You are on page 1of 14


Hamstring strength and exibility after hamstring

strain injury: a systematic review and meta-analysis
Nirav Maniar,1 Anthony J Shield,2 Morgan D Williams,3 Ryan G Timmins,1
David A Opar1

Additional material is ABSTRACT expense of convalescence.27 Consequently, there has

published online only. To view Objective To systematically review the evidence base been much interest recently in observations of ham-
please visit the journal online
( related to hamstring strength and exibility in previously string structure and function in previously injured
bjsports-2015-095311). injured hamstrings. legs compared with control data.2834 Despite the
1 Design Systematic review and meta-analysis. possible limitation of this approach, it is often
School of Exercise Sciences,
Australian Catholic University, Data sources A systematic literature search was agreed that decits that exist in previously injured
Melbourne, Victoria, Australia conducted of PubMed, CINAHL, SPORTDiscus, Cochrane hamstrings could be a maladaptive response to
School of Exercise and Library, Web of Science and EMBASE from inception to injury.35 As such, these decits that persist beyond
Nutrition Sciences and Institute August 2015. return to play could provide markers to better
of Health and Biomedical
Innovation, Queensland Inclusion criteria Full-text English articles which monitor athletes during and/or at the completion
University of Technology, included studies which assessed at least one measure of of rehabilitation.35
Brisbane, Queensland, hamstring strength or exibility in men and women with Which parameters are the best markers to monitor
Australia prior hamstring strain injury within 24 months of the an athletes progress during rehabilitation?
School of Health, Sport and
testing date. Conventional clinical practice focuses on measures
Professional Practice, University
of South Wales, Pontypridd, Results Twenty-eight studies were included in the of strength and exibility; however, the evidence is
Wales, UK review. Previously injured legs demonstrated decits based on predominantly retrospective observations
across several variables. Lower isometric strength was of strength,28 29 3642 strength ratios36 37 39 40 43 44
Correspondence to found <7 days postinjury (d=1.72), but this did not and exibility26 28 42 4549 in previously injured ath-
Nirav Maniar, School of
Exercise Sciences, Australian persist beyond 7 days after injury. The passive straight letes. These studies are limited in reporting single or
Catholic University, 17 Young leg raise was restricted at multiple time points after isolated measures with methodologies and popula-
Street, Fitzroy 3065, VIC, injury (<10 days, d=1.12; 1020 days, d=0.74; tions that differ from study to study. To advance
Australia; Nirav.Maniar@acu. 2030 days, d=0.40), but not after 4050 days knowledge, we aimed to systematically review the
postinjury. Decits remained after return to play in evidence base related to hamstring strength and
Accepted 19 March 2016 isokinetically measured concentric (60/s, d=0.33) and exibility in previously injured hamstrings.
Published Online First Nordic eccentric knee exor strength (d=0.39). The
13 April 2016 conventional hamstring to quadricep strength ratios
were also reduced well after return to play (60:60/s,
Literature search
d=0.32; 240:240/s, d=0.43) and functional
A systematic literature search was conducted of
(30:240/s, d=0.88), but these effects were
PubMed, CINAHL, SPORTDiscus, Cochrane
inconsistent across measurement methods.
Library, Web of Science and EMBASE from incep-
Conclusions After hamstring strain, acute isometric
tion to August 2015. Keywords (table 1) were chosen
and passive straight leg raise decits resolve within 20
in accordance with the aims of the research.
50 days. Decits in eccentric and concentric strength and
Retrieved references were imported into EndNote X7
strength ratios persist after return to play, but this effect
(Thomson Reuters, New York City, New York, USA),
was inconsistent across measurement methods. Flexibility
with duplicates subsequently deleted. To ensure all
and isometric strength should be monitored throughout
recent and relevant references were retrieved, citation
rehabilitation, but dynamic strength should be assessed
tracking was performed via Google Scholar and refer-
at and following return to play.
ence list searches were also conducted.

Selection criteria
INTRODUCTION Selection criteria were developed prior to searching to
Hamstring strain injuries (HSIs) are the most common maintain objectivity when identifying studies for inclu-
non-contact injuries in Australian rules football,15 sion. To address the aims, included papers had to:
soccer,610 rugby union,1114 track and eld1517 and Assess at least one parameter of hamstring
American football.18 HSIs result in time away from strength (maximum strength, associated strength
competition,9 nancial burden9 19 and impaired per- ratios and angle of peak torque) or exibility in
formance on return to competition.20 humans with a prior HSI within 24 months
Further to this, recurrent HSI often leads to a from the time of testing;
greater severity of injury than the initial insult.10 14 Have control data for comparison (whether it
The most commonly cited risk factor for future HSI was a contralateral uninjured leg or an uninjured
To cite: Maniar N, is a previous HSI.2124 The high recurrence rates of group);
Shield AJ, Williams MD, HSI10 14 are proposed to result from incomplete Have the full-text journal article in English
et al. Br J Sports Med recovery and/or inadequate rehabilitation25 26 available (excluding reviews, conference
2016;50:909920. because of pressure for early return to play at the abstracts, case studies/series);
Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311 1 of 14

