Journal of Biomechanics 82 (2019) 142–148

Contents lists available at ScienceDirect

Journal of Biomechanics
journal homepage: www.elsevier.com/locate/jbiomech
www.JBiomech.com

Influence of altered geometry and material properties on tissue stress

distribution under load in tendinopathic Achilles tendons – A subject-
specific finite element analysis
Vickie B. Shim b,⇑, Wencke Hansen a, Richard Newsham-West a, Leila Nuri a, Steven Obst c,
Claudio Pizzolato a, David G. Lloyd a, Rod S. Barrett a
a
Menzies Health Institute Queensland, Griffith University, Gold Coast Campus, Australia
b
Auckland Bioengineering Institute, University of Auckland, Auckland, New Zealand
c
School of Health, Medical and Applied Sciences, Central Queensland University, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Achilles tendon material properties and geometry are altered in Achilles tendinopathy. The purpose of
Accepted 20 October 2018 this study was to determine the relative contributions of altered material properties and geometry to free
Achilles tendon stress distribution during a sub-maximal contraction in tendinopathic relative to healthy
tendons. Tendinopathic (n = 8) and healthy tendons (n = 8) were imaged at rest and during a sub-
Keywords: maximal voluntary isometric contraction using three-dimensional freehand ultrasound. Images were
Ultrasound manually segmented and used to create subject-specific finite element models. The resting cross-
Tendinopathy
sectional area of the free tendon was on average 31% greater for the tendinopathic compared to healthy
Young’s modulus
Cross-sectional area
tendons. Material properties for each tendon were determined using a numerical parameter optimisation
Strain approach that minimised the difference in experimentally measured longitudinal strain and the strain
von Mises predicted by the finite element model under submaximal loading conditions for each tendon. The mean
Young’s modulus for tendinopathic tendons was 53% lower than the corresponding control value. Finite
element analyses revealed that tendinopathic tendons experience 24% less stress under the same sub-
maximal external loading conditions compared to healthy tendons. The lower tendon stress in
tendinopathy was due to a greater influence of tendon cross-sectional area, which alone reduced tendon
stress by 30%, compared to a lower Young’s modulus, which alone increased tendon stress by 8%. These
findings suggest that the greater tendon cross-sectional area observed in tendinopathy compensates for
the substantially lower Young’s modulus, thereby protecting pathological tendon against excessive stress.
Ó 2018 Published by Elsevier Ltd.

1. Introduction properties (Arya and Kulig, 2010; Helland et al., 2013). For the
Achilles tendon, stiffness was reduced by 20% and Young’s modu-
Achilles tendinopathy is a common tendon injury which is lus by 50% relative to healthy tendons (Arya and Kulig, 2010). For
prevalent in, but not exclusive to, athletic populations a tendon of a given length, a lower stiffness would result in greater
(Kongsgaard et al., 2005; Rolf and Movin, 1997). Up to 25% of tendon strain under the same external load. A further adaptation in
affected patients will eventually require operative treatment and Achilles tendinopathy is tendon thickening, which increases the
20% of those undergo further surgery (Alfredson, 2003). Achilles tendon cross-sectional area (CSA) by 20–66% (Arya and Kulig,
tendinopathy is most commonly located in the mid-portion of 2010; Helland et al., 2013; Leung and Griffith, 2008; Nuri et al.,
the free tendon, 2–6 cm from the calcaneal insertion (Astrom and 2017b, 2018; Obst et al., 2018). Tendon thickening may therefore
Rausing, 1995; Maffulli et al., 2004; Rolf and Movin, 1997) and is be an adaptation that at least partially compensates for the reduc-
characterised by collagen bundle disruption, hyper-cellularity, tion in tendon stiffness, thereby protecting the tendon from high
hyper-vascularity and altered collagen, glycosaminoglycan and strains. For a given external load, an increased tendon CSA would
fluid content (Khan et al., 1999). Tendinopathy has also been also be expected to decrease the net tendon stress. In the study
reported to cause a reduction in tendon mechanical and material by Arya and Kulig (2010), peak Achilles tendon stress (force/CSA)
was approximately half in tendinopathic compared to healthy
⇑ Corresponding author. Achilles tendons, primarily due to much larger CSA in tendinopa-
E-mail address: [email protected] (V.B. Shim). thy. However, it is currently unclear how altered subject-specific

https://doi.org/10.1016/j.jbiomech.2018.10.027
0021-9290/Ó 2018 Published by Elsevier Ltd.
 V.B. Shim et al. / Journal of Biomechanics 82 (2019) 142–148 143

