Review

European Journal of Preventive

Cardiology

Best mode of inspiratory muscle training 2018, Vol. 25(16) 1691–1701

! The European Society of
Cardiology 2018
in heart failure patients: a systematic Article reuse guidelines:
sagepub.com/journals-permissions
review and meta-analysis DOI: 10.1177/2047487318792315
journals.sagepub.com/home/ejpc

Zahra Sadek1,2, Ali Salami1, Wissam H Joumaa1,

Charifa Awada1, Said Ahmaidi2 and Wiam Ramadan1,3

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024

Abstract
Objectives: The objective of this study was to evaluate the effects of inspiratory muscle training on inspiratory muscle
strength, functional capacity and dyspnoea for patients with chronic heart failure, by summarising the published research
on the effects of inspiratory muscle training. To identify the best mode of intervention in terms of: the load of maximal
inspiratory pressure; the frequency of sessions; and the total duration of intervention.
Methods: A relevant literature research using the PubMed database, Cochrane and references of published studies,
from 1998 to 2016, was conducted. Out of 65 randomised controlled trials, seven were considered as potentially
relevant and were retrieved for detailed analysis. The methodological quality of each randomised controlled trial was
rated using the physiotherapy evidence database scale.
Results: The included seven studies contained data on 203 patients. Typical training protocols involved training three, six
or seven times per week with intensity ranging from 30% to 60% and for a duration ranging from 6 to 12 weeks. Maximal
inspiratory pressure, walking distance and dyspnoea were improved in all studies and especially in those who set a load of
60% in their maximal inspiratory pressure, and have trained patients six times per week for 12 weeks.
Conclusion: In chronic heart failure patients, inspiratory muscle training results in a marked improvement in inspiratory
muscle strength, walking distance and dyspnoea, notably when training patients at 60% of maximal inspiratory pressure,
six times per week and for 12 weeks. A small number of studies and heterogeneity among studies may limit the findings
of the present study.

Keywords
Chronic heart failure, inspiratory muscle training, dyspnoea, quality of life, maximal inspiratory pressure
Received 22 March 2018; accepted 12 July 2018

Introduction abnormalities, decreased oxidative enzyme activities,

Chronic heart failure (CHF) patients often suffer from muscle atrophy, in addition to an increase in pro-inflam-
increased work of breathing and dyspnoea during phys- matory cytokines and markers of oxidative stress.
ical activity.1 First, the researchers considered haemo-
dynamic disturbances as a major determinant of these
1
symptoms, because of the well-known ability of the Rammal Hassan Rammal Research Laboratory, Physio-toxicity (PhyTox),
Lebanese University, Faculty of Sciences, Nabatieh, Lebanon
ineffective heart pump to increase pulmonary and sys- 2
Laboratory EA-3300: APERE. Adaptations Physiologiques à l’Exercice et
temic venous pressures, and to reduce cardiac output. Réadaptation à l’Effort. Picardie Jules Verne University, Amiens, F-80025,
However, this hypothesis has been rejected and France
3
shifted towards the muscle hypothesis, which suggests Lebanese Institutes for Biomedical Research and Application (LIBRA),
abnormalities of skeletal muscle as the source of Lebanese International University, Lebanon
CHF symptoms,2 due to the fact that the main contribu-
Corresponding author:
tor to exercise intolerance is musculoskeletal Zahra Sadek, Laboratory EA3300, APERE, Picardie de Jule Vernes
pathology,3 associated with an increased ergoreflex: a University, Amiens, F-80025, France.
muscle reflex stimulated by work,4 mitochondrial Email: [email protected]
1692 European Journal of Preventive Cardiology 25(16)

It is therefore likely that these changes are not limited strategies involving the querying of two online data-
to the musculature of the lower limbs, but are widespread bases (PubMed database and Cochrane), followed by
and may affect respiratory muscles. Observations showed evaluation of the bibliographies of published studies
a significantly lower percentage of type II and type IIa related to this subject for the identification of related
muscle fibres and a greater percentage of type I muscle relevant publications. We also use back-tracking to find
fibres in the costal diaphragm, fibre atrophy and abnorm- earlier relevant sources, and forward-tracking in Web
alities in ventilation and/or perfusion.5 of Science to find the literature that refers to our sub-
It has also been suggested that the symptoms of fati- ject. As a starting point for this study, we developed a
gue and dyspnoea usually reported in CHF patients series of key search terms in order to outline the scope
may be partially attributed to decreased strength and of the literature to include in the review. So, we used the
endurance of the inspiratory muscles,6 assessed, following research term categories in combination:

