Journal Logo


Is the Soleus a Sentinel Muscle for Impaired Aerobic Capacity in Heart Failure?


Author Information
Medicine & Science in Sports & Exercise: March 2015 - Volume 47 - Issue 3 - p 498-508
doi: 10.1249/MSS.0000000000000431


Impaired exercise capacity is a hallmark symptom of chronic heart failure (CHF) (8,9). The important contribution of peripheral, as against cardiac, factors in the functional limitation, which characterizes CHF has long been recognized and is reinforced by the finding that ejection fraction is poorly correlated with exercise capacity (9). Moreover, it seems that dysfunctional skeletal muscles have a direct influence on the reduced exercise capacity evident in patients with CHF (8) and may contribute to progressive deterioration in clinical status. These findings resulted in the “muscle hypothesis” of CHF (8), which proposes that skeletal muscle dysfunction is a key factor responsible for reduced exercise capacity in this group.

Consistent with the “muscle hypothesis”, muscle wasting is common in patients with CHF (8,24), especially in advanced disease. It has been found that limb muscle mass is correlated with reduced exercise capacity (peak oxygen consumption (V˙O2peak)) (7). Indeed, a recent analysis of 200 patients with CHF (12) revealed that reduced muscle mass is independently associated with lower absolute V˙O2peak (mL·min−1) when age, sex, ejection fraction, and comorbidities are accounted for and that muscle mass is directly linked to disease progression.

Skeletal muscle wasting is therefore highly relevant to the progression and treatment of CHF. However, although overall reduction in skeletal muscle mass in CHF is evident, a more detailed understanding of the morphology behind muscle wasting and the relation between muscle morphology, muscle functional capacity, (strength), and exercise capacity in CHF remains unexplored. For example, it is not known whether loss of mass occurs evenly across lower limb muscles or whether certain muscles exhibit proportionately more or less muscle wasting. It is conceivable that predominantly slow-twitch muscles may be more markedly affected, given the reported increase in the proportion of Type II fibers (8), the reduced blood flow, and vascular transport capacity in CHF (31) and reports that Type I fiber atrophy is related to severity of the disease (10). Muscle wasting may also be more prominent in the distal leg muscles, given their greater functional role in generating the mechanical work and power for walking and for the maintenance of posture (25). The “architectural” changes that underlie the loss of muscle mass are also poorly understood. Loss of muscle mass may occur through decreased fiber length, with physiological cross-sectional area (PCSA) of the muscle remaining unaffected, or through loss of PCSA, with fiber length remaining unaffected, or a combination of these. The distinction is important, given that whole-muscle architecture is a strong predictor of skeletal muscle mechanical capacity (e.g., strength) (21) and may also effect muscle energetics (35). Finally, the mechanical function of a muscle is also dependent on in-series tendon properties. Tendons function by storing and releasing elastic energy during locomotion, thereby influencing the contractile behavior of muscle, including work and power production and efficiency (20). Tendon mechanics are known to be closely associated with muscle wasting in aging and chronic unloading (32,38). To the best of our knowledge, a detailed structural analysis combining muscle volume, fiber length, and pennation angle along with tendon properties has not previously been undertaken in human CHF, nor have they been analyzed in relation to muscle strength or exercise capacity in this group.

The aims of the present study were therefore as follows: 1) to assess whether muscle wasting in CHF is constant across different muscles of the lower limb, particularly across synergist muscles known to differ in fiber composition and function, 2) to assess the detailed architecture of synergist skeletal muscle and tendon properties in patients with CHF, compared with healthy age-matched control participants (with similar adiposity and exercise activity); 3) to assess how skeletal muscle architecture, muscle strength, and exercise capacity are related among patients with CHF and control participants. To address these questions, we analyzed lower limb lean mass and performed a detailed architectural assessment of the triceps surae (calf muscles) and Achilles tendon, along with functional strength measurements of this muscle group. The triceps surae (soleus (SOL), medial gastrocnemius (MG), and lateral gastrocnemius (LG)) was chosen because it offers a highly relevant comparison between synergist muscles known to possess different fiber types (predominantly slow twitch in the SOL vs mixed fibers in the gastrocnemius) (13). Furthermore, the triceps surae muscles are functionally important, representing the main source of power during walking (25), and have been shown to be a major locus of gait impairment in aging (11,34).



We recruited 11 subjects with CHF (seven men and four women; New York Heart Association II–IV; ejection fraction, 30% ± 10%; age, 61.8 ± 10.0 yr; height, 1.68 ± 0.10 m; weight, 72.8 ± 18.0 kg; body mass index, 25.6 ± 4.8 kg·m−2). Criteria for exclusion included severe renal (creatinine >250 mM or estimated Glomerular Filtration Rate <30 mL·min−1 per 1.73 m2) or hepatic (bilirubin >50 mM) dysfunction, unexplained anemia (hemoglobin <100 g·L−1) or thrombocytopenia (platelets <100 × 109 per liter), unstable angina or exercise-induced ischemia at low exercise levels, severe aortic stenosis, severe mitral or aortic regurgitation, or hypertrophic cardiomyopathy. All subjects with CHF were medically stable at the time of testing (no cardiac-related admissions to the hospital in the 3 months before testing), and their medications (see Table, Supplemental Digital Content 1,, which describes the medications taken by the CHF participants) were clinically optimized by medical staff (L. D.). The group with CHF was attending supervised exercise training 2–3 times per week for approximately 1 h per session (treadmill walking, cycle ergometry, and resistance training) as part of their multidisciplinary care. The control group consisted of 15 healthy subjects recruited from the local community (nine men and six women; age, 60.7 ± 6.2 yr; height, 1.72 ± 0.07 m; weight, 69.9 ± 8.6 kg; body mass index, 23.5 ± 2.5 kg·m−2). Eleven of these control subjects were recruited because they matched the group with CHF for age, sex, and adiposity and had levels of exercise similar to those of the patients with CHF and were used in comparative statistics. The additional four controls were recruited specifically to validate the ultrasound measurement of muscle volume (six subjects participated for ultrasound validation). All subjects provided a written informed consent before participating in the study. All procedures were approved by the human research ethics committee at the University of Western Australia (approval ID, RA/4/1/2533) and Royal Perth Hospital (approval ID, 2011/019).