material properties, localised geometry, and their interaction affect ramped isometric contraction over 3–5 s with 1–2 min of rest.
the stress distribution of tendinopathic Achilles tendons. Three-dimensional freehand ultrasound (Barber et al., 2009) was
Localised stress of healthy Achilles tendons have been assessed used for tendon imaging at rest and during submaximal contractions
using subject-specific finite element (FE) models as the relative held for approximately 30 s. The contraction intensity was 70%MVIC
effects of variations in tendon geometry and material properties for participants with healthy tendons. A contraction intensity of 50%
on localised tendon stress cannot be assessed experimentally. Shim MVIC was chosen for participants with tendinopathic tendons to
and colleagues (Shim et al., 2014) combined three-dimensional ensure they could complete the task without excessive discomfort.
reconstructions of the Achilles tendon geometry from transverse Real-time visual feedback of torque was provided to ensure the
ultrasound images with corresponding load-elongation data of ankle torque was kept within ±5% of the target load.
fresh-frozen cadaver tendons and reported that tendon stress distri- The three-dimensional ultrasound system consisted of a two-
butions during loading were more sensitive to variations in tendon dimensional ultrasound (SonixTouch, Ultrasonix, Richmond, Bri-
geometry than variations in material properties. Similarly, Hansen tish Columbia, Canada) combined with a five-camera optical track-
et al. (2017) reported that in-vivo free Achilles tendon stress magni- ing system (Optitrack V100:R2, Tracking Tools v2.5.2, NaturalPoint,
tude was more influenced by tendon geometry than material prop- Corvallis, OR, USA) with tracking accuracy <1 mm. All scans were
erties. Shim et al. (Shim et al., 2018) further reported that the performed by a single investigator using a single transverse sweep
Achilles tendon may experience non-uniform tissue deformation between the osteotendinous insertion and the musculotendinous
between tendon regions when in vivo measures of sub-tendon junction. Ultrasound images were acquired using a 58-mm linear
geometry and tendon twist were incorporated in their FE model. transducer with a central frequency of 10 MHz, sampling fre-
Taken together, the abovementioned FE studies suggest that the quency of 40 Hz and standardised B-mode image settings. The
mechanical behaviour of the healthy Achilles tendon under load is two-dimensional images were spatially aligned using Stradwin
particularly sensitive to tendon geometry. However, these results software (Version 5.2) to form a reconstructed image stack. The
cannot be generalised to tendinopathic Achilles tendons, which three-dimensional point accuracy of spatial calibrations using an
are characterised by alteration in both geometry and material optical tracking system has been reported to be less than 1 mm
properties. (Hsu et al., 2006; Treece et al., 2003). Prior studies have established
The purpose of this study was to determine the relative contribu- the validity and repeatability of the three-dimensional ultrasound
tion of altered geometry and material properties to tendon stress in system for in vivo measurement of Achilles free tendon morphol-
tendinopathic compared to healthy free Achilles tendons. We ogy (length, volume, CSA, AP diameter and ML diameter) under
hypothesised that the larger CSA in tendinopathic Achilles tendons passive and active loading conditions (Obst et al., 2014a; Obst
would have a greater influence on tissue stress distribution than the et al., 2014b). Free Achilles tendon volume reconstructions were
reduction in stiffness, thereby resulting in an overall stress performed based on at least 6 segmented cross-sections using
reduction. the volume algorithms provided in Stradwin (Treece et al., 1999)
(Fig. 1). The average CSA of the free tendon was obtained from
2. Methods the ratio of measured free tendon volume to free tendon length.