potential confounders that may affect strength or exibility out-

Table 1 Summary of keyword grouping employed during database
comes. The major confounders include other lower limb injuries
likely to affect strength and exibility, interventions and rehabili-
Muscle group Injury Time tation programmes performed. Furthermore, insufcient evidence
exists regarding the interaction between gender and HSI; thus,
Hamstring* Injur* Past
mixed-gender cohorts were considered as a potential confounder.
Semitendinosus Strain* Prior
Semimembranosus Tear Retrospective*
Data analysis
Biceps Femoris Rupture* Previous*
Although objectively synthesising evidence via a meta-analysis is
Posterior Thigh Pull* Recent*
often desirable, this technique was not able to be applied to all
Thigh Trauma Histor*
the evidence retrieved in this review owing to insufcient
reporting of data (ie, two or more studies or subgroups with
*Truncation. Boolean term OR was used within categories, while AND was used mean, SD and participant numbers for contralateral leg compar-
between categories.
isons) or methodological variations between studies.
When sufcient data were available, meta-analysis and graph-
Not include hamstring tendon or avulsion injuries. ical outputs were performed using selected packages5355 on R
The titles and abstracts of each article were scanned by one (R Development Core Team. R: A language and environment
author (NM) and removed if information was clearly inappropriate. for statistical computing. Vienna, Austria: R Foundation for
Selection criteria were then independently applied to the remaining Statistical Computing, 2010). Standardised mean differences
articles by three authors (NM, RGT and DAO). Full text was (Cohens d) facilitated the comparison of studies reporting vari-
obtained for remaining articles, with selection criteria reapplied by ables in different units, with effect estimates and 95% CIs sum-
one author (NM) and cross-referenced by another author (DAO). marised in forest plots. A random-effects model was used to
determine the overall effect estimate of all studies within the vari-
Analysis able or subgroup as appropriate, with variance estimated through
Assessing bias and methodological quality a restricted maximum likelihood (REML) method. The magni-
Risk-of-bias assessment was performed independently by two tudes of the effect size were interpreted as small (d=0.20), moder-
examiners. We used a modied version of a checklist by Downs ate (d=0.50) and large (d=0.80), according to thresholds
and Black.50 The original checklist contained 27 items, however proposed by Cohen.56 Where studies reported multiple types of
many were relevant only to intervention studies. Since the data (eg, multiple isokinetic velocities, multiple subgroups or mul-
majority of the papers in this review were of a retrospective tiple time points), these data were analysed as subgroups to avoid
nature, items 4, 8, 9 13, 14, 15, 17, 19, 22, 23, 24 and 26 were biasing the weighting of the data. These time points were dictated
excluded as they were not relevant to the aims of the review. by the data available. Where data were available in the acute stages
Of the remaining items, 1, 2, 3, 5, 6, 7 and 10 assessed (prior to return to play), time bands were kept at <10 days as it
factors regarding the reporting of aims, methods, data and would be expected that decits would change relatively rapidly
results, while items 16, 18, 20, 21 and 25 assessed internal val- during this time owing to ongoing rehabilitation and recovery.
idity and bias. Item 27 was not suitable to the context of the Data presented for participants at or after return to play were
current review and was modied to address power calculations. pooled for two reasons: (1) no included study reported any
Two new items (items 28 and 29) relating to injury diagnosis ongoing rehabilitation after return to play and (2) many of these
and rehabilitation/interventions were added to more appropri- studies had variable time from injury until testing between in-
ately assess the risk of bias; thus, the modied checklist con- dividual participants. Where a study had multiple time points
tained 17 items (see online supplementary table S1). that t within post return to play time band (eg, at return to
Fourteen of the items were scored 0 if the criterion was not play and follow-up), the earlier option was chosen as there was
met or it was unable to be determined, while successfully met expected to be a lower chance of bias due to other uncontrolled
criteria were scored 1 point. The other three items (items 5, 28 or unmonitored activities. For the purposes of meta-regression
and 29) were scored 0, 1 or 2 points, as dictated by the criteria (employed to assess the effects of time since injury), studies with
presented in online supplementary table S1. This resulted in a multiple time points were pooled to provide the best assessment
total of 20 points available for each article. of the effect of time on the given variable. By considering each
Similarly, modied versions of this checklist have been used in time point as a subgroup, sufcient data (>10 subgroups) were
previous systematic reviews investigating factors leading to heel available for meta-regression analysis57 providing that time
pain51 and risk factors associated with hamstring injury.52 The from injury until testing was reported. Funnel plots were visu-
risk-of-bias assessment was conducted by two authors (NM and ally inspected for asymmetry to assess publication bias.
DAO), with results expressed as a percentage. In the case of dis- Heterogeneity was determined by the I2 statistic and can be
agreement between assessors, an independent individual was interpreted via the following thresholds:57
consulted with consensus reached via discussion if necessary. In 040%: might not be important
situations where one of the assessors (DAO) was a listed author 3060%: may represent moderate heterogeneity
on a study included for review, the independent individual com- 5090%: may represent substantial heterogeneity
pleted the risk-of-bias assessment in their place. 75100%: considerable heterogeneity.
In situations where it was deemed that reported data (ie, mean,
Data extraction SD, participant numbers for contralateral leg comparisons) were
Relevant data were extracted including the participant numbers, insufcient for meta-analysis and could not be obtained via
population and sampling details, diagnosis technique, severity of supplementary material or from contacting the corresponding
injury, time from injury to testing (in days assuming 30.4 days/ author, a best evidence synthesis58 was employed. The level of
month, 365 days/year), variables investigated and how these were evidence was ranked according to criteria consistent with previ-
tested, results including statistical analysis and, where appropriate, ously published systematic reviews59 60 as outlined below:
2 of 14 Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311

Strong: two or more studies of a high quality and generally con- RESULTS
sistent ndings (75% of studies showing consistent results); Search results
Moderate: one high-quality study and/or two or more low- The search strategy consisted of six steps (gure 1). The initial
quality studies and generally consistent ndings (75% of search yielded 7805 items (Cochrane Library=131;
studies showing consistent results); PubMed=2407, CINAHL=604; SPORTDiscus=640; Web of
Limited: one low-quality study; Science=1049; EMBASE=2974) from all databases. After
Conicting: inconsistent ndings (<75% of studies showing duplicates were removed, 4306 items remained. Title and
consistent results); abstract screening resulted in 92 remaining articles, reference list
None: no supportive ndings in the literature. hand searching and citation tracking resulted in the addition of
A high-quality study was dened as a risk-of-bias assessment six articles. Independent application of the selection criteria
score of 70%, whereas a low-quality study had a risk of bias yielded 28 articles to be included in the review, 22 of which
assessment score of <70%.57 were included in meta-analysis.

Figure 1 Flow diagram outlining

steps for study inclusion/exclusion.

Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311 3 of 14


Risk-of-bias assessment Outcomes

Risk-of-bias assessment of each article is displayed in table 2. It The strength variables examined were concentric, eccentric and
is important to note that the risk-of-bias assessment was not the isometric strength (absolute and normalised to body mass),
basis of exclusion. Included articles ranged from a score of 8 to strength ratios (usually hamstring to quadricep (H:Q)) and angle
18 of a possible 20 (4090%). of peak torque. The ve exibility variables examined were
passive straight leg raise, active straight leg raise, passive knee
Description of studies extension, active knee extension and the sit and reach. All ve
Participants strength variables (concentric, eccentric, isometric, strength
A sample of 898 participants (n=802 male, n=96 female; age ratios and angle of peak torque) and three exibility variables
range, 1547 years) were examined across the included studies. A (passive straight leg raise, active knee extension and passive
total of 17 studies included only male participants,29 34 36 37 3943 knee extension) were included for meta-analysis. Sufcient data
45 46 48 49 6164
10 studies had mixed gender,26 28 33 47 6570 were available to run meta-regression analysis for isometric
while only 1 exclusively studied females.71 Participants were gen- strength and the passive straight leg raise. The best evidence syn-
erally considered recreationally active at a minimum. thesis method was applied to remaining variables for which
insufcient data were available for meta-analysis. The best evi-
Injury dence synthesis is summarised in table 3.
Methods of diagnosis varied between studies, with some studies
using multiple methods of diagnosis. A total of 12 studies used
clinical criteria,26 28 33 34 36 37 42 48 6669 10 used Strength
MRI,26 28 29 33 34 62 65 6769 5 had medical or health practi- Concentric strength
tioner diagnosis,39 41 43 48 7 used a questionnaire or self- Data for all studies which examined concentric strength can be
report,40 46 47 49 58 63 71 2 used ultrasound36 37 and 2 had found in online supplementary table S2.
unclear methods of diagnosis.45 70 Description of severity of Meta-analysis. Concentric strength was measured isokinetically
injury varied signicantly between studies, with the most at 60,29 40 48 6163 66 67 71 18029 40 61 71 and 300/s.39 40 62 71
common being time to return to play26 28 29 4043 48 49 63 67 A statistically signicant small effect for lower concentric strength
and grade (IIII) of injury.29 31 33 39 62 66 68 69 70 Description of at 60/s was found in previously injured legs (effect size 0.33;
time from injury to testing varied signicantly between studies 95% CI 0.53 to 0.13; I2 0%), but no signicant effects were
(range, 2690 days). found at 180 or 300/s (gure 2).

