Journal Logo


Eccentric Hamstring Strength and Hamstring Injury Risk in Australian Footballers

Opar, David A.1; Williams, Morgan D.2; Timmins, Ryan G.1; Hickey, Jack3; Duhig, Steven J.4; Shield, Anthony J.4,5

Author Information
Medicine & Science in Sports & Exercise: April 2015 - Volume 47 - Issue 4 - p 857-865
doi: 10.1249/MSS.0000000000000465
  • Free


Australian football is a dynamic game that shares characteristics with soccer (high aerobic running demands), rugby (upper limb tackling), and Gaelic football (punting kicking), although it is unique with regard to a large field size and a high number of player interchanges allowed (20). Like many other sports (5,15,25), Australian football imposes a risk of injury to its participants (20). Hamstring strain injury (HSI) has been the predominant cause of lost playing time at the elite level of Australian football for more than 20 years (20) and results in significant financial loss to teams via athlete unavailability (12). A number of nonmodifiable risk factors for HSI in Australian footballers have been identified previously, most prominently, increasing age and previous injury (10,19,24). However, in recent times, a greater emphasis has been placed on modifiable risk factors (2), which have the scope to be altered with appropriate interventions and can lead to reductions in an athlete’s risk of injury (2). Of these modifiable risk factors, eccentric hamstring strength and between-limb imbalances in strength have received the most attention (6,22,26); however, the role of hamstring strength in the etiology of HSI in elite Australian football remains controversial (3,14,18). Although eccentric strength, between-limb strength imbalance, and hamstring-to-quadriceps ratios have shown an association with the future incidence of HSI in professional soccer (6) and subelite sprinters (22), a recent meta-analysis found that isokinetically derived eccentric hamstring strength was not a risk factor for future injury (8). In prospective studies that examine eccentric hamstring strength and strength ratios as risk factors for future HSI, isokinetic dynamometry has been the chosen strength testing methodology (3,6,22,26). Although isokinetic dynamometry is considered the gold standard tool for assessing eccentric hamstring strength, its wide spread application is limited because of the device being largely inaccessible and expensive to purchase (16) and its use as a predictor of future HSI risk is questionable (8). Further to this, the time taken to complete an assessment of an individual athlete (up to 20 min), normally at an off-site location, is prohibitive in elite sporting environments (16).

We have recently developed a novel field testing device for the assessment of eccentric hamstring strength to overcome the limitations of isokinetic dynamometry (16). Using the commonly employed Nordic hamstring exercise, the device is able to record maximal eccentric hamstring strength and between-limb imbalances, with an assessment time of less than 2 min per athlete. Although this device is a reliable measure of eccentric knee flexor forces during the Nordic hamstring exercise (16), there is currently no literature examining whether measures derived from this device are predictive of an athlete’s risk of future HSI. The Nordic hamstring exercise is the best supported exercise in the literature for the prevention of HSI (1,21), and it might reasonably be expected that the measurement of eccentric hamstring strength during this exercise could provide some information as to the risk of future HSI. Recently, it has also been suggested that the interaction between previously identified nonmodifiable risk factors, such as increasing age and previous history of injury, and modifiable factors, such as strength, needs to be considered (23). A greater understanding of how various risk factors interact with one another is likely to give practitioners a better understanding of an athlete’s risk of injury compared with univariate or independent risk factors.

The purpose of this study was to determine whether eccentric hamstring strength and between-limb imbalances in eccentric hamstring strength, derived from the aforementioned field testing device, was predictive of future HSI risk in elite Australian footballers across the preseason and in-season periods. Furthermore, the study also aimed to examine the interrelationship between previously identified risk factors, namely, increasing age and previous HSI, and eccentric hamstring strength when determining the risk of future HSI. The primary hypothesis was that the athletes who sustained an HSI would display lower levels of eccentric hamstring strength and larger between-limb imbalances in eccentric hamstring strength compared with their uninjured counterparts. In addition, we explored the interactions between increasing age, previous HSI, and eccentric strength and how these factors combined influence the probability of future HSI.