2.1. Participant characteristics 2.3. Mesh generation

Eight participants with mid-portion Achilles tendinopathy (6 The surface geometries of each free AT were imported from
males, 2 females; age: 37 ± 11 years, height: 177 ± 9 cm, weight: Stradwin into 3matics 9.0 (Materialise, Leuven, Belgium). The lon-
78 ± 13 kg) and 8 age-matched controls with normal, healthy gitudinal axes of inertia were aligned with the x-axis of the global
Achilles tendons (5 males, 3 females; age: 29 ± 4 years, height: coordinate system and point cloud data subsequently exported to
180 ± 6 cm, weight: 75 ± 14 kg) participated in the study. The eligi- the FE analysis software Continuum Mechanics Image analysis, Sig-
bility criteria for the tendinopathy group were symptomatic mid- nal processing and System Identification (CMISS) (www.cmiss.org,
potion Achilles tendinopathy (self-reported tendon pain, tender- freely available for academic use).
ness to palpation, focal thickening), minimum duration of symp- A point cloud describing the free AT surface geometry for a sep-
toms of 3 months, and a Victorian Institute of Sports Assessment– arate healthy participant (female, 32 years, 177 cm, 64 kg) was
Achilles tendon (VISA-A) score of less than 80 points (Robinson generated from a transverse, high-accuracy magnetic resonance
et al., 2001). Potential participants were excluded if they had inser- imaging scan (Philips Ingenia 3T, slice thickness 3.5 mm, voxel size
tional tendinopathy, neurological or musculoskeletal conditions 0.6 mm  0.6 mm, 50 slices, slice gap 0.3 mm) using the software
affecting the lower limbs, or had undergone lower limb surgery. Mimics (Research 17.0, Materialise, Leuven, Belgium). The point
The study was approved by the Institutional Human Research Ethics cloud was used to create a generic finite element mesh made up
Committee and all relevant ethical guidelines, including provision of three-dimensional cubic Hermite element in CMISS, which was
of a written informed consent prior to participation in the study, then morphed to match the subject-specific mesh of each partici-
were followed. pants Achilles tendon using the freeform deformation method
(Shim et al., 2016). The average root mean square (RMS) error
2.2. Experimental data collection and analysis procedures between the subject-specific point cloud and the morphed mesh
was 0.21 ± 0.07 mm. Element distortions occurred during morph-
All data were collected from the right Achilles tendon of each par- ing were removed via manual adjustment. An average free tendon
ticipant, who were positioned prone with knee extended and ankle geometry was computed for all healthy (n = 8) and tendinopathic
in neutral position rigidly fixed to an isometric dynamometer tendons (n = 8) and used for subsequent analysis.
(TFF600, Futek, Irvine, CA, USA). Particular care was taken in order
to align ankle and dynamometer axes of rotation. Tendon was pre- 2.4. Boundary conditions, material properties and finite element
conditioned following a standardised procedure, which included analysis
two sub-maximal contractions followed by four maximal contrac-
tions (Nuri et al., 2017a). Maximum voluntary isometric contraction Force boundary conditions were derived from the measured
(MVIC) torque was then determined as the greatest ankle plan- ankle plantarflexion torque divided by a generic ankle joint
tarflexion torque generated from 3 MIVC trials, which involved moment arm (Kongsgaard et al., 2011). The direction of the force
144 V.B. Shim et al. / Journal of Biomechanics 82 (2019) 142–148

Fig. 1. (Left) Reconstructed three-dimensional ultrasound image of the free Achilles tendon showing (Top) segmentation lines for six trans-axial image slices, and (Bottom)
fully rendered free Achilles tendon. (Right) Final finite element meshes of tendinopathic (Top) and healthy (Bottom) tendons.

vector was derived from individual muscle volumes of the Triceps mulation has five coefficients (c1–c5). Values c2 and c3 were fixed.
Surae based on one subject’s magnetic resonance imaging scan As the ground substance matrix was considered a Neo-Hookean
(Fig. 2). The resulting boundary load was equally distributed across material, c2 was set to 0. The scaling parameter of the strain energy
the 11 nodes of the proximal face. The nodes in the distal face were functionc3 was obtained from literature(Shim et al., 2014) and set
fixed. to 19.4. The remaining three parameters were optimised. c1
The free Achilles tendon was treated as incompressible, trans- describes the scaling of ground substance, c4 describes the rate of
versely isotropic and hyperelastic tissue (Hansen et al., 2017; the collagen fibre loading and c5 represents the Young’s modulus.
Shim et al., 2014) based on the exponential strain energy function While c4 describes the tendon behaviour in the toe region, c5 influ-
for hyperelastic material developed by Weiss et al. (1996). This for- ences the behaviour in the tendon’s elastic, linear region. The mate-

Fig. 2. (Left) A diagram showing the applied boundary and loading conditions. Applied force vectors were computed from individual force vectors of Soleus (S),
Gastrocnemius Medialis (GM) and Gastrocnemius Lateralis (GL). The force was assigned to the nodes of the proximal face (green) and the displacement boundary condition
was assigned to the nodes of the distal face (blue). (Right) Von Mises stress [MPa] for tendinopathic tendons under subject-specific 50% MVIC load with subject-specific
geometry and material properties (Top row) and healthy tendons under subject-specific 70% MVIC load with subject-specific geometry and material properties. Red (blue)
indicates the greatest (lowest) stress. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
 V.B. Shim et al. / Journal of Biomechanics 82 (2019) 142–148 145

