relative MLSSc , which Overall-ANOVA revealed that the percentage of training in the
increased significantly by 13.3 ± 5.5% (P = 0.008, d = 1.56) HVT-zone was significantly higher (P = 0.003), the percentage of
(Figures 2C,D). There was a high correlation of relative MLSSc training in the THR-zone (P = 0.02) and HIT-zone (P = 0.008)
and the relative power output at MLSSP (r = 0.94, P < 0.001). was significantly lower in the classical 3-zone model compared to

Frontiers in Physiology | www.frontiersin.org 158 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

FIGURE 2 | Relative maximal oxygen uptake (VO2max ) (A), relative peak power output (PPO) (B), relative power output at lactate minimum (MLSSP ) (C), and
relative power output at calculated maximal lactate steady state (MLSSC ) (D). Numbers within the bars represent the absolute values. *Significant difference; (*) values
nearly reached statistical significance (P = 0.06).

the power specific 3-zone model during each of the 6 months of DISCUSSION
training.
The aim of the present study was to analyze the TID for a
Race Data Race Across America Team with two different approaches and
Individual race data (time, distance, altitude difference, to examine the corresponding development of physiological
average relative power output, average power output, average parameters as well as the actual performance over the course
power output in percent of MLSSP , average power output of the preparation for the race, where the team finished in
in percent of LT1, and average cadence) are shown in second place. Over the training period of 6 months, the five
Table 3. athletes reduced their body fat significantly. Moderate increases

Frontiers in Physiology | www.frontiersin.org 159 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

FIGURE 3 | Relative peak power output in the sprint test (Sprint PPO) (A), relative mean power output in the sprint test (Sprint MPO) (B), maximal lactate
production rate (dLa/dtmax ) (C), and drop in heart rate after 7 min of passive recovery after maximal exhaustion (Delta HR Recovery) (D). Numbers within the bars
represent the absolute values. *Significant difference; (*) values nearly reached statistical significance (P = 0.06).

were shown in relative VO2max and moderate decreases in on how to analyze power data have been suggested (Allen and
dLa/dtmax . Power output at MLSS (MLSSP and MLSSC ) increased Coggan, 2010). Allen and Coggan (2010) proposed using an
significantly together with performance as indicated by the exponentially weighted averaging process to represent the data.
calculated CP. As an additional indication of increased fitness We decided to use an alternative approach, eliminating time
levels, moderate effects were shown by the drop in heart rate with zero or very low power output, based on the assumption
within 7 min of passive recovery. The TID of the established 3- that a certain threshold intensity needs to be reached to result
zone model reflected a PYR TID (zone 1: 63 ± 16%, zone2: 28 in an effective training stimulus (Davies and Knibbs, 1971).
± 13% and zone 3: 9 ± 4%). Deleting training time with power According to Meyer et al. (1999) and Wolpern et al. (2015),
output below 50% of MLSSP , resulted in a THR TID (zone 1: 48 we decided to use a lactate threshold orientated lower fix point
± 13%, zone 2: 39 ± 10%, zone 3: 13 ± 4%). for the determination of zone 1 and set the minimum power
The different results of our two approaches concerning TID, output necessary to elicit a stimulus for training adaptation to
underline that the inherent variability in power output during 50% of MLSSP . Applying the classic 3-zone model in the present
training raises several challenges when attempting to evaluate study, weekly training time is similar to a report of athletes
the exact nature of a given training session. Different strategies participating in a RAAM qualifying race (Knechtle et al., 2012b).

Frontiers in Physiology | www.frontiersin.org 160 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

FIGURE 4 | Relative mean maximal power output in the field tests over 2 min (A), 5 min (B) and 10 min (C) and the calculated relative critical power (CP) (D).
Numbers within the bars represent the absolute values. *Significant difference.

