Whether it occurs during activities of daily living, recreation, or work, muscle fatigue is a major limiting factor to sustained physical effort. Muscle fatigue has also been implicated as a predisposing factor in sporting injuries (6), in particular, hamstring muscle strains (19,23). Hamstring muscle strains are incurred frequently by athletes competing in explosive activities such as sprinting and jumping events (18,25). This injury is also common in team sports involving maximal or near maximal sprints that are repeated over an extended period. For example, hamstring muscle strains are the most common injury sustained by Australian Rules Football (ARF) players, accounting for 13% of all injuries and 16% of total time missed due to injury (15). Seward and Patrick (16) noted that hamstring muscle strains tended to occur in the first 10 min of the first and second quarters and toward the end of the third and fourth quarters of ARF matches. On the basis of these findings, the authors suggested that inadequate warm-up and player fatigue were factors associated with hamstring muscle strains (16).
As a predisposing factor in hamstring injuries and a limitation to sustained physical effort, extensive research has focused on muscle fatigue. There has also been an emerging interest in recent years into the effects of fatigue on the performance of explosive activities such as sprinting (11–14,17,20). This research has shown fatigue induced alterations to the biomechanics of sprint running, including decreased hip flexion and thigh angular velocity and increased knee extension during the swing phase of the stride cycle (17,20). The magnitude of myoelectric activity in select lower limb muscles has also been shown to increase during fatigued sprint running (11,13,14). Furthermore, long-duration fatiguing tasks have been shown to increase the duration of myoelectric activity of the involved muscles (9).
However, most of this research has utilized single maximal efforts to induce fatigue. No research was located which investigated alterations in sprinting biomechanics caused by fatigue induced by repeated dynamic sprint efforts, efforts which are more typical of most team sports compared with single effort sprints. Therefore, the purpose of the present study was to determine the effects of hamstring fatigue induced by repeated maximal efforts on hamstring muscle function during maximal sprint running. It was hypothesized that after hamstring fatigue, subjects would display a reduced efficiency of the hamstring muscles, reflected by changes in leg and thigh motion, increased duration of hamstring activity, and alterations to hamstring/quadriceps synchronization relative to initial foot-ground contact during the late swing phase of the stride cycle.
Twelve skilled, male football players (mean age = 22.3 ± 2.5 yr; height = 1.82 ± 0.08 m; mass = 86.7 ± 14.2 kg), selected from three ARF teams (N = 4) and one Rugby League (RL) club (N = 8), who had no history of major lower limb injury or disease participated in the study. Subjects were selected from these team sports as they involve repeated sprint running throughout a game and players from these sports suffer a high incidence of hamstring muscle strains. Written informed consent was obtained from each subject before participating in the study and all testing was conducted according to the National Health & Medical Research Council Statement on Human Experimentation.
The experimental procedure is summarized in Figure 1. Each player’s hamstring extensibility was assessed before testing using active and passive knee extension tests (5) to eliminate inadequate hamstring extensibility as a confounding factor that may influence a subject’s response to the fatigue protocol (23). After adequate warm-up, subjects initially performed three maximal 40-m sprints during which time high-speed film of the subject’s sprint action and electromyographic (EMG) data of five lower extremity muscles were recorded (nonfatigued condition, NFC). Each subject’s knee flexion and extension strength were then assessed before they completed specific and general hamstring fatigue tasks. Subjects then completed three final 40-m sprints (fatigued condition, FC) during which time high-speed film of the subjects’ sprint action and EMG of the same muscles were again recorded. During all tests, the subjects’ dominant limb (preferred kicking limb) was assessed.
Hamstring and Quadriceps Strength Assessment
To obtain baseline strength data and quantify decrements in muscle strength resulting from a subsequent specific hamstring fatigue task, hamstring and quadriceps peak muscle torques (N·m) were assessed using a calibrated Cybex® II+ isokinetic dynamometer (Huntsville, AL). Strength assessments were performed at velocities of 60°·s−1 and 180°·s−1, through a range of 90° knee flexion and isometrically at 60° of knee flexion following the protocol of Kannus (8). Mean peak hamstring muscle torques recorded for the subjects at 60°·s−1 and 180°·s−1 were 1.61 ± 0.12 Nm·kg−1 and 1.23 ± 0.22 Nm·kg−1, respectively whereas mean peak quadriceps muscle torques recorded at the same two test velocities were 2.74 ± 0.24 Nm·kg−1 and 2.27 ± 0.33 Nm·kg−1, respectively. As subjects in the present study displayed lower limb strength characteristics similar to those of other skilled, uninjured athletes (24) it was assumed that any alterations in the subjects’ hamstring muscle function as a result of fatigue in this study were not influenced by deficits in knee extension or knee flexion strength.