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024

respectively, by measuring maximal inspiratory pres- ‘congestive heart failure’, ‘cardiac failure’, ‘CHF’,
sure (PImax) and by the ability to sustain pressure in ‘IMT’, ‘respiratory muscle training’, ‘exercise training’
time.7,8 and ‘maximal inspiratory pressure’. All clinical trials
A multimodal approach is required to treat CHF published in English were included. The second step
patients; it involves a combination of non-pharmacolo- for limiting and ameliorating the first research was to
gical and pharmacological treatment.6 In this review, we include only those articles on patients with ‘heart fail-
will emphasise the non-pharmacological management ure’ or ‘congestive heart failure’.
and especially inspiratory muscle training (IMT) because We have screened all titles and abstracts and made
it has been shown to benefit patients by an improvement decisions regarding potential eligibility after full-text
in inspiratory muscle strength, inspiratory muscle endur- review.
ance and quality of life, in addition to a greater sense of
wellbeing, and a better functional capacity. From a
therapeutic point of view, the beneficial effects of select-
Eligibility criteria
ive respiratory muscle training, in improving respiratory We included all randomised controlled trials (RCTs)
muscle strength and overall aerobic capacity in CHF, assessing IMT in the rehabilitation of CHF. We
were first demonstrated by Mancini et al. with further included studies that compared IMT to control
data being published over the past decade.9 groups (placebo IMT: the same regimen as the IMT
In fact, IMT uses progressive resistance to provide group but with minimum inspiratory load or with no
loading to the inspiratory muscles to achieve a training effects or any other intervention such as an
strengthening effect.10 The devices used have evolved education programme) and that evaluate the following
considerably, and they are: the resistive type of muscle outcome measures: PImax, distance walked in the 6-
trainer, the pressure threshold muscle trainer and the minute walking test (6MWT) or 12-minute walking
isocapnia hyperpnoea manoeuvre. Actually, the most test (12MWT) and dyspnoea.
commonly used is the second one, which is a commer- Some studies were excluded based on the examin-
cially available spring-loaded threshold valve that can ation of the title or abstract, and other exclusion cri-
easily be adjusted to provide incremental resistance.11 teria were as follows: (a) studies that include subjects
Moreover, all studies published used different proto- other than CHF patients; (b) studies that focus on gen-
cols of training in terms of the load of PImax and the eral training and not IMT; (c) studies with the absence
duration of intervention. So we have variability in out- of a control group; (d) studies with intervention dur-
comes. So far, the best mode of training for CHF ation less than 6 weeks; and (e) non-randomised con-
patients remains unclear. trolled trials and animal-based studies. Actually, seven
Accordingly, the aim of this study was to shed light on studies were considered as potentially relevant and were
the positive effect of IMT in patients with CHF by sum- retrieved for detailed analysis (Figure 1).
marising the published research on the effects of IMT and
subsequently to identify the best mode of intervention in
terms of: (a) the load of PImax; (b) the frequency of ses-
Data extraction
sions; and (c) the total duration of intervention. Demographic data such as age, gender, New York
Heart Association (NYHA) functional class, and ejec-
tion fraction (EF) were recorded to characterise the
Methods trials and to determine the homogeneity of participants
among groups and trials. The target intensity of PImax,
Literature search
duration, frequency, the total period of time for the
A literature search of published studies, from 1998 to exercise training programme and the nature of the con-
2016, was conducted. We used standard research trol group were also recorded.
Sadek et al. 1693

Medical database
(Pubmed/
cochrane)

65 citations 33 excluded (based on

received from the the reading of
database (RCTs) abstract)

25 excluded (do not meet inclusion criteria):

32 articles 1 was a guideline article

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024

1 included COPD patients and not HF patients
2 involved an intervention for less than 6 weeks
3 were not RCTs
13 were not focusing on IMT intervention
7 articles used in 5 were review articles covering the topic of HF
the study

Figure 1. Flow chart for identifying eligible articles.

COPD: chronic obstructive pulmonary disease; HF: heart failure; RCT: randomised controlled trial; IMT: inspiratory muscle training.

which values above 25% and 50% were considered

Data analysis indicative of moderate and high heterogeneity, respect-
To calculate the degree of the intervention effects, we ively.1 All analyses were conducted using the software
determined the effect sizes for the following outcome comprehensive meta-analysis ‘V3’.
variables: inspiratory muscle strength, walking test dis-
tance and dyspnoea. The effect size (d) of treatment (pre
Methodological quality
vs. post or experimental vs. control) was calculated
using equation (1): The physiotherapy evidence database (PEDro) is used
in order to estimate the value of the methodological
 e  c quality of each RCT.13 It is an instrument utilised to
d¼ ð1Þ
Sp assess the methodological quality of RCTs in physical
therapy and exercise studies. Eleven items (see Table 1)
where d is Cohen’s effect size, e is the mean of the on the PEDro scale were derived from a Delphi con-
experimental group, c is the mean of the control sensus procedure.14 Its scale scores can range from 0 to
group, and Sp is the pooled estimate of variance calcu- 10, with 10 associated with a better methodological
lated as follows: quality. A total score is derived from the summation
of responses to items 2–11, and item 1 relates to exter-
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
nal validity. The reliability of this scale was evaluated
ðn1  1ÞS21 þ ðn2  1ÞS22
Spool ¼ ð2Þ with acceptable good results in interclass correlation
n1 þ n2  2
coefficients 0.56–0.91.13,15

