Professional Documents
Culture Documents
Accepted Article
Article type : Original Article
Cut-offs of isokinetic strength ratio and hamstring strain prediction in professional soccer players
Authors:
1,2
Dauty Marc
1,2
Menu Pierre
1, 2
Fouasson-Chailloux Alban
1
Nantes Hospital University, Physical Medicine and Rehabilitation, Hôpital Saint Jacques, 85 rue Saint
2
INSERM UMR 1229-RMES, "Regenerative Medicine and Skeleton". Team STEP "Skeletal
physiopathology and joint regenerative medicine". Nantes University Hospital, France. School Of
Tel: (33)2 40 84 62 11
Fax: (33)2 40 84 61 91
alban.fouassonchailloux@chu-nantes.fr
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1111/sms.12890
are controversial to predict hamstring strain in professional soccer players. We aimed to predict
hamstring strain in accordance to cut-offs of isokinetic knee strength ratios. Bilateral, conventional
and functional isokinetic strength ratios were calculated in 194 professional soccer players at the
beginning of 15 consecutive seasons. 36 soccer players presented a moderate hamstring strain and
158 were not injured. The different calculated isokinetic ratios were compared with the right and left
limb of the uninjured population. Different usual cut-offs were tested: at 0.85 and 0.90 for the
bilateral concentric and eccentric hamstring-to-hamstring ratio, at 0.60 and 0.47 for the conventional
hamstring-to-quadriceps ratio and at 0.80 and 1 for the mixed hamstring-to-quadriceps ratio. The
specific ratios for the studied population were also determined by the 10th percentile and then
tested. Hamstring strain prediction was established in term of odds ratios. No cut-off with bilateral,
conventional or functional isokinetic strength ratio was predictive of hamstring strain after univariate
analysis. Specific cut-offs determined from the studied population were not more predictive. Very
few injured soccer players presented values under the cut-offs at 0.47 for the conventional ratio and
at 0.80 for the mixed ratio. Regardless of their values, cut-offs of isokinetic strength ratios were not
predictive of hamstring injuries. The use of isokinetic cut-offs is not recommended to predict
Key Terms: muscle balance, sport injury, prediction, isokinetic ratio, cut-off
competition and training. This risk appears unacceptable according to the accepted criteria of the
occupational health setting (Drawer & Fuller, 2002). This type of injury is recorded as the most
common injury accounting for 12% of all injuries in elite soccer players and can be responsible for
substantial time loss from sport (Hawkins et al., 2001; Petersen et al., 2010). So, hamstring injury
prevention appears as an essential goal. From different literature reviews, some risk factors have
been identified but understanding injury causes is limited. Risk factors are traditionally divided into
two categories, internal (athlete) and external (environmental) risk factors (Bahr & Holme, 2003).
There is some evidence to suggest that there are modifiable risk factors for hamstring muscle injury,
such as muscle strength imbalance, fatigue, and flexibility (Freckleton & Pizzari, 2013; McCall et al.,
2015). The most consistent risk factor is a prior hamstring muscle injury. Knee isokinetic strength
ratios represent a usual assessment of muscle balance and are currently used to predict hamstring
injuries (Coombs & Garbutt, 2002; Dauty et al., 2016; van Dyk et al., 2016). Poor hamstring strength
and low hamstring-to-quadriceps (H/Q) strength ratio may represent a weak risk factor (van Dyk et
al., 2016). The knee muscle imbalance was arbitrary defined by a minimum bilateral hamstring-to-
hamstring concentric and eccentric ratio at 0.85 (Orchard et al., 1997) and/or a conventional
concentric H/Q ratio at 0.60 at the angular speed of 60°/sec to prevent muscle injury in soccer
players (Heiser et al., 1984). However, to assess knee muscle balance, this conventional ratio is
limited: the hamstring peak torque is measured during knee flexion and quadriceps peak torque
during knee extension. The angle of peak torque will be different for every athlete. So, the agonist-
to-antagonist function is better evaluated by reporting the hamstring eccentric maximal peak torque
to the quadriceps concentric maximal peak torque (Dvir et al., 1989; Aagaard et al., 1995). A cut-off
superior to 1 indicates that hamstring muscle could break the concentric quadriceps muscle action
and be predictive of knee injuries (Coombs & Garbutt, 2002). More recently, Croisier et al., have used
a functional mixed ratio at different angular speeds at 30 °/sec in an eccentric mode for hamstring
0.85 decreased the occurrence of hamstring injury after interventional study (Croisier et al., 2008).