Body composition measurements

Overall body composition was determined using dual energy x-ray absorptiometry (DXA) (Luna Prodigy, encore 2004; GE Medical Systems, Madison, WI) on each subject including a measurement of total fat-free mass, total fat mass, and bone mineral density. Lean mass for the lower limbs was computed separately by selecting a region of interest from the greater trochanter to the pubic symphysis and included the leg and foot.

Triceps surae volume calculation

Because of the presence of internal cardioverter defibrillators in 10 of the 11 subjects with CHF, rendering magnetic resonance imaging (MRI) unsuitable, muscle volume was computed using a three-dimensional ultrasound technique (3DUS) on the basis of combination of B-mode ultrasound imaging and 3D motion data (2). Subjects kneeled with their right leg in a custom-designed plastic water bath equipped with an adjustable metal footplate designed to secure the subject’s foot and set the ankle joint angle to 0° (i.e., the foot perpendicular to the tibia), with the knee in 10° flexion (near-maximal extension). The water bath was maintained within a temperature range of 24°C–25°C to ensure that temperature-dependent variations in the speed of sound were minimized.

Ultrasound images (Echoblaster128; Telemed, Lithuania; 7.5-MHz linear-array probe) and 3D marker trajectories of a probe-mounted marker cluster that defined the probe position in space (five-camera Vicon MX system, 250 Hz; Oxford Metrics, United Kingdom) were used to calculate muscle volume. The ultrasound and motion capture systems were synchronized using a 3-V square pulse generated at the beginning of the images collection by the ultrasound device. The subjects’ lower leg was scanned longitudinally from the popliteal cave to the calcaneus. Three sweeps were performed to scan the whole triceps surae, starting at the lateral side and ending medially so that each sweep overlapped by a minimum of 20 mm. A customized MATLAB script (MathWorks, Inc.) was then used to combine ultrasound images and 3D position data and to convert them to Stradwin format (Medical Imaging Research Group, Cambridge University Engineering Department, United Kingdom) (2). Manual segmentation of the muscles was performed in Stradwin software by a single investigator (F. A. P.) on approximately 30 slices covering the length of the muscle group (interslice gap, approximately 15 mm). Muscle volumes for the three muscles under investigation (SOL, MG, and LG) were computed. The segmentation procedure and analysis of muscle volumes were conducted in a blinded manner to avoid operator bias. To account for the possible variability during data collection and manual segmentation, the collection procedure described previously was repeated three times for each subject and an average volume for the SOL, MG, and LG was computed.

We conducted comparisons between our ultrasound volume measurements and those obtained from axial plane MRI scans of the lower leg in a subsample (n = 6) of control subjects who were able to undergo this form of imaging. MRI scans were collected using a Magnetom Espree 1.5-T scanner (Siemens, Erlangen, Germany) with the subjects’ leg position matched to that of the 3DUS experiment (see Text, Supplemental Digital Content 2,, which describes the full details of the MRI procedures). The accuracy of the cross-sectional area (CSA) of the Achilles tendon was also assessed by comparing ultrasound- and MRI-derived images.

Fascicle length, pennation angle, and muscle PCSA

The ultrasound probe described previously was first placed over the middle of the selected muscle belly (SOL, MG, and LG) following the guidelines of Rubenson et al. (36), with its longitudinal axis aligned with the orientation of the fascicle. Fascicle length and pennation angle at neutral ankle (0°) and 10° knee flexion joint angles (approximating the angle where passive fascicle forces and net joint torque approach zero [1,36]) were calculated in ImageJ, as per Rubenson et al. (36). For better characterization of the muscle anatomy, fascicle lengths of the SOL and MG muscles were also calculated in the proximal, medial, and distal locations of the muscle, with approximately 4-cm spacing between each location.

The assessment of regional muscle lengths was performed only on the MG and SOL to shorten the testing time and avoid discomfort of prolonged experimentation in the group with CHF. For each muscle, PCSA was calculated as follows:

Fascicle length and pennation angles used for the PCSA calculation were obtained at the midbelly position of the muscle from images collected in the 3DUS water bath experiment. PCSA was expressed in square centimeters and subsequently normalized to body surface area (BSA) calculated as per Mosteller (30).

Tendon length and CSA

A scaled subject-specific musculoskeletal model in OpenSim 2.0.2 (1) was used to estimate the muscle–tendon unit lengths for each subject at 0° of plantarflexion and 10° knee flexion. Tendon length, defined as the free tendon plus aponeurosis (series elastic element), was subsequently calculated as the difference between the estimated muscle–tendon unit length of each of the SOL, MG, and LG obtained by the scaled OpenSim model and the estimated muscle length from ultrasound of the respective muscles (fascicle length × cos(&thetas;)) at an ankle angle of 0° (Fig. 1). An average of the three tendon lengths was obtained with this procedure (see Text, Supplemental Digital Content 3,, which describes the triceps surae tendon length estimates).