Hamstring Fatigue Protocol
Specific hamstring fatigue task.
To induce specific hamstring fatigue each subject performed three sets of 20 repetitions (1-min recovery between sets) from a standardized position on the Cybex® leg curl machine (Model 4110) at 50% of 1 repetition maximum. The level of hamstring fatigue that resulted from the specific fatigue task was assessed using the Cybex® II+ isokinetic dynamometer immediately after completion of the hamstring curls. The decrement in each subject’s hamstring peak muscle torque was then taken as representative of hamstring fatigue (2). The specific hamstring fatigue task was found in pilot testing to decrease peak hamstring muscle torque by 30% in five subjects and was therefore considered appropriate.
General hamstring fatigue task.
The general hamstring fatigue task consisted of 10 maximal 40-m sprints with 30-s recovery between each sprint. Time to complete each sprint was recorded using an infrared triggered photoelectric timing cell system accurate to 0.01 s (Department of Biomedical Science, University of Wollongong, Wollongong, NSW, Australia). Verbal encouragement was provided to assist subjects achieve maximal effort during all sprints. The increase in sprint times relative to the fastest sprint was determined (4) and taken as an indication of fatigue. Repeated maximal sprint running was used in this study as the general fatiguing activity due to its ability decrease repeat sprint performance by 5% in pilot testing, its dynamic nature, and the significant amount of hamstring activity employed during the task (21).
The sprinting action of each subject in the plane of progression was recorded using a LOCAM 16-mm high-speed camera (Model 51–0002) fitted with Canon TV Zoom Lens (12.5 to 75 mm; 1:1.8). The camera operated at a nominal speed of 200 Hz (1/600 s exposure time) and was leveled on a tripod 10 m from the subjects and perpendicular to the direction of motion (1-m scale reference within the camera field of view). The camera was time-synchronized with the EMG data by marking the film with an ultrabright current-limited LED system placed in the camera’s field of view. All filming was performed at the 30-m mark of the 40-m sprint, to allow the subjects to reach maximum sprint velocity, and at wind speeds less than 8 km·h−1 (4).
Three representative trials per test condition per subject were analyzed from 10 frames before the frame representing foot-ground contact (FGC) of the dominant limb until 10 frames after the next FGC of the ipsilateral limb. Two-dimensional coordinates for one reference point and 15 landmarks on each subject were digitized (200-Hz sampling rate) per frame using a GP9 sonic digitizer (Science Accessories Corporation, Southport, CT). After manual correction of any gross errors incurred during the digitizing process, a fourth-order zero-phase-shift Butterworth digital low-pass filter (22) was applied to the data to filter out high-frequency noise. A common cut-off frequency of 10 Hz was determined using the residual analysis method (22).
Kinematic variables calculated from the processed digitized data during the stride cycle included: stride length (m); stride rate (Hz); duration of the stance, braking, propulsion, and swing phases of the stride cycle (s); average horizontal velocity of the total body center of gravity (TBCG, m·s−1); timing of toe-off (TO), maximum thigh flexion during the swing phase (max TF), and maximum knee extension during the swing phase (max KE) in the stride cycle (expressed as a percentage of the total stride cycle); hip, knee, trunk, thigh, and leg angles (°) and angular velocities (°·s−1) at specified positions of the stride cycle (see Fig. 2); angular displacement of the trunk, thigh, and leg (°) throughout the stride cycle; and horizontal displacement from the point of first FGC to the TBCG (TBCG-to-FGC; m) (see Fig. 2). These variables were selected for analysis as they provided an indication of hamstring muscle function during the stride cycle.