Table 2 Itemised scoring of study quality using a modified (see online supplementary table S1) Downs and Black checklist (50)
First author, year 1 2 3 5 6 7 10 11 12 16 18 20 21 25 27 28 29 Total Per cent Quality

Arumugam, 2015 0 1 1 1 1 1 1 0 0 1 1 1 1 0 0 0 0 10 50 Low

Askling, 2006 1 1 1 1 1 1 0 0 0 1 1 1 1 1 0 2 1 14 70 High
Askling, 2010 1 1 1 1 1 1 0 0 0 1 1 1 1 0 0 2 0 12 60 Low
Brockett, 2004 1 1 1 1 1 0 1 0 0 1 1 1 1 0 0 1 1 12 60 Low
Croisier, 2000 1 1 1 1 1 1 0 0 0 1 1 1 0 1 0 1 0 11 55 Low
Croisier, 2002 1 1 1 1 1 1 0 0 0 1 0 1 1 1 0 2 0 12 60 Low
Dauty, 2003 1 1 0 1 1 1 0 0 0 1 1 1 1 1 0 1 0 11 55 Low
Doherty, 2012 1 1 1 1 0 1 1 0 0 1 1 1 1 1 0 0 0 11 55 Low
Hennessey, 1993 1 1 0 0 1 1 0 0 0 1 1 1 1 0 0 0 0 8 40 Low
Jonhagen, 1994 1 1 1 1 1 1 1 0 0 1 1 1 1 1 0 0 0 12 60 Low
Lee, 2009 1 1 1 1 1 1 1 0 0 1 1 0 1 1 0 1 0 12 60 Low
Lowther, 2012 1 1 1 1 1 1 1 0 0 1 1 1 1 0 0 0 0 11 55 Low
Mackey, 2010 1 1 1 1 1 1 1 0 0 1 1 1 1 0 0 0 0 11 55 Low
Opar, 201341 1 1 0 1 1 1 1 0 0 1 1 1 1 1 0 2 0 13 65 Low
Opar, 201329 1 1 1 2 1 1 1 0 0 1 1 1 1 1 0 1 1 15 75 High
Opar, 2015 1 1 1 2 1 1 1 0 0 1 1 1 1 1 1 2 0 16 80 High
OSullivan, 2009 1 1 1 1 1 1 1 0 0 1 1 1 1 1 0 0 0 12 60 Low
OSullivan, 2009 1 1 1 1 1 1 1 0 0 1 1 1 1 1 0 0 0 12 60 Low
OSullivan, 2008 1 1 1 1 1 1 0 0 0 1 1 1 1 1 0 0 0 11 55 Low
Reurink, 2015 1 1 1 2 1 1 1 0 0 1 1 1 1 1 1 2 2 18 90 High
Reurink, 2013 1 1 1 2 1 1 1 0 0 1 1 1 1 1 1 2 2 18 90 High
Sanfilippo, 2013 1 1 1 2 1 1 1 0 0 1 1 1 1 1 0 2 2 17 85 High
Silder, 2010 1 1 1 1 1 1 1 0 0 1 1 1 0 1 0 2 1 14 70 High
Silder, 2013 1 1 1 2 1 1 1 0 0 1 1 1 1 1 0 2 2 17 85 High
Sole, 2011 1 1 1 1 1 1 1 0 0 1 1 1 0 1 0 1 0 12 60 Low
Timmins, 2015 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 2 1 16 80 High
Tol, 2014 1 1 0 1 1 1 1 0 0 1 1 1 1 1 0 2 2 15 75 High
Worrell, 1991 1 1 1 1 1 1 0 0 0 1 1 1 1 1 0 0 0 11 55 Low
A high-quality study was defined as a risk-of-bias assessment score of 70%, whereas a low-quality study had a risk-of-bias assessment score of <70%.

4 of 14 Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311


Table 3 Best evidence synthesis data for all major categories of outcome variables assessed in individuals with a prior hamstring strain injury
not included in the meta-analysis.
Consistency (%)

No. of No Quality Level of evidence

Variable Testing method studies Decrease change Increase (meanSD) of difference

Concentric strength* Seated isokinetic (240/s) 3 100 0 0 6716 Moderate

Eccentric strength Seated isokinetic (30/s) 4 75 25 0 583 Moderate
Eccentric strength Seated isokinetic (120/s) 2 100 0 0 584 Moderate
Isometric Strength Hip 0; knee 90 1 100 0 0 90 Moderate
Concentric strength Seated isokinetic (270/s) 1 100 0 0 60 Limited
Eccentric strength Seated isokinetic (230/s) 1 100 0 0 60 Limited
Eccentric strength Seated isokinetic (300/s) 1 100 0 0 60 Limited
Eccentric hamstring:hip flexor peak Seated/standing isokinetic 1 100 0 0 60 Limited
torque ratio (300/s)
Eccentric angle of peak torque Seated isokinetic (30/s) 1 0 0 100 55 Limited
Flexibility Passive knee extension 3 67 33 0 573 Conflicting
Flexibility Active straight leg raise 2 50 50 0 5014 Conflicting
Consistency refers to the percentage of studies showing a particular outcome.
Outcome variables that had no supporting evidence are not included in this table.
*One study67 showed deficit present at return to play and 6 months postinjury.
Deficit present at initial evaluation and 7-day follow-up.
Deficit assessed post return to play.

Figure 2 Forest plot of concentric strength measured at (A) 60/s, (B) 180/s and (C) 300/s.