Participants and study design

This prospective cohort study was approved by the Queensland University of Technology Human Research Ethics Committee and was completed during the 2013 Australian Football League season, during the preseason (November 2012 to February 2013) and in-season (March 2013 to September 2013) periods. Five of the six professional teams invited to participate elected to take part in the study. All members of the playing squad for each team (approximately 42–45 athletes per team) were approached and provided with a plain language statement explaining the study. In total, 210 elite male Australian footballers gave informed written consent to participate. Before the commencement of data collection, the team medical staff completed a previous injury questionnaire that detailed, for all athletes, the history of hamstring, quadriceps, and calf strain injuries and chronic groin pain within the preceding 12 months and the history of anterior cruciate ligament (ACL) injury at any stage in the athlete’s career. These previous injury reports were completed with information from each club’s internal medical recording system. Athletes had their eccentric hamstring strength assessed at three time points throughout the season: start of preseason training (November 2012), end of preseason training (February 2013), and during the middle of the competitive season (June/July 2013). All players deemed fit by the team medical staff to complete testing at each time point were assessed. For athletes who experienced an HSI during the study period, a standard hamstring injury report form was completed by the club medical staff. At the conclusion of the season, club staff also reported on each athlete’s primary position on the field. If the athletes played multiple positions, the club was asked to identify a single position that was most representative of the athletes’ match play demands. As per previous work (4), these playing positions were defined as follows: center-half back, center-half forward, full back, full forward, small back, small forward, midfielder, and ruckman.

Eccentric hamstring strength assessment

The assessment of eccentric hamstring strength using the Nordic hamstring exercise field testing device has been reported previously (16). Participants knelt on a padded board, with the ankles secured immediately superior to the lateral malleolus by individual ankle braces that were attached to custom-made uniaxial load cells (Delphi Force Measurement, Gold Coast, Australia) with wireless data acquisition capabilities (Mantracourt, Devon, UK). The ankle braces and load cells were secured to a pivot, which allowed the force to always be measured through the long axis of the load cells. After a warm-up set, participants performed one set of three maximal repetitions of the bilateral Nordic hamstring exercises. The instructions to players were to gradually lean forward at the slowest possible speed while maximally resisting this movement with both legs while keeping the trunk and hips held in a neutral position throughout and the hands held across the chest (16). Participants were loudly exhorted to provide maximal effort throughout each repetition. A trial was deemed acceptable when the force output reached a distinct peak (indicative of maximal eccentric strength), followed by a rapid decline in force, which occurred when the athlete was no longer able to resist the effects of gravity acting on the segment above the knee joint.

Prospective HSI reporting

For the purposes of this study, an HSI was defined as acute pain in the posterior thigh that caused immediate cessation of exercise and damage to the muscle and or tendon, which was later confirmed with magnetic resonance imaging examination. Reports were not completed for injuries that did not fulfill these criteria. For all HSI that meet these inclusion criteria in the study period, the team doctor or physiotherapist completed a standard injury report form that detailed which limb was injured (dominant/nondominant, left/right), the muscle injured (biceps femoris long head/biceps femoris short head/semimembranosus/semitendinosus), the location of injury (proximal/distal, muscle belly/muscle–tendon junction), the activity type performed at time of injury (i.e., running and kicking), the grade of injury (I, II, or III), and the number of days taken to return to full participation in training/competition. At the conclusion of the competitive season, these reports were forwarded to the investigators.

Data analysis

Eccentric hamstring force data for each limb were transferred to a personal computer at 100 Hz through a wireless USB base station receiver (Mantracourt). Subsequently, the peak force for the three repetitions for each limb (left and right) was determined using LabChart 7.3 (ADInstruments, New South Wales, Australia). Eccentric hamstring strength, reported in absolute terms (N) and relative to body mass (N·kg−1), was determined as the average of the peak forces from the three repetitions for each limb, resulting in a left and right limb measure of eccentric strength (16).

The between-limb imbalance in eccentric hamstring force was calculated as a left/right limb ratio for the uninjured group and as an uninjured/injured limb ratio in the injured group. The between-limb imbalance ratio was converted to the percentage difference as recommended (13) using log-transformed raw data followed by back transformation. Negative percentage imbalances indicate that the left limb was stronger than the right limb in the uninjured group or that the injured limb was stronger than the uninjured limb in the injured group. For athletes who remained free of HSI, the eccentric hamstring strength measurements from the left and right limbs were averaged, because the limbs did not differ (P >0.05), to give a single control group strength score. If an athlete sustained an HSI, their eccentric hamstring strength was not considered at future time points.