Muscle Activation Patterns
Electromyographic activity of semitendinosus (ST), biceps femoris (BF), rectus femoris (RF), vastus lateralis (VL), and the medial head of gastrocnemius (G) were recorded (1000 Hz) during the three sprints in the NFC and in the FC. After standard preparation (1) to reduce electrical impedance of the skin, bipolar silver-silver chloride Red Dot® Infant disposable electrodes (St. Paul, MN) were fixed to the skin over the relevant muscle bellies and aligned parallel to the underlying muscle fiber direction (interdetection-surface spacing of 20 mm). A common reference electrode was placed on the medial tibial condyle of the test limb. Myoelectric potentials were relayed from a Telemyo 8/16 battery powered transmitter (Noraxon, U.S.A., Scottsdale, AZ) strapped firmly to the subject’s lower back, to a Telemyo 8/16 Receiver. The analog output for the five muscles from the receiver (± 5 V for full scale) was recorded in real-time at 1000 Hz (0–340 Hz bandwidth). Signals from a custom-designed foot switch (Department of Biomedical Science, University of Wollongong, Wollongong, NSW, Australia) were recorded via the EMG system to indicate initial FCG during the sprinting tasks and to enable the EMG data to be synchronized to the subjects’ gait cycle. The footswitch (mass = 4 g) was attached to the anterior-superior aspect of each subject’s shoe at the distal portion of the shoe’s tongue.
After removing signal offset, the raw EMG signals were filtered using a fourth-order zero-phase-shift Butterworth digital high-pass filter (cut-off = 20 Hz) to eliminate any movement artifact. To quantify temporal characteristics of the muscle bursts, the filtered EMG data were full-wave rectified and then filtered using a fourth-order zero-phase-shift Butterworth digital low-pass filter (cut-off = 20 Hz) to create linear envelopes (m·V).
To determine the onset and offset of muscle activity, a threshold detector (10% of the maximum amplitude of the muscle burst selected for analysis) was applied to the linear envelope of the EMG signal. The threshold detector was used in conjunction with visual inspection of the raw signals to confirm accuracy of the quantitative evaluation of the onset and offset times for each muscle burst (3). A 1% threshold was applied to the linear envelope representing the footswitch signal to determine the time of initial FGC during the stride cycle of the dominant limb for each subject. Three trials for each muscle and footswitch signal were analyzed and the average taken to determine: onset and offset times for each of the five muscles (RF, VL, ST, BF, and G) as a percentage of the stride cycle; and duration of each muscle burst (ms).
Means and standard deviations for the dependent variables for the two test conditions (NFC and FC) were calculated. After confirming normality (Kolmogorov-Smirnov test with Lilliefors’ correction), the data were analyzed using a correlated groups t-tests to determine whether there were any significant differences between the means of the dependent variables in the NFC compared to the FC. A Pearson product-moment correlation matrix was also generated for the kinematic and EMG variables at each phase of the stride cycle. For those variables that were not highly correlated (r < 0.8), univariate statistics were deemed appropriate to analyze the data, and the critical alpha level was set at P < 0.05. For the few highly correlated variables (r > 0.8), the alpha level was adjusted using a modified Bonferroni technique to reduce the possibility of finding significant differences between group means by chance (Type I error).
Hamstring Fatigue Tasks
A significant decrease (24.07 ± 11.07%) in the mean peak hamstring muscle torque was noted between the NFC (1.23 ± 0.22 Nm·kg−1) and the FC (0.94 ± 0.24 Nm·kg−1;t = 8.155;P < 0.05). There was also a 4.25 ± 2.44% decrease in repeat sprint ability (RSA) observed during the general hamstring fatigue task and the time taken to complete each sprint increased from the NFC (5.57 ± 0.25 s) to the FC (6.01 ± 0.27 s;t = −7.149;P < 0.05). Therefore, it was considered that the combined effects of the specific and general fatigue tasks produced desired fatigue levels of the hamstring muscles and satisfactory decrements in sprint performance.
Results for those kinematic parameters displayed during sprinting that were found to be significantly different in the NFC compared to the FC are shown in Table 1. There was a significant increase in the mean duration of the whole stride when fatigued which was attributed to a small but significant increase in the mean duration of the swing phase of the stride cycle. The increase in stride duration corresponded to a significant decrease in stride rate from the NFC to the FC. It was also noted that maximum knee extension in the swing phase occurred later in the FC compared with the NFC, although fatigue had no significant effect on TBCG-to-FGC distance in the NFC (mean = 0.24 ± 0.04 m) compared to the FC (mean = 0.23 ± 0.06 m).
