Share this article on:

Effects of Eccentric Exercise on Passive Mechanical Properties of Human Gastrocnemius in vivo


Medicine & Science in Sports & Exercise: May 2007 - Volume 39 - Issue 5 - pp 849-857
doi: 10.1249/MSS.0b013e318033499b

Introduction: In this study, we used a newly developed method for measuring passive length-tension relations of a single human muscle in vivo to quantify changes in the mechanical properties of the human gastrocnemius after eccentric exercise.

Methods: Twelve subjects performed eccentric exercise on the right leg for 1 h by walking backward downhill on a treadmill. Passive ankle torque was measured as the ankle was rotated within its available range, with the knee in eight different angles. Subjects were studied before exercise, 1 h after exercise, and 24 h later, with further measurements at 48 h and at 1 wk in a subset of six subjects. Subjects also rated the level of perceived muscle soreness on a 10-point scale during walking on flat ground. We examined passive tension in the gastrocnemius at a standard length before and at various times after exercise.

Results: Muscle tension increased significantly at this length 1 h after exercise (34.7 ± 7.3%; mean ±SEM), peaked at 24 h (88.4 ± 12.6%), declined at 48 h (45.5 ± 4.4%), and returned to the control level at 1 wk. Stiffness of the gastrocnemius in the sitting and standing postures (i.e., at short and long lengths) was derived from passive length-tension relations. Stiffness increased after exercise, and the relative changes in muscle stiffness were similar in both positions. There was no apparent correlation between stiffness and subjective reports of muscle soreness during walking.

Conclusion: This study provides the first specific measurements of the increase in stiffness of the human gastrocnemius in vivo after a single bout of eccentric exercise. The increase peaks at 24 h and is nearly fully resolved within 1 wk.

1School of Physiotherapy, Faculty of Health Sciences, University of Sydney, AUSTRALIA; and 2Prince of Wales Medical Research Institute and University of New South Wales, AUSTRALIA

Address for correspondence: Robert D. Herbert, Ph.D., School of Physiotherapy, University of Sydney, East Street, Lidcombe, NSW 2141 Australia; E-mail:

Submitted for publication March 2006.

Accepted for publication December 2006.

Relaxed muscles resist lengthening and develop a passive tension when stretched beyond their rest length. The relationship between the length and passive tension is an offset exponential function (10,11). According to current evidence, weakly bound cross-bridges and titin seem to be the main structures that generate passive tension in the muscles. Weakly bound cross-bridges influence passive mechanical properties of resting muscles at short lengths (24), and titin, a large polypeptide linking the thick filaments and the Z-lines, determines passive properties of muscles at longer lengths (13). The tendon also contributes to the mechanical properties of the whole muscle-tendon unit, even at low forces experienced by resting muscles (8). Together, these mechanisms determine the passive mechanical behaviors of muscles.

Studies on animal muscles in situ have shown that after eccentric contractions, in which the muscles are excited as they are lengthened, the passive stiffness of the muscles increases (28). Observations of passive torque-angle relations of human muscles in vivo provide indirect evidence that the passive stiffness of the muscles increases after eccentric exercise (2,14,16). Other studies have reported decreases in resting elbow joint angle that are thought to reflect increases in the passive stiffness of human elbow flexors after eccentric exercise (4,5). Changes in muscle hardness after eccentric exercise also have been used to infer altered muscle stiffness (23). Direct measures of muscle passive stiffness and passive length-tension relations after eccentric contraction have been made in animal studies (27) but not yet in human muscle in vivo.

A method has been devised recently to measure the passive length-tension relations of a single human muscle in vivo (11). The method, which is based on an approach first developed by Herzog and ter Keurs (9), allows reliable measurement of passive length-tension relations of the gastrocnemius muscle in individual subjects. The method requires measurement of the passive torque-angle relation at the ankle for a wide range of knee angles so that the absolute contribution of the gastrocnemius, which crosses both the knee and ankle, can be extracted. The main aim of this study was to specifically measure changes in the passive length-tension properties of the human gastrocnemius after eccentric exercise. Unless otherwise stated, the term gastrocnemius is used to denote the medial as well as the lateral gastrocnemius muscles, which are treated as a functional unit. A secondary aim was to examine the relation between specifically measured changes in muscle passive mechanical properties and exercise-induced muscle soreness. Finally, we provide functionally relevant measures of gastrocnemius stiffness when the body is in two common postures, sitting and standing.