However, the definition of muscle imbalance is not consensual (McCall et al., 2015) because the cut-
offs have been established by case-control studies and not by prospective cohort studies, without
The objective of this cohort study was to predict hamstring injury in accordance to usual cut-offs of
METHODS
Populations
All the professional soccer players from the same professional team, who played more than 20
professional matches per season in the French Premier and Second League were systematically
evaluated at the beginning of 15 consecutive seasons (from 1999 to 2014) without perceiving any
financial advantage. The players who presented previous hamstring injuries recurrence were
excluded because of a durable loss of strength that these various injuries could have engendered
(Askling, Karlsson, & Thorstensson, 2003). All players were followed-up during one sport season to
Individual playing positions and anthropometric data were reported. The intrinsic hamstring injury
was defined as an acute occurrence of a “physical complaint in the region of the posterior thigh
sustained during a soccer match or training, without contact, irrespective of the need for medical
attention or time loss from soccer activities” with confirmation by ultrasound and recorded from the
professional sport injury register (Petersen et al., 2011). Only moderate hamstring injuries were
(time loss > 28 days) were excluded because that time could depend on external factors such as the
choice of trainer or the season calendar for example (Shrier et al., 2015). Other significant previous
Ethics approval was obtained from the internal review board of the soccer team and the study was
performed in accordance with ethical standards in Sport Medicine (Dunn et al., 2007). After players’
consents, all the performance data were anonymized before analysis to ensure player confidentiality.
After a 10-minute cycloergometer warm up, isokinetic strength tests were performed using a Cybex
Norm® dynamometer (Lumex Inc. Ronkoma, NY, USA). The quadriceps and hamstring torques were
gravity corrected and the dynamometer calibration was performed in accordance with the
manufacturer’s instructions. The two knees were evaluated in a random order after instruction and
with verbal encouragements and visual feedback. After familiarization with the isokinetic movement,
the soccer players were tested over 3 repetitions of concentric knee extension and flexion at 60°/sec
knee flexion at 30°/sec. The time of recovery between each series was 30 seconds of rest. The
maximal eccentric peak torque was validated when the Clasp-knife-reflex was obtained (Garrett,
1990). This reflex corresponds to muscle inhibition, depending on muscle stretch when initial force
Isokinetic strength ratios were calculated from measurements of maximal quadriceps and hamstring
peak torques at different angular velocity and muscle contraction mode. Bilateral concentric
hamstring-to-hamstring ratios (H/Hcon) were calculated at 60°/sec and eccentric ratios (H/Hecc) at
Two different cut-offs were considered at 0.47 and 0.60 (Heiser et al., 1984; Croisier et al., 2002). The
at 30°/sec relative to concentric quadriceps strength at 240°/sec (Croisier et al., 2008). Two different
cut-offs were considered at 0.80 and 1 (Croisier et al., 2002; Fousekis et al., 2011). Because of the
specificity of the studied population - professional soccer players-, the specific cut-offs for all the
calculated ratios were tested after determination of the 10th percentile, so that 10% of the
All these ratios were calculated on the left and right limbs for the uninjured population. For the
injured population, the bilateral ratios were calculated by dividing the injured limbs from the
uninjured ones. For the conventional and functional ratios, only the injured limbs were compared.
Moderate relative reliability of 0.85 was established for strength ratio by intraclass correlation
Statistical analysis
From this cohort that we followed during one sport season after isokinetic tests, we identified two
populations according to the occurrence or not of hamstring injuries. Statistical analysis were
performed using the SPPS 23.0 software (IBM corp. Dublin, Ireland). Univariate analysis (independent
t-test) and Chi-2 test were used to compare quantitative and qualitative data of the injured and
uninjured populations. The results were considered significant at the 5% critical level (p<0.05). The
prediction of hamstring injury occurrence (hamstring injury vs. no hamstring injury) was assessed
using odds-ratios, where the effect of each risk factors (strength ratio) was tested separately as
categorical variable (under or upper the cut-off) (Bahr & Holme, 2003).