Best evidence synthesis. Of the dynamic strength variables strength in the previously injured hamstrings. Concentric strength
which were not included in the meta-analysis, one (seated isokin- at 270/s in a seated position42 had limited evidence and concentric
etic at 240/s)36 37 67 had moderate evidence for a decrease in strength at 60/s in a prone position49 had no supporting evidence.
Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311 5 of 14

Eccentric strength signicant effect for time since injury for isometric strength
Data for all studies which examined eccentric strength can be (intercept 0.92, p=0.002; coefcient 0.003, p=0.292).
found in online supplementary table S3. Best evidence synthesis. One study68 assessed isometric strength
Meta-analysis. Eccentric strength measured during the Nordic in a short muscle length (hip 0, knee 90). This study did not stat-
hamstring exercise34 41 64 and isokinetically at 6029 48 62 63 70 istically test for differences between muscles, but based on effect
and 180/s29 70 were included in the meta-analysis. Signicant size and CIs, isometric strength was reduced at the initial evalu-
decits in previously injured legs were found for eccentric ation (effect size 0.74; 95% CI 1.07 to 0.41) and at the
strength measured via the Nordic hamstring exercise (effect size 7-day follow-up (effect size 0.39; 95% CI 0.71 to 0.07),
0.39; 95% CI 0.77 to 0.00; I2 0%), but not for any other but not at the 26-week follow-up (effect size 0.12; 95% CI
method (gure 3). 0.45 to 0.20).
Best evidence synthesis. Eccentric isokinetic strength measured
at 3036 37 42 61 and 120/s36 37 had moderate evidence, indicat-
ing lower strength in previously injured hamstrings, whereas H:Q torque ratio
measures at 23042 and 300/s39 had limited evidence. The Data for all studies which examined H:Q ratios can be found in
measurement of eccentric strength at 60/s in a prone position49 online supplementary tables S5 and S6.
had no supporting evidence. Meta-analysis. The conventional H:Q ratio, whereby peak
torque of each muscle group is assessed during concentric iso-
Isometric strength kinetic contraction, was assessed at 60:60,36 37 40 43 48 61 70 71
Data for all studies which examined isometric strength can be 180:180,40 61 70 71 240:24036 37 and 300:300/s39 40 71 (gure 6).
found in online supplementary table S4. A statistically signicant small effect for a lower conventional H:
Meta-analysis. Isometric strength measured at long muscle Q ratio was found in previously injured legs compared with the
lengths (hip 0; knee 015) was included in the uninjured contralateral legs at 60:60 (effect size 0.32; 95% CI
meta-analysis.28 34 68 Measures were taken at multiple time 0.54 to 0.11; I2=0%) and 240:240/s (effect size 0.43;
points (<7, 714, 21, 42 and >180 days) postinjury; thus, sub- 95% CI 0.83 to 0.03; I2 0%), but not at 180:180 and
groups were analysed (gure 4) and meta-regression was per- 300:300/s. Meta-analysis of the functional H:Q (fH:Q),
formed. A large effect for lower long-length isometric strength whereby the hamstring group is assessed eccentrically, but the
was statistically signicant in previously injured legs compared quadricep group is assessed concentrically, included isokinetic
with the uninjured contralateral legs <7 days postinjury (effect velocities 30:24036 37 67 and 60:60/s43 48 63 70 (gure 7). A
size 1.72; 95% CI 3.43 to 0.00; I2 91%), but not at any large effect size for a lower fH:Q ratio was found in previously
other time point. Meta-regression analysis (gure 5) revealed no injured legs at 30:240/s (effect size 0.88; 95% CI 1.27 to

Figure 3 Forest plot of eccentric strength measured at (A) 60/s, (B) 180/s and (C) during the Nordic hamstring exercise. Note that one study70
had two subgroups: Doherty 2012a, Division I athletes; Doherty 2012b, Division III athletes.

6 of 14 Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311


Figure 4 Forest plot of isometric strength assessed at (A) <7 days postinjury, (B) 714 days postinjury, (C) 21 days postinjury, (D) 42 days
postinjury and (E) >180 days postinjury. Note that one study28 had two subgroups: Askling 2006a, Sprinters; Askling 2006b, Dancers.

0.48; I2 0%), but there were no signicant differences between Angle of peak torque
injured and uninjured legs at 60:60/s. Data for all studies which examined optimal angle of peak
Best evidence synthesis. One study which examined H:Q torque can be found in online supplementary table S7.
(60:60/s)49 was not included in the meta-analysis owing to the Meta-analysis. The optimal angle of peak torque (concentric
prone and supine position in which knee exor and quadricep 60/s) had sufcient data61 66 67 for meta-analysis. No signicant dif-
strength, respectively, were assessed. This study found no signi- ferences between injured and uninjured legs were found (gure 8).
cant difference between injured and uninjured legs. No support- Best evidence synthesis. Limited evidence was found for the
ing evidence was found for the fH:Q strength ratio at eccentric angle of peak torque to occur at signicantly shorter
180:180,70 30:60 and 30:180/s,61 and limited evidence was muscle lengths in the injured legs compared with the uninjured
found for 300:300/s.39 The eccentric H:Q, whereby both knee contralateral legs at 30/s.61 No differences were found for angle of
exor and quadricep strength are assessed via eccentric contrac- peak torque between legs/groups at 24067 and 300/s39 concentric-
tions, was assessed isokinetically in prone/supine49 position. ally or 300/s39 eccentrically measured angle of peak torque.
This study found no differences between previously injured and
uninjured legs. Limited evidence was found for eccentric knee Flexibility
exor torque to concentric hip exor torque ratio decits in Passive straight leg raise
previously injured legs (effect size 0.9) compared with unin- Data for all studies which examined the passive straight leg raise
jured contralateral legs.39 can be found in online supplementary table S8.
Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311 7 of 14

1.12; 95% CI 1.76 to 0.48; I2 81%), a moderate effect

between 10 and 20 days (effect size 0.74; 95% CI 1.38 to
0.09; I2 76%) and a small effect between 20 and 30 days
(effect size 0.40; 95% CI 0.78 to 0.03; I2 4%) since the
time of injury, with no signicant effect found after 40 days
since the time of injury (gure 9). Meta-regression analysis
(gure 10) revealed a signicant effect for time since injury
(intercept 0.81, p<0.0001; coefcient 0.006, p=0.019), indi-
cating that the magnitude of the range of motion decit decreases
with increasing time from injury.