However, when using the power specific 3-zone model in our parameter. In contrast, 7 months of a PYR TID with slightly more
study, total training time is reduced by 104 ± 7 h per athlete, training per week (+ ∼1 h) than above-mentioned studies, led to
leading to a THR TID. Our athletes, therefore, performed a significant improvements in relative VO2max of 10% in fourteen
lot of training between LT1 and MLSS, which was shown to male young top-level national road cyclists (Zapico et al., 2007).
be a rather ineffective training intensity when the goals are to In any way, lactate values for a given sub-maximal workload
gain performance improvements and physiological adaptations during a ramp test decreased after the 7 months of training,
(Esteve-Lanao et al., 2007; Yu et al., 2012; Stöggl and Sperlich, suggesting an increased reliance on oxidative metabolism (Luciá
2014). et al., 2000), MMP over 5 min increased significantly (Aagaard
Luciá et al. (2000) and Aagaard et al. (2011) studied et al., 2011) and MLSS increased by 15% (Zapico et al., 2007),
professional cyclists and young top-level national cyclists for which is in line with our results. Since our cyclists were not top
periods of seven (PYR TID) and four (unknown TID) months, levels athletes, it is very unlikely that they reached a ceiling of
respectively. Despite higher weekly training volumes compared improvement for VO2max . Nonetheless, there were only small
to our athletes, they could not show significant changes in increases in relative VO2max , which can mainly be attributed to
VO2max . The authors concluded that the very high fitness reductions in body weight. It can be speculated that the amount
level of their athletes prevented further improvements for this of HIT/SIT, which was shown to be most effective in improving

Frontiers in Physiology | www.frontiersin.org 161 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

FIGURE 5 | Training intensity distribution according to the 3 zone model (top) and the power specific 3 zone model (bottom) over the time course of 6
months. Numbers in boxes represent the percentage of total training time spent in each zone per month for each athlete. Vertical arrows assign performance tests.
Symbols represent each individual athlete. The included table shows the TID of each athlete for the whole preparation period.

VO2max (McKenna et al., 1997; Dawson et al., 1998; MacDougall output underline the importance to document changes in these
et al., 1998; Breil et al., 2010; Gibala et al., 2012; Etxebarria parameters, besides VO2max , in order to gain more insight into
et al., 2014; Milanović et al., 2015) was too low in the present physiological changes, leading to performance improvements
study to increase VO2max substantially, especially in the context (e.g., the rise in CP).
of the limited regular training volume. However, the decrease Additionally, Zapico et al. (2007) observed parallel to an
in dLa/dtmax and the corresponding increase in MLSSP power increase in PPO, a significant reduction in maximal heart rate

Frontiers in Physiology | www.frontiersin.org 162 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

TABLE 3 | Competition data of each individual racer (colored symbols) during the Race Across America.

Athlete Time [hh:mm] Distance [km] Altitude Average power Average Power [% MLSS] Power [% LT1] Average
difference [m] [W·kg−1 ] power [W] cadence [rpm]

29:55 890 7411 2.3 186 85 141 78

34:41 1081 9566 2.6 199 69 100 88
45:16 1628 9528 2.8 214 69 91 88
37:08 1184 9100 2.7 187 69 92 90