rial parameter optimization was performed using the non-linear slopes of the regression between von Mises tendon stresses at all
least squares lsqnonlin algorithm in Matlab’s optimisation tool 4000 Gauss points were influenced by each combination of
box (The MathWorks, USA), as per our previous studies (Hansen subject-specific versus average geometry and material properties.
et al., 2017; Shim et al., 2014). The optimization routine iteratively
altered the values of parameters c1, c4, and c5 for the unloaded AT
3. Results
mesh to minimise the difference between FE predicted shape of
the AT and the experimental AT shape obtained from ultrasound
There were no significant group differences in the free AT length
images during the 50%MVIC contraction. Multiple initial parameter
(Healthy AT: 62.4 ± 12.7 mm, Tendinopathic AT: 68.5 ± 16.1,
guesses were used to ensure a global minima was found for each
t = 0.84, p = 0.42) or volume (Healthy AT: 3617 ± 984 mm3,
tendon. Details of the parameter optimization process can be found
Tendinopathic AT: 5341 ± 2239 mm3, t = 1.99, p = 0.07). However
in our previous works (Hansen et al., 2017; Shim et al., 2014).
the average resting CSA was significantly higher for the tendino-
pathic compared to healthy ATs (Healthy AT: 57.5 ± 10.7 mm2,
2.5. Stress analysis Tendinopathic AT: 75.2 ± 17.8 mm2, t = 2.41, p = 0.03).
The maximum voluntary isometric ankle plantarflexion torque
Internal tissue stress was initially computed at more than 4000 was significantly higher for the healthy compared to tendinopathic
internal data points located at numerical integration points called tendons (Healthy AT: 141 ± 37 Nm, Tendinopathic AT: 103 ± 35 Nm,
Gauss points for each subject-specific tendon geometry and corre- t = 0.84, p = 0.049). The free AT strain measured using freehand
sponding subject-specific material properties at a load correspond- three-dimensional ultrasound was 5.9 ± 1.3% at 70%MVIC for the
ing to the 70%MVIC condition for the healthy ATs (Fig. 2) measured healthy tendons, and 7.1 ± 2.9% at 50%MVIC for the tendinopathic
from the experiments. The average values for measured torques for tendons.
healthy and tendinopathic AT are given in Table 1. High stresses in The subject-specific boundary loads applied in the optimisation
the regions where boundary conditions were applied were not to determine the free AT material properties were 2288 ± 605 N for
included in accordance with Saint-Venant’s principle (Horgan the healthy tendons (70%MVIC) and 1197 ± 406 N for the tendino-
and Knowles, 1983). pathic tendons (50%MVIC). The average subject-specific material
von Mises stresses were subsequently computed for different parameter values were significantly higher for the healthy tendons
combinations of average healthy and average tendinopathic geom- compared to the tendinopathic tendons for c1 (Healthy AT:
etry and material properties at a load corresponding to the 70% 36 ± 10, Tendinopathic AT: 11 ± 7, t = 5.80, p < 0.01) and c4
MVIC condition for the healthy ATs. The influence of tendinopathic (Healthy AT: 17 ± 11, Tendinopathic AT: 4 ± 2, t = 3.19, p = 0.02).
material properties on the internal stress distribution of the free The Young’s modulus (c5 ) was significantly lower for the tendino-
tendon was examined by plotting stress distributions based on pathic tendon (555 ± 176 MPa) compared to the healthy tendon
average tendinopathic material properties against stress distribu- (1184 ± 342 MPa) (t = 1.02, p < 0.01).
tions from average healthy material properties both using average The stress distributions for the tendinopathic tendons under
healthy geometry. The influence of tendinopathic geometry on load (Fig. 2) showed a high degree of variability in peak stress
internal stress distribution was examined by plotting stress distri- magnitude and stress location between tendons. The peak stress
butions from average tendinopathic geometry against average magnitude was significantly lower for the tendinopathic tendons
healthy geometry both using average healthy material properties. (28.9 ± 12.3 MPa) compared to the healthy tendons
The combined effect of average tendinopathic geometry and mate- (70.2 ± 13.2 MPa) (t = 6.50, p < 0.01). No differences in peak stress
rial properties on localised free AT stress was also investigated by location were detected between groups (Tendinopathy: 56 ± 16%
plotting against stress distributions from average healthy geome- length from calcaneus to soleus muscle-tendon junction (MTJ);
try and material properties. A further analysis was performed to Healthy: 67 ± 22% length from calcaneus to soleus MTJ) (t = 1.13,
examine how subject-specific compared to average geometry, p = 0.28).
and subject-specific versus average material properties altered The change in local stress caused by the introduction of patho-
localised AT stresses for tendinopathic tendons. logical properties to the healthy Achilles tendon is displayed in
Fig. 3. Replacing the healthy material properties with tendino-
2.6. Statistical analysis pathic material properties resulted in increased overall tendon
stress as indicated by the slope of the regression line which was
Independent sample t-tests (Assuming Unequal Variances) were greater than 1 (Fig. 3A). In contrast, replacing the average healthy
used to compare outcome measures between groups. A two-tailed geometry with the average tendinopathic geometry resulted in an
p-value < (a = 0.05) indicated a significant difference between the overall decrease in tendon stress as indicated by the regression
cohorts. Linear regression analysis was performed using the Lin- slope which was below 1 (Fig. 3B). When the average geometry
earModel.fit algorithm in MATLAB R2013a to determine how the and material properties for the healthy tendon were replaced by
average values for tendinopathic tendon, the slope of the regres-
sion line remained below 1, indicating that tendinopathic geome-
Table 1 try has a greater influence on tendon stress than the
Experimental measures of Achilles tendon resting morphology and strain under load
tendinopathic material properties.
for healthy and tendinopathic groups (n = 8 per group). *indicates significantly
different from healthy control (p = 0.03). For the tendinopathic tendons, tendon stress was found to be
more sensitive to variations in geometry assessed at each individ-
Healthy Tendinopathy
uals subject-specific material properties (Fig. 4A) than variations in
At rest material properties when assessed at each individuals subject-
Length [mm] 62.4 ± 12.7 68.5 ± 16.1
specific tendon geometry (Fig. 4B).
Volume [mm3] 3617 ± 984 5341 ± 2239
Average cross-sectional area [mm2] 57.5 ± 10.7 75.2 ± 17.8*
Under load 4. Discussion
Maximum voluntary isometric 140.5 ± 37.2 102.5 ± 35.1
ankle plantarflexor torque [Nm]
Longitudinal strain [%] 5.9 ± 1.3 7.1 ± 2.9
Previous studies reported that mid portion Achilles tendinopa-
thy is characterised by tendon thickening (Arya and Kulig, 2010;
146 V.B. Shim et al. / Journal of Biomechanics 82 (2019) 142–148