Back to Top | Article Outline


Back to Top | Article Outline


The right legs of 12 healthy subjects (five females and seven males; mean age 36 yr) were studied. Subjects had no history of significant orthopedic problems in the lower limbs, and none were currently involved in a physical training program. All subjects gave written consent to the experiment protocol, which had been approved by the human research ethics committees of the University of New South Wales and the University of Sydney. The protocol conformed to the Helsinki Declaration on Human Experimentation.

Back to Top | Article Outline

Testing equipment.

The testing apparatus consisted of a specially designed inclined board that enabled both the knee and the ankle to move freely within their available ranges without moving the lower leg (Fig. 1A). A potentiometer mounted at the hinge of the inclined board recorded knee angles. The ankle could be passively plantarflexed and dorsiflexed to the end of its ranges using a handle attached to the footplate. The position of the footplate could be adjusted for differences in leg length. A calibrated force transducer (XTran, Melbourne, Australia; linear to 250 N) and another potentiometer were attached to the footplate for recording passive ankle torque and ankle angle, respectively. The analog outputs from the force transducer and potentiometers were sampled at 50 Hz. Before each experiment, the footplate was rotated cyclically without the foot to record the torque-angle relation caused by the weight of the footplate. This torque was later subtracted from the measured ankle torque to account for the weight of the footplate (see Data analysis).

Surface electromyography (EMG) was used to monitor relaxation during ankle movement. Bipolar surface electrodes (Ag-AgCl, 10-mm diameter) were placed over the muscle bellies of the gastrocnemius, soleus, and tibialis anterior with an interelectrode distance of 3 cm. EMG signals were amplified (×1000) and bandpass filtered at 100-1000 Hz (Grass, IP 511, West Warwick, RI) and sampled at 2000 Hz. All signals were monitored at high gain during recording. Data were recorded on a laboratory computer via a CED 1401plus interface with Spike2 software (Cambridge Electronic Design, Cambridge, UK).

Back to Top | Article Outline

Eccentric exercise.

We used an exercise protocol similar to that described by Whitehead et al. (28) because this protocol increased muscle passive stiffness (measured as torque-angle relations) and produced muscle soreness. After a short demonstration, all subjects performed eccentric exercise for 1 h by walking backwards and downhill on a treadmill (Treadmaster, Tetley Technologies) as follows. The back of the treadmill was raised approximately 13° from horizontal. The subjects stood facing the back of the treadmill and were asked to step on the moving belt as far backward as possible with the right leg in a toe-to-heel manner while keeping the knee straight. After each step with the right leg, the left leg stepped back in a normal manner. By this time, the belt had moved the subject forward and upward, ready for the next step. The speed was adjusted between 3 and 3.2 km·h−1 (depending on the height of the subjects) so that the step rate was between 32 and 35 steps per minute. Although this protocol is less intensive than similar protocols used by others (28) in that the subject did not wear a 15-kg backpack, it produced muscle soreness and local tenderness in all subjects except one immediately after exercise, with the soreness and tenderness peaking 1-2 d later. The response in one subject was more severe; she had difficulty walking for 3 d, beginning 24 h after the exercise. There was marked tenderness, and ultrasound investigation confirmed a grade II muscle tear in the medial gastrocnemius. Consequently, this participant's data were not included in the analysis. The subject recovered uneventfully after 2 wk. Our requirement to step back as far as possible is likely to have maximized the lengthening and, hence, the damage (17).

Back to Top | Article Outline


Passive ankle torque-angle relations were measured at eight knee angles (0, 10, 20, 50, 60, 70, 90, and 100° in random order; 0° represents full knee extension). With the subject's right foot firmly strapped to the footplate so that the lateral malleolus was aligned with the center of the potentiometer, the footplate was manually, slowly rotated to fully plantarflex and dorsiflex the ankle at approximately 0.11 Hz (average angular speed of ~12°·s−1) via the handle attached to the footplate. Experiments in our laboratory confirmed that this angular speed of the ankle is too slow to elicit a significant stretch reflex response. During ankle rotation, passive ankle torque, ankle angle, and knee angle were recorded. This procedure was repeated at each knee angle. Subjects were asked to remain as relaxed as possible throughout.

Because there was some time-dependent deformation in the first few cycles of ankle plantarflexion-dorsiflexion, data were sampled for 10 cycles or more, depending on EMG recordings. Data in the first four cycles after the sixth cycle in which there was no obvious EMG activity were used for analysis. It was rare for more than 15 cycles to be required. Electromyographic evidence of muscle contraction was most common during measurements performed 1-2 d after exercise. If the muscles were not relaxed, further cycles were conducted until satisfactory data were obtained.