during a ramp test after 7 months of training, which is in contrast the origin and interpretation of MLSS (Mader and Heck,
to our athletes. It has to be taken into account, however, that the 1986).
subjects in our study were ∼25 years older and ∼11 kg heavier, The present study is the first to describe changes in the
which might partly explain the overall lower values, especially in dLa/dtmax over a long training period, and also its correlation
VO2max , as well as the smaller improvements. The TIDs, although with performance (CP, MLSS, MPO). Based on the model of
being PYR in both studies, differ in so far, that our subjects Mader and Heck (1986) only both together—the moderate
trained fewer hours in total (−105 h) and to a less extent in zone increase in VO2max (7.1%) and to a larger proportion the decrease
1 (−9.5%), and more in zone 2 (+6.5%). It can be speculated in dLa/dtmax (16.3%)—can explain the increase in performance
that the higher amount of THR training might be one reason for (MLSS, CP) as measured in the present study. In accordance
the lower gains in the physiological variables of the present study with the decrease in dLa/dtmax is the loss of anaerobic power,
(Laursen, 2010). This assumption is supported by the study of also reflected by the decrease in PPO and MPO within the
Neal et al. (2011), who monitored 10 triathletes over a 6 month sprint test. However, we found no significant changes in the
training period with a mean training volume of 203 ± 71 h. strength assessments for five athletes. This is partly in line with
Similar to our athletes, these triathletes performed a PYR TID the findings of Aagaard et al. (2011), where cyclists’ maximal
with a high amount of THR training, resulting in no changes in isometric strength, tested via leg press, also remained unchanged,
power output at lactate turn point. The lower effectiveness of a after performing only endurance training over a period of 4
THR TID is further supported by another study of Neal et al. months.
(2013), which resulted in greater improvements in 40-km time The analysis of our race data showed, that three athletes used
trial and LT1 and lactate turn point in a group of cyclists training similar relative power outputs in relation to their MLSSP and
POL when compared to a THR TID. LT1. The race intensity of our athletes is in agreement with the
The LMT and the calculation of the MLSS (MLSSC ) using data of Laursen et al. (1999), who described a mean intensity
VO2max and dLa/dtmax were shown to be valid to determine slightly below the ventilatory threshold of a 4 person RAAM
MLSS (Knöpfli-Lenzin and Boutellier, 2011; Hauser et al., 2014a). team. A solo racer during the RAAM performed at slightly lower
The fact that both independent methods to determine the power intensities (77% of LT1) (Schumacher et al., 2011). The weakest
output at MLSS showed a high correlation underlines that the subject in our study performed more intense with regards to
increases in MLSS were accurately determined. Additionally, % of MLSSP and % of LT1 during the race, which might be
the present study is the first to document changes in MLSS due to the lower distance and altitude differences that he had
over the course of a training period for cyclists using these two to cover compared to the other riders. Another explanation
approaches. Only four previous studies investigated the effects might be that this athlete had the highest relative amounts of
of training on the lactate minimum. Carter et al. (1999) showed training time in zone 2 and 3, compared to the stronger riders.
that the LMT is not sensitive to identify longitudinal effects It has been reported before, that the cycling speed during the
of endurance training in sports students over 6 weeks, despite training units was significantly and negatively related to race
significant improvements in VO2max . Similar to our results, three time (Knechtle et al., 2011, 2012b). Nevertheless, differences
other studies found significant increases in the LM intensity, in race performances in ultra-endurance events can also result
investigating elite and youth soccer players (da Silva et al., 2007; from various other influencing factors like race tactics, weather
Miranda et al., 2013) and youth swimmers (Campos et al., conditions, motivation (Lahart et al., 2013), sleep deprivation
2014), supporting that the LMT can be performed to identify (Knechtle et al., 2012a), nutrition (Stewart and Stewart, 2007;
longitudinal training effects. Again, this is further supported by Hulton et al., 2010; Bescós et al., 2012; Lahart et al., 2013; Paulin
the high correlation of the two methods we used to determine et al., 2015) and so on, which are not taken into account in this
MLSS. publication.
According to Hauser et al. (2014b), the MLSS is mainly
influenced by VO2max and dLa/dtmax . While VO2max and its CONCLUSION
influence on endurance performance has been focused on in
most endurance studies (Coyle et al., 1988; Schumacher and The present study shows a THR orientated TID and only
Mueller, 2002; Støren et al., 2012), dLa/dtmax has been neglected moderate (Cohen, 1988) increases in physiological parameters.
and appears to be an underestimated parameter in terms of The resulting TID, however, may largely depend on the kind

