Content uploaded by Leslie L Nicholson
Author content
All content in this area was uploaded by Leslie L Nicholson
Content may be subject to copyright.
Generalized Joint Hypermobility
and Risk of Lower Limb Joint Injury
During Sport
A Systematic Review With Meta-Analysis
Verity Pacey,
*
yz
GradDip (Sports Physiotherapy), Leslie L. Nicholson,
y
PhD,
Roger D. Adams,
y
PhD, Joanne Munn,
y
PhD, and Craig F. Munns,
§||
MBBS, PhD, FRACP
From the
y
Discipline of Physiotherapy, The University of Sydney, Sydney, New South Wales,
Australia, the
z
Physiotherapy Department, and
§
Department of Endocrinology, The Children’s
Hospital at Westmead, Westmead, New South Wales, Australia, and
||
Discipline of Pediatrics and
Child Health, The University of Sydney, Sydney, New South Wales, Australia
Background: Generalized joint hypermobility is a highly prevalent condition commonly associated with joint injuries. The current
literature has conflicting reports of the risk of joint injury in hypermobile sporting participants compared with their nonhypermobile
peers. Systematic reviews have not been conclusive and no meta-analysis has been performed.
Purpose: This review was undertaken to determine whether individuals with generalized joint hypermobility have an increased
risk of lower limb joint injury when undertaking sporting activities.
Study Design: Systematic review with meta-analysis.
Methods: Studies were identified through a search without language restrictions of PubMed, CINAHL, Embase, and SportDiscus
databases from the earliest date through February 2009 with subsequent handsearching of reference lists. Inclusion criteria for
studies were determined before searching and all included studies underwent methodological quality assessment by 2 indepen-
dent reviewers. Meta-analyses for joint injury of the lower limb, knee, and ankle were performed using a random effects model.
The difference in injury proportions between hypermobility categories was tested with the zstatistic.
Results: Of 4841 identified studies, 18 met all inclusion criteria with methodological quality ranging from 1 of 6 to 5 of 6. A variety
of tests of hypermobility and varied cutoff points to define the presence of generalized joint hypermobility were used, so the
authors determined a standardized cutoff to indicate generalized joint hypermobility. Using this criterion, a significantly increased
risk of knee joint injury for hypermobile and extremely hypermobile participants compared with their nonhypermobile peers was
demonstrated (P\.001), whereas no increased risk was found for ankle joint injury. For knee joint injury, a combined odds ratio of
4.69 (95% confidence interval, 1.33-16.52; P5.02) was calculated, indicating a significantly increased risk for hypermobile
participants playing contact sports.
Conclusion: Sport participants with generalized joint hypermobility have an increased risk of knee joint injury during contact
activities but have no altered risk of ankle joint injury.
Keywords: hypermobility; knee; ankle; joint; sports injury; meta-analysis
Generalized joint hypermobility (GJH) is a condition in
which most of an individual’s synovial joints move beyond
the ‘‘normal’’ limits, with the age, gender, and ethnic
background of the individual taken into account.
14
Because
of differing definitions and case identifications, the preva-
lence of GJH in published reports varies from 5% to 43% in
adults
7,20
and 2% to 55% in children.
30
Generalized joint
hypermobility is also a recognized feature of many her-
itable disorders of connective tissue, such as Ehlers-Danlos
syndrome,Osteogenesis Imperfecta, and Marfansyndrome.
45
Many of these disorders are associated with symptoms of
chronic fatigue and widespread musculoskeletal pain,
which may result from the GJH.
45
The increased connective tissue flexibility in GJH is
considered to be of primarily genetic origin, given its com-
mon autosomal dominant presentation.
15
Although many
M
*
Address correspondence to Verity Pacey, GradDip (Sports Physio-
therapy), Physiotherapy Department, The Children’s Hospital at West-
mead, Locked Bag 4001, Westmead, NSW 2145, Australia (e-mail:
verityp@chw.edu.au).
The authors declared that they had no conflicts of interests in their
authorship and publication of this contribution.
The American Journal of Sports Medicine, Vol. 38, No. 7
DOI: 10.1177/0363546510364838
Ó2010 The Author(s)
1487
Winner of the 2009 Systematic Review Competition
of the genes responsible for the monogenetic disorders
associated with GJH have been identified,
45
the cause of
idiopathic GJH, including those with benign joint hyper-
mobility syndrome, requires further investigation.