[A] [B] [C]

Local stresses with tendinopathic mat. Prop (MPa)

Local stresses with tendinopathic geometry (MPa)

Local stresses with tendinopathic geometry &

m=1.079 m=0.698 m=0.763

Tendinopathic mat. prop.(MPa)

Local stresses with average healthy geometry and healthy mat. prop. (MPa)

Fig. 3. Scatterplots and linear regression lines of local stresses (MPa) in the healthy tendon with healthy material properties compared to (A) with tendinopathic material
properties, (B) with tendinopathic geometry and (C) with tendinopathic geometry and tendinopathic material properties. The regression line and its slope (m) are also
displayed in each panel. The grey identity line represents perfect agreement.

[A] Same subject-specific tendinopathic [B] Same subject-specific tendinopathic

mat. prop. geometries
Local stresses with subj.-specific

Local stresses with subj.-specifc

tendinopathic geometry (MPa)

mat. Prop (MPa).

tendinopathic

m= 0.92 ± 0.39 m= 1.03 ± 0.12

Local stress with average tendinopathic Local stresses with average tendinopathic
geometry (MPa) mat. Prop. (MPa)

Fig. 4. Scatterplots and linear regression lines of local stresses (MPa) for combinations of subject-specific versus average geometry and material properties for tendinopathic
tendons. Comparisons are for (A) tendinopathic tendon with subject-specific versus average geometry at the same subject-specific material properties, and (B) for subject
specific tendinopathic tendons with subject-specific versus average material properties.

Nuri et al., 2017b, 2018) and altered material properties (Arya and been shown to be a critical determinant of tendon stress in healthy
Kulig, 2010). This study quantified the relative contributions of ten- Achilles tendons (Hansen et al., 2017).
don thickening and decreased material properties on tissue stress Our FE analysis based on subject-specific estimates of free AT
under load in tendinopathic compared to healthy Achilles tendons. morphology and material properties revealed that tendon stress
Subject-specific FE analyses revealed that tendinopathic Achilles computed at over 4000 Gauss points across the tendon was on
tendons experienced less stress at submaximal loading compared average 24% lower in the tendinopathic compared to the healthy
to healthy tendons. Linear regression analysis showed that tendino- tendons. The lower tendon stress in tendinopathy was due to a
pathic alterations in tendon CSA had a greater influence on local greater influence of altered tendon CSA, which reduced tendon
tendon stress than changes in Young’s modulus. These findings sug- stress by 30%, compared to material parameters, which increased
gest that the greater tendon CSA observed in tendinopathy compen- tendon stress by 8%. These findings indicate that tendon thickening
sates for the substantially lower Young’s modulus, and thereby in tendinopathy is a favourable biological adaptation that compen-
protects pathological tendon against excessive stress. sates for reduced tendon mechanical and material properties and
The resting free tendon’s CSA in the present study was on aver- thereby protects the tendon from excessive localised stress
age 31% greater for the tendinopathic compared to the healthy ten- (Docking and Cook, 2016).
dons, which is within the 20–66% range reported in other studies Hansen et al. (2017) and Shim et al. (2014) previously demon-
(Arya and Kulig, 2010; Nuri et al., 2017b, 2018). The mean free strated that stress distribution in healthy Achilles tendon is more
Achilles tendon volume was similarly greater in the tendinopathy sensitive to geometry than material properties. In the present
group in the present study, although this volume difference rela- study, subject-specific variations in tendon geometry were found
tive to the healthy tendons only approached statistical significance to have a greater effect on localised tendon stress computed at
(p = 0.07). We also found, via numerical optimisation, that the each point across the tendon than subject-specific variation in
mean Young’s modulus for pathological tendons was 53% lower material properties. When considered together with the findings
than the corresponding value for healthy tendon. Arya and Kulig from Hansen et al. (2017) and Shim et al. (2014), the findings of
(2010) similarly reported a Young’s modulus in Achilles tendinopa- the present study suggest that the stress experienced by the free
thy that was approximately half that for healthy tendons. It was Achilles tendon is highly dependent on subject-specific tendon
also notable from our study that there was a substantial degree geometry, and therefore that personalised three-dimensional ten-
of inter-subject variability in free Achilles tendon geometry and don geometry should be an essential feature of efforts to model
material properties within the tendinopathy group, which has Achilles tendon stress.
 V.B. Shim et al. / Journal of Biomechanics 82 (2019) 142–148 147