Frontiers in Physiology | www.frontiersin.org 163 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

of analysis of power data. The amount of training below 50% estimate of long-term endurance performance. Future studies,
of MLSS in our amateur athletes is remarkably high (104 h therefore, should consider measuring dLa/dtmax and MLSS in
(∼28% of total training time)). Researchers, coaches, and athletes, addition to VO2max and performance.
either analyzing TID retrospectively or planning training in
advance, should pay close attention to the different ways of AUTHOR CONTRIBUTIONS
interpreting training data, to gain realistic insights in TID and the
corresponding improvements in performance and physiological CM: performed tests and statistics, acquired data, and wrote the
parameters. In matters of physiological parameters determined paper. JM: wrote paper. WK: acquired data. PW: performed tests
during training periods, power output at LT1 seems to be a good and statistics, acquired data, and wrote the paper.

REFERENCES with competition performance. Med. Sci. Sports Exerc. 37, 496–504.
doi: 10.1249/01.MSS.0000155393.78744.86
Aagaard, P., Andersen, J. L., Bennekou, M., Larsson, B., Olesen, J. L., Crameri, R., Etxebarria, N., Anson, J. M., Pyne, D. B., and Ferguson, R. A. (2014). High intensity
et al. (2011). Effects of resistance training on endurance capacity and muscle cycle interval training improves cycling and running performance in triathletes.
fiber composition in young top-level cyclists. Scand. J. Med. Sci. Sports 21, Eur. J. Sport Sci. 14, 521–529. doi: 10.1080/17461391.2013.853841
e298–e307. doi: 10.1111/j.1600-0838.2010.01283.x Faria, E. W., Parker, D. L., and Faria, I. E. (2005). The science of
Allen, H., and Coggan, A. R. (2010). Training and Racing with a Power Meter. cycling: physiology and training – part 1. Sports Med. 35, 285–312.
Boulder, CO: VeloPress. doi: 10.2165/00007256-200535040-00002
Bagger, M., Petersen, P. H., and Pedersen, P. K. (2003). Biological variation in Gibala, M. J., Little, J. P., Macdonald, M. J., and Hawley, J. A. (2012). Physiological
variables associated with exercise training. Int. J. Sports Med. 24, 433–440. adaptations to low-volume, high-intensity interval training in health and
doi: 10.1055/s-2003-41180 disease. J. Physiol. 590, 1077–1084. doi: 10.1113/jphysiol.2011.224725
Bescós, R., Rodríguez, F.-A., Iglesias, X., Knechtle, B., Benítez, A., Marina, M., et al. Guellich, A., Seiler, K. S., and Emrich, E. (2009). Training methods and intensity
(2012). Nutritional behavior of cyclists during a 24-hour team relay race: a field distribution of young world-class rowers. Int. J. Sports Physiol. Perform. 4,
study report. J. Int. Soc. Sports Nutr. 9:3. doi: 10.1186/1550-2783-9-3 448–460. doi: 10.1123/ijspp.4.4.448
Billat, V., Demarle, A., Slawinski, J., Paiva, M., and Koralsztein, J. P. (2001). Halson, S. L. (2014). Monitoring training load to understand fatigue in athletes.
Physical and training characteristics of top-class marathon runners. Med. Sci. Sports Med. 44(Suppl. 2), 139–147. doi: 10.1007/s40279-014-0253-z
Sports Exerc. 33, 2089–2097. Hauser, T., Adam, J., and Schulz, H. (2014a). Comparison of calculated and
Bowen, R. L., Adams, J. H., and Myburgh, K. H. (2006). Nausea and high serum experimental power in maximal lactate-steady state during cycling. Theor. Biol.
osmolality during a simulated ultraendurance adventure race: a case-control Med. Model. 11, 1–12. doi: 10.1186/1742-4682-11-25
study. Int. J. Sports Physiol. Perform. 1, 176–185. doi: 10.1123/ijspp.1.2.176 Hauser, T., Adam, J., and Schulz, H. (2014b). Comparison of selected lactate
Breil, F. A., Weber, S. N., Koller, S., Hoppeler, H., and Vogt, M. (2010). threshold parameters with maximal lactate steady state in cycling. Int. J. Sports
Block training periodization in alpine skiing: effects of 11-day HIT Med. 35, 517–521. doi: 10.1055/s-0033-1353176
on VO2max and performance. Eur. J. Appl. Physiol. 109, 1077–1086. Heck, H., Schulz, H., and Bartmus, U. (2003). Diagnostics of anaerobic