A significant difference was found in the mean thigh segment angle at maximal knee extension of the swing phase, indicating less thigh flexion was evident during the FC compared with the NFC (see Table 1). Furthermore, the mean knee flexion angle at maximal knee extension was significantly larger in the NFC when compared with the FC. A significant difference was found in the angular displacement of the thigh segment from maximum thigh flexion to maximum knee extension in the swing phase in the FC compared to the NFC. That is, the subjects displayed a significantly smaller range of thigh movement during the FC compared with the NFC. Furthermore, subjects displayed significantly less angular displacement of the trunk, thigh, and the leg segments from maximum knee extension in the swing phase to FGC in the FC. There was also a significant increase in thigh angle at TO from the NFC to the FC (see Table 1). A significant decrease in mean leg angular velocity immediately before FGC was also evident in the FC compared to the NFC. No significant differences were observed in any of the other kinematic parameters calculated at any other part of the stride cycle.
Muscle Activation Patterns
The general pattern of myoelectric activity for all muscles was consistent during both the NFC and the FC (see Fig. 3). ST and BF exhibited one continuous period of activation beginning before FGC at approximately 65% of the stride cycle, remaining active throughout the stance phase before turning off approximately 27% into the next stride. RF displayed two distinctive bursts during the stride cycle. The first RF burst (RF1) occurred during the swing phase, beginning at approximately 30% of the stride and ceasing activity at about 65% of the stride. The second RF burst (RF2) began at approximately 88% of the stride cycle and remained active through to about 16% of the following stride. VL and GA both began activation during the late forward swing phase at approximately 70% of the stride cycle and continued to be active throughout the stance phase, with VL ceasing activity at toe-off (30% of the following stride) and GA turning off just before toe-off (20% of the following stride).
Duration of ST muscle activation increased significantly with ST activity beginning earlier (61.1 ± 4.0%) and finishing later (35.2 ± 20.0% of the following stride) in the FC when compared with the NFC (63.1 ± 6.4% and 29.3 ± 13.0% of the following stride, respectively). When expressed as a percentage of the stride duration, RF1 ceased activation significantly earlier in the FC when compared with the NFC, whereas RF2 turned on significantly earlier in the FC (see Table 2).
The decrease in stride rate in the FC noted in the present study was attributed to a small but significant increase in swing phase duration. The increase in swing phase duration was thought to be related to alterations to the kinematics of the lower limb, originating from greater thigh extension at TO. A more extended thigh would place the lower limb further behind the TBCG, thus requiring the limb to travel through a greater range of motion to prepare for the following FGC. This greater thigh extension at TO could also explain the significant decrease in thigh flexion at maximum knee extension in the swing phase, whereby subjects were not able to move the thigh as far forward in preparation for FGC when fatigued compared to in the NFC. Furthermore, the earlier cessation of RF1 and the trend for earlier onset of ST and BF may have also contributed to the decrease in thigh flexion during fatigued sprint running as these muscles assist in controlling thigh motion about the hip.
Decreased thigh and knee flexion noted in the FC in the present study were consistent with previous studies (17,20). Tupa et al. (20) claimed that decreased thigh flexion during the swing phase of fatigued sprint running resulted in the lower leg being “whipped” through causing greater knee extension in preparation for FGC. In agreement with this notion, maximum knee extension in the present study was larger in the FC compared with the NFC. However, there was no significant difference in leg angular velocities late in the forward swing phase (corrected critical alpha = 0.025), although there was a trend (P = 0.047) for leg angular velocity to decrease from the NFC (mean = 965 ± 72°·s−1) to the FC (mean = 925 ± 58°·s−1) rather than to increase as would be expected if “whipping” of the leg was evident. Based on these results, it is suggested that the increase in knee extension seen in the present study may have occurred as a result of a decreased ability of the fatigued hamstring muscles to limit the end point of forward leg rotation rather than a “whipping” through of the leg. Decreased angular displacement of the trunk, thigh, and leg segments in the late swing phase evident in the FC of the present study was also attributed to this decrease in thigh and knee flexion at maximum knee extension in the swing phase.