because of previous lower limb injuries or because they played less than 20 matches at professional
level. One hundred ninety four players were followed and thirty-six of them (18.5%) presented a
hamstring injury during the followed season. The anthropometric parameters and the playing
positions on the field were not different between the injured population and the uninjured one
(Table 1). The hamstring injuries occurred in the right limb for 22 players and in the left limb for 14
players. The time loss was of 24.5 +- 22 days [8-106]. The mean duration between isokinetic tests and
The different ratios are shown in Table 2 and we found no difference between the injured population
and the uninjured one. The specific cut-offs for every calculated ratios determined from the 10th
percentile are shown in Table 3. Because the 95%-confident interval includes the value 1, there is no
significant association between the strength ratio and the occurrence of hamstring injury. No cut-off
can predict the occurrence of hamstring injury. The cut-offs at 0.47 and 0.80 for the conventional and
mixed ratios respectively seem to be too low to be used as a reference because only one injured
player out of 36 (2.7%) had been detected thanks to these cut-offs (Table 3).
DISCUSSION
The hamstring injury prediction by low hamstring muscle strength is still controversial. A recent
literature analysis including four studies (195 participants) did not support hamstring concentric peak
torque as a risk factor for hamstring muscle strain injuries (Freckleton & Pizzari, 2013). After
hamstring eccentric isokinetic strength assessment, Bennell et al. reported the same conclusion in
Australian Rules football (Bennell et al., 1998). From a large soccer player cohort; van Dyk et al.
reported weak risk factors of hamstring injury after measuring lower concentric and eccentric
When isokinetic strength ratios are used and cut-offs taken as references, the prediction depends on
isokinetic assessment. Orchard et al. were the first authors to assess prediction by a canonical
discriminant analysis of athletes as units of measurement (Orchard et al., 1997). From 6 injured
Australian footballers, 4 presented values in the lowest quartile for both calculated ratios at 60°/s
(bilateral < 0.92 and conventional < 0.60) and the other two presented values in the lowest quartile
for one of each. In competitive sprinters, if conventional ratio at 180°/s is less than 0.60, it increases
the risk of hamstring injury by 17 (OR: 17.4; 95%IC: 1.3-231) (Yeung, Suen, & Yeung, 2009).
Moreover, the conventional ratio at 60°/s was not predictive in NFL players (Zvijac et al., 2013). After
meta-analysis, the conventional hamstring-to-quadriceps ratio at 60°/sec was not predictive in 216
Concerning the bilateral hamstring-to-hamstring ratio at 60°/sec less than 0.85, a prediction in the
eccentric mode in soccer players was reported (OR: 3.88; 95%IC: 1.13-13.23) (Fousekis et al., 2011).
For Dauty et al., this bilateral concentric hamstring-to-hamstring ratio presented a good specificity of
82% but poor sensibility of 32% (Dauty et al., 2016). However, no prediction can be done individually
in soccer players during a one season follow-up from the cut-offs (Dauty, Potiron-Josse, &
Rochcongar, 2003).
For the mixed ratio, no prediction was found for a cut-off less than 1 (Fousekis et al., 2011). But a risk
factor was reported by Croisier et al. in association with another isokinetic strength deficit in soccer
players (Croisier et al., 2008). The occurrence of hamstring injuries was 4 times more likely if no
correction of muscle imbalance was realized and controlled to normalize isokinetic strength under
cut-offs.
However, the challenge is to define cut-offs from the concept of muscle balance (McCall et al., 2015).
the first to propose the cut-off value of 0.60 in football players (Heiser et al., 1984). After
compensating training to correct this ratio, only 1% of the players presented primary hamstring
injuries with no recurrence against 7.6% with 31.7% of recurrence if compensating training was not
done. Conventional ratios were similar between 14 hamstring injuries and 29 non-hamstring injuries
in professional soccer players (Paton et al., 1989). In the same way, because 47% of Australian Rules
footballers presented a bilateral ratio less than 0.90 and 77% a conventional ratio less than 0.60, the
risk of hamstring injury could not be predicted by these cut-offs (Bennell et al., 1998). More recently,
Croisier et al. proposed different cut-offs depending on the dynamometer used (Croisier et al., 2008).