Three measurements were obtained for each of 12 subjects: before and 1 h after exercise, and 24 h later. Additional measures were performed on six of the subjects 48 h and 7 d after exercise. At 24 h and for some subsequent measures, it was usually not possible to dorsiflex the ankle fully because muscle soreness was exacerbated by the passive lengthening. Hence, the footplate was moved to the angle at which subjects indicated that they did not want the ankle further dorsiflexed because of discomfort.

To measure the length of the lower leg (ls), the reference length of the gastrocnemius (lref), and the length of the foot (lf) of each subject, skin markers were placed on the middle of the lateral femoral epicondyle, the center of the lateral malleolus, and the head of the second metatarsal of the right foot. The distances between these markers were measured with a tape measure, with the knee and ankle angles at 90°.

In each session after exercise, subjects were asked to rate the level of muscle soreness during normal walking on flat ground. This was assessed with a 10-point scale, with 0 for no soreness and 9 for maximal soreness. To obtain this value 1 h after exercise, subjects walked in the laboratory for 3-5 min before torque-ankle angle measurements were resumed.

Back to Top | Article Outline

Data analysis.

The analysis derived passive length-tension relations of gastrocnemius using a procedure that has been presented in detail previously (11). Key features of the measurement and analytical approach are repeated here.

The method assumes that the passive torque measured at the ankle joint depends on torque caused by (i) single-joint structures such as single-joint muscles and ligaments that cross the plantar aspect of the ankle joint but not the knee joint, (ii) single-joint structures that cross the dorsal aspect of the ankle joint but not the knee joint, and (iii) the two-joint muscle, the gastrocnemius, which crosses the plantar aspect of the ankle and the knee (Fig. 1B). Therefore, differences in the passive ankle torques measured at different knee angles are caused by changes in the length of the gastrocnemius. The contributions to the passive ankle torque from other two-joint structures such as the plantaris muscle, nerves, and blood vessels were assumed to be negligible. This approach was used by Herzog and ter Keurs (9) to experimentally determine length-active tension relations of the intact human gastrocnemius in vivo. We treated the torque-angle properties of the single-joint structures and length-tension properties of the gastrocnemius as simple offset exponential functions above slack length. Under this assumption, the total passive torque measured at the ankle joint is:

where τankle{θa,θk} is the passive torque at the ankle that is a function of both ankle and knee joint angles;

is the torque attributable to single-jointstructures on the plantar aspect of the ankle;

is the torque attributable to single-jointstructures on the dorsal aspect of the ankle;

is torque attributable to the gastrocnemius; θa and θk are ankle angle and knee angle, respectively; and ap, kp, ad, kd, ag, and kg are parameters relating to the stiffness of structures that cross the plantar aspect of the ankle (ap, kp), the dorsal aspect of the ankle (ad and kd), and the gastrocnemius (ag, kg). θP and θD are ankle angles at which ankle plantar flexors and dorsiflexors are slack, respectively. mg is the moment arm of gastrocnemius at the ankle. lg is the length of the gastrocnemius, and lG is the slack length of the gastrocnemius.

In equation 1, two variables are known from direct measurement: τankle and θa. Length of the gastrocnemius (lg) was derived from knee and ankle angles and published anthropometric data (7), and the moment arm of the gastrocnemius (mg) at the ankle was calculated by differentiating change in length of the gastrocnemius with respect to change in ankle angle.

The remaining parameters in equation 1 (ag, kg, lG, ap, kp, θP, ad, kd, and θD) are unknown and were estimated during the analysis process. The three parameters of interest were ag, kg, and lG because they determine gastrocnemius length-tension properties. The reproducibility of the passive length-tension curves generated from the three parameters has been reported (11). The curves were reproducible with an average root mean square error of 3 and 6% of maximal passive tension for pairs of measurements made on the same day and a week apart, respectively.

Before analysis, raw data were transformed as follows. First, the torques attributable to the weight of the footplate and estimated weight of the foot (both functions of ankle angle) were subtracted from the measured ankle torques. Second, data from the plantarflexion half of each cycle of dorsiflexion and plantarflexion (Fig. 2A) were discarded (convention: 90° when the foot is in neutral position; the angle reduces when the foot is in plantarflexion and increases when the foot is in dorsiflexion). Use of the dorsiflexion portion of the curve mimics clinical examination of the ankle to assess the plantarflexor muscles and is the direction that provokes muscle pain near the extremes of dorsiflexion. It does not materially alter the variables we derived. The weight of the foot was estimated using anthropometric data (29). Therefore, the data used for analysis included transformed passive torque-ankle angle relations during stretching of the ankle plantarflexor muscles, and the calculated changes in length and moment arms of the gastrocnemius across the ranges of ankle and knee angles that were covered during the experiment.