A primary function of the hamstring muscles during sprinting is to act eccentrically to decelerate forward motion of the thigh and leg late in the swing phase of the stride cycle in preparation for foot-ground contact (18). Therefore, the decrease in angular displacement of the lower limb, brought about by alterations to lower limb muscle activity, may be a protective mechanism employed to reduce the rapid lengthening of the hamstring muscles during fatigued sprint running. Furthermore, it is suggested that the earlier onset and longer duration of hamstring activity during the FC may have been necessary to compensate for the decreased force generation capacity of the fatigued muscles. A decreased ability of muscle to generate force is thought to reduce energy absorption capabilities of the muscle which, in turn, can increase potential for musculotendinous injuries (6). If earlier hamstring activation did not occur during rapid forward extension of the leg in the late swing phase, the rate of energy absorption by the hamstrings may have increased the injury potential of the fatigued muscles. As hamstring muscle strains in sprinting often occur during the late swing phase when the hamstring muscles are acting eccentrically to decelerate forward leg motion (18), it is recommended that the relationship between hamstring fatigue and hamstring muscle strains at this point in the stride cycle during sprinting be further investigated.
Tupa et al. (20) noted a significant increase in the TBCG-to-FGC distance displayed by elite male sprinters from nonfatigued (0.034 ± 0.06 m) to fatigued sprint running (0.40 ± 0.06 m). This increased TBCG-to-FGC distance has been associated with larger extension moments about the hip joint, whereby larger hip extension moments were thought to increase stress placed on the hamstring muscles during early stance and have been linked to hamstring muscle strains (7,10). In contrast, hamstring fatigue in the present study did not alter TBCG-to-FGC distance and, subsequently, did not appear to alter the magnitude of hip extension moments. However, as muscle moments were not directly calculated in the present study, the effect of fatigue in hip extension moments warrants further investigation.
According to Wiemann and Tidow (21), rapid backward motion of the leg segment before FGC is a major determinant of forward propulsion velocity during sprinting. Therefore, the decrease in angular velocity of the leg segment noted immediately before FGC in the FC compared with the NFC may indicate a decreased efficacy of the fatigued hamstring muscles to generate sufficient force to pull the leg backward before FGC during the fatigued sprinting efforts. Sprague and Mann (17) concluded that the fatigue induced alterations to kinematic variables before FGC resulted in greater stress being placed on lower limb muscles during the initial support phase of the stride cycle. As hamstring muscle strains are also evident during early stance phase of the sprinting stride when the hamstring muscles are concentrically contracting to rapidly extend the thigh (18), it is recommended that the relationship between fatigue induced kinematic variables before FGC evident in the present study and hamstring muscle strains are further investigated.
Repeated dynamic efforts and specific hamstring muscle fatiguing tasks performed by skilled football players were shown to result in significantly increased duration of hamstring activity and earlier cessation of RF activity during the stride cycle of maximum sprint efforts. The changes noted in muscle activation patterns were also reflected in changes to the kinematics of fatigued sprint running, which included decreased hip and knee flexion during the swing phase, decreased leg angular velocity immediately before FCG, and decreased angular displacement of the trunk, thigh, and leg segments during the late swing phase. As the hamstring muscles act eccentrically during the late swing phase to control forward motion of the thigh and leg segments, these changes in sprint technique appeared to serve as protective mechanisms to reduce stress on the hamstring muscles at critical phases of the stride cycle. These protective mechanisms, in turn, may have been elicited in an attempt to compensate for a reduced force generating capacity of the fatigued hamstring muscles and may decrease potential for hamstring muscle strain injury. However, further research is recommended to ascertain whether the protective mechanisms identified in the present study can decrease the incidence of hamstring injury.
1. Basmajian, J. V., and C. J. De Luca. Muscles Alive: Their Functions Revealed by Electromyography
(5th Ed.). Baltimore: Williams & Wilkins, 1985, pp. 561.
2. Bilcheck, H. M., W. J. Kraemer, C. M. Maresh, and M. A. Zito. The effects of isokinetic fatigue on recovery of maximal isokinetic concentric and eccentric strength in women. J. Strength Condit. Res. 7:43–50, 1993.
3. Di Fabio, R. P. Reliability of computerized surface electromyography
for determining the onset of muscle activity. Phys. Ther. 67:43–48, 1987.