The cut-offs were 0.85, 0.47 and 0.80 for the bilateral, conventional and mixed ratios respectively
when the Cybex® dynamometer was used and 0.85, 0.45 and 0.89 with the Biodex® dynamometer.
From these cut-offs, more than 46% of professional soccer players presented muscle imbalance. This
result was comparable with those of Daneshjoo et al. who reported 41% of young soccer players with
muscle imbalance but with a cut-off at 0.90 for the bilateral ratio (Daneshjoo et al., 2013). However,
in 2015, these cut-offs might be very low because only 24% of professional soccer players presented
muscle imbalance and no player presented a conventional ratio less than 0.47 or a mixed ratio less
than 0.80 (Ardern et al., 2015). These results are comparable with ours because only one injured
player presented values less than the cut-off of each ratio in our cohort study. But other explanations
could be given. Firstly, the inclusion of soccer players, with previous histories or recurrences of
hamstring injuries, could explain very low ratios in eccentric and concentric because hamstring
muscle strength was reduced. After previous injuries, 64% of soccer players presented muscle
imbalance against 33% in uninjured players (Lehance et al., 2009). In Irish Gaelic football, the
conventional ratio at 60°/s after hamstring injury was less than those measured without injury (0.61
+- 0.1 vs 0.69 +- 0.1) (O’Sullivan et al., 2008). Secondly, the cut-offs could be different according to
the practiced sport (Cheung, Smith, & Wong, 2012). Collegiate soccer players presented a significant
pick torque is rarely described. If the Clasp-knife-reflex is not reached, a poor evaluation is at risk and
could explain a very low eccentric hamstring pick torque and consequently a very low bilateral and
mixed ratio. This reflex corresponds to a muscle inhibition which persists beyond the termination of
stretch during eccentric contraction. It is neither due to Golgi tendon organ nor to secondary spindle
afferents (Cleland & Rymer, 1990). This explains why we have determined the specific cut-offs of the
studied sport cohort. Based on the 10th percentile, the specific cut-offs were different for the
conventional ratio of 0.55 and bilateral eccentric ratio less than 0.83. The best cut-off for the mixed
ratio was near 1. However, no more hamstring injury prediction was reported.
Limitations
The number of hamstring injuries was limited to calculate injury prediction, despite the duration of
the study. We can explain that by the strict inclusion criteria of soccer players to reduce biases
related to a non-homogeneous population. Bahr and Holmes suggested that 30 to 40 injured players
would be needed to detect strong to moderate associations and 200 cases are needed to find small
to moderate associations (Bahr & Holme, 2003). The mechanism of hamstring injury was not
reported and it might have been interesting to know if the injury had occurred in surpassing the
tolerance of hamstring muscle to resist to knee extension (Coombs & Garbutt, 2002). Indeed, the
mixed ratio was developed to approach this mechanism. The period between isokinetic tests and the
occurrence of hamstring injuries was not considered. Dauty et al. have shown a prediction
probability limited beyond 3 months (Dauty et al., 2016). From a statistical point of view, the
interaction with isokinetic ratios remains unknown because all the potential risks to explain
hamstring injuries, such as age, fatigue, and flexibility, have not been studied (Freckleton & Pizzari,
Perspectives
hamstring strength. Isokinetic cut-offs have been developed to define knee muscle imbalance and so
the risk of knee muscle injuries (Heiser et al., 1984; Dvir et al., 1989; Coombs & Garbutt, 2002;
Croisier et al., 2008). Low hamstring isokinetic strength represents a modifiable risk factor in
professional soccer players (Orchard et al., 1997). From this report, we have calculated isokinetic
ratios to investigate knee muscle balance. Different cut-offs usually used in the literature were tested
to verify if they were predictive of intrinsic hamstring injury. Isokinetic strength ratios inferior to the
cut-offs of bilateral, conventional and mixed ratios were not predictive of hamstring injuries in
continuous isokinetic muscle strength values and not by isokinetic cut-offs (van Dyk et al., 2016).
References
Ardern CL, Pizzari T, Wollin MR, Webster KE. Hamstrings strength imbalance in professional football
(soccer) players in Australia. J Strength Cond Res 2015: 29: 997–1002.