One important practical implication of these findings is that Arya, S., Kulig, K., 2010. Tendinopathy alters mechanical and material properties of
the Achilles tendon. J. Appl. Physiol. 1985 (108), 670–675.
generic rehabilitation exercises commonly prescribed to treat
Astrom, M., Rausing, A., 1995. Chronic Achilles tendinopathy. A survey of surgical
tendinopathy, such as eccentric heel drops (Obst et al., 2016), are and histopathologic findings. Clin. Orthop. Relat. Res., 151–164
likely to result in very different Achilles tendon stresses and strains Barber, L., Barrett, R., Lichtwark, G., 2009. Validation of a freehand 3D ultrasound
for different patients. Tendon strain is an important mechanical system for morphological measures of the medial gastrocnemius muscle. J.
Biomech. 42, 1313–1319.
trigger for the biological cascade regulating tendon structural Bogaerts, S., Desmet, H., Slagmolen, P., Peers, K., 2016. Strain mapping in the
adaptation (Mehdizadeh et al., 2017). The findings of this study Achilles tendon – a systematic review. J. Biomech. 49, 1411–1419.
therefore point to the need for personalised training and rehabili- Chatzistergos, P., Maganaris, C., Chockalingam, N., 2016. Sensitivity of a numerical
model to detect regional differences in mechanical properties of tendons. J. Foot
tation for the Achilles tendon that takes account of the substantial Ankle Surg. 22, 15.
variation in tendon geometry and material properties between Choi, R.K., Smith, M.M., Martin, J.H., Clarke, J.L., Dart, A.J., Little, C.B., Clarke, E.C.,
individuals to ensure an optimal loading stimulus is provided that 2016. Chondroitin sulphate glycosaminoglycans contribute to widespread
inferior biomechanics in tendon after focal injury. J. Biomech. 49, 2694–2701.
maximises positive tissue adaptation for healthy and tendino- Docking, S.I., Cook, J., 2016. Pathological tendons maintain sufficient aligned fibrillar
pathic tendons (Pizzolato et al., 2017a). structure on ultrasound tissue characterization (UTC). Scand. J. Med. Sci. Sports
There are several limitations to our study. Firstly, we assumed 26, 675–683.
Hansen, W., Shim, V.B., Obst, S., Lloyd, D.G., Newsham-West, R., Barrett, R.S., 2017.
that the Achilles tendon material properties were homogeneous, Achilles tendon stress is more sensitive to subject-specific geometry than
which does not reflect the possibility that tendon material proper- subject-specific material properties: a finite element analysis. J. Biomech. 56,
ties could be anisotropic (Bogaerts et al., 2016) and thereby influ- 26–31.
Helland, C., Bojsen-Moller, J., Raastad, T., Seynnes, O.R., Moltubakk, M.M., Jakobsen,
ence the local stress-strains, particularly in the region of the lesion
V., Visnes, H., Bahr, R., 2013. Mechanical properties of the patellar tendon in
(Chatzistergos et al., 2016). However as the spatial variations of elite volleyball players with and without patellar tendinopathy. Br. J. Sports
Achilles tendon material properties are as yet unknown, and may Med. 47, 862–868.
even be homogenous in tendinopathy (Choi et al., 2016), this Horgan, C.O., Knowles, J.K., 1983. Recent developments concerning Saint-Venant’s
Principle. In: Hutchinson, J.W., Wu, T.Y. (Eds.), Advances in Applied Mechanics.
assumption seemed justified in the first instance. Secondly, Elsevier, pp. 179–269.
although we were able to match global strains from our model to Hsu, P.W., Prager, R.W., Gee, A.H., Treece, G.M., 2006. Rapid, easy and reliable
experimental data, validation of local strain estimates remains calibration for freehand 3D ultrasound. Ultrasound. Med. Biol. 32, 823–835.
Khan, K.M., Cook, J.L., Bonar, F., Harcourt, P., Astrom, M., 1999. Histopathology of
challenging (Maganaris et al., 2017). We also assumed that the common tendinopathies. Update and implications for clinical management.
boundary force applied to our finite element model could be repre- Sports Med. 27, 393–408.
sented by the measured ankle joint plantarflexion moment divided Kongsgaard, M., Aagaard, P., Kjaer, M., Magnusson, S.P., 2005. Structural Achilles
tendon properties in athletes subjected to different exercise modes and in
by the Achilles tendon moment arm, which neglects the effects of Achilles tendon rupture patients. J. Appl. Physiol. 1985 (99), 1965–1971.
antagonistic co-contraction and assumes the calculated force is Kongsgaard, M., Nielsen, C.H., Hegnsvad, S., Aagaard, P., Magnusson, S.P., 2011.
fully transmitted to the Achilles tendon. Future studies may need Mechanical properties of the human Achilles tendon, in vivo. Clin. Biomech.
(Bristol, Avon) 26, 772–777.
to include subject-specific estimates of Achilles tendon moment Lersch, C., Grotsch, A., Segesser, B., Koebke, J., Bruggemann, G.P., Potthast, W., 2012.