doi: 10.1007/s00421-010-1455-1 power and capacity. Eur. J. Sport Sci. 3, 1–23. doi: 10.1080/174613903000
Brown, L. E., and Weir, J. P. (2001). ASEP procedures recommendation I: accurate 73302
assessment of muscular strength and power. J. Exerc. Physiol. 4, 1–21. Hulton, A. T., Lahart, I. M., Williams, K. L., Godfrey, R., Charlesworth, S., Wilson,
Buchheit, M. (2014). Monitoring training status with HR measures: do all roads M., et al. (2010). Energy expenditure in the Race Across America (RAAM). Int.
lead to Rome? Front. Physiol. 5:73. doi: 10.3389/fphys.2014.00073 J. Sports Med. 31, 463–467. doi: 10.1055/s-0030-1251992
Campos, E. Z., Nordsborg, N. B., da Silva, A. S. R., Zagatto, A. M., Gerosa Neto, Ice, R. G., Millman, P. L., Ice, D. C., and Camp, J. C. (1988). “A physiological
J., Andrade, V. L. D., et al. (2014). The response of the lactate minimum profile of the 1984–1986 race across America winner,” in Medical and Scientific
test to a 12-week swimming training. Motriz. Rev. Educ. Fis. 20, 286–291. Aspects of Cycling, eds E. R. Burke and M. M. Newsom (Champaign, IL: Human
doi: 10.1590/S1980-65742014000300007 Kinetics), 173–180.
Carter, H., Jones, A. M., and Doust, J. H. (1999). Effect of 6 weeks of Jacobs, R. A., Rasmussen, P., Siebenmann, C., Díaz, V., Gassmann, M., Pesta, D.,
endurance training on the lactate minimum speed. J. Sports Sci. 17, 957–967. et al. (2011). Determinants of time trial performance and maximal incremental
doi: 10.1080/026404199365353 exercise in highly trained endurance athletes. J. Appl. Physiol. 111, 1422–1430.
Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Hillsdale, doi: 10.1152/japplphysiol.00625.2011
NJ: Erlbaum. Knechtle, B., Enggist, A., and Jehle, T. (2005). Energy turnover at the Race
Coyle, E. F., Coggan, A. R., Hopper, M. K., and Walters, T. J. (1988). Determinants Across AMerica (RAAM) – a case report. Int. J. Sports Med. 26, 499–503.
of endurance in well-trained cyclists. J. Appl. Physiol. 64, 2622–2630. doi: 10.1055/s-2004-821136
da Silva, A. S. R., Bonette, A. L., Santhiago, V., and Gobatto, C. A. (2007). Effect of Knechtle, B., Knechtle, P., Rust, C. A., Rosemann, T., and Lepers, R. (2011).
soccer training on the running speed and the blood lactate concentration at the Finishers and nonfinishers in the ‘Swiss Cycling Marathon’ to qualify
lactate minimum test. Biol. Sport 24, 105–114. for the ‘Race Across America’. J. Strength Cond. Res. 25, 3257–3263.
Davies, C. T., and Knibbs, A. V. (1971). The training stimulus. The effects of doi: 10.1519/JSC.0b013e31821606b3
intensity, duration and frequency of effort on maximum aerobic power output. Knechtle, B., Wirth, A., Knechtle, P., Rüst, C. A., and Rosemann, T. (2012b). A
Int. Z. Angew. Physiol. 29, 299–305. comparison of ultra-endurance cyclists in a qualifying ultra-cycling race for
Dawson, B., Fitzsimons, M., Green, S., Goodman, C., Carey, M., and Cole, K. Paris-Brest-Paris and Race Across America-Swiss cycling marathon. Percept.
(1998). Changes in performance, muscle metabolites, enzymes and fibre types Mot. Skills 114, 96–110. doi: 10.2466/05.PMS.114.1.96-110
after short sprint training. Eur. J. Appl. Physiol. Occup. Physiol. 78, 163–169. Knechtle, B., Wirth, A., Knechtle, P., Rust, C. A., Rosemann, T., and Lepers,
doi: 10.1007/s004210050402 R. (2012a). No improvement in race performance by naps in male ultra-
Esteve-Lanao, J., Foster, C., Seiler, K. S., and Luci,á, A. (2007). Impact of training endurance cyclists in a 600-km ultra-cycling race. Chin. J. Physiol. 55, 125–133.
intensity distribution on performance in endurance athletes. J. Strength Cond. doi: 10.4077/CJP.2012.BAA022
Res. 21, 943–949. doi: 10.1519/R-19725.1 Knöpfli-Lenzin, C., and Boutellier, U. (2011). Lactate minimum is valid to estimate
Esteve-Lanao, J., San Juan, A. F., Earnest, C. P., Foster, C., and Luciá, maximal lactate steady state in moderately and highly trained subjects. J.
A. (2005). How do endurance runners actually train? Relationship Strength Cond. Res. 25, 1355–1359. doi: 10.1519/JSC.0b013e3181d6dbf4