Statistical analysis

All statistical analysis was performed using JMP 10.02 (SAS Institute, Inc., Cary, NC). The mean and SD of age, height, and body mass at the start of the study and the eccentric hamstring strength and strength imbalance at all three time points were determined. Univariate analysis was performed to compare age, height, weight, percentage between-limb imbalance between the injured and uninjured groups, eccentric hamstring strength of the injured limb, the contralateral uninjured limb, and the average of the left and right limbs from the uninjured group using Student’s t-test at all three time points, with Bonferroni corrections performed to account for multiple comparisons. To determine univariate relative risk (RR) and 95% confidence intervals (CI) of future HSI, athletes were grouped according to the following:

  • with or without prior
    •  ○ hamstring
    •  ○ calf
    •  ○ quadriceps
    •  ○ ACL
    •  ○ groin injury
  • average eccentric knee flexor strength thresholds above or below
    •  ○ 256 N at the start of the preseason
    •  ○ 279 N at the end of the preseason
      •   ▪ which were determined using receiver operator characteristic curves on the basis of the strength threshold that maximized the difference between sensitivity and 1 − specificity
  • limbs above or below a 10%, 15%, and 20% between-limb strength imbalance
    •  ○ at the start of the preseason
    •  ○ at the end of the preseason
  • athletes above the age cutoffs
    •  ○ 18.9 yr
    •  ○ 20.1 yr
    •  ○ 22.6 yr
    •  ○ 25.5 yr
    •  ○ 28.9 yr
  • athletes between the height and weight cutoffs defined previously by Gabbe et al. (10).

HSI rates from these groups were then compared and RR determined and significance assessed via a two-tailed Fisher exact test. In addition, univariate logistic regression was conducted with prospective HSI occurrence as a dichotomous dependent variable and average eccentric knee flexor strength at the start and end of the preseason as continuous independent variables in separate analyses. These data are reported as odds ratios (OR) and 95% CI per 10-N increment in force.

To improve the understanding of risk from the univariate analysis and remove possible confounding effects, two multivariate logistic regression models were built using risk factors identified in previously published evidence (10,19,24) and the findings from the current research. The first model included mean (average of both limbs) eccentric strength from the start of the preseason and history of HSI and their interaction. The mean eccentric strength of both limbs was used following data screening that showed that right leg eccentric peak force was highly related (r = 0.75) to the left leg peak force. The second model again included mean (average of both limbs) eccentric strength from the start of the preseason, but this time, it included the participant’s age. Although a third model including both age and history of HSI was built, it was not reported because it confirmed previous work that suggested age is a confounder to history of HSI with regard to predicting injury risk because those who are older are more likely to have a history of HSI. Significance was set at P <0.05 for all analyses, and where appropriate, Cohen d was used to calculate the effect size. For univariate analysis, the difference between limbs/groups is expressed as mean differences and 95% CI.


Post hoc power calculations

Using eccentric strength data from the start of the preseason testing time point, power was calculated as 0.97 for the use of two-tailed independent t-test to compare groups (input parameters, effect size = 0.80; alpha = 0.05, sample size group 1 = 182, sample size group 2 = 27) using G*Power (version 3.1.7).