arm which may also be assessed using freehand three- Influence of calcaneus angle and muscle forces on strain distribution in the
dimensional ultrasound (Obst et al., 2017) and include more human Achilles tendon. Clin. Biomech. (Bristol, Avon) 27, 955–961.
Leung, J.L., Griffith, J.F., 2008. Sonography of chronic Achilles tendinopathy: a case-
detailed muscle force representations via musculoskeletal mod-
control study. J. Clin. Ultrasound. 36, 27–32.
elling approaches (Pizzolato et al., 2015; Pizzolato et al., 2017b). Maffulli, N., Sharma, P., Luscombe, K.L., 2004. Achilles tendinopathy: aetiology and
The effect of more detailed representation of tendon microstruc- management. J. R. Soc. Med. 97, 472–476.
ture including tendon tissue architecture and tendon twist Maganaris, C.N., Chatzistergos, P., Reeves, N.D., Narici, M.V., 2017. Quantification of
internal stress-strain fields in human tendon: unraveling the mechanisms that
(Pekala et al., 2017; Shim et al., 2018), as well as the effect of cal- underlie regional tendon adaptations and mal-adaptations to mechanical loading
caneal rotation (Lersch et al., 2012), may also need to be included and the effectiveness of therapeutic eccentric exercise. Front. Physiol. 8, 91.
in future models. It is also conceivable that such models could Mehdizadeh, A., Gardiner, B.S., Lavagnino, M., Smith, D.W., 2017. Predicting
tenocyte expression profiles and average molecular concentrations in Achilles
operate in near real-time and help guide tendon training programs tendon ECM from tissue strain and fiber damage. Biomech. Model.
that seek to use targeted mechanical loading as a way of optimis- Mechanobiol. 16, 1329–1348.
ing training adaptation (Pizzolato et al., 2017a; Pizzolato et al., Nuri, L., Obst, S.J., Newsham-West, R., Barrett, R.S., 2017a. Regional three-
dimensional deformation of human Achilles tendon during conditioning.
2017c). Scand. J. Med. Sci. Sports 27, 1263–1272.
In conclusion, this study demonstrated that tendon thickening Nuri, L., Obst, S.J., Newsham-West, R., Barrett, R.S., 2017b. The tendinopathic
can be viewed as a positive biological adaptation in tendinopathy Achilles tendon does not remain iso-volumetric upon repeated loading: insights
from 3D ultrasound. J. Exp. Biol. 220, 3053–3061.
that compensates for diminished tendon material properties and Nuri, L., Obst, S.J., Newsham-West, R., Barrett, R.S., 2018. Three-dimensional
helps protect the tendon against excessive stress. Our overall find- morphology and volume of the free Achilles tendon at rest and under load in
ings also confirm the strong influence of tendon geometry on ten- people with unilateral mid-portion Achilles tendinopathy. Exp. Physiol. 103,
358–369.
don stress distribution and reaffirm the importance of
Obst, S.J., Barber, L., Miller, A., Barrett, R.S., 2017. Reliability of achilles tendon
incorporating subject-specific tendon geometry in efforts to esti- moment arm measured in vivo using freehand three-dimensional ultrasound. J.
mate Achilles tendon stress. Appl. Biomech. 33, 300–304.
Obst, S.J., Heales, L.J., Schrader, B.L., Davis, S.A., Dodd, K.A., Holzberger, C.J., Beavis, L.
B., Barrett, R.S., 2018. Are the mechanical or material properties of the achilles
Conflicts of interest and patellar tendons altered in tendinopathy? A systematic review with meta-
ANALYSIS. Sports Med. 48, 2179–2198.
Obst, S.J., Newsham-West, R., Barrett, R.S., 2014a. In vivo measurement of human
The authors have no conflicts of interest related to this study. achilles tendon morphology using freehand 3-D ultrasound. Ultrasound. Med.
Biol. 40, 62–70.
Obst, S.J., Newsham-West, R., Barrett, R.S., 2016. Changes in Achilles tendon
Acknowledgement
mechanical properties following eccentric heel drop exercise are specific to the
free tendon. Scand. J. Med. Sci. Sports 26, 421–431.
This work was funded by the Australian Research Council Link- Obst, S.J., Renault, J.B., Newsham-West, R., Barrett, R.S., 2014b. Three-dimensional
deformation and transverse rotation of the human free Achilles tendon in vivo
age Grant (LP110100581)
during isometric plantarflexion contraction. J. Appl. Physiol. 1985 (116), 376–
384.
References Pekala, P.A., Henry, B.M., Ochala, A., Kopacz, P., Taton, G., Mlyniec, A., Walocha, J.A.,
Tomaszewski, K.A., 2017. The twisted structure of the Achilles tendon
unraveled: a detailed quantitative and qualitative anatomical investigation.
Alfredson, H., 2003. Chronic midportion Achilles tendinopathy: an update on
Scand. J. Med. Sci. Sports 27, 1705–1715.
research and treatment. Clin. Sports Med. 22, 727–741.
148 V.B. Shim et al. / Journal of Biomechanics 82 (2019) 142–148