Frontiers in Physiology | www.frontiersin.org 164 December 2016 | Volume 7 | Article 642

Manunzio et al. RAAM Participants’ TID and Physiology

Lahart, I. M., Lane, A. M., Hulton, A. T., Williams, K. L., Godfrey, R., Pedlar, Schumacher, Y. O., Ahlgrim, C., Prettin, S., and Pottgiesser, T. (2011). Physiology,
C., et al. (2013). Challenges in maintaining emotion regulation in a sleep and power output, and racing strategy of a Race Across America finisher. Med. Sci.
energy deprived state induced by the 4800Km ultra-endurance bicycle race; the Sports Exerc. 43, 885–889. doi: 10.1249/MSS.0b013e3181fec009
Race Across AMerica (RAAM). J. Sports Sci. Med. 12, 481–488. Schumacher, Y. O., and Mueller, P. (2002). The 4000-m team pursuit cycling world
Laursen, P. B. (2010). Training for intense exercise performance: high-intensity record: theoretical and practical aspects. Med. Sci. Sports Exerc. 34, 1029–1036.
or high-volume training? Scand. J. Med. Sci. Sports 20(Suppl. 2), 1–10. doi: 10.1097/00005768-200206000-00020
doi: 10.1111/j.1600-0838.2010.01184.x Seiler, K. S., and Kjerland, G. Ø. (2006). Quantifying training intensity distribution
Laursen, P. B., Rhodes, E. C., and Buchanan, J. M. (1999). Physiological analysis of in elite endurance athletes: is there evidence for an “optimal” distribution?
a high intensity ultraendurance event. Strength Cond. J. 21, 26–38. Scand. J. Med. Sci. Sports 16, 49–56. doi: 10.1111/j.1600-0838.2004.00418.x
Luciá, A., Hoyos, J., Pardo, J., and Chicharro, J. L. (2000). Metabolic and Stewart, I. B., and Stewart, K. L. (2007). Energy balance during two
neuromuscular adaptations to endurance training in professional cyclists: days of continuous stationary cycling. J. Int. Soc. Sports Nutr. 4:15.
a longitudinal study. Jpn. J. Physiol. 50, 381–388. doi: 10.2170/jjphysiol. doi: 10.1186/1550-2783-4-15
50.381 Stöggl, T. L., and Sperlich, B. (2014). Polarized training has greater impact on key
MacDougall, J. D., Hicks, A. L., MacDonald, J. R., McKelvie, R. S., Green, H. J., and endurance variables than threshold, high intensity, or high volume training.
Smith, K. M. (1998). Muscle performance and enzymatic adaptations to sprint Front. Physiol. 5:33. doi: 10.3389/fphys.2014.00033
interval training. J. Appl. Physiol. 84, 2138–2142. Stöggl, T. L., and Sperlich, B. (2015). The training intensity distribution
Mader, A., and Heck, H. (1986). A theory of the metabolic origin of “anaerobic among well-trained and elite endurance athletes. Front. Physiol. 6:295.
threshold.” Int. J. Sports Med. 7(Suppl. 1), 45–65. doi: 10.3389/fphys.2015.00295
McKenna, M. J., Heigenhauser, G. J., McKelvie, R. S., Obminski, G., MacDougall, Støren, Ø., Bratland-Sanda, S., Haave, M., and Helgerud, J. (2012). Improved
J. D., and Jones, N. L. (1997). Enhanced pulmonary and active skeletal muscle VO2max and time trial performance with more high aerobic intensity interval
gas exchange during intense exercise after sprint training in men. J. Physiol. training and reduced training volume: a case study on an elite national cyclist.