Figure 5 Meta-regression plot (with 95% CI) for isometric strength. Passive knee extension
Intercept 0.92, p=0.002; coefcient 0.003, p=0.292. Data for all studies which examined the passive knee extension
can be found in online supplementary table S9.
Meta-analysis. No signicant differences were found for the
Meta-analysis. Quantitative analysis of the passive straight leg passive knee extension measure at either time point subgroup
raise26 28 62 68 revealed signicantly reduced range of motion in analysed (<10 and 2030 days postinjury; gure 11A,B).
previously injured legs compared with the uninjured contralat- Best evidence synthesis. A subset of the passive knee extension
eral leg. A large effect was found within 10 days (effect size (insufcient data for subgroup meta-analysis, unable to be

Figure 6 Forest plot of conventional H:Q ratio assessed at (A) 60:60/s, (B) 180:180/s, (C) 240:240/s and (D) 300:300/s. Note that one study70
had two subgroups: Doherty 2012a, Division I athletes; Doherty 2012b, Division III athletes. H:Q, hamstring to quadricep.
8 of 14 Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311

Figure 7 Forest plot of the fH:Q ratio assessed at (A) 30:240/s and (B) 60:60/s. Note that one study70 had two subgroups: Doherty 2012a,
Division I athletes; Doherty 2012b, Division III athletes. fH:Q, functional hamstring to quadricep.

pooled with acute data) showed conicting evidence across the DISCUSSION
three studies46 47 49 that conducted this assessment post return Our systematic review revealed that after HSI, isometric strength
to play. and passive straight leg raise decits normalised within 20
50 days. Decits at or after return to play, if they did exist, mani-
fested during dynamic strength measures (eccentric and concen-
Active knee extension
tric strength and their associated H:Q strength ratios).
Data for all studies which examined the active knee extension
We only included research articles that contained data from
can be found in online supplementary table S9.
participants who had previously sustained a HSI (between 2 and
Meta-analysis. No signicant differences were found for the
690 days prior). As a result, we cannot determine whether the
active knee extension measure at either time point subgroup ana-
reported decits were the cause of injury or the result of injury.
lysed (<10, 1030 and >100 days postinjury; gure 11CE).
Given the increased risk of future HSI in those with an injury
history,2124 the characteristics that exist in these legs should be
Active straight leg raise given consideration by the clinicians responsible for rehabilita-
Data for all studies which examined the active straight leg raise tion and clearance to return to play.
can be found in online supplementary table S8.
Best evidence synthesis. Conicting evidence was found for
Strength and exibility decits after hamstring injury
decits in the active straight leg raise.45 65 Of note, the one
Conventional rehabilitation practice traditionally focuses on
study65 which did nd decits in previously injured legs per-
restoring isometric strength and range of motion.72 The
formed the active straight leg raise in a rapid manner (Askling-H
meta-analysis revealed that decits in long length (hip 0; knee
test), and as such this study could not be appropriately pooled
015) isometric strength and the passive straight leg raise are
with the other data for meta-analysis purposes.
resolved 2050 days postinjury. This provides support for the
use of the passive straight leg raise and isometric strength mea-
Sit and reach sures during rehabilitation.72 Furthermore, decits in isometric
Best evidence synthesis. No evidence for differences in the sit strength and range of motion (as measured by the active knee
and reach was found between healthy and previously injured extension test) just after return to play are independent predic-
participants.48 63 tors of reinjury,73 suggesting that these variables likely also have

Figure 8 Forest plot for angle of peak torque assessed during 60/s concentric contraction.

Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311 9 of 14


Figure 9 Forest plot of the passive straight leg raise at (A) <10 days postinjury, (B) 1020 days postinjury, (C) 2030 days postinjury and (D) >40 days
postinjury. Note that two studies26 28 had two subgroups: Askling 2006a, Sprinters; Askling 2006b, Dancers; Silder 2013a, progressive agility and trunk
stabilisation (PATS) rehabilitation protocol; Silder 2013b, progressive running and eccentric strengthening (PRES) rehabilitation protocol.

value in criterion-based rehabilitation progressions. However, likelihood that the demands of high-force musculotendinous
where evidence of decits were found beyond return to play, lengthening, such as during the terminal swing phase of running,
these were during measures of dynamic strength. exceed the mechanical limits of the tissue.74 It may be that lower
The evidence supporting decits in eccentric strength in those eccentric strength in previously injured hamstrings is at least partly
with prior HSI is mixed.29 34 36 37 39 4143 48 63 64 70 Lower responsible for the greater risk of recurrent hamstring strain.75
levels of eccentric hamstring strength are proposed to increase the Other measures of dynamic strength, including concentric
strength29 33 36 37 40 48 6163 66 67 71 and both
conventional33 36 37 39 40 43 48 61 66 70 71 and
functional36 37 39 43 48 61 63 67 70 H:Q strength ratios, also show
conicting ndings, with measures at some testing velocities
showing lower strength in previously injured legs, but others
showing no differences. The reasons for these discrepancies are
unclear, but it may be due to inherent differences in groups
studied and/or methodological issues. For example, studies which
included females tended to observe slightly higher strength in pre-
viously injured legs.70 71 Insufcient data were available to assess
this observation via regression analysis; thus, more research is
needed to investigate any potential gender-specic responses to
HSI. The particulars of the rehabilitation performed could also
Figure 10 Meta-regression plot (with 95% CI) for the passive straight explain the disparity, as differing rehabilitation strategies would
leg raise. Intercept 0.81, p<0.0001; coefcient 0.006, p=0.019. result in differing adaptations. Rehabilitation was rarely controlled
10 of 14 Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311

Figure 11 Forest plot for the knee extension assessments of range of motion at (A) passive, <10 days postinjury, (B) passive, 2030 days
postinjury, (C) active, <10 days postinjury, (D) active, 1030 days postinjury and (E) active, >100 days postinjury. Note that one study26 had two
subgroups: (A) progressive agility and trunk stabilisation (PATS) rehabilitation protocol; (B) progressive running and eccentric strengthening (PRES)
rehabilitation protocol.

in the included studies, suggesting that more studies should aim to not occur during concentric contractions;29 48 thus, further
control rehabilitation to limit this potential confounder. research is needed to understand why dynamic strength decits
tend to persist beyond return to play.
Mechanisms that may explain long-term dynamic muscle
strength decits Clinical implications
There is the possibility that chronic decit/s in dynamic strength The data presented in this review have implications for practi-
in previously hamstring strain injured legs is a downstream tioners who rehabilitate and return athletes to play following HSIs.
outcome of prolonged neuromuscular inhibition.35 Reduced The supplementary result tables provide practitioners a detailed
activation of previously injured hamstrings has been associated resource of data for almost all strength and exibility measures that
with maximal eccentric contractions,29 30 48 76 particularly at have been assessed in athletes with a prior HSI. These data can be
long muscle lengths.29 48 What remains to be seen, however, is used to compare individual athlete/patient data. It should also
whether or not these decits are associated with increased risk enable practitioners to select measures to monitor in their injured
of injury or reinjury, and what the most appropriate intervention athletes which are known to be in decit despite successful return
is to ameliorate these decits. However, activation decits do to play. The presented evidence justies the use of the passive
Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311 11 of 14