Pizzolato, C., Lloyd, D.G., Barrett, R.S., Cook, J.L., Zheng, M.H., Besier, T.F., Saxby, D.J., Shim, V.B., Fernandez, J.W., Gamage, P.B., Regnery, C., Smith, D.W., Gardiner, B.S.,
2017a. Bioinspired technologies to connect Musculoskeletal Mechanobiology to Lloyd, D.G., Besier, T.F., 2014. Subject-specific finite element analysis to
the person for training and rehabilitation. Front. Comput. Neurosci. 11, 96. characterize the influence of geometry and material properties in Achilles
Pizzolato, C., Lloyd, D.G., Sartori, M., Ceseracciu, E., Besier, T.F., Fregly, B.J., Reggiani, tendon rupture. J. Biomech. 47, 3598–3604.
M., 2015. CEINMS: a toolbox to investigate the influence of different neural Shim, V.B., Besier, T.F., Lloyd, D.G., Mithraratne, K., Fernandez, J.F., 2016. The
control solutions on the prediction of muscle excitation and joint moments influence and biomechanical role of cartilage split line pattern on tibiofemoral
during dynamic motor tasks. J. Biomech. 48, 3929–3936. cartilage stress distribution during the stance phase of gait. Biomech. Model.
Pizzolato, C., Reggiani, M., Modenese, L., Lloyd, D.G., 2017b. Real-time inverse Mechanobiol. 15, 195–204.
kinematics and inverse dynamics for lower limb applications using OpenSim. Shim, V.B., Handsfield, G.G., Fernandez, J.W., Lloyd, D.G., Besier, T.F., 2018.
Comput. Methods Biomech. Biomed. Eng. 20, 436–445. Combining in silico and in vitro experiments to characterize the role of
Pizzolato, C., Reggiani, M., Saxby, D.J., Ceseracciu, E., Modenese, L., Lloyd, D.G., fascicle twist in the Achilles tendon. Sci. Rep. 8, 13856.
2017c. Biofeedback for gait retraining based on real-time estimation of Treece, G.M., Gee, A.H., Prager, R.W., Cash, C.J., Berman, L.H., 2003. High-definition
tibiofemoral joint contact forces. IEEE Trans. Neural. Syst. Rehabil. Eng. 25, freehand 3-D ultrasound. Ultrasound. Med. Biol. 29, 529–546.
1612–1621. Treece, G.M., Prager, R.W., Gee, A.H., Berman, L., 1999. Fast surface and volume
Robinson, J.M., Cook, J.L., Purdam, C., Visentini, P.J., Ross, J., Maffulli, N., Taunton, J.E., estimation from non-parallel cross-sections, for freehand three-dimensional
Khan, K.M., Victorian Institute Of Sport Tendon Study, G., 2001. The VISA-A ultrasound. Med. Image Anal. 3, 141–173.
questionnaire: a valid and reliable index of the clinical severity of Achilles Weiss, J.A., Maker, B.N., Govindjee, S., 1996. Finite element implementation of
tendinopathy. Br. J. Sports Med. 35, 335–341. incompressible, transversely isotropic hyperelasticity. Comput. Methods Appl.
Rolf, C., Movin, T., 1997. Etiology, histopathology, and outcome of surgery in Mech. Eng. 135, 107–128.
achillodynia. Foot Ankle Int. 18, 565–569.