Data analysis involved two steps. First, the parameters in equation 1 were estimated using the quasi-Newton algorithm in Statistica version 6 (StatSoft, Inc., Tulsa, OK) (for details, see Hoang et al. (11). Second, the estimated values of the three gastrocnemius constants ag, kg, and lG were used to construct passive length-tension curves of the gastrocnemius using the formula:

To examine changes in the passive mechanical properties of the gastrocnemius, length-tension curves obtained before and after exercise were rescaled as percentages of the maximum tension values measured before exercise. Changes in muscle passive tension were analyzed using repeated-measures ANOVA and post hoc pairwise comparisons with Bonferroni correction. Hence, statistical significance was set at P < 0.0125.

To assess the physiological consequences of the changes in behavior of the gastrocnemius after eccentric exercise, two additional measures were derived: stiffness of the gastrocnemius in the standing position (knee angle at 0° and ankle angle at 90°), and in the sitting position (knee and ankle angles at 90°). Stiffness was calculated by differentiation of equation 2 with respect to the length of the gastrocnemius. Data are presented as mean ± SE. We also sought correlations (Spearman rank coefficient) at each time point between our measures of passive tension and stiffness (N and N·m−1) and the subjective scores for pain.

Back to Top | Article Outline


Passive torque-angle curves for ankle dorsiflexion were obtained at eight different knee angles spanning a wide range from 0 to 100°, before and after 1 h of walking backward downhill. This exercise damaged the plantarflexor muscles eccentrically. From the curves we derived passive length-tension curves for the gastrocnemius muscle. This derivation is possible because the gastrocnemius crosses both the ankle and knee (11).

Raw ankle torque-angle data from one subject measured 24 h after eccentric exercise with the knee at 10° are shown in Figure 2A. Four cycles of movement without EMG activity were used for analysis. Figure 2B shows passive torque-angle curves for ankle dorsiflexion obtained at four of the eight knee angles for the same subject.

Back to Top | Article Outline

Changes in length-tension properties of the gastrocnemius after eccentric exercise.

Measurements were performed on 12 subjects before exercise, within 1 h of exercise, and 24 h after exercise, with further measurements at 48 h and 1 wk after exercise in a subgroup of six subjects. Soreness commonly limited passive dorsiflexion of the ankle, especially at 24 h. Hence, length-tension properties were compared over the same absolute range of lengths for the gastrocnemius in each subject, with the longest length being the maximal comfortable length reached in the study 24 h after exercise.

An example of passive length-tension curves for the gastrocnemius before exercise and at different times after exercise for one subject is shown in Figure 3. For this subject, compared with before exercise, passive tension of gastrocnemius at the maximal comfortable length at 24 h increased 25% immediately after exercise, 78% at 24 h, and 54% at 48 h, and had returned within 3% of the control value at 1 wk. All subjects showed an immediate rise in passive tension after eccentric exercise and a time course similar to that shown in Figure 3.

For the group of 12 subjects, 1 h after exercise and 24 h later, the mean passive tension (at the longest length at 24 h) was significantly higher than before exercise (F2,11 = 25.773, P < 0.001). The mean increase was 34.7 ± 7.3% (P = 0.001) and 88.4 ± 12.6% (P < 0.001) for the 1- and 24-h values, respectively (Fig. 4). For the subgroup of six subjects, the mean passive tension increase was 24.7 ± 7.6% (not significant) 1 h after exercise and 85.8 ± 10.2% at 24 h (P < 0.001) (Fig. 4). At 48 h after exercise, the mean passive tension remained increased (45.5 ± 4.4%, P < 0.001). Tension approached control values after 1 wk.

Twenty-four hours after exercise, when the increase in passive tension was greatest, the mean value of ag had reduced from 0.0400 ± 0.0492 to 0.0296 ± 0.0213 N, kg had decreased from 65.83 ± 25.97 to 65.06 ± 20.25m−1, and lG had decreased from 0.3359 ± 0.0419 to 0.3200 ± 0.0382 m. These changes were not statistically significant (P > 0.05).