where n1 and n2 are pre and post-intervention sample

sizes, or experimental versus control group sample Results
sizes; S1 and S2 are pre and post-intervention standard
deviations, or experimental versus control group stand-
Methodological quality
ard deviations. According to Cohen, an effect size of 0.2 PEDro scores ranged from 5 to 8 points, with a
represents a small effect, 0.5 represents a moderate mean score of 6.14 (Table 1). A high quality
effect and 0.8 represents a large effect.12 of the chosen trials has been assured because all of
For meta-analysis, calculations were done using a the selected studies scored 5 or more. Moreover, all
random effect method. The level of significance was of the studies specified the eligibility criteria, and
set at P < 0.05. Statistical heterogeneity of the treat- the subjects were randomly allocated to groups
ment effects among studies was assessed using except in two;16,17 all were blinded to allocation and
Cochran’s Q test and the inconsistency I2 test, in showed the similarities at baseline. In fact, one of
1694 European Journal of Preventive Cardiology 25(16)

Table 1. Methodological classification of studies using the PEDro scale and clinical characteristics of patients randomised to training
group (TG) or control group (CG) in the different selected studies.

First author PEDro scale N Mean age EF NYHA PImax (cmH2O)

Johnson, 199818 6 TG ¼ 8 70  4.6 12 in II and 6 in III 70  33

CG ¼ 8 63.4  4.5 84  18
Weiner, 199920 6 TG ¼ 10 66  4 24.7  1.6 II–III 46.5  4.7
CG ¼ 10 63  4 22.9  2.4 II–III 50.7  4.2
Laoutaris, 200416 5 TG ¼ 20 57.6  2.3 23.4  1.5 60% II and 40% III 82.8  5.7
CG ¼ 15 60  2.6 25.7  2.1 47% II and 53% III 78.4  6.9
Dall’Ago, 20067 7 TG ¼ 16 54  33 39  3 60.2  9

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024

CG ¼ 16 58  2 38  3 58.1  1
Laoutaris, 200717 5 TG ¼ 15 53  2 28  3 10 in II and 5 in III 79.8  4.7
CG ¼ 23 59  2 27  2 12 in II and 11 in III 80.2  5
Padula, 200919 6 TG ¼ 15 76 (51–89) 30 5 in II and 7 in III 48  25
CG ¼ 17 73 (32–95) 33 9 in II and 6 in III 52  27
Bosnak-Guclu, 20118 8 TG ¼ 16 69.50  7.96 33.43  7.23 11 in II and 5 in III 62.00  33.57
CG ¼ 14 65.71  10.52 36.11  7.62 9 in II and 5 in III 78.64  35.95
EF: ejection fraction; NYHA, New York Heart Association; PImax: maximal inspiratory pressure.