Schematic illustration of the procedure for obtaining triceps surae tendon length. A. Subject marker set used for motion capture of skeletal elements. B. Subject-specific scaled OpenSim model depicting the MG and SOL muscle tendon units. C. Ultrasound measurement of muscle fascicle length and pennation angle. D. Prediction of tendon length (see Text, Supplemental Digital Content 2,, which describes the tendon length estimate).

Using the same setup and settings described for the determination of 3D muscle volume, ultrasound images of Achilles tendon CSA were taken at the level of the medial malleolus to standardize the measurements across participants. To obtain tendon CSA, images were manually digitized in ImageJ.


Strength measurements were assessed using a robotic dynamometer (M3; Biodex, Shirley, NY). Subjects sat with their right foot strapped to a custom-built footplate designed to align the center of rotation of the ankle with that of the Biodex and to minimize heel movement during the contraction. Subjects performed three maximum voluntary contractions with the knee extended (10° flexion) and the ankle dorsiflexed to 10°, approximating the angle at which maximum torque is produced at the ankle joint (1,36). The participant’s joint angles were obtained using a three-camera OptiTrack motion capture system (NaturalPoint, Corvallis, OR; 100 Hz). A minimum of 2 min rest was observed between repetitions. Peak torque was calculated as the difference between the maximum torque during contraction and the torque measured at rest arising primarily from the weight of the rig and foot and small passive muscle forces (36). Analog data were collected using a Cambridge Electronic Design data acquisition system running Spike2 version 7 software (Micro1401-3; Cambridge Electronic Design, Cambridge, United Kingdom; 2000 Hz) and processed using a custom-written program in MATLAB (MathWorks, Natick, MA). For each subject, the trial with the highest peak torque was reported and used for comparisons.

Exercise capacity

V˙O2peak was assessed using a purpose-designed incremental walking protocol on a motorized treadmill starting at a speed 20% slower than each individual’s preferred speed with stepwise increases in speed and grade every 3 min until the participant reached volitional exhaustion. Indirect calorimetry was conducted using a Vmax Encore gas analysis system (Sensormedics, Yorba Linda, CA), which enabled the measurement of expired gas concentrations and volumes. Absolute V˙O2peak was expressed in milliliters per minute and normalized to body mass (mL·kg−1·min−1). All tests were performed by an exercise physiologist (A. J. M.) experienced with undertaking cardiopulmonary exercise testing in patients with CHF (22).


Differences in muscle volume (absolute and lean body mass-normalized) and PCSA (absolute and BSA-normalized) were analyzed using a two-way between-groups ANOVA with group (CHF/control) and muscle synergist (SOL/MG/LG) as fixed factors. Differences in muscle fascicle length were analyzed using a two-way between-groups ANOVA with group (CHF/control) and muscle synergist (SOL/MG/LG) as fixed factors for the midbelly location. If a significant interaction effect between group and muscle was observed (P < 0.05), a one-way ANOVA with group as a fixed factor was performed for each muscle. A one-way ANOVA was used to determine differences in pennation angle, lower limb lean mass, strength, and tendon properties between groups. ANOVA were performed in SPSS using a Bonferroni post hoc analysis (SPSS Statistics 21; IBM ).

Linear regression was used to determine correlations between V˙O2peak and a) lower limb lean mass, b) muscle volume, c) muscle PCSA (total triceps surae and single muscles), and d) strength and between the muscle PCSA and tendon cross-section. A correlation analysis was also performed between 3DUS- and MRI-derived measurements of muscle volume, and the agreement between the two techniques was assessed using the limits of agreement method (6). Correlation coefficients (r) and significance level (P < 0.05) were determined in SPSS. Values presented are mean ± SD unless otherwise stated.


Subject characteristics

No significant differences in age, height, mass, body mass index, BSA or other anthropometric characteristics were observed between the control group and group with CHF; leg length was 0.38 ± 0.04 m in the group with CHF and 0.40 ± 0.02 m in the control group (P = 0.14); adiposity (fat body mass) was 21.3 ± 11.3 kg in the group with CHF and 17.8 ± 7.2 kg in the control group (P = 0.38) V˙O2peak was 15.5 ± 3.0 mL·kg−1·min−1 in CHF and 35.6 ± 8.3 mL·kg−1·min−1 in control (P < 0.0001). It should be noted that one patient with CHF and two control participants did not undergo strength testing because of discomfort, and in one subject with CHF, image degradation prevented an accurate SOL volume.