4. Fitzsimons, M., B. Dawson, D. Ward, and A. Wilkinson. Cycling and running tests of repeated sprint ability. Aust. J. Sci. Med. Sport 25 (4):82–87, 1993.
5. Gajdosik, R. L., M. A. Rieck, D. K. Sullivan, and S. E. Wightman. Comparison of four clinical tests for assessing hamstring muscle length. J. Orthop. Sports Phys. Ther. 18:614–618, 1993.
6. Garrett, W. E. Muscle strain injuries: clinical and basic aspects. Med. Sci. Sports Exerc. 22:436–443, 1990.
7. Iiboshi, A., M. Ae, M. Suenaga, and K. Miyashita. Torque patterns of sprinters who had suffered from hamstring muscles strain. In: Biomechanics XI-B
, G. de Groot, A. P. Hollander, P. A. Huijing, and G. J. van Ingen Schenau (Eds.),. Amsterdam: Free University Press, 1987, pp. 605–609.
8. Kannus, P. Normality, variability and predictability of work, power and torque acceleration energy with respect to peak torque in isokinetic muscle testing. Int. J. Sports Med. 13:249–256, 1992.
9. Lucidi, C. A. and S. L. Lehman. Adaptation to fatigue of long duration in human wrist movements. J. Appl. Physiol. 73:2596–2603, 1992.
10. Mann, R., and P. Sprague. A kinetic analysis of the ground leg during sprint running. Res. Q. 51:334–348, 1980.
11. Mero, A., and E. Peltola. Neural activation fatigued and non-fatigued conditions of short and long sprint running. Biol. Sport 6 (1):43–57, 1989.
12. Mero, A., P. Luhtanen, P. Komi, and P. Susanka. Kinematics of top sprint (400m) running in fatigued conditions. Track Field Q. Rev. 88:42–45, 1988.
13. Nummela, A., H. Rusko, and A. Mero. EMG activities and ground reaction forces during fatigued and nonfatigued sprinting
. Med. Sci. Sports Exerc. 26:605–609, 1994.
14. Nummela, A., T. Vuorimaa, and H. Rusko. Changes in force production, blood lactate and EMG activity in the 400-m sprint. J. Sports Sci. 10:217–228, 1992.
15. Seward, H., J. Orchard, H. Hazard, and D. Collinson. Football injuries in Australia at the elite level. Med. J. Aust. 159:298–301, 1993.
16. Seward, H. G., and J. Patrick. A three year survey of Victorian Football League injuries. Aust. J. Sci. Med. Sport 24 (2):51–54, 1992.
17. Sprague, P., and R. V. Mann. The effects of muscular fatigue on the kinetics of sprint running. Res. Q. 54 (1):60–66, 1983.
18. Stanton, P., and C. Purdam. Hamstring injuries in sprinting
: the role of eccentric exercise. J. Orthop. Sports Phys. Ther. 10:343–349, 1989.
19. Sutton, G. Hamstrung by hamstring strains: a review of the literature. J. Orthop. Sports Phys. Ther. 5 (4):184–195, 1984.
20. Tupa, V., F. Gusenov, and I. Mironenko. Fatigue influenced changes to sprinting
technique. Modern Athlete and Coach 33 (3):7–10, 1995.
21. Wiemann, K., and G. Tidow. Relative activity of hip and knee extensors in sprinting
: implications for training. New Studies in Athletics 10 (1):29–49, 1995.
22. Winter, D. A. Biomechanics and Motor Control of Human Movement
(2nd Ed.). Ontario, Canada: John Wiley & Sons, Inc., 1990.
23. Worrell, T. W. Factors associated with hamstring injuries: an approach to treatment and preventative measures. Sports Med. 17:338–345, 1994.
24. Worrell, T. W., D. H. Perrin, B. M. Gansneder, and J. H. Gieck. Comparison of isokinetic strength and flexibility measures between hamstring injured and noninjured athletes. J. Orthop. Sports Phys. Ther. 13 (3):118–125, 1991.
25. Yamamoto, T. Relationship between hamstring strains and leg muscle strength: a follow-up study of collegiate track and field athletes. J. Sports Med. Phys. Fitness 33:194–199, 1993.