the trials had blinded therapists: one of them had

blinded participants5 and two had blinded assessors.4,5
IMT methods
All trials had retention rates of 85% or greater In this study, two IMT methods were analysed: inspira-
and none met the intention-to-treat analysis criteria. tory resistive training and threshold loading. The pres-
Statistical analysis was applied to group differences sure threshold device uses a spring-loaded one-way
in all the studies as well as reporting point estimates valve that places consistent resistance on the inspira-
and measurements of variability. None of the tory muscles.
studies was excluded on the basis of their methodo- However, resistive training provides an endurance
logical quality. training effect as the patient is exercising at a percent-
The initial research yielded 65 RCTs. Of these, 33 age of sustained PImax (SMIP) (pressure–time product).
studies were excluded based on the examination of the The mode of IMT in most studies was threshold
title or abstract, and 32 were considered as potentially training (Philips Respironics, Andover, MA, USA) at
relevant and retrieved for detailed analysis. an intensity ranging from 30% to 60% of PImax, dur-
An additional 25 studies were excluded for different ation from 15 to 30 minutes for an average of 10–12
reasons (Figure 1). The results of each review have been weeks and 6–7 days per week. In the literature, two
favourable regarding many of the pathophysiological studies performed by the same investigators16,17 had
manifestations of CHF and the mechanisms and effects administered IMT using the test of incremental respira-
of IMT. tory endurance (TIRE) by means of the TRAINAIR
Overall, a total of seven articles with 203 patients device (Project Electronics Ltd., Kent, UK) at 60% of
were included in the systematic review. PImax/SMIP with an increasing work to rest ratio for 20
The sample size for all studies was small, with an to 30 minutes, three times per week. TRAINAIR IMT
average of 29 and the mean age of subjects ranged was performed to the point of exhaustion.16,17
from 53 to 76 years, which means that all studies Although all studies used different protocols of
focused on middle-aged and older adults because training in terms of PImax, frequency and duration, stat-
CHF affects older people in particular. istically significant increases in PImax values were
Concerning systolic dysfunction, EF was less than detected in all studies. Almost all of them showed
40% and this is obvious as the main abnormality in improvements in dyspnoea and walking test (Table 2).
CHF is impaired left ventricular function and a low
cardiac output.
Meta-analysis
In addition, PImax values ranged from 48 to
82 cmH2O and this has been arbitrarily defined by Figures 2, 3 and 4 show that there was a significant
respiratory muscle weakness that is prevalent in CHF improvement in PImax, walking test and dyspnoea
patients. when comparing IMT versus all control groups
 Sadek et al.

Table 2. Study measurements and outcomes of different studies of inspiratory muscle training in heart failure patients randomly assigned to training group (TG) or control group
(CG).

Mode of the intervention

First author N Load (device, duration) Study measurements Study outcomes used for this review

Johnson, 199818 TG ¼ 8 30% PImax Threshold for 8 weeks (15 min, PImax, exercise duration, Treatment group:
CG ¼ 8 15% PImax 2*/d, 7*/wk) 100 m walking test,  PImax (cmH2O): þ 25.4 (11.2)
QOL  Treadmill test time: þ 152 s (144)y
Weiner, 199920 TG ¼ 10 60% PImax Threshold for 12 weeks (30 min, PImax, PEmax, IME, PFTs, Treatment group:
CG ¼ 10 No load 6*/wk) peak VO2, 12MWT,  PImax (cmH2O): 46.5 (4.7) to 63.6 (4)z
dyspnoea  12MWT (m): 458 (29) to 562 (32)z
 dyspnoea index: 1.7 (0.2) to 2.7 (0.2)z
Laoutaris, 200416 TG ¼ 20 60% SMIP TRAINAIR for 10 weeks (3*/wk PImax, SMIP, peak VO2, Treatment group:
CG ¼ 15 15% SMIP for an uncertain duration/ 6MWT, dyspnoea,  PImax (cm H2O): 82.8 (5.7) to 111.9 (6.8)^
training to exhaustion) QOL  6MWD (m): 367.1 (22.3) to 433.4 (16.9)^
 Borg dyspnoea scale after 6-min walking test:
10.5 (0.7) to 9 (0.5)^
Dall’Ago, 20067 TG ¼ 16 30% PImax Threshold for 12 weeks (30 min, PImax, IME, 6MWT, dys- Treatment group:
CG ¼ 16 No load 7*/wk) pnea, QOL, VO2  Pthmax (%PImax) : 57 (9) to 66 (7)z
 6MWD (m): 449 (17) to 550 (17)z
 Dyspnoea during 6MWT: 3.7(2) to 1.5 (1.4)z
 VO2: 17% increase
Laoutaris, 200717 TG ¼ 15 60% SMIP TRAINAIR for 10 weeks (3*/wk PImax, SMIP, PFTs, peak High intensity group:
CG ¼ 23 15% SMIP for an uncertain duration/ VO2, 6MWT, dyspnoea  PImax (cmH2O): 79.8 (4.7) to 105.1(4.9)^
training to exhaustion)  6MWD (m): 378.2 (10.4) to 404.3 (11.9)z
 Dyspnea: Borg scale 9.2 (0.4) to 8.0 (0.4)z
Padula, 200919 TG ¼ 15 30% of PImax Threshold for 12 weeks (10– PImax, dyspnoea, QOL Treatment group:
CG ¼ 17 Education on 20 min, 7*/wk)  PImax (cmH2O): 48 (25) to 78 (37)^
self-efficacy  Decreased dyspnoea (Borg scale)y
Bosnak-Guclu, 20118 TG ¼ 16 40% PImax Threshold for 6 weeks (30 min, PImax, PEmax, 6MWT, dys- Treatment group
CG ¼ 14 15% PImax 7*/wk) pnea, PFTs, depression,  PImax (cmH2O): 62 (33) to 97 (32)z*
QOL  6MWD(m): 418(123) to 478(131)^*
 D Borg Dyspnea scale: 2.42(1.7) to 1.42(1.3)y
y
: < 0.05; z: < 0.01; ^: < 0.001; *: between groups; each value in the table was expressed as mean (standard deviation).
6MWT: 6-minute walking test; 6MWD: 6-minute walking distance; 12MWT: 12-minute walking test; HR: heart rate; IME: inspiratory muscle endurance; IMT: inspiratory muscle training; PEmax: maximal
expiratory pressure; PImax: maximal inspiratory pressure; PFTs: pulmonary function tests; QOL: quality of life; SMIP: sustained maximal inspiratory pressure.
1695