Having demonstrated that eccentric exercise induces changes in the passive properties of the gastrocnemius, we compared the changes in the properties of the gastrocnemius muscle with changes in the properties of single-joint structures and of the whole ankle. The coefficients in equation 1 were used to plot ankle torque, torque attributable to the gastrocnemius muscle-tendon unit, and torque attributable to the single-joint structures crossing the plantar aspect of the ankle against ankle angle. For the gastrocnemius plot, the knee angle was fixed at 60°, near the middle of its range. We generated these plots for each subject at two points in time: before exercise, and 24 h after exercise when the increase in stiffness was at its peak. The change from before exercise to 24 h after exercise was quantified by calculating the percent increase in area under the curve. The mean increase in area under the curve (N = 12) was 60% (± 24%, P < 0.001) for the whole ankle, 55% (± 40%, P < 0.001) for the single-joint plantarflexor structures, and 111% (± 43%, P < 0.001) for the gastrocnemius muscle-tendon unit.

To provide a functional measure of the altered length-tension properties of the gastrocnemius after exercise, changes in muscle stiffness were calculated for the standing and sitting postures (Fig. 5). In standing, both the full group (N = 12) and subgroup (N = 6) showed significant increases in stiffness of the gastrocnemius at 24 h after exercise (F2,11 = 38.6, P < 0.001 and F4,5 = 12.9, P < 0.001, respectively). For the subgroup, the stiffness remained high but not quite significant (P = 0.013) at 48 h and returned to close to the control value after 1 wk. Similarly, in sitting, the full group showed a significant increase (F2,11 = 12.6, P < 0.001) in stiffness of the gastrocnemius at 24 h but not 1 h after exercise. For the subgroup, there were increases in mean stiffness of the gastrocnemius in the sitting position up to 48 h after exercise, but the increases were not significant (perhaps because of low statistical power attributable to the small number of subjects).

Back to Top | Article Outline

Changes in passive tension and muscle soreness.

After 1 h of eccentric exercise, all subjects reported mild soreness in the leg during walking on flat ground about 5 min after the end of exercise (mean score of 2 out of 10). On subsequent days, in all but one subject, muscle soreness increased and peaked at 48 h (mean score was 3.3 and 3.5 at 24 and 48 h, respectively). In all but one subject, soreness was minimal at 1 wk (Fig. 6). Both soreness and passive tension increased immediately after exercise, continued to increase for 1-2 d, and had recovered by 1 wk. However, a direct link between them was not established. For individual subjects, there was no correlation between perceived muscle soreness and measures of muscle passive stiffness. In addition, there was a tendency for soreness to peak at 48 h while passive stiffness was decreasing.

Back to Top | Article Outline


Our main aim was to measure specific changes in the passive length-tension relations of the human gastrocnemius produced by eccentric exercise. Although it was likely that a bout of unaccustomed eccentric exercise would increase muscle passive stiffness, previous studies could only infer increases in muscle stiffness from changes in joint torque-angle relations (3,14,28) or changes in resting joint angles (5,17). These measures reflect tension produced by all structures crossing a joint. Our method provides a measure of the passive length-tension properties of the two-joint component of the ankle plantarflexors (i.e., the gastrocnemius isolated from single-joint muscles such as the soleus). A secondary aim was to explore the relationship between perceived muscle soreness and changes in muscle passive stiffness produced by eccentric exercise.

Muscle passive stiffness increased within 1 h of exercise in all subjects, peaked typically at 24 h, declined over the next day, and returned close to control levels after a week. Studies on upper-limb muscles have shown more profound damage. Howell and colleagues (14) report that passive elbow torque increased by 150% immediately after exercise and stayed about 50% higher than the preexercise level for 10 d. Similarly, Chleboun et al. (3) have shown that elbow flexor stiffness increased by 60% immediately after exercise and remained elevated for 5 d. The difference in the size of changes in the muscle passive stiffness found between this study and studies of the elbow flexor muscles could be explained by the differences in the exercise protocols, the muscle group studied, and the measuring method. Also, from a functional viewpoint, daily activities of upper-limb muscles produce less intense eccentric contractions than occur for the antigravity muscles of the lower limb during locomotion (12), so upper-limb muscles may be less well adapted and more prone to damage than lower-limb muscles, leading to greater changes in muscle stiffness.

Increases in the area under the gastrocnemius torque-angle curve (111%) were about twice the increase in the area under the ankle torque-angle curve (60%) and the single-joint plantarflexor torque-angle curve (55%). This suggests that backward downhill walking increases the stiffness of the gastrocnemius more than other plantarflexor muscles, and it shows that measures of ankle joint stiffness underestimate changes in stiffness of the gastrocnemius.