straight leg raise and isometric strength measures to monitor pro- acknowledge that the search strategy may not have captured all
gression through rehabilitation, while additional measures of relevant literature. However, reference list searching and cit-
dynamic strength may have more value at and after return to play. ation tracking were also performed to enhance article retrieval.
In addition, the present review would also question the use of
commonly recommended74 77 and employed markers for suc- CONCLUSION
cessful rehabilitation, such as knee exor angle of peak torque. In conclusion, the meta-analysis found that decits in isometric
The use of angle of peak knee exor torque, particularly during strength and exibility (as measured by the passive straight leg raise)
concentric contraction, in athletes with prior HSI has been resolve within 2050 days following HSI. Decits that were present
popularised following the seminal paper;66 however, the beyond return to play were found for dynamic measures of strength
ensuing evidence is generally conicting,33 39 61 67 suggesting (concentric and eccentric strength, and conventional and functional
that the value of this measure should be questioned. H:Q strength ratios). This evidence suggests that clinicians monitor
isometric strength and the passive straight leg raise throughout
Limitations rehabilitation, while dynamic measures of strength may hold more
The primary limitation of this review is that the retrospective value at/after return to play. Furthermore, it may behove clinicians
nature of the data makes it impossible to determine if decits are and patients to continue rehabilitation after return to play.
the cause or result of injury. For example, eccentric strength de-
cits could be the result of uncorrected strength deciency that
may have caused injury, as higher levels of eccentric strength and What are the ndings?
eccentric training are associated with a reduction in new and
recurrent HSI.73 78 79 Furthermore, the majority of the included
studies did not control rehabilitation, and this introduces another
After HSI,
potential source of bias. For example, a study in which partici-
Isometric strength returns to the level of the contralateral
pants focused heavily on eccentric exercise as part of rehabilita-
uninjured leg within 20 days.
tion may show no evidence of signicant eccentric strength
Range of motion measured by the passive straight leg raise
decits post HSI. Consequently, the effect of these interventions
returns to the level of the contralateral uninjured leg within
on strength and exibility outcomes remains an area for future
50 days.
research. Ideally, researchers should control rehabilitation to min-
Lower dynamic strength (concentric, eccentric and associated
imise confounding, and where this is not possible, collect and
strength ratios) in previously injured legs compared with the
report details of rehabilitation protocols. Inconsistent time from
uninjured contralateral legs persists beyond return to play,
injury until testing between studies also introduces bias. We ana-
but this is inconsistent across measurement technique.
lysed data in time bands and performed meta-regression analysis
where possible to assess and adjust for this potential confounder,
but we also acknowledge that this approach was limited by
within-study variability, variability between studies within the How might it impact on clinical practice in the future?
time band subgroups and insufcient data for regression analysis.
Future research should investigate the effect of time since injury Isometric strength and the passive straight leg raise provide
on decits, particularly prior to return to play, as strength and a measure of progression during rehabilitation.
exibility appear to change rapidly during this period. Dynamic strength (concentric/eccentric hamstring strength
One of the difculties of this review was the numerous methods and associated hamstring to quadricep strength ratios) may
employed by different studies to assess a given parameter. For also be helpful in monitoring progress through rehabilitation
strength testing, it appeared that lower isokinetic velocities (<60/ and return-to-play decisions.
s) were the most sensitive to decits; however, there are insufcient This review adds weight to the argument that rehabilitation
data at higher velocities to draw denitive conclusions. Similarly, a should continue after return to play if the goal is to achieve
number of different measures of exibility (passive26 28 42 65 and symmetry in strength and range of motion.
active45 65 straight leg raise, passive26 46 47 49 and active knee
extension,26 48 sit and reach test48 63) have been assessed in previ-
ously injured athletes, with inconsistent ndings among studies.
Twitter Follow Nirav Maniar @niravmaniar91, Anthony Shield @das_shield,
Indeed, within each variable, the meta-analysis revealed signicant Morgan Williams @drmorgs, Ryan Timmins @ryan_timmins and David Opar
heterogeneity as determined by the I2 statistic in certain measures, @davidopar
particularly in the initial days following injury. Acknowledgements The primary authors position was supported through the
To address these issues as far as possible, we performed sensi- Australian Governments Collaborative Research Networks (CRN) program. The
tivity analysis (see online supplementary table S10) to examine authors would also like to sincerely thank Professor Geraldine Naughton for acting
the inuence of individual studies on effect estimates and hetero- as an independent assessor for the risk-of-bias assessment.
geneity where moderate (30%) heterogeneity57 may have been Contributors NM conducted the search, risk-of-bias and criteria assessments,
present. While high heterogeneity often impairs the validity of extracted the data, performed all analysis and drafted the manuscript. AJS and
MDW contributed to interpretation of results and the manuscript. RGT conducted
synthesised data, the low number of studies in many of these criteria assessments and contributed to the manuscript. DAO conducted risk-of-bias
subgroups precludes condence in the precision in these I2 esti- and criteria assessments and contributed to the interpretation of results and the
mates, suggesting that more studies are needed to properly manuscript.
interpret heterogeneity estimates. These studies should also Funding Australian Governments Collaborative Research Networks (CRN).
take care to accurately describe diagnostic procedures, injury Competing interests DAO and AJS are listed as coinventors on an international
severity and other lower limb injuries likely to confound patent application lled for the experimental device (PCT/AU2012/001041.2012)
results. The data reported in this review may also have limited used in three of the included studies in this review. The authors declare no other
application to female athletes, as majority of the data were competing interests.
obtained from male only or predominately male cohorts. We Provenance and peer review Not commissioned; externally peer reviewed.