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024

1696 European Journal of Preventive Cardiology 25(16)

IMT Control
PImax Mean SD Total Mean SD Total

Bosnak 2011 35 32.5 16 12.4 33.2 14

Johnson 1998 25.4 11.2 8 12.3 12.1 8

Laoutaris 2004 29.1 6.3 20 8.2 6.6 15

Laoutaris 2007 25.3 4.8 15 10.1 7.4 23

Weiner 1999 17.1 4.4 10 0 4.2 10

Padula 2009 29.8 31.9 15 0.3 27.8 17

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024

Dallago 2006 9 8.1 16 1 14.5 16

Study name Statistics for each study Difference in means and 95% CI
Difference Lower Upper Relative
in means limit limit weight
Bosnak 2011 22,600 –0,946 46,146 2,03
Johnson 1998 13,100 1,675 24,525 7,39
Laoutaris 2004 20,900 16,596 25,204 24,34
Laoutaris 2007 15,200 10,963 19,437 24,63
Weiner 1999 17,100 13,330 20,870 26,70
Padula 2009 29,500 8,821 50,179 2,59
Dallago 2006 8,000 –0,138 16,138 12,31
16,575 13,129 20,021
–55,00 –27,50 0,00 27,50 55,00
Favours control Favours IMT

Heterogeneity: tau2 =7.877,chi2 =10.615,df= 6 (P = 0.001);I 2 =43.478%

Test for overall effect: Z = 9.427(P<0.00001)

Figure 2. Maximal inspiratory pressure for treatment with inspiratory muscle training versus control groups.
CI: confidence interval; IMT: inspiratory muscle training; SD: standard deviation.

(P < 0.00001, P ¼ 0.001 and P < 0.00001, respectively). because it is performed only three times a week
The chi-square test for heterogeneity was significant for (Table 3).
PImax (P=0.001) and for dyspnoea and Walking test
(P< 0.00001). The I2 value was approximately 43% for
PImax, 96% for the walking test and 85% for dyspnoea.
PImax
These quantitative results suggest there was study vari- For PImax, there was a remarkable improvement in all
ability (i.e. heterogeneity). studies discussed when comparing IMT versus all con-
trol groups (not shown here). However, a greater bene-
fit was shown for a high intensity–high duration
Outcomes of the different categories
programme (4.088 cmH2O) when compared to
We have divided the studies into four programmes the three other programmes (3.132 for high intensity–
according to their load of PImax and duration. moderate duration, 0.929 for low intensity–high
In fact, we have considered a high intensity pro- duration and 2.27 for low intensity–low duration)
gramme for those who provide training of more than (Figure 5).
50% PImax, and a low intensity programme for those
who provide training of less than 50% PImax.
Dyspnoea
Moreover, low duration was attributed to 8 weeks max-
imum, and high duration to 12 weeks. And as for the 10 As for dyspnoea, the most marked improvement
weeks training, it is considered as moderate duration was also for the high intensity–high duration
Sadek et al. 1697

IMT Control
Dyspnoea Mean SD Total Mean SD Total

Bosnak 2011 –1 1.52 16 –0.87 1.17 14

Laoutaris 2004 –1.9 0.61 20 –0.1 0.8 15

Laoutaris 2007 –1.2 0.4 15 0.5 0.6 23

Weiner 1999 –1 0.2 10 0 0.2 10

Dallago 2006 –2.2 1.73 16 –0.1 1.35 16

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024

Study name Statistics for each study Difference in means and 95% CI

Difference Lower Upper Relative

in means limit limit weight
Bosnak 2011 –0,130 –1,112 0,852 13,84
Laoutaris 2004 –1,800 –2,267 –1,333 22,47
Laoutaris 2007 –1,700 –2,048 –1,354 24,47
Weiner 1999 –1,000 –1,175 –0,825 26,62
Dallago 2006 –2,100 –3,175 –1,025 12,60
–1,369 –1,888 –0,850

–3,50 –1,75 0,00 1,75 3,50

Favours IMT Favours control

Heterogeneity: tau2 =0.255,chi2 =26.649,df= 4 (P < 0.00001);I 2 =84.990%

Test for overall effect: Z = – 5.174 (P< 0.00001)

Figure 3. Dyspnoea with inspiratory muscle training versus control groups.