DXA lean mass and triceps surae muscle volume and PCSA

Total body lean mass (CHF, 49.0 ± 10.6 kg; control, 49.2 ± 8.3 kg; P = 0.95) and lower limb lean mass (CHF, 7.6 ± 1.9 kg; control, 8.3 ± 1.5 kg; P = 0.4) as assessed by DXA did not significantly differ between groups. The two-way ANOVA determined a main effect of both group and muscle on absolute volume (P = 0.015 and P < 0.0001, respectively) and volume normalized to lean body mass (P = 0.007 and P < 0.0001, respectively). Post hoc analysis found that volumes collapsed over groups were statistically different between all muscles (P < 0.01). The combined muscle volume of the triceps surae was, on average, 27.2% higher in the control group compared with that in the group with CHF (632 ± 169 vs 497 ± 155 mL) (Fig. 2A). A significant interaction effect between group and muscle was found in the lean mass-normalized volume (P = 0.03). Post hoc analysis revealed a significant difference in the normalized SOL muscle volume between the CHF and control group (20.5% smaller) but not in the MG or LG (Fig. 2B). A trend toward an interaction effect was also observed between group and muscle in the absolute muscle volume (P = 0.08) and was largely due to the small difference in LG volume between the CHF and control group (8.7% smaller) compared with the SOL (27.5% smaller) and the MG (20.9% smaller).

Muscle volume and PCSA values for the total triceps surae and individual plantarflexor muscles. A. Absolute volume. B. Lean body mass-normalized volume. C. Absolute PCSA. D. BSA-normalized PCSA. Data are means ± SD. # Indicates a main effect (P < 0.05) (ANOVA) of group (CHF/control; volume and PCSA collapsed over individual triceps surae muscles). * Indicates a statistical difference (one-way ANOVA) between groups (CHF/control). A one-way ANOVA was performed only when an interaction effect between group and muscle was present. TS, triceps surae).

From the volume renderings of the triceps surae muscle group (Fig. 3A), an average difference of 20 ± 46 mL was observed between 3DUS and MRI techniques for the combined triceps surae muscles. This difference amounted to 2.5% ± 6.3% difference between measurements of the total triceps surae volume between the two techniques. The average difference in volume for each individual muscle was 4.2%. The overall correlation coefficient between 3DUS and MRI was 0.989 (Fig. 3B), and the 95% confidence interval ranged between +47 and −33 mL (Fig. 3C).

3DUS- and MRI-derived muscle volume. A. Example of 3D volume rendering of the triceps surae created with 3DUS and with MRI from the same individual. B. The correlation between volume calculation using 3DUS imaging and MRI (equation: y = 0.998x + 9.365; r = 0.988; P < 0.001). The solid line represents the line of best fit, and the dotted lines represent the 95% prediction interval. C. Bland–Altman plot of the difference between 3DUS and MRI versus the average of the MRI and 3DUS values. The horizontal lines on the plot represent the mean difference between MRI and 3DUS and the upper and lower 95% limits of agreement.

No main effect of group on fascicle length or interaction effect between group and muscle was observed in any ANOVA. However, a main effect of muscle on fascicle length was observed for the comparison in the distal muscle location (P = 0.04) (Table 1). A main effect of both group and muscle was found for absolute PCSA (P < 0.0001 for both). A main effect of both group and muscle was also found for BSA-normalized PCSA (P < 0.0001 for both). Post hoc analysis found that PCSA collapsed across groups were statistically different between all muscles (P < 0.01). The PCSA of triceps surae in the control group was, on average, 23.7% higher than in the group with CHF (Fig. 2C). Interaction effects between muscle and group were found in the absolute PCSA (P = 0.042) and in the BSA-normalized PCSA (P = 0.024). Post hoc analysis found significant differences between the SOL and MG absolute PCSA from the group with CHF and the control group but not for the LG. The average absolute PCSA of the SOL and MG were 22.2% and 25.3% smaller in the group with CHF than that in the control group, respectively (Fig. 2C). Similar results were observed for the BSA-normalized PCSA, whereby only the SOL and MG were significantly different between groups (Fig. 2D).

Muscle fascicle length and pennation angle of the SOL, MG, and LG in the group with CHF and the control group.

A significant correlation between total lean lower limb mass and absolute V˙O2peak (mL·min−1) (r = 0.82, P = 0.004) (Fig. 4A) and total triceps surae volume, and absolute V˙O2peak (mL·min−1) (r = 0.93, P < 0.0001) (Fig. 4B) was observed in the group with CHF. When analyzed on an individual muscle level, only the SOL body mass-normalized volume was found to correlate with the body mass normalized V˙O2peak (mL·kg−1·min−1) in the group with CHF (r = 0.72, P = 0.018) (Fig. 4D). The control group exhibited weaker correlations between lean lower limb mass and absolute V˙O2peak and triceps surae volume and absolute V˙O2peak (r = 0.63, P = 0.06; and r = 0.58, P = 0.08, respectively) (Fig. 4A and B). Unlike the group with CHF, body mass-normalized data were not correlated in the control group (Fig. 4C and F). Muscle PCSA was independent of V˙O2peak in all muscles in both the group with CHF and the control group.

Relation between V˙O2peak and leg muscle size. Relation between absolute V˙O2peak (mL·min−1) and lower limb lean mass (kg) (A) and triceps surae volume (mL) (B). Relation between body mass-normalized V˙O2peak and individual muscle volume in the combined triceps surae (C), SOL (D), MG (E), and LG (F). Subjects with CHF are displayed in gray triangles (▵), and control subjects, in open circles (○).

Tendon architectural parameters

Achilles tendon CSA was significantly smaller in patients with CHF compared with that in the control group (59.2 ± 16.1 vs 73.4 ± 20.0 mm2, respectively, P = 0.046). However, the estimate of tendon length was not different between the group with CHF (27.4 ± 2.7 cm) and the control (28.6 ± 2.3 cm) group (P = 0.2). A trend toward a positive correlation between tendon CSA and the triceps surae PCSA existed for the group with CHF (r = 0.65, P = 0.06) but not for the control group (r = 0.35, P = 0.3).