501(Pt 3), 703–716. J. Strength Cond. Res. 26, 2705–2711. doi: 10.1519/JSC.0b013e318241deec
Meyer, T., Gabriel, H. H., and Kindermann, W. (1999). Is determination of exercise Støren, Ø., Ulevåg, K., Larsen, M. H., Støa, E. M., and Helgerud, J. (2013).
intensities as percentages of VO2max or HRmax adequate? Med. Sci. Sports Physiological determinants of the cycling time trial. J. Strength Cond. Res. 27,
Exerc. 31, 1342–1345. 2366–2373. doi: 10.1519/JSC.0b013e31827f5427
Milanović, Z., Sporiš, G., and Weston, M. (2015). Effectiveness of High-Intensity Tønnessen, E., Sylta, Ø., Haugen, T. A., Hem, E., Svendsen, I. S., and Seiler,
Interval Training (HIT) and continuous endurance training for VO2max K. S. (2014). The road to gold: training and peaking characteristics in the
improvements: a systematic review and meta-analysis of controlled trials. Sports year prior to a gold medal endurance performance. PLoS ONE 9:e101796.
Med. 45, 1469–1481. doi: 10.1007/s40279-015-0365-0 doi: 10.1371/journal.pone.0101796
Miranda, R., Antunes, H., Pauli, J. R., Puggina, E. F., and da Silva, A. S. R. (2013). Vincent, W. J. (2005). Statistics in Kinesiology. Champaign, IL: Human Kinetics.
Effects of 10-week soccer training program on anthropometric, psychological, Wahl, P., Güldner, M., and Mester, J. (2014). Effects and sustainability of a 13-day
technical skills and specific performance parameters in youth soccer players. high-intensity shock microcycle in soccer. J. Sports Sci. Med. 13, 259–265.
Sci. Sports 28, 81–87. doi: 10.1016/j.scispo.2012.02.005 Wahl, P., Sanno, M., Ellenberg, K., Frick, H., Böhm, E., Haiduck, B., et al. (2016).
Monod, H., and Scherrer, J. (1965). The work capacity of a synergic muscular Aqua cycling does not affect recovery of performance, damage markers and
group. Ergonomics 8, 329–338. doi: 10.1080/00140136508930810 sensation of pain. J. Strength Cond. Res. doi: 10.1519/JSC.0000000000001462.
Muñoz, I., Seiler, K. S., Bautista, J., España, J., Larumbe, E., and Esteve-Lanao, J. [Epub ahead of print].
(2014). Does polarized training improve performance in recreational runners? Wolpern, A. E., Burgos, D. J., Janot, J. M., and Dalleck, L. C. (2015). Is a threshold-
Int. J. Sports Physiol. Perform. 9, 265–272. doi: 10.1123/ijspp.2012-0350 based model a superior method to the relative percent concept for establishing
Neal, C. M., Hunter, A. M., Brennan, L., O’Sullivan, A., Hamilton, D. L., individual exercise intensity? A randomized controlled trial. BMC Sports Sci.
de Vito, G., et al. (2013). Six weeks of a polarized training-intensity Med. Rehabil. 7:16. doi: 10.1186/s13102-015-0011-z
distribution leads to greater physiological and performance adaptations Yu, H., Chen, X., Zhu, W., and Cao, C. (2012). A quasi-experimental study
than a threshold model in trained cyclists. J. Appl. Physiol. 114, 461–471. of Chinese top-level speed skaters’ training load: threshold versus polarized