CI: confidence interval; IMT: inspiratory muscle training; SD: standard deviation.

programme (5) whereas less improvement was shown depending on the load of PImax adjusted, the frequency,
for the other programmes (3.166 for high intensity– and the duration of the intervention.
moderate duration, 1.427 for low intensity–high dur-
ation and 0.139 for low intensity–low duration)
Low intensity–low/high duration
(Figure 6).
Starting with the two low intensity programmes, in the
studies by Johnson18 and Padula et al.,19 the improve-
Walking tests ment in PImax was not important as much as it was in
With respect to the distance covered in the walking tests other studies because maybe a higher percentage than
(6MWT and 12MWT), it was improved in all studies 30% was required for these patients. In addition, we
but with additional benefits for the low intensity–high must consider the possibility that the sham training
duration programme (4.514 m), while it was improved performed by the control group had a training effect
for 0.907 m in the high intensity–high duration pro- which caused no significant difference between both
gramme, 3.217 m for the high intensity–moderate train- groups.18
ing programme and 0.36 m for the low intensity–low According to the study by Bosnak-Guclu et al.
duration programme (Figure 7). (2011)8 (low intensity–low duration), the improve-
ment in all parameters was small when compared
to other studies, and this was due to the short dur-
Discussion
ation of intervention and the low intensity of
Meta-analysis demonstrated that IMT results have training.
marked an improvement in PImax, dyspnoea and walk- Moreover, we noticed that Dall’Ago et al.,7 (low
ing tests, but they were varying between studies intensity–high duration) who had performed IMT on
1698 European Journal of Preventive Cardiology 25(16)

IMT Control
Walking test Mean SD Total Mean SD Total

Bosnak 2011 59.97 127.51 16 13.68 134.69 14

Laoutaris 2004 66.3 19.79 20 8.4 23.49 15

Laoutaris 2007 26.1 11.18 15 8 13.64 23

Weiner 1999 104 30.54 10 0 31 10

Dallago 2006 101 17 16 –21 51.39 16

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024

Study name Statistics for each study Difference in means and 95% CI

Difference Lower Upper Relative

in means limit limit weight

Bosnak 2011 46,290 –47,596 140,176 10,82

Laoutaris 2004 57,900 43,548 72,252 22,90
Laoutaris 2007 18,100 9,813 26,387 23,32
Weiner 1999 104,000 77,029 130,971 21,45
Dallago 2006 122,000 95,477 143,523 21,51
71,042 29,018 113,066

–150,00 –75,00 0,00 75,00 150,00

Favours control Favours IMT

Heterogeneity: tau2 =1953.857,chi2 =91.047,df= 4 (P < 0.00001);I 2 =95.607 %

Test for overall effect: Z = 3.313(P = 0.001)

Figure 4. Distance walked in the walking test with inspiratory muscle training versus control groups.
CI: confidence interval; IMT: inspiratory muscle training; SD: standard deviation.

Table 3. Distribution of the studies into different programmes.

Programme First author Load Duration

High intensity–moderate duration Laoutaris, 200416 60% 10 weeks, 3 times/week

Laoutaris, 200717
High intensity–high duration Weiner, 199920 60% 12 weeks, 6 times/week
Low intensity–low duration Johnson, 199818 30% 8 weeks, 7 times/week
Bosnak-Guclu, 20118 40% 6 weeks, 7 times/week
Low intensity–high duration Dall’Ago, 20067 30% 12 weeks, 7 times/week
Padula, 200919

patients with inspiratory muscle weakness for 12 weeks,

and had a control group that received placebo IMT
High intensity–moderate/high duration
without load inspiratory, showed greater gains from For those who have set a load of 60% PImax, Weiner
the distance in the 6MWT. This might be caused by et al.20 has shown the best results in terms of PImax and
the greater degree of impairment of the patients in dyspnoea, and this was related to the long duration of
these trials, the longer duration of the intervention, intervention and the frequency of sessions that the
and the fact that we had a comparison between the authors had taken.
IMT group and the control group that received the However, Weiner et al.20 did not find any increase in
placebo IMT without inspiratory load. the aerobic capacity when compared to Laoutaris and
Sadek et al. 1699

time, because the opening valve closes at pressures below

Plmax (cmH20)
4.088 the set threshold usually at the mid to end inspiratory
3.132
range, inhibiting patients from training at the full
inspiratory range (especially at higher training intensi-
2.27
ties). However, IMT using electronic systems of resistive
loading (e.g. TRAINAIR) with pressure-time profiles
0.929
may provide training at %PImax/SMIP, enabling
Low intensity Low intensity
patients to ‘sustain’ pressure in time and train to the
High intensity High intensity
high duration moderate high duration low duration end-inspiratory range even at high training intensities.
duration So, increasing the endurance of respiratory muscles is
equally if not more important than increasing the