Cohort and prospective injury details

Two-hundred and ten athletes (age, 23.3 ± 3.7 yr; height, 188.0 ± 7.2 cm; mass, 87.3 ± 8.2 kg) were assessed on at least one occasion, of which 121 were assessed at all three time points, 156 were assessed at the start and end of the preseason, and 186 were assessed at the start of the preseason. From the total sample of 210, 182 athletes did not sustain an HSI (age, 23.2 ± 3.7 yr; height, 188.4 ± 7.2 cm; mass, 87.9 ± 7.9 kg) and 28 did (age, 23.9 ± 3.6 yr; height, 185.1 ± 6.2 cm; mass, 84.3 ± 5.6 kg). The athletes that went on to be injured displayed lesser height and mass compared with the uninjured athletes (P < 0.05). Twenty-eight first-time HSI were sustained (19 left limb and 9 right limb), and of these, six injuries reoccurred within the study period. Of the 28 initial HSI, 78.6% were in the biceps femoris long head, 17.9% in the semimembranosus, and 3.6% in the semitendinosus. High-speed running was the primary mechanism of injury (60.7%) followed by kicking (17.9%) and running while bent over to collect the ball (7.1%). No injuries were sustained during the Nordic hamstring exercise testing sessions. Of the 28 initial HSI, five occurred in the preseason (between the start of the preseason and end of the preseason strength assessments), 16 occurred during the early part of the in-season period (between the end of the preseason and midseason strength assessments), and seven occurred during the late in-season period (after the midseason strength assessment). The number of athletes that had their eccentric knee flexor strength assessed at each time point was as follows: start of the preseason, 186 (27 went on to sustain HSI); end of the preseason, 184 (17 went on to sustain HSI); in-season period, 155 (2 went on to sustain HSI). The distribution of player positions in the subsequently injured group (center-half back, 4%; center-half forward, 4%; full back, 4%; full forward, 4%; small back, 18%; small forward, 29%; midfielder, 36%; ruckman, 4%) compared with the uninjured group (center-half back, 12%; center-half forward, 6%; full back, 7%; full forward, 5%; small back, 18%; small forward, 14%; midfielder, 31%; ruckman, 8%) suggested that center-half backs and ruckmen were underrepresented in the subsequently injured group, whereas small forwards were overrepresented.

Univariate analysis

Eccentric hamstring strength and between-limb imbalances at all three time points for the subsequently injured and uninjured limbs from the injured group and the average of both limbs from the uninjured group can be found in Table 1. Using absolute strength measures, the subsequently injured limbs were significantly weaker at the start (mean difference, 55 N; 95% CI, 21–89 N; P = 0.002; d = 0.67) and end of the preseason (mean difference, 46 N; 95% CI, 9–83 N; P = 0.014; d = 0.61) compared with the average of both limbs in the uninjured group. There were no significant differences (P > 0.05) in absolute eccentric hamstring strength between the injured limb and contralateral uninjured limb at any time point. Eccentric strength from the uninjured limb from the injured group did not differ from the average of both limbs from the uninjured group at the start of the preseason (P = 0.108), but it was lower at the end of the preseason (mean difference, 39 N; 95% CI, 2–75 N; P = 0.038; d = 0.53). Eccentric hamstring strength relative to body mass showed similar differences, with the subsequently injured limb weaker than the average of both limbs in the control group at the start (mean difference, 0.77 N·kg−1; 95% CI, 0.34–1.20 N·kg−1; P = 0.001; d = 0.76) and end (mean difference, 0.68 N·kg−1; 95% CI, 0.21–1.14 N·kg−1; P = 0.005; d = 0.73) of the preseason. Similarly, the uninjured limb from the injured group was weaker than the average of both limbs in the control group at the start (mean difference, 0.45 N·kg−1; 95% CI, 0.01–0.88 N·kg−1; P = 0.046; d = 0.36) and end (mean difference, 0.58 N·kg−1; 95% CI, 0.12–1.04 N·kg−1; P = 0.014; d = 0.66) of the preseason. Between-limb imbalance in eccentric hamstring strength did not differ between the subsequently injured and uninjured groups at the start of the preseason (mean difference, −3.0%; 95% CI, −11.7% to 5.7%; P = 0.498; d = −0.13) or the end of the preseason (mean difference, −2.6%; 95% CI, −7.6% to 2.4%; P = 0.306; d = −0.27). There was no difference between the age of the subsequently injured and uninjured athletes (mean difference, −0.7 yr; 95% CI, −2.2 to 0.8 yr; P = 0.250; d = −0.19); however, height (mean difference, 3.3 cm; 95% CI, 0.5–6.1 cm; P = 0.024; d = 0.49) and weight (mean difference, 3.6 kg; 95% CI, 0.5–6.7 kg; P = 0.021; d = 0.53) were significantly higher in the uninjured group. Average absolute eccentric knee flexor strength at the start (OR = 0.937; 95% CI, 0.888–0.990; P = 0.020) and end (OR = 0.914; 95% CI, 0.845–0.989; P = 0.026) of the preseason, using univariate logistic regression, was found to have a significant inverse relationship with the incidence of prospectively occurring HSI. As such, for every 10-N increases in eccentric knee flexor strength, the risk of HSI was reduced by 6.3% (early preseason) and 8.9% (late preseason).