Strength measurements

Body mass-normalized peak torque (N·m·kg−1) was 13.3% greater in the control group vs the group with CHF, although this difference was not statistically different (P = 0.3) (Fig. 5A). Peak torque normalized by the triceps surae PCSA (N·m·cm−2) was 3.6% smaller (nonsignificant) in the control group vs the group with CHF (P = 0.8) (Fig. 5B). A strong correlation between body mass-normalized peak torque and V˙O2peak was found in the group with CHF (r = 0.75, P < 0.01), but there was no such relation in the control group (r = 0.25, P = 0.5) (Fig. 5C).

Peak plantarflexor torque values and their relation with V˙O2peak. Peak plantarflexor torque normalized to body mass (A) and triceps surae PCSA (B). Data are means ± SD. C. Relation between body mass-normalized V˙O2peak and body mass-normalized peak plantarflexor torque. Note that one additional subject was added to the group with CHF that was tested in a parallel study (the correlation based on 11 participants is r = 0.762, with P < 0.01). Subjects with CHF are displayed in gray triangles (▵), and control subjects, in open circles (○).


There is growing evidence that skeletal muscle is adversely affected in patients with CHF and that it contributes to the reduced exercise capacity characteristic of this condition. Skeletal muscle wasting, in particular, has been shown to be strongly correlated with exercise capacity and disease progression in CHF (12), yet little data exist on the specifics of wasting between muscles or the involvement of whole muscle–tendon morphology in muscle wasting. The present study investigated lower limb lean mass and detailed muscle–tendon architecture and strength of the triceps surae muscles (calf muscles), an important functional muscle group for maintaining posture and for locomotion (25). We assessed architectural and strength differences between subjects with CHF and control subjects as well as relations with exercise capacity. Although there were no detectable differences in lower limb lean mass as measured by DXA, a clear reduction in the size of the SOL and to a less extent, the MG, and a reduction in Achilles tendon CSA were observed in the CHF compared with control subjects. However, these differences did not translate to statistically significant differences in strength. Furthermore, SOL muscle volume and plantarflexor strength were found to strongly correlate with exercise capacity in the group with CHF but not in the control group. These findings offer the possibility that the distal lower limb muscles, and the SOL in particular, may be key skeletal muscles determining exercise capacity and function in patients with CHF.

Is muscle wasting in CHF muscle specific?

Although several studies have identified generalized muscle wasting in CHF (12,24), little is known about the specificity of muscle loss. Interestingly, our 3DUS data revealed that muscle wasting was not uniform across the three muscles of the triceps surae. Although the overall reductions in triceps surae muscle volume and PCSA observed in this study are comparable with those in one previous MRI-based report assessing triceps surae volume in CHF (24), our novel observation is that the reductions in these parameters are seen only in the SOL and (with respect to volume) to a somewhat less extent in the MG. The lack of atrophy in the LG might be related to a difference in muscle mechanical function (16) or possibly a consequence of its relatively small size and limited contribution to posture (16). The more pronounced reduction in muscle volume in the SOL compared with the MG in patients with CHF (Fig. 2) (particularly when normalized to lean body mass) could suggest that muscle wasting is sensitive to muscle aerobic capacity and fiber composition, given the known differences in Type I vs Type II fiber distribution between these muscles (13). Greater CHF-induced atrophy in Type I compared with that in Type II fibers has been reported for the human vastus lateralis (40), although the opposite has been found in other work reporting fiber type-specific fiber atrophy in CHF (23). In a rat model of CHF, Type I fiber atrophy has been correlated to the severity of the disease (10) and, recently, significant atrophy in the rat SOL (predominantly Type I fibers) but not the plantaris has been observed (28). Alternatively, the SOL volume may be more affected in CHF because of reduced blood flow (31) or possibly because the SOL has been shown to provide the majority of mechanical work among the lower limb muscles during walking (25), and thus, the effect of CHF may be augmented in this functionally important muscle.

Whereas ultrasound-based measurements of triceps surae volume revealed marked reduction in muscle size (approximately 30%), DXA measurements of total lower limb lean mass were, surprisingly, only slightly lower (nonsignificant) in patients with CHF compared with those in the control group. The approximately 0.7-kg (7.8%) reduction in total lower limb lean mass is similar to that reported by Toth et al. (39), who also found nonsignificant differences in lean mass between CHF and an age-matched control group. When taken together, the high-fidelity 3DUS of the calf muscles and whole-limb DXA recording suggest that the SOL and MG may be particularly prominent and possibly early sites of muscle loss in CHF. Indeed, after normalizing muscle volume to lower limb lean mass, a significant difference in the triceps surae volume remains between the CHF and the control group, indicating proportionately greater wasting in these muscles. Fülster et al. (12) found that 20% of patients with CHF (New York Heart Association class II–III) exhibited muscle wasting compared with a healthy reference group. Using their criteria, only 30% of participants in the current study would be diagnosed with generalized lower limb muscle wasting, yet these patients exhibited muscle wasting in the SOL and MG. It seems likely that muscle-specific changes in the triceps surae may have gone undetected in previous analyses of overall lower limb lean mass using DXA. This may relate to the relatively small size of the triceps surae muscles (approximately 500 g) compared with the lower limb lean mass (approximately 7.5 kg) and the lower sensitivity of DXA (approximately 300-g sensitivity in detecting muscle mass (19) compared with that of 3DUS (approximately 35-mL sensitivity) (Fig. 3C)).