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024

Figure 5. Improvement of maximal inspiratory pressure strength because SMIP was demonstrated to be more
(cmH2O) in the four programmes. significantly depressed compared to PImax in CHF.21
We observed that IMT carried out for 12 weeks
improved PImax more than training carried out for 6–
8 weeks when compared to controls, and according to
Dyspnea
Weiner et al.,20 six times per week of training was the
High intensity, High intensity Low intensity, Low intensity,
high duration moderate duration high duration low duration
most desirable.
Therefore, training for 12 weeks and six times per
week was proved to be the best protocol.
–0.139
–1.427 Finally, the studies included samples of different
sizes for short/long training duration, utilisation of
–3.166 relatively low/high intensity IMT, treated/untreated
control groups, inclusion of patients with/without
–5 inspiratory muscle weakness, blinded/unblinded stu-
dies, etc., but despite the variety of studies, gains were
Figure 6. Improvement of dyspnoea (Borg scale 6–20) in the higher in the high-intensity training group. So, we can
four programmes. consider that 60% is the best intensity to apply.
Overall, six sessions per week with an IMT set at
60% PImax, for 30 minutes and for 12 weeks is the opti-
mal programme so far.
Walk test (m)
4.514 Limitations
3.217 The improvement identified in the control group was
the result of the training effect performed by the control
group, because four studies have used the threshold
0.907
0.36 loading of 15% PImax as a sham therapy.
High intensity High intensity
Thus, all studies should be designed in a way that the
Low intensity Low intensity
high duration moderate high duration low duration control group performs without any inspiratory load.
duration In addition, studies included a variety of patients,
some of them suffered from inspiratory muscle weakness
Figure 7. Improvement of the distance covered in the walking and the others did not. Therefore, this fact can contrib-
test (m) in the four programmes. ute to a risk of bias because the weakness of respiratory
muscles has contributed to exercise intolerance and
could have influenced the results of the training.
colleagues16,17 as evidenced by the unchanged peak Moreover, this review must include a greater number
oxygen consumption (17.3  0.9 to 19.4  0.4) ml/kg/ of RCTs in order to obtain a more generalised conclu-
minute versus (17.3  0.9 to 19.4  0.4) ml/kg/minute sion that covers the effect of IMT on several outcomes.
in Laoutaris and colleagues.16,17 The heterogeneity found in this meta-analysis can be
This may be related to the different training devices attributed to between-study differences in patient selec-
used. In fact, when IMT is performed using a mechanical tion, study design, clinical setting and supervised and
pressure-threshold trainer (e.g THRESHOLD), patients non-supervised IMT.
are required abruptly to develop a pressure above a fixed Despite these restrictions, it is still useful to explore
threshold value (%PImax) and not to ‘sustain’ pressure in sources of heterogeneity and incorporate it explicitly in
1700 European Journal of Preventive Cardiology 25(16)

the analysis in order to explain it and remove it in fur- 4. Clark A. Origin of symptoms in chronic heart failure.
ther studies. Besides, selecting no workloads as a sham Heart 2006; 92: 12–16.
therapy will/would be better while investigating the 5. Gibbs C. ABC of heart failure: non-drug management.
effects of IMT to avoid the training effect. BMJ 2000; 320: 366–369.
6. Plentz R. Inspiratory muscle training in patients with
heart failure: meta-analysis of randomized trials. Arq
Conclusion Bras Cardiol 2012; 99: 762–771.
7. Dall’Ago P, Chiappa G, Guths H, et al. Inspiratory
In conclusion, this study has provided evidence that muscle training in patients with heart failure and inspira-
training the inspiratory muscles at 60% PImax, six tory muscle weakness. JACC J Am Coll Cardiol 2006; 47:
times per week for 12 weeks is the best protocol used 757–763.
in order to improve inspiratory muscle strength, walk- 8. Bosnak-Guclu M, Arikan H, Savci S, et al. Effects of