A limitation of this study is that we could not compare the passive tension of the gastrocnemius before and after exercise at the longest muscle length. At 24-48 h, it was usually not possible to dorsiflex the ankle fully, because of muscle soreness. Therefore, we compared the passive tension at the longest muscle length that subjects were comfortable with 24 h after exercise. This length was about 1-2 cm shorter than that measured before exercise. From the slopes of the length-tension curves (Fig. 3), we would predict that the increase in passive tension of the gastrocnemius after exercise at the longest physiological length measured before exercise would be higher than described here, by about 25-30%. Another limitation was that the measuring method does not allow division of the length-tension relation into the contributions from the medial and lateral parts of the gastrocnemius. Changes in muscle passive stiffness after eccentric exercise might not be the same in the lateral and medial parts of the muscle, because of different recruitment patterns of the two parts during exercise (30).

There are potential sources of error in the measuring method used here, although the test-retest reliability of the method is established (11). We took care to minimize movement of the foot in the testing apparatus. Because we repeated measurements in the same subjects, any systematic error attributable to uncertainty in the moment arm of the gastrocnemius or anthropometric variables should not have affected our main conclusions.

The model described with equation 1 only permits certain changes in the shape of the torque-angle curve with knee angle: the effect of an increment in knee angle can only be to increase the exponent of the gastrocnemius component of the ankle torque-angle relationship. Two observations suggest that the properties of the ankle are not fully accounted for in this model. According to the model, the right ends of the torque-angle curves obtained at different knee angles (Figure 2B) should progressively diverge at larger ankle angles. This seems to happen up to about 125°, but for the last 10 or 15° of ankle dorsiflexion, there is no further divergence. In addition, the left ends of the ankle torque-angle curves obtained at different knee angles should progressively converge at smaller ankle angles. The curves do converge at small ankle angles, but, contrary to the model, they cross over at ankle angles below about 70°. We are currently exploring additional refinements to the method to explain these behaviors.

In the present study, the gastrocnemius was passively stretched within an intact posterior compartment of the lower leg, including triceps surae and other muscles as well as surrounding structures. Studies on rat muscles in situ with intact connective tissues (15,25) suggest that an important feature of a muscle that is not isolated from its surrounding tissue is the extramuscular myofascial force transmission for both active and passive forces-that is, force transmitted via connective tissues such as compartmental fascia and connective tissue tracts containing bundles of nerves and blood vessels. Evidence for this pathway of transmission is provided by the consistent findings of a length-dependent difference between the measured forces at the distal and proximal ends of the muscle. In the present study, we derived the passive length-tension curves of the gastrocnemius from torques measured at the ankle only. However, if systematic errors occurred because of lateral force transmission, it should not have affected our main conclusions, because we performed repeated measures, enabling within-subject comparison. Moreover, in subsequent experiments (Hoang and Herbert, 2006, unpublished observations), we have consistently observed the same relationship between ultrasonographically measured changes in muscle fiber length and the length of whole gastrocnemius muscle-tendon units at different knee angles. This strongly suggests that extramuscular myofascial force transmission has negligible effects on passive properties of human gastrocnemius muscles in vivo, at least in the context of the testing protocols we employed. Further studies are needed to assess changes in any such transmission after the physiological damage produced by the eccentric exercise used here.

Although this study did not attempt to explain the mechanisms that produce the rise in muscle passive stiffness after eccentric exercise, it is possible to speculate on a chain of events that may have occurred. A potential mechanism involves a rise in resting Ca2+ levels in muscle fibers; this has been demonstrated in several studies on mammalian muscles (1). Whitehead and colleagues (28) postulate that a rise in Ca2+, caused by membrane damage accompanying sarcomere disruption during eccentric exercise (22), triggers low-level activation and produces "contracture clots," which increase muscle passive tension. Howell and colleagues (14) suggest that stretch-activated Ca2+ release might be exaggerated in injured muscles, leading to increased muscle stiffness. Irrespective of the precise cause, any damage-induced rise in intracellular Ca2+ levels in muscle fibers after eccentric contractions can increase cross-bridge numbers and lead to a high resting stiffness.

It also has been suggested that part of the increased passive stiffness at the long muscle lengths after eccentric exercise reflects the swelling accompanying muscle damage (14,19). However, because the passive tension of muscle increased within 1 h after exercise, whereas muscle swelling did not increase significantly until 48 h after exercise, others suggest that swelling is likely to play only a minor role in the increase in passive stiffness in the first 24-48 h after exercise (3,28).