Nordic hamstring exercise force variables from hamstring strain injured and uninjured elite Australian footballers.


The effect of prior HSI and ACL injury and calf, quadriceps, and chronic groin pain on the RR of future HSI can be found in Table 2. Athletes with average limb strength below the receiver-operator-curve-determined thresholds of 256 N (area under the curve = 0.65; sensitivity = 0.63; 1 – specificity = 0.35) at the start of the preseason and 279 N (area under the curve = 0.67; sensitivity = 0.65; 1 – specificity = 0.26) at the end of the preseason had 2.7 (RR, 2.7; 95% CI, 1.3–5.5; P = 0.006) and 4.3 (RR, 4.3; 95% CI, 1.7–11.0; P = 0.002) times greater risk of subsequent HSI, respectively. Similar RR values were seen for eccentric strength normalized to body mass; however, no measure of between-limb imbalance lead to a statistically significant increase in RR (Table 2).

Univariate RR to sustain a future HSI using eccentric strength and imbalance, previous injury, and demographic data as risk factors.

Multivariate logistic regression

Details of both logistic regression models can be found in Table 3 and Figures 1 and 2. Although models were significant (model 1, previous HSI and eccentric strength at the start of the preseason, P = 0.021; model 2, age and eccentric strength at the start of the preseason, P = 0.009), only the interaction of age and eccentric strength reached significance (P = 0.025). The interaction between previous HSI and eccentric strength was not significant (P = 0.406). For both models, it was eccentric strength at the start of the preseason that made a significant contribution to the model.

Logistic regression model outputs and receiver operator characteristic curve data using previous HSI, age, and eccentric strength at the start of the preseason as input variables.
The nonsignificant interaction between eccentric hamstring strength at the start of preseason training, history of HSI, and probability of future HSI (error bars indicate 95% CI).
The interaction between eccentric hamstring strength at the start of preseason training, age, and probability of future HSI. The ages are representative of the 10th, 25th, 50th, 75th, and 90th percentile of the cohort. The additional likelihood of future HSI in older athletes is pronounced at low levels of eccentric hamstring strength; however, this risk can be offset with increasing eccentric hamstring strength. Note that data have been offset (to the left or right) on the x-axis to allow for the visibility of error bars for all age groups. The data points and error bars are reflective of data at 100, 200, 300, 400, and 500 N for all groups.


The purpose of this study was to determine whether elite Australian footballers with lower levels of eccentric hamstring strength and larger between-limb strength imbalances were at an elevated risk of future HSI. The key unique findings from this study were that 1) limbs that went on to sustain an HSI were significantly weaker than the limbs of uninjured athletes at the start and end of the preseason, 2) between-limb imbalance in eccentric hamstring strength did not differ between the uninjured or injured groups, nor did an imbalance of 10%, 15%, or 20% increase the risk of future HSI, and 3) the interaction of age and history of HSI with eccentric strength provided additional information on athlete risk profile compared with age and history of HSI alone.

Lower levels of eccentric hamstring strength, and not between-limb imbalance, assessed during the Nordic hamstring exercise, increased the risk of subsequent HSI in elite Australian footballers. Given that close to two thirds of the HSI reported in the current study occurred during high-speed running, low levels of eccentric strength might suggest a reduced ability of the hamstrings to decelerate the forward moving leg during the terminal swing phase of gait (17), which may lead to an acute injury, although when in the gait cycle, the hamstrings are most susceptible to strain injury, which remain controversial. The relationship seen in the present study between eccentric hamstring weakness and HSI differs from the one other major study examining eccentric strength and HSI risk in Australian football (3). However, this previous study looked at isokinetic strength measures in a smaller sample (n = 102) with a mixture of elite and nonelite athletes, and this may account for the divergent findings. The different testing methodologies is the most likely explanation of the disparate findings between the current study and the work from Bennell et al. (3). First, the Nordic hamstring test is bilateral compared with the isokinetic testing, which involves unilateral strength assessments. Second, the external torque around the knee joint, which an athlete is required to resist via forceful knee flexor contraction during the Nordic hamstring exercise, increases as the athlete progresses toward the ground, whereas isokinetic testing involves maximal effort throughout the range of motion. The increasing demands of the Nordic hamstring exercise throughout the range of motion might indicate that stronger individuals are also able to progress further through the exercise, and this might also be a surrogate indicator of more strength at longer muscle length.