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024

ing distance and dyspnoea for patients with CHF. inspiratory muscle training in patients with heart failure.
These positive effects advocate for application of IMT Respir Med 2011; 105: 1671–1681.
as a safe and effective intervention in cardiac rehabili- 9. Mancini D, Henson D, Manca J, et al. Benefit of selective
tation programmes. respiratory muscle training on exercise capacity in
patients with chronic congestive heart failure.
Moreover, healthcare professionals should receive
Circulation 1995; 91: 320–329.
more guidance in order to put a specific treatment pro-
10. Bisset B, Leditschke I, Paratz J, et al. Respiratory dys-
gramme for each single patient into practice, learn IMT function in ventilated patients: can inspiratory muscle
theory and use it properly. training help? Anaesth Intens Care 2012; 40: 236–246.
For this reason, a digital training and decision sup- 11. Lin S. Inspiratory muscle training in patients with heart
port system for exercise prescription in cardiovascular failure: a systematic review. Cardiopulm Phys Ther J
disease patients was recently developed: the European 2012; 23: 29–36.
Association of Preventive Cardiology Exercise 12. Hurley WL, Denegar CR and Hertel J. Research
Prescription in Everyday Practice and Rehabilitative Methods: A Framework for Evidence-based Clinical
Training (EXPERT) tool. Practice, 1st ed. Baltimore, MD: Wolters Kluwer/
Lippincott Williams & Wilkins, 2011, p.426.
13. Maher C and Sherrington C. Reliability of the PEDro
Author contribution scale for rating quality of randomized controlled trials.
Zahra Sadek and Charifa Awada contributed to the conception Phys Ther 2003; 83: 713–721.
and design of the work. Ali Salami contributed to the analysis 14. Verhagen AP, de Vet HC, de Bie RA, et al. The Delphi
and interpretation of data for the work. Zahra Sadek and list: a criteria list for quality assessment of randomized
Charifa Awada drafted the manuscript. Wissam H Joumaa, clinical trials for conducting systematic reviews developed
Wiam Ramadan and Said Ahmaidi critically revised the manu- by Delphi consensus. J Clin Epidemiol 1998; 51:
script. All authors gave final approval and agree to be account- 1235–1241.
able for all aspects of the work ensuring integrity and accuracy. 15. Foley N, Bhogal S and Teasell R. Estimates of quality
and reliability with the physiotherapy evidence-based
Declaration of conflicting interests database scale to assess the methodology of randomized
controlled trials of pharmacological and nonpharmacolo-
The author(s) declared no potential conflicts of interest with gical interventions. Phys Ther 2006; 86: 817.
respect to the research, authorship, and/or publication of this 16. Laoutaris I, Dritsas A, Brown M, et al. Inspiratory
article. muscle training using an incremental endurance test alle-
viates dyspnea and improves functional status in patients
Funding with chronic heart failure. Eur J Cardiovasc Prev Rehabil
The author(s) received no financial support for the research, 2004; 11: 489–496.
authorship, and/or publication of this article. 17. Laoutaris I, Dritsas A, Brown M, et al. Immune response
to inspiratory muscle training in patients with chronic
heart failure. Eur J Cardiovasc Prev Rehabil 2007; 14:
References 679–685.
1. Johnson M, McDonagh T, Harkness A, et al. Morphine 18. Johnson P. A randomized controlled trial of inspiratory
for the relief of breathlessness in patients with chronic muscle training in stable chronic heart failure. Eur Heart
heart failure-a pilot study. Eur J Heart Fail 2002; 4: J 1998; 19: 1249–1253.
753–756. 19. Padula C, Yeaw E and Mistry S. A home-based nurse-
2. Clark A, Poole-Wilson P and Coats A. Exercise limitation coached inspiratory muscle training intervention in heart
in chronic heart failure: central role of the periphery. J Am failure. Appl Nurs Res 2009; 22: 18–25.
Coll Cardiol 1996; 5: 1092–1102. 20. Weiner P, Magadle R, Berar-Yanay N, et al. The effect of
3. Nilsson K, Duscha B, Hranitzky P, et al. Chronic heart specific inspiratory muscle training on the sensation of
failure and exercise intolerance: the hemodynamic para- dyspnea and exercise tolerance in patients with congestive
dox. Curr Cardiol Rev 2008; 2: 92–100. heart failure. Clin Cardiol 1999; 22: 727–732.
Sadek et al. 1701

21. Laoutaris I, Adamopoulos S, Manginas A, et al. Prescription in Everyday Practice and Rehabilitative
Inspiratory work capacity is more severely depressed Training (EXPERT) tool: a digital training and decision
than inspiratory muscle strength in patients with heart support system for optimized exercise prescription in car-
failure: novel applications for inspiratory muscle training. diovascular disease. Concept, definitions and construc-
Int J Cardiol 2016; 221: 622–626. tion methodology. Eur J Prev Cardiol 2017; 24:
22. Hansen D, Dendale P, Coninx K, et al. The European 1017–1031.
Association of Preventive Cardiology Exercise

Downloaded from https://academic.oup.com/eurjpc/article/25/16/1691/5926433 by guest on 01 November 2024