Finally, at an ultrastructural level, Z-disc streaming and smearing occur after eccentric damage and are considered morphological hallmarks of muscle injury (6). Given that titin anchors myosin to the Z-disc, these observations suggest that titin could be damaged after eccentric exercise. Recent studies of human muscle biopsies taken after eccentric exercise have shown changes in the contents of titin and other myofibrillar protein components such as alpha-actin and desmin (26,31). These changes were thought to reflect myofibrillar remodeling rather than myofibril damage. On the other hand, previous studies have suggested that the passive force within the sarcomere is largely determined by a unique sequence within the giant titin molecule, the PEVK segment (20), and that this segment can increase stiffness in a calcium-dependent way (18). Therefore, it is conceivable that rise in intracellular Ca2+ after eccentric exercise could affect the mechanical behavior of the PEVK segment and, hence, increase passive tension.

Although muscle passive stiffness was elevated for at least 48 h after eccentric exercise, it is unlikely that this increase is directly related within an individual subject to the muscle soreness experienced when walking. There was no correlation between muscle tension or stiffness and the level of soreness reported by an individual subject at any time point after exercise. Furthermore, muscle stiffness seemed to be declining at 48 h, whereas soreness ratings were still increasing (a trend evident in other studies). This result, if confirmed, has practical implications: it means that subjects reporting high levels of soreness may not necessarily have the most severe muscle damage, at least as measured by muscle passive stiffness, and that pain scales may not be the best indicators of muscle recovery from damage after eccentric exercise.

In addition, although eccentric exercise is known as an effective training option for muscle strengthening, clinicians may need to take care when selecting the load for eccentric exercise. In this study, one subject incurred a grade II muscle tear-signs and symptoms appeared 24 h after performing the exercise protocol as described. We also observed similar muscle injuries in two subjects in another study with a similar exercise protocol (Johnson and Herbert, 2006, unpublished). Therefore, we argue that eccentric exercise, if prescribed, should start at a relatively low load to make use of the repeated bout effect-that is, the adaptation whereby a single bout of eccentric exercise protects against muscle damage from subsequent eccentric bouts (21).

In summary, we have used a new, more specific method to quantify the increase in the passive properties of "isolated" human gastrocnemius muscles in vivo after a single bout of eccentric exercise. The increase peaks within 24 h and is nearly fully resolved within a week. The increase in muscle stiffness was greater at stretched lengths. There was no apparent correlation between increased muscle passive stiffness and subjective reported muscle soreness after exercise.

This work was supported by the National Health and Medical Research Committee and Multiple Sclerosis Ltd. We are grateful to Hilary Carter for construction of the apparatus and to Drs Trevor Allen, Gabrielle Todd, and Robert Gorman for comments on a draft manuscript.