doi: 10.1152/japplphysiol.00652.2012 model. Int. J. Sports Physiol. Perform. 7, 103–112. doi: 10.1123/ijspp.7.2.103
Neal, C. M., Hunter, A. M., and Galloway, S. D. (2011). A 6-month Zapico, A. G., Calderón, F. J., Benito, P. J., González, C. B., Parisi, A., Pigozzi, F.,
analysis of training-intensity distribution and physiological adaptation in et al. (2007). Evolution of physiological and haematological parameters with
Ironman triathletes. J. Sports Sci. 29, 1515–1523. doi: 10.1080/02640414.2011.5 training load in elite male road cyclists: a longitudinal study. J. Sports Med. Phys.
96217 Fitness 47, 191–196.
Padilla, S., Mujika, I., Cuesta, G., and Goiriena, J. J. (1999). Level ground and
uphill cycling ability in professional road cycling. Med. Sci. Sports Exerc. 31, Conflict of Interest Statement: The authors declare that the research was
878–885. conducted in the absence of any commercial or financial relationships that could
Paulin, S., Roberts, J., Roberts, M., and Davis, I. (2015). A case study be construed as a potential conflict of interest.
evaluation of competitors undertaking an antarctic ultra-endurance event:
nutrition, hydration and body composition variables. Extrem. Physiol. Med. 4:3. Copyright © 2016 Manunzio, Mester, Kaiser and Wahl. This is an open-access article
doi: 10.1186/s13728-015-0022-0 distributed under the terms of the Creative Commons Attribution License (CC BY).
Sandbakk, Ø., Hegge, A. M., Losnegard, T., Skattebo, Ø., Tønnessen, E., and The use, distribution or reproduction in other forums is permitted, provided the
Holmberg, H.-C. (2016). The physiological capacity of the world’s highest original author(s) or licensor are credited and that the original publication in this
ranked female cross-country skiers. Med. Sci. Sports Exerc. 48, 1091–1100. journal is cited, in accordance with accepted academic practice. No use, distribution
doi: 10.1249/MSS.0000000000000862 or reproduction is permitted which does not comply with these terms.

Frontiers in Physiology | www.frontiersin.org 165 December 2016 | Volume 7 | Article 642

Advantages 90

of publishing OPEN ACCESS

Articles are free to read
FAST PUBLICATION
Around 90 days

in Frontiers for greatest visibility

and readership
from submission
to decision

HIGH QUALITY PEER-REVIEW TRANSPARENT PEER-REVIEW

Rigorous, collaborative, Editors and reviewers
and constructive acknowledged by name
peer-review on published articles

Frontiers
Avenue du Tribunal-Fédéral 34
1005 Lausanne | Switzerland

Visit us: www.frontiersin.org

Contact us: [email protected] | +41 21 510 17 00 REPRODUCIBILITY OF DIGITAL PUBLISHING
RESEARCH Articles designed
Support open data for optimal readership
and methods to enhance across devices
research reproducibility

FOLLOW US IMPACT METRICS EXTENSIVE PROMOTION LOOP RESEARCH NETWORK

@frontiersin Advanced article metrics Marketing Our network
track visibility across and promotion increases your
digital media of impactful research article’s readership