Contrary to the hypothesis and a previous finding in elite soccer (7), between-limb imbalances in eccentric strength of 10%, 15%, and 20% did not increase the risk of HSI. This finding corroborates previous work in Australian football (3) and recent findings in American football using concentric isokinetic measures (27), which have found no relation between incidence of HSI and between-limb knee flexor strength imbalance. It may be that excessively large between-limb imbalance (perhaps >25%) increases the risk of injury (particularly given that the typical error as a percentage of coefficient of variation reported for left-to-right limb imbalance measure using the Nordic testing device is 6.0%) (16); however, a study with a larger sample size would be required to explore this further. It should also be acknowledged that the test of between-limb imbalance in this study, the Nordic hamstring exercise, was bilateral in nature. Other tests of imbalance typically involve unilateral maximal efforts (3,7,27), and as such, the results are not directly comparable. We chose to employ the traditional bilateral Nordic hamstring exercise because we have previously shown that it has greater reliability (intraclass correlation coefficient = 0.85) than a unilateral Nordic test (intraclass correlation coefficient = 0.55) (16). Further work is required on the validity of a bilateral test as a measure of between-limb imbalance.

The examination of the interaction between eccentric strength at the start of the preseason with age and previous HSI, respectively, also provides novel information on risk profiles for HSI in elite Australian footballers. Age and previous HSI have both been identified as independent risk factors for future HSI in Australian football (10,19,24). The current data (Fig. 1 and 2) indicate that the elevated probability of HSI in older athletes or those with a previous HSI can be offset by higher levels of eccentric hamstring strength. In light of these findings, it should be considered that increasing age and previous HSI only elevate the likelihood of injury if the athlete also has relatively low levels of eccentric hamstring strength. Such evidence, if replicated in future studies, might cause a shift in the understanding of risk factor analysis for injury. The idea that nonmodifiable risk of future injury can be modulated via alterations in modifiable factors is worthy of further exploration for HSI in other sports and other pathologies. Indeed, risk of recurrent HSI has been significantly reduced after eccentric strength intervention (21), suggesting that, in this case, the term nonmodifiable risk factor is a misnomer. Despite the lack of difference in age and injury history variables between the injured and uninjured groups, logistic regression revealed a significant interrelationship between age, previous HSI (nonmodifiable risk factors), and eccentric hamstring strength (modifiable risk factor). Most notably, older players who display lower levels of eccentric hamstring strength at the start of the preseason were far more likely to sustain an HSI compared with younger players with the same level of strength. As an example, a 33-yr old with eccentric hamstring strength of 159 N at the start of the preseason has a 78% chance of sustaining a future HSI compared with a 20% chance of an HSI for a 22-yr old with the same level of strength.

It may be tempting to suggest that the differences in height and weight noted between the uninjured and uninjured groups could confound measures of eccentric hamstring strength, given the load experience around the knee joint during the Nordic hamstring exercise is influenced by height and weight (16). However, correlation analysis revealed r2 values of 0.02 (height vs strength) and 0.04 (weight vs strength), suggesting that these anthropometric variables do not affect eccentric strength measures using the device from the present study. This was further confirmed by the use of eccentric knee flexor strength normalized to body mass, which was also lower in the injured compared with uninjured athletes. Examination of the distribution of player position data did identify that small forwards were overrepresented in the subsequently injured group when compared with the uninjured group, whereas center-half backs and ruckmen where underrepresented. This distribution may have influenced the anthropometric characteristics of the groups, because small forwards were on average 181.5 cm tall and weighed 82.1 kg, which is shorter and lighter than the cohort average. Why these specific positional players were overrepresented in the subsequently injured group remains to be seen; however, it should be noted that in Australian football, small forwards often play hybrid roles in both the forward line and the midfield, and these extensive running demands and this exposure may impose additional risk. It is likely that the underrepresentation of center-half backs and ruckmen is due to lesser running demands of these positions (4). Given the complex nature of Australian football, with unlimited substitutions (although this rule changed in the following season) and with athletes asked to play multiple positions, simple data on primary athlete position are most likely not sufficient to elucidate the relation between the player’s role in the game and HSI risk.