Back to Top | Article Outline


1. Balnave, C. D., D. F. Davey, and D. G. Allen. Distribution of sarcomere length and intracellular calcium in mouse skeletal muscle following stretch-induced injury. J. Physiol. 502:649-659, 1997.
2. Chleboun, G. S., J. N. Howell, H. L. Baker, et al. Intermittent pneumatic compression effect on eccentric exercise-induced swelling, stiffness, and strength loss. Arch. Phys. Med. Rehabil. 76:744-749, 1995.
3. Chleboun, G. S., J. N. Howell, R. R. Conatser, and J. J. Giesey. Relationship between muscle swelling and stiffness after eccentric exercise. Med. Sci. Sports Exerc. 30:529-535, 1998.
4. Clarkson, P. M., K. Nosaka, and B. Braun. Muscle function after exercise-induced damage and rapid adaptation. Med. Sci. Sports Exerc. 30:529-535, 1992.
5. Cleak, M. J., and R. G. Eston. Muscle soreness, swelling, stiffness and strength loss after intense eccentric exercise. Br. J. Sports Med. 26:267-272, 1992.
6. Friden, J., and R. L. Lieber. Segmental muscle fiber lesions after repetitive eccentric contractions. Cell Tissue Res. 293:165-171, 1998.
7. Grieve, D. P. S., and P. R. Cavanagh. Prediction of gastrocnemius length from knee and ankle joint posture. In: Biomechanics-VI-AE, E. Asmussen and K. Jorgensen (Eds.). International Series of Biomechanics Vol. 2A, Baltimore, MD: University Park Press, pp. 405-412, 1978.
8. Herbert, R. D., A. M. Moseley, J. E. Butler, and S. C. Gandevia. Change in length of relaxed muscle fascicles and tendons with knee and ankle movement in humans. J. Physiol. 539:637-645, 2002.
9. Herzog, W., and E. ter Keurs. A method for the determination of the force-length relation of selected in-vivo human skeletal muscles. Eur. J. Physiol. 411:637-641, 1988.
10. Hill, D. K. Tension due to interaction between the sliding filaments in resting striated muscle. The effect of stimulation. J. Physiol. 199:637-684, 1968.
11. Hoang, P. D., R. B. Gorman, G. Todd, S. C. Gandevia, and R. D. Herbert. A new method for measuring passive length-tension properties of human gastrocnemius muscle in vivo. J. Biomech. 38:1333-1341, 2005.
12. Hof, A. L. Muscle mechanics and neuromuscular control. J. Biomech. 36:1031-1038, 2003.
13. Horowits, R. The physiological role of titin in striated muscle. Rev. Physiol. Biochem. Pharmacol. 138:57-96, 1999.
14. Howell, J. N., G. Chleboun, and R. Conatser. Muscle stiffness, strength loss, swelling and soreness following exercise-induced injury in humans. J. Physiol. 464:183-196, 1993.
15. Huijing, P. A., and G. C. Baan. Myofascial force transmission: muscle relative position and length determine agonist and synergist muscle force. J. Appl. Physiol. 94:1092-1107, 2003.
16. Jones, C., T. Allen, J. Talbot, D. L. Morgan, and U. Proske. Changes in the mechanical properties of human and amphibian muscle after eccentric exercise. Eur. J. Appl. Physiol. 76:21-31, 1997.
17. Jones, D. A., D. J. Newham, and P. M. Clarkson. Skeletal muscle stiffness and pain following eccentric exercise of the elbow flexors. Pain 30:233-242, 1987.
18. Labeit, D., K. Watanabe, C. Witt, et al. Calcium-dependent molecular spring elements in the giant protein titin. Proc. Natl. Acad. Sci. U. S. A. 100:13716-13721, 2003.
19. Leger, A. B., and T. E. Milner. Passive and active wrist joint stiffness following eccentric exercise. Eur. J. Appl. Physiol. 82:472-479, 2000.
20. Linke, W. A., M. Ivemeyer, P. Mundel, M. R. Stockmeier, and B. Kolmerer. Nature of PEVK-titin elasticity in skeletal muscle. Proc. Natl. Acad. Sci. U. S. A. 95:8052-8057, 1998.
21. McHugh, M. P. Recent advances in the understanding of the repeated bout effect: the protective effect against muscle damage from a single bout of eccentric exercise. Scand. J. Med. Sci. Sports 13:88-97, 2003.
22. Morgan, D. L., and D. G. Allen. Early event in stretch-induced muscle damage. J. Appl. Physiol. 67:2007-2015, 1999.
23. Murayama, M., K. Nosaka, T. Yoneda, and K. Minamitani. Changes in hardness of the human elbow flexor muscles after eccentric exercise. Eur. J. Appl. Physiol. 82:361-367, 2000.
24. Proske, U., and D. L. Morgan. Do cross-bridges contribute to the tension during stretch of passive muscle? J. Muscle Res. Cell Motil. 20:433-442, 1999.
25. Rijkelijkhuizen, J. M., G. C. Baan, A. de Haan, C. J. de Ruiter, and P. A. Huijing. Extramuscular myofascial force transmission for in situ rat medial gastrocnemius and plantaris muscles in progressive stages of dissection. J. Exp. Biol. 208:129-140, 2005.
26. Trappe, T. A., J. A. Carrithers, F. White, C. P. Lambert, W. J. Evans, and R. A. Dennis. Titin and nebulin content in human skeletal muscle following eccentric resistance exercise. Muscle Nerve 25:289-292, 2002.
27. Whitehead, N. P., J. E. Gregory, D. L. Morgan, and U. Proske. Passive mechanical properties of the medial gastrocnemius muscle of the cat. J. Physiol. 536:893-903, 2001.
28. Whitehead, N. P., N. S. Weerakkody, J. E. Gregory, D. L. Morgan, and U. Proske. Changes in passive tension of muscle in humans and animals after eccentric exercise. J. Physiol. 533:593-604, 2001.
29. Winter, D. A. Biomechanics of Human Movement. 2nd ed. New York, NY: John Wiley & Sons, Inc., 1990.
30. Yanagisawa, O., M. Niitsu, H. Yoshioka, K. Goto, and Y. Itai. MRI determination of muscle recruitment variations in dynamic ankle plantar flexion exercise. Am. J. Phys. Med. Rehabil. 82:760-765, 2003.
31. Yu, J. G., D. O. Furst, and L. E. Thornell. The mode of myofibril remodelling in human skeletal muscle affected by DOMS induced by eccentric contractions. Histochem. Cell Biol. 119:383-393, 2003.


©2007The American College of Sports Medicine