Askling C, Karlsson J, Thorstensson A. Hamstring injury occurrence in elite soccer players after
preseason strength training with eccentric overload. Scand J Med Sci Sports 2003: 13: 244–
250.
Bahr R, Holme I. Risk factors for sports injuries--a methodological approach. Br J Sports Med 2003:
37: 384–392.
Cheung RTH, Smith AW, Wong DP. H:q ratios and bilateral leg strength in college field and court
sports players. J Hum Kinet 2012: 33: 63–71.
Cleland CL, Rymer WZ. Neural mechanisms underlying the clasp-knife reflex in the cat. I.
Croisier JL, Forthomme B, Namurois MH, Vanderthommen M, Crielaard JM. Hamstring muscle strain
recurrence and strength performance disorders. Am J Sports Med 2002: 30: 199–203.
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: 1469–1475.
Daneshjoo A, Rahnama N, Mokhtar AH, Yusof A. Bilateral and unilateral asymmetries of isokinetic
strength and flexibility in male young professional soccer players. J Hum Kinet 2013: 36: 45–
53.
Dauty M, Collon S. Incidence of injuries in French professional soccer players. Int J Sports Med 2011:
32: 965–969.
Drawer S, Fuller CW. Evaluating the level of injury in English professional football using a risk based
assessment process. Br J Sports Med 2002: 36: 446–451.
Dunn WR, George MS, Churchill L, Spindler KP. Ethics in sports medicine. Am J Sports Med 2007: 35:
840–844.
Dvir Z, Eger G, Halperin N, Shklar A. Thigh muscle activity and anterior cruciate ligament insufficiency.
Clin Biomech Bristol Avon 1989: 4: 87–91.
van Dyk N, Bahr R, Whiteley R, Tol JL, Kumar BD, Hamilton B, Farooq A, Witvrouw E. Hamstring and
Quadriceps Isokinetic Strength Deficits Are Weak Risk Factors for Hamstring Strain Injuries: A
4-Year Cohort Study. Am J Sports Med 2016: 44: 1789–1795.
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: 351–358.
Hawkins RD, Hulse MA, Wilkinson C, Hodson A, Gibson M. The association football medical research
programme: an audit of injuries in professional football. Br J Sports Med 2001: 35: 43–47.
Heiser TM, Weber J, Sullivan G, Clare P, Jacobs RR. Prophylaxis and management of hamstring muscle
injuries in intercollegiate football players. Am J Sports Med 1984: 12: 368–370.
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: 113–119.
Lehance C, Binet J, Bury T, Croisier JL. Muscular strength, functional performances and injury risk in
professional and junior elite soccer players. Scand J Med Sci Sports 2009: 19: 243–251.
McCall A, Carling C, Davison M, Nedelec M, Le Gall F, Berthoin S, Dupont G. Injury risk factors,
screening tests and preventative strategies: a systematic review of the evidence that
underpins the perceptions and practices of 44 football (soccer) teams from various premier
leagues. Br J Sports Med 2015: 49: 583–589.
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: 81–85.
O’Sullivan K, O’Ceallaigh B, O’Connell K, Shafat A. The relationship between previous hamstring injury
and the concentric isokinetic knee muscle strength of Irish Gaelic footballers. BMC
Musculoskelet Disord 2008: 9: 30.
Paton RW, Grimshaw P, McGregor J, Noble J. Biomechanical assessment of the effects of significant
hamstring injury: an isokinetic study. J Biomed Eng 1989: 11: 229–230.
Petersen J, Thorborg K, Nielsen MB, Hölmich P. Acute hamstring injuries in Danish elite football: a 12-
month prospective registration study among 374 players. Scand J Med Sci Sports 2010: 20:
588–592.
Shrier I, Matheson GO, Boudier-Revéret M, Steele RJ. Validating the three-step return-to-play
decision model. Scand J Med Sci Sports 2015: 25: e231-239.
Yeung SS, Suen AMY, 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:
Goalkeepers 3 22
Defenders 7 30
Midfielders 16 55
Forwards 10 75
In bold, specific cut-off calculated from the 10th percentile of the studied population