We are unaware of previous studies describing muscle-specific atrophy in CHF. Our data offer the intriguing possibility that calf muscle volume, and the SOL in particular, may be a sentinel muscle for early muscle wasting in this disease. Greater distal muscle atrophy has previously been documented in aging (27) and in other conditions of unloading, such as bed rest (18,32) and space flight (32). However, in these studies, a more prominent reduction in total lower limb lean mass was observed compared with the present study. We propose that a distal-to-proximal gradation in muscle loss may occur because of the prominent role of the distal muscles in body support and propulsion (25). Muscle wasting may be amplified, especially in the major plantarflexor muscles compared with the other leg muscles, including the LG (16), which has lower contribution to these functions in activities of daily living and particularly in walking. The SOL and MG play a key role in supporting the trunk during single-leg stance and pre-swing (33), and the SOL is fundamental for the horizontal acceleration in the late stance phase, which is necessary to provide propulsion to the trunk (33). Because of their importance to gait mechanics, it can be argued that an approximately 25% reduction in normalized muscle volume and PCSA in the SOL will have a negative effect on function. It is reasonable to hypothesize, therefore, that alterations in gait and movement mechanics and fatigue in CHF will result as a consequence of the reduction in size of the triceps surae and especially the SOL.

The relation between muscle–tendon architecture and wasting in CHF

Muscle architecture analyses revealed that the loss of muscle volume is primarily a result of a loss of muscle PCSA, with muscle length and pennation angles remaining relatively unchanged. Age-related muscle wasting, on the other hand, has been reported to result partly from reduction in muscle length (38), although recent work suggests that the active muscle lengths in the MG may be unaffected by age (3). Our findings therefore suggest a pattern of CHF muscle atrophy that is distinct from that associated with aging per se. Given that volume (17) and PCSA (21) influence a muscle’s mechanical work and force capacity, respectively, our findings suggest a biomechanical basis for impaired function of the calf muscles in patients with CHF. It was somewhat surprising, therefore, that a more pronounced reduction in strength was not observed between the CHF and control groups in the present study. Interestingly, a lack of statistically significant differences in ankle strength is consistent with previous measurements of voluntary plantarflexor torque in CHF (14) and possibly reflects variability in voluntary torque measurements arising from factors such as neural drive, cocontraction, muscle moment arm lengths (4), and/or fiber operating lengths (36). Despite the lack of statistical differences in absolute strength, it is also worth noting that the differences between groups are minimized when normalized for PCSA.

Similar to other studies addressing triceps surae muscle loss (18), the current study also showed that the Achilles tendon undergoes a concurrent reduction in CSA in patients with CHF. Tendon remodeling occurs in response to the mechanical loading stimulus (37), and it follows, therefore, that the reduced triceps surae muscle size in the patients with CHF may be associated with reduced habitual in vivo muscle forces.

The relation between muscle architecture, strength, and exercise capacity in CHF

Previous studies have reported significant correlations between estimates of skeletal muscle mass and functional capacity in CHF (12,24). Indeed, several studies have reported correlations between total lower limb lean mass or the size of specific muscle groups and V˙O2peak (typically expressed as absolute V˙O2peak (mL·min−1)). Our findings suggest, for the first time, that a principal reason for such leg muscle mass correlations may relate specifically to the relation between SOL muscle volume and V˙O2peak . We observed correlations between SOL volume and V˙O2peak (absolute and body mass-normalized) in CHF, but no such correlation between V˙O2peak and the MG or LG. Although similar correlations between the combined calf muscle volume and V˙O2peak have been observed (24), our finding reinforces the suggestion that the SOL muscle may be of particular importance as a determinant of functional capacity in patients with CHF. This may relate to the SOL being an important determinant of aerobic potential because of its high oxidative capacity (13) and its key functional role in gait (25). Indeed, the size of the gastrocnemius, a muscle known to have lower oxidative capacity (13), did not correlate with V˙O2peak in our study. It should be noted that similar to the present study, Harrington et al. (15) reported correlations between both quadriceps and thigh CSA and absolute V˙O2peak in CHF. However, the strong correlations observed in this study were likely due, partly, to the high covariance between body mass and muscle size and absolute V˙O2peak. When thigh and quadriceps CSA were correlated to body mass-normalized V˙O2peak, relatively weak correlations were observed compared with those of the SOL in the present study. Furthermore, it should be noted that Harrington et al. (15) assessed quadriceps anatomical CSA from a single axial plane image and thus may not represent the average PCSA of the muscle (typically requiring volume and fiber length and pennation angle measurements). Interestingly, our results also suggest that it may be the reduced muscle volume per se, more so than PCSA, that is linked with reduced V˙O2peak in CHF.

The link between the SOL function and V˙O2peak in CHF is also supported by the strong correlation between plantarflexor torque, for which the SOL is the major contributor, and V˙O2peak in these patients. We are unaware of previous studies correlating plantarflexor strength and V˙O2peak in CHF, although the strength of quadriceps has been previously shown to correlate with V˙O2peak in CHF (26). That both the SOL muscle size and plantarflexor strength are correlated to V˙O2peak only in the group with CHF supports the notion that peripheral determinants of V˙O2peak may be present to a greater extent in patients with CHF compared with that in healthy adults (8).