We acknowledge some limitations in this study. First, the lack of exposure data does not allow for the determination of injury incidence between groups relative to time of exposure. Future work should examine total exposure and exposure to high speed running to examine the influence of this on injury risk. Second, the study was performed only on elite Australian footballers, and generalizing these findings to other athletic groups should be done with caution. Similar studies examining athletes from other sports are warranted to determine whether the same relation exists in a wider population. Third, the eccentric strength measures were measured as a force output and not converted to a joint torque. Athletes with a longer lower leg lever (distance from knee joint axis of rotation to the ankle strap) would produce more torque than athletes with a comparatively smaller lever, despite the same force output. Despite this, the force measure still provided useful information as to an athlete’s risk of HSI. Whether a torque measure would be a more sensitive measure remains to be seen. Finally, the knee flexor measures were not made relative to an anterior muscle group such as the knee extensors or the hip flexors, which might have allowed for the determination of a hamstring-to-quadriceps ratio, or something similar. It might be argued that an index of knee flexor to knee extensor (6) or hip flexor (26) strength might increase the specificity and/or sensitivity of the measures derived from the current strength assessment. Despite this, the eccentric strength measures taken at the start and end of the preseason provided information as to the risk of future injury, and this information is valuable for practitioners looking to minimize the risk of HSI.

In conclusion, elite Australian footballers who displayed low levels of eccentric hamstring strength during the Nordic hamstring exercise when assessed at the start and end of the preseason displayed significantly greater risk of future HSI compared with stronger athletes. However, a larger between-limb imbalance in eccentric strength of the hamstrings did not significantly increase the risk of HSI. The interrelationship between eccentric hamstring strength and previously identified risk factors for HSI should assist in better assessing an individual’s risk of future injury.

The authors wish to thank A/Prof. Kristian Thorborg for reviewing the manuscript before submission.

This study was funded by a proof of concept grant from Qutbluebox. Dr. David Opar and Dr. Anthony Shield are listed as coinventors on an international patent application filled for the experimental device (PCT/AU2012/001041.2012). The results of the present study do not constitute endorsement by the American College of Sports Medicine.