12 of 14 Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311


REFERENCES 31 Silder A, Heiderscheit BC, Thelen DG, et al. MR observations of long-term

1 Gabbe B, Finch C, Wajswelner H, et al. Australian football: injury prole at the musculotendon remodeling following a hamstring strain injury. Skeletal Radiol
community level. J Sci Med Sport 2002;5:14960. 2008;37:11019.
2 Orchard J, Seward H. Epidemiology of injuries in The Australian Football League, 32 Silder A, Reeder SB, Thelen DG. The inuence of prior hamstring injury on
seasons 19972000. Br J Sports Med 2002;36:3944. lengthening muscle tissue mechanics. J Biomech 2010;43:225460.
3 Orchard J, Seward H. Injury report 2009: Australian football league. Sport Health 33 Silder A, Thelen DG, Heiderscheit BC. Effects of prior hamstring strain injury on
2010;28:10. strength, exibility, and running mechanics. Clin Biomech (Bristol, Avon)
4 Orchard J, Seward H. Injury Report 2008: Australian Football League. Sport Health 2010;25:6816.
2009;27:29. 34 Timmins RG, Shield AJ, Williams MD, et al. Biceps femoris long head architecture: a
5 Seward H, Orchard J, Hazard H, et al. Football injuries in Australia at the elite level. reliability and retrospective injury study. Med Sci Sports Exerc 2015;47:90513.
Med J Aust 1993;159:298301. 35 Fyfe JJ, Opar DA, Williams MD, et al. The role of neuromuscular inhibition in
6 Ekstrand J, Gillquist J. Soccer injuries and their mechanisms: a prospective study. hamstring strain injury recurrence. J Electromyogr Kinesiol 2013;23:52330.
Med Sci Sports Exerc 1983;15:26770. 36 Croisier JL, Crielaard JM. Hamstring muscle tear with recurrent complaints: an
7 Ekstrand J, Hgglund M, Waldn M. Injury incidence and injury patterns in isokinetic prole. Isokinet Exerc Sci 2000;8:17580.
professional football: the UEFA injury study. Br J Sports Med 2011;45:5538. 37 Croisier JL, Forthomme B, Namurois MH, et al. Hamstring muscle strain recurrence
8 Hawkins R, Hulse M, Wilkinson C, et al. The association football medical research and strength performance disorders. Am J Sports Med 2002;30:199203.
programme: an audit of injuries in professional football. Br J Sports Med 38 Freckleton G, Cook J, Pizzari T. The predictive validity of a single leg bridge test for
2001;35:437. hamstring injuries in Australian Rules Football Players. Br J Sports Med
9 Woods C, Hawkins R, Hulse M, et al. The Football Association Medical Research 2014;48:71317.
Programme: an audit of injuries in professional football-analysis of preseason 39 Lee MJ, Reid SL, Elliott BC, et al. Running biomechanics and lower limb strength
injuries. Br J Sports Med 2002;36:43641. associated with prior hamstring injury. Med Sci Sports Exerc 2009;41:194251.
10 Woods C, Hawkins R, Maltby S, et al. The Football Association Medical Research 40 OSullivan K, OCeallaigh B, et al. The relationship between previous hamstring
Programme: an audit of injuries in professional footballanalysis of hamstring injury and the concentric isokinetic knee muscle strength of Irish Gaelic footballers.
injuries. Br J Sports Med 2004;38:3641. BMC Musculoskelet Disord 2008;9:30.
11 Brooks JH, Fuller CW, Kemp SP, et al. Epidemiology of injuries in English 41 Opar DA, Piatkowski T, Williams MD, et al. A novel device using the Nordic
professional rugby union: part 1 match injuries. Br J Sports Med 2005;39: hamstring exercise to assess eccentric knee exor strength: a reliability and
75766. retrospective injury study. J Orthop Sports Phys Ther 2013;43:63640.
12 Brooks JH, Fuller CW, Kemp SP, et al. Epidemiology of injuries in English 42 Jnhagen S, Nmeth G, Eriksson E. Hamstring injuries in sprinters. The role of
professional rugby union: part 2 training Injuries. Br J Sports Med 2005;39:76775. concentric and eccentric hamstring muscle strength and exibility. Am J Sports Med
13 Brooks JH, Fuller C, Kemp S, et al. A prospective study of injuries and training 1994;22:2626.
amongst the England 2003 Rugby World Cup squad. Br J Sports Med 43 Dauty M, Potiron-Josse M, Rochcongar P. Identication of previous hamstring
2005;39:28893. muscle injury by isokinetic concentric and eccentric torque measurement in elite
14 Brooks JH, Fuller CW, Kemp SP, et al. Incidence, risk, and prevention of hamstring soccer player. Isokinet Exerc Sci 2003;11:13944.
muscle injuries in professional rugby union. Am J Sports Med 2006;34:1297306. 44 Houweling TAW, Head A, Hamzeh MA. Validity of isokinetic testing for
15 Sugiura Y, Saito T, Sakuraba K, et al. Strength decits identied with concentric previous hamstring injury detection in soccer players. Isokinet Exerc Sci
action of the hip extensors and eccentric action of the hamstrings predispose to 2009;17:21320.
hamstring injury in elite sprinters. J Orthop Sports Phys Ther 2008;38:45764. 45 Hennessey L, Watson AW. Flexibility and posture assessment in relation to hamstring
16 Yeung SS, Suen AM, Yeung EW. A prospective cohort study of hamstring injuries in injury. Br J Sports Med 1993;27:2436.
competitive sprinters: preseason muscle imbalance as a possible risk factor. Br J 46 Lowther D, OConnor A, Clifford AM, et al. The relationship between lower limb
Sports Med 2009;43:58994. exibility and hamstring injury in Male Gaelic footballers. Physiother Ireland
17 Bennell KL, Crossley K. Musculoskeletal injuries in track and eld: incidence, 2012;33:228.
distribution and risk factors. Aust J Sci Med Sport 1996;28:6975. 47 OSullivan K, Murray E, et al. The effect of warm-up, static stretching and dynamic
18 Feeley BT, Kennelly S, Barnes RP, et al. Epidemiology of National Football League stretching on hamstring exibility in previously injured subjects. BMC Musculoskelet
training camp injuries from 1998 to 2007. Am J Sports Med 2008;36:1597603. Disord 2009;10:37.
19 Hickey J, Shield AJ, Williams MD, et al. The nancial cost of hamstring strain 48 Sole G, Milosavljevic S, Nicholson HD, et al. Selective strength loss and
injuries in The Australian Football League. Br J Sports Med 2014;48:72930. decreased muscle activity in hamstring injury. J Orthop Sports Phys Ther
20 Verrall GM, Kalairajah Y, Slavotinek JP, et al. Assessment of player performance 2011;41:35463.
following return to sport after hamstring muscle strain injury. J Sci Med Sport 49 Worrell TW, Perrin DH, Gansneder BM, et al. Comparison of isokinetic strength and
2006;9:8790. exibility measures between hamstring injured and noninjured athletes. J Orthop
21 Arnason A, Sigurdsson SB, Gudmundsson A, et al. Risk factors for injuries in Sports Phys Ther 1991;13:11825.
football. Am J Sports Med 2004;32(1 Suppl):5S16S. 50 Downs SH, Black N. The feasibility of creating a checklist for the assessment of the
22 Orchard JW. Intrinsic and Extrinsic Risk Factors for Muscle Strains in Australian methodological quality both of randomised and non-randomised studies of health
Football Neither the author nor the related institution has received any nancial care interventions. J Epidemiol Community Health 1998;52:37784.
benet from research in this study. Am J Sports Med 2001;29:3003. 51 Irving DB, Cook JL, Menz HB. Factors associated with chronic plantar heel pain:
23 Hgglund M, Waldn M, Ekstrand J. Previous injury as a risk factor for injury in elite a systematic review. J Sci Med Sport 2006;9:1122; discussion 3-4.
football: a prospective study over two consecutive seasons. Br J Sports Med 52 Freckleton G, Pizzari T. Risk factors for hamstring muscle strain injury in sport:
2006;40:76772. a systematic review and meta-analysis. Br J Sports Med 2013;47:3518.
24 Verrall GM, Slavotinek JP, Barnes PG, et al. Clinical risk factors for hamstring 53 Schwarzer G. Meta: general package for meta-analysis. 2015.
muscle strain injury: a prospective study with correlation of injury by magnetic 54 Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat
resonance imaging. Br J Sports Med 2001;35:4359; discussion 40. Softw 2010;36:148.
25 Connell DA, Schneider-Kolsky ME, Hoving JL, et al. Longitudinal study comparing 55 Wickham H. ggplot2: elegant graphics for data analysis. Springer Science &
sonographic and MRI assessments of acute and healing hamstring injuries. AJR Am Business Media, 2009.
J Roentgenol 2004;183:97584. 56 Cohen J. A power primer. Psychol Bull 1992;112:155.
26 Silder AMY, Sherry MA, Sanlippo J, et al. Clinical and morphological changes 57 Higgins J, Green S. Cochrane handbook for systematic reviews of interventions
following 2 rehabilitation programs for acute hamstring strain injuries: a randomized version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011.
clinical trial. J Orthop Sports Phys Ther 2013;43:28499. 58 Slavin RE. Best evidence synthesis: An intelligent alternative to meta-analysis. J Clin
27 Orchard J, Best TM. The management of muscle strain injuries: an early return Epidemiol 1995;48:918.
versus the risk of recurrence. Clin J Sport Med 2002;12:35. 59 de Visser HM, Reijman M, Heijboer MP, et al. Risk factors of recurrent hamstring
28 Askling C, Saartok T, Thorstensson A. Type of acute hamstring strain affects injuries: a systematic review. Br J Sports Med 2012;46:12430.
exibility, strength, and time to return to pre-injury level. Br J Sports Med 60 Serner A, van Eijck CH, Beumer BR, et al. Study quality on groin injury
2006;40:404. management remains low: a systematic review on treatment of groin pain in
29 Opar DA, Williams MD, Timmins RG, et al. Knee exor strength and bicep femoris athletes. Br J Sports Med 2015;49:813.
electromyographical activity is lower in previously strained hamstrings. 61 Mackey C, OSullivan K, et al. Altered hamstring strength prole in Gaelic
J Electromyogr Kinesiol 2013;23:696703. footballers with a previous hamstring injury. Isokinetics Exerc Sci 2011;19:4754.
30 Opar DA, Williams MD, Timmins RG, et al. Rate of torque and electromyographic 62 Tol JL, Hamilton B, Eirale C, et al. At return to play following hamstring injury the
development during anticipated eccentric contraction is lower in previously strained majority of professional football players have residual isokinetic decits. Br J Sports
hamstrings. Am J Sports Med 2013;41:11625. Med 2014;48:13649.

Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311 13 of 14

63 Arumugam A, Milosavljevic S, Woodley S, et al. Effects of external pelvic 71 OSullivan K, Burns S. Comparing concentric isokinetic thigh muscle strength in
compression on isokinetic strength of the thigh muscles in sportsmen with and female gaelic football players with and without previous hamstring injury. Physiother
without hamstring injuries. J Sci Med Sport 2015;18:2838. Ireland 2009;30:3944.
64 Opar DA, Williams MD, Timmins RG, et al. The effect of previous hamstring strain 72 Heiderscheit BC, Sherry MA, Silder A, et al. Hamstring strain injuries:
injuries on the change in eccentric hamstring strength during preseason training in recommendations for diagnosis, rehabilitation and injury prevention. J Orthop Sports
elite Australian footballers. Am J Sports Med 2015;43:37784. Phys Ther 2010;40:6781.
65 Askling CM, Nilsson J, Thorstensson A. A new hamstring test to complement the 73 De Vos RJ, Reurink G, Goudswaard GJ, et al. Clinical ndings just after return to
common clinical examination before return to sport after injury. Knee Surg Sports play predict hamstring re-injury, but baseline MRI ndings do not. Br J Sports Med
Traumatol Arthrosc 2010;18:1798803. 2014;48:137784.
66 Brockett CL, Morgan DL, Proske U. Predicting hamstring injury in elite athletes. Med 74 Croisier JL. Factors associated with recurrent hamstring injuries. Sports Med
Sci Sports Exerc 2004;36:37987. 2004;34:68195.
67 Sanlippo JL, Silder A, Sherry MA, et al. Hamstring strength and morphology 75 Opar DA, Williams MD, Timmins RG, et al. Eccentric hamstring strength and
progression after return to sport from injury. Med Sci Sports Exerc 2013;45: hamstring injury risk in Australian footballers. Med Sci Sports Exerc 2015;47:85765.
44854. 76 Bourne MN, Opar DA, Williams MD, et al. Muscle activation patterns in the Nordic
68 Reurink G, Goudswaard GJ, Moen MH, et al, Dutch HIT-study Investigators. hamstring exercise: impact of prior strain injury. Scand J Med Sci Sports 2015.
Rationale, secondary outcome scores and 1-year follow-up of a randomised trial of 77 Mendiguchia J, Brughelli M. A return-to-sport algorithm for acute hamstring
platelet-rich plasma injections in acute hamstring muscle injury: the Dutch injuries. Phys Ther Sport 2011;12:214.
Hamstring Injection Therapy study. Br J Sports Med 2015;49:120612. 78 Askling C, Karlsson J, Thorstensson A. Hamstring injury occurrence in elite soccer
69 Reurink G, Goudswaard GJ, Oomen HG, et al. Reliability of the active and passive players after preseason strength training with eccentric overload. Scand J Med Sci
knee extension test in acute hamstring injuries. Am J Sports Med 2013;41:175761. Sports 2003;13:24450.
70 Doherty J, Van Lunen BL, Ismaeli ZC, et al. Hamstring strength measurements in 79 Petersen J, Thorborg K, Nielsen MB, et al. Preventive effect of eccentric training on
collegiate athletes with a history of hamstring injury. Athl Train Sports Health Care acute hamstring injuries in mens soccer a cluster-randomized controlled trial. Am J
2012;4:3844. Sports Med 2011;39:2296303.

14 of 14 Maniar N, et al. Br J Sports Med 2016;50:909920. doi:10.1136/bjsports-2015-095311