The present study compared patients with CHF with control subjects matched for age and with similar adiposity. We cannot rule out the possibility that reduced activity levels in the patients with CHF contributed, at least partly, to the reduced muscle and tendon size and the relation between muscle volume, strength, and exercise capacity observed in this study. There are several factors, however, that led us to conclude that any influence of activity levels on our findings is likely to have been minimal. Firstly, all patients with CHF were engaged in an exercise rehabilitation program and were undertaking supervised exercise 2–3 times weekly, similar to the activity of the control participants. Secondly, our finding that the reduction in muscle volume and PCSA as well as the relation between muscle volume and V˙O2peak are highly muscle specific would not be expected as a result of disuse alone, which should logically be expressed generically. For example, although disuse studies have found muscle loss in distal leg muscles (18), they also reported loss in the total lower limb lean mass (18). In contrast, the present study only found a clear loss of muscle in the SOL, with no differences in total lower limb lean mass from DXA. Thirdly, the SOL has been reported to have smaller reduction in size (or no reduction) compared with the MG and LG as a result of aging/disuse (3,29), whereas in the present study on CHF, the SOL presents the largest reduction in size.


The SOL is a key muscle involved in postural control and locomotion. It is classically considered to possess a higher proportion of Type I muscle fibers than other muscles of the triceps surae and lower limb. Our findings indicate that SOL wasting is particularly marked in CHF, even when compared with other lower limb muscles, in these individuals. Furthermore, the SOL volume correlated strongly with exercise capacity, whereas other individual leg muscle volumes did not, and is largely responsible for the correlation between plantarflexor strength and V˙O2peak. For these reasons, we propose that the SOL is a key muscle reflecting loss of function and exercise capacity in CHF and may thus be a sentinel skeletal muscle in patients with CHF. Finally, our results offer an evidence base for including calf muscle-specific exercise training as an important component of whole body training to help restore functional capacity in CHF.

The authors would like to acknowledge Mr. Tony Roby for building the water bath used for three-dimensional ultrasound measurements, Mr. Jack Liddell and Tim Verbeek for their help during data collection, and all of the participants who volunteered for this study.

This work was supported by a Grant-in-Aid (G09P 4469) from the National Heart Foundation of Australia to J. R., D. J. G., A. J. M., and D. G. L. and a thesis dissertation grant from the International Society of Biomechanics to F. A. P.

None of the authors involved in the present study have any conflict of interest, financial, personal, or otherwise, which would influence this research. The results do not constitute endorsement by the American College of Sports Medicine.