1. Arnason A, Andersen TE, Holme I, Engebretsen L, Bahr R. Prevention of hamstring strains in elite soccer: an intervention study. Scand J Med Sci Sports. 2008; 18( 1): 40–8.
2. Bahr R, Holme I. Risk factors for sports injuries—a methodological approach. Br J Sports Med. 2003; 37( 5): 384–92.
3. Bennell K, Wajswelner H, Lew P, et al. Isokinetic strength testing does not predict hamstring injury in Australian Rules footballers. Br J Sports Med. 1998; 32( 4): 309–14.
4. Brewer C, Dawson B, Heasman J, Stewart G, Cormack S. Movement pattern comparisons in elite (AFL) and sub-elite (WAFL) Australian football games using GPS. J Sci Med Sport. 2010; 13( 6): 618–23.
5. Brooks JH, Fuller CW, Kemp SP, Reddin DB. Epidemiology of injuries in English professional rugby union: part 1 match injuries. Br J Sports Med. 2005; 39( 10): 757–66.
6. Croisier JL, Ganteaume S, Binet J, Genty M, Ferret JM. Strength imbalances and prevention of hamstring injury in professional soccer players: a prospective study. Am J Sports Med. 2008; 36( 8): 1469–75.
7. Fousekis K, Tsepis E, Poulmedis P, Athanasopoulos S, Vagenas G. Intrinsic risk factors of non-contact quadriceps and hamstring strains in soccer: a prospective study of 100 professional players. Br J Sports Med. 2011; 45( 9): 709–14.
8. Freckleton G, Pizzari T. Risk factors for hamstring muscle strain injury in sport: a systematic review and meta-analysis. Br J Sports Med. 2013; 47( 6): 351–8.
9. Fyfe JJ, Opar DA, Williams MD, Shield AJ. The role of neuromuscular inhibition in hamstring strain injury recurrence. J Electromyogr Kinesiol. 2013; 23( 3): 523–30.
    10. Gabbe BJ, Bennell KL, Finch CF, Wajswelner H, Orchard JW. Predictors of hamstring injury at the elite level of Australian football. Scand J Med Sci Sports. 2006; 16( 1): 7–13.
    11. Gabbe BJ, Finch CF, Bennell KL, Wajswelner H. How valid is a self reported 12 month sports injury history? Br J Sports Med. 2003; 37( 6): 545–7.
      12. Hickey J, Shield AJ, Williams MD, Opar DA. The financial cost of hamstring strain injuries in the Australian Football League. Br J Sports Med. 2014; 48( 8): 729–30.
      13. Impellizzeri FM, Bizzini M, Rampinini E, Cereda F, Maffiuletti NA. Reliability of isokinetic strength imbalance ratios measured using the Cybex NORM dynamometer. Clin Physiol Funct Imaging. 2008; 28( 2): 113–9.
      14. Morgan DL. New insights into the behavior of muscle during active lengthening. Biophys J. 1990; 57( 2): 209–21.
      15. 15. Opar DA, Drezner J, Shield A, et al. Acute hamstring strain injury in track-and-field athletes: a 3-year observational study at the Penn Relay Carnival. Scand J Med Sci Sports. 2014; 24( 4): e254–9.
      16. Opar DA, Piatkowski T, Williams MD, Shield AJ. A novel device using the Nordic hamstring exercise to assess eccentric knee flexor strength: a reliability and retrospective injury study. J Orthop Sports Phys Ther. 2013; 43( 9): 636–40.
      17. Opar DA, Williams MD, Shield AJ. Hamstring strain injuries: factors that lead to injury and re-injury. Sports Med. 2012; 42( 3): 209–26.
      18. Orchard J, Marsden J, Lord S, Garlick D. Preseason hamstring muscle weakness associated with hamstring muscle injury in Australian footballers. Am J Sports Med. 1997; 25( 1): 81–5.
      19. Orchard JW. Intrinsic and extrinsic risk factors for muscle strains in Australian football. Am J Sports Med. 2001; 29( 3): 300–3.
      20. Orchard JW, Seward H, Orchard JJ. Results of 2 decades of injury surveillance and public release of data in the Australian Football League. Am J Sports Med. 2013; 41( 4): 734–41.
      21. Petersen J, Thorborg K, Nielsen MB, Budtz-Jorgensen E, Holmich P. Preventive effect of eccentric training on acute hamstring injuries in men’s soccer: a cluster-randomized controlled trial. Am J Sports Med. 2011; 39( 11): 2296–303.
      22. Sugiura Y, Saito T, Sakuraba K, Sakuma K, Suzuki E. Strength deficits identified with concentric action of the hip extensors and eccentric action of the hamstrings predispose to hamstring injury in elite sprinters. J Orthop Sports Phys Ther. 2008; 38( 8): 457–64.
      23. Thorborg K. What are the most important risk factors for hamstring muscle injury? Clin J Sport Med. 2014; 24( 2): 160–1.
      24. Verrall GM, Slavotinek JP, Barnes PG, Fon GT, Spriggins AJ. Clinical risk factors for hamstring muscle strain injury: a prospective study with correlation of injury by magnetic resonance imaging. Br J Sports Med. 2001; 35( 6): 435–9.
      25. Woods C, Hawkins R, Hulse M, Hodson A. The Football Association Medical Research Programme: an audit of injuries in professional football-analysis of preseason injuries. Br J Sports Med. 2002; 36( 6): 436–41.
      26. Yeung SS, Suen AM, Yeung EW. A prospective cohort study of hamstring injuries in competitive sprinters: preseason muscle imbalance as a possible risk factor. Br J Sports Med. 2009; 43( 8): 589–94.
      27. Zvijac JE, Toriscelli TA, Merrick S, Kiebzak GM. Isokinetic concentric quadriceps and hamstring strength variables from the NFL Scouting Combine are not predictive of hamstring injury in first-year professional football players. Am J Sports Med. 2013; 41( 7): 1511–8.


      © 2015 American College of Sports Medicine