1. Arnold EM, Ward SR, Lieber RL, Delp SL. A model of the lower limb for analysis of human movement. Ann Biomed Eng. 2010; 38 (2): 269–79.
2. Barber L, Barrett R, Lichtwark G. Validation of a freehand 3D ultrasound system for morphological measures of the medial gastrocnemius muscle. J Biomech. 2009; 42 (9): 1313–9.
3. Barber LA, Barrett RS, Gillett JG, Cresswell AG, Lichtwark GA. Neuromechanical properties of the triceps surae in young and older adults. Exp Gerontol. 2013; 48 (11): 1147–55.
4. Baxter JR, Piazza SJ. Plantarflexor moment arm and muscle volume predict torque-generating capacity in young men. J Appl Physiol (1985). 2014; 116 (5): 538–44.
5. Besier TF, Sturnieks DL, Alderson JA, Lloyd DG. Repeatability of gait data using a functional hip joint centre and a mean helical knee axis. J Biomech. 2003; 36 (8): 1159–68.
6. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986; 1 (8476): 307–10.
7. Cicoira M, Zanolla L, Franceschini L, et al. Skeletal muscle mass independently predicts peak oxygen consumption and ventilatory response during exercise in noncachectic patients with chronic heart failure. J Am Coll Cardiol. 2001; 37 (8): 2080–5.
8. Clark AL, Poole-Wilson PA, Coats AJ. Exercise limitation in chronic heart failure: central role of the periphery. J Am Coll Cardiol. 1996; 28 (5): 1092–102.
9. Cohn JN, Johnson GR, Shabetai R, et al. Ejection fraction, peak exercise oxygen consumption, cardiothoracic ratio, ventricular arrhythmias, and plasma norepinephrine as determinants of prognosis in heart failure. Circulation. 1993; 87 (6S): 15–6.
10. Delp MD, Duan C, Mattson JP, Musch TI. Changes in skeletal muscle biochemistry and histology relative to fiber type in rats with heart failure. J Appl Physiol (1985). 1997; 83 (4): 1291–9.
11. DeVita P, Hortobagyi T. Age causes a redistribution of joint torques and powers during gait. J Appl Physiol (1985). 2000; 88 (5): 1804–11.
12. Fülster S, Tacke M, Sandek A, et al. Muscle wasting in patients with chronic heart failure: results from the studies investigating comorbidities aggravating heart failure (SICA-HF). Eur Heart J. 2013; 34 (7): 512–9.
13. Gollnick PD, Sjödin B, Karlsson J, Jansson E, Saltin B. Human soleus muscle: comparison of fiber composition and enzyme activities with other leg muscles. Pflugers Arch. 1974; 348: 247–55.
14. Harridge SD, Magnusson G, Gordon A. Skeletal muscle contractile characteristics and fatigue resistance in patients with chronic heart failure. Eur Heart J. 1996; 17 (6): 896–901.
15. Harrington D, Anker SD, Chua TP, et al. Skeletal muscle function and its relation to exercise tolerance in chronic heart failure. J Am Coll Card. 1997; 30 (7): 1758–64.
16. Héroux ME, Dakin CJ, Luu BL, Inglis JT, Blouin JS. Absence of lateral gastrocnemius activity and differential motor unit behavior in soleus and medial gastrocnemius during standing balance. J Appl Physiol (1985). 2014; 116: 140–8.
17. Josephson RK. Dissecting muscle power output. J Exp Biol. 1999; 202: 3369–75.
18. LeBlanc AD, Schneider VS, Evans HJ, Pientok C, Rowe R, Spector E. Regional changes in muscle mass following 17 weeks of bed rest. J Appl Physiol (1985). 1992; 73 (5): 2172–8.
19. Levine JA, Abboud L, Barry M, Reed JE, Sheedy PF, Jensen MD. Measuring leg muscle and fat mass in humans: comparison of CT and dual-energy x-ray absorptiometry. J Appl Physiol (1985). 2000; 88 (2): 452–6.
20. Lichtwark GA, Wilson AM. Is Achilles tendon compliance optimised for maximum muscle efficiency during locomotion? J Biomech. 2007; 40 (8): 1768–75.
21. Lieber RL, Fridén J. Functional and clinical significance of skeletal muscle architecture. Muscle Nerve. 2000; 23 (11): 1647–66.
22. Maiorana A, O’Driscoll G, Cheetham C, et al. Combined aerobic and resistance exercise training improves functional capacity and strength in chronic heart failure. J Appl Physiol (1985). 2000; 88 (5): 1565–70.
23. Mancini DM, Coyle E, Coggan A, et al. Contribution of intrinsic skeletal muscle changes to 31P NMR skeletal muscle metabolic abnormalities in patients with chronic heart failure. Circulation. 1989; 80 (5): 1338–46.
24. Mancini DM, Walter G, Reichek N, et al. Contribution of skeletal muscle atrophy to exercise intolerance and altered muscle metabolism in heart failure. Circulation. 1992; 85 (4): 1364–73.
25. McGowan CP, Kram R, Neptune RR. Modulation of leg muscle function in response to altered demand for body support and forward propulsion during walking. J Biomech. 2009; 42 (7): 850–6.
26. Minotti JR, Christoph I, Oka R, Weiner MW, Wells L, Massie BM. Impaired skeletal muscle function in patients with congestive heart failure. Relationship to systemic exercise performance. J Clin Invest. 1991; 88 (6): 2077–82.
27. Mitchell WK, Williams J, Atherton P, Larvin M, Lund J, Narici M. Sarcopenia, dynapenia, and the impact of advancing age on human skeletal muscle size and strength; a quantitative review. Front Physiol. 2012; 3: 260.
28. Moreira JB, Bechara LR, Bozi LH, et al. High-versus moderate-intensity aerobic exercise training effects on skeletal muscle of infarcted rats. J Appl Physiol (1985). 2013; 114 (8): 1029–41.
29. Morse CI, Thom JM, Birch KM, Narici MV. Changes in triceps surae muscle architecture with sarcopenia. Acta Physiol Scand. 2005; 183 (3): 291–8.
30. Mosteller RD. Simplified calculation of the body-surface area. N Engl J Med. 1987; 317: 1098.
31. Musch TI, Terrell JA. Skeletal muscle blood flow abnormalities in rats with a chronic myocardial infarction: rest and exercise. Am J Physiol. 1992; 262 (2): 411–9.
32. Narici MV, de Boer MD. Disuse of the musculo-skeletal system in space and on earth. Eur J Appl Physiol. 2011; 111 (3): 403–20.
33. Neptune RR, Kautz SA, Zajac FE. Contributions of the individual ankle plantar flexors to support, forward progression and swing initiation during walking. J Biomech. 2001; 34 (11): 1387–98.
34. Panizzolo FA, Green DJ, Lloyd DG, Maiorana AJ, Rubenson J. Soleus fascicle length changes are conserved between young and old adults at their preferred walking speed. Gait Posture. 2013; 38 (4): 764–9.
35. Rubenson J, Henry HT, Dimoulas PM, Marsh RL. The cost of running uphill: linking organismal and muscle energy use in guinea fowl (Numida meleagris). J Exp Biol. 2006; 209 (13): 2395–408.
36. Rubenson J, Pires NJ, Loi HO, Pinniger GJ, Shannon DG. On the ascent: the soleus muscle is conserved to the ascending limb of the force-length curve across gait mechanics in humans. J Exp Biol. 2012; 215 (20): 3539–51.
37. Smith DW, Rubenson J, Lloyd D, et al. A conceptual framework for computational models of Achilles tendon homeostasis. Wiley Interdiscip Rev Syst Biol Med. 2013; 5 (5): 523–38.
38. Stenroth L, Peltonen J, Cronin NJ, Sipilä S, Finni T. Age-related differences in Achilles tendon properties and triceps surae muscle architecture in vivo. J Appl Physiol (1985). 2012; 113 (10): 1537–44.
39. Toth MJ, Shaw AO, Miller MS, et al. Reduced knee extensor function in heart failure is not explained by inactivity. Int J Cardiol. 2010; 143 (3): 276–82.
40. Williams AD, Selig S, Hare DL, et al. Reduced exercise tolerance in CHF may be related to factors other than impaired skeletal muscle oxidative capacity. J Card Fail. 2004; 10 (2): 141–8.


Supplemental Digital Content

© 2015 American College of Sports Medicine