Dislocations, subluxations, and sprains are commonly
reported in individuals with GJH
1
and it is assumed
that the risk of such injuries is magnified during activi-
ties that are more physically challenging, particularly
where the lower limbs are involved. However, reports in
the current literature are inconclusive as to whether the
risk of lower limb joint injury during sport is greater in
hypermobile participants compared with their nonhyper-
mobile peers. Conflicting evidence of the relationship
between hypermobility and joint injuries has been
reported among ballet dancers
16,22
and gridiron play-
ers.
21,32
To date, systematic reviews have also been
unable to definitively determine any difference in the
risk of lower limb joint injury sustained by hypermobile
sporting participants
18,29
and no meta-analysis has been
performed.
This situation has led to varying recommendations
from clinicians and researchers advising individuals
with GJH on the risks incurred by sports participation.
Advising caution when giving advice on sports participa-
tion for those with GJH, recommendations include
participation in noncontact activities only, such as swim-
ming, pilates, and tai chi
40
; or greater caution so that ‘‘rel-
atively lax individuals should avoid physical exertion at
a higher than normal rhythm.’’
13
Others suggest that
hypermobile participants can undertake sporting activi-
ties such as netball, stipulating that they should do so
with the use of strapping and supports in order to limit
injury.
41
Conversely, Murray
30
recommends full involve-
ment in sporting activities for pain-free hypermobile
individuals.
Physical activity is routinely prescribed for patients
with chronic conditions; however, patients with GJH
need a statement of the best information available on the
risks associated with such participation. Accordingly, the
aim of the current study is to systematically review the lit-
erature to determine whether people with GJH are at sig-
nificantly greater risk of lower limb joint injury if they par-
ticipate in sporting activities.
METHODS
Identification of Studies
Eligible studies were identified through a search without
language restrictions of PubMed, CINAHL, Embase, and
SportDiscus databases from the earliest date through
February 2009. The search strategy (Table 1) was formed
by the authors in conjunction with an experienced medi-
cal librarian. Handsearching of reference lists of all
included studies and relevant review papers was also
performed.
All studies identified by the search were screened by the
first author using the inclusion criteria and checked by
a second author as required.
Inclusion Criteria
Studies were included if they used a prospective design with
an objective scale to measure GJH and included participants
who were undertaking any type of sports activity. To be
included in the review, studies also had to be peer reviewed
and have objective, quantitative injury data. Injury defini-
tion included injury diagnosis by a health professional, self-
reported dysfunction, or time lost to athletic participation.
Methodological Quality Assessment
Methodological quality was assessed using the 6-point
scale developed for prognostic studies by Pengel et al,
35
TABLE 1
PubMed Search Strategy
a
1. Causal*[tw]
2. Causation[tw]
3. Pred*[tw]
4. Risk*[tw]
5. Risk assessment[mh]
6. Risk factors[mh]
7. Odds ratio[mh]
8. Clinical trials[mh]
9. Prevalence[mh]
10. Prevalence studies[mh]
11. Incidence studies[mh]
12. Incidence[mh]
13. Epidemiological studies[mh]
14. Or/1-13
15. Hip injuries[mh]
16. Foot injuries[mh]
17. Knee injuries[mh]
18. Athletic injuries[mh]
19. Injur*[tw]
20. Sublux*[tw]
21. Disloc*[tw]
22. Sprain*[tw]
23. Sprains or strains[mh]
24. Dislocations[mh]
25. Ankle injuries[mh]
26. Or/15-25
27. Joint instability[mh]
28. Hypermob*[tw]
29. Hyper-mob*[tw]
30. Joint laxity[tw]
31. Joint instability*[tw]
32. Ligament*AND instability*[tw]
33. Ligament*AND laxity[tw]
34. Beighton[tw]
35. Hyperflex*[tw]
36. Hyperextens*[tw]
37. Or/27-36
38. 14 AND 26 AND 37
a
This search strategy was modified for searches of other
databases.
Database search terms: tw, textword; mh, MeSH term (biomed-
ical term assigned to describe the subject of indexed articles);
*wildcard (search for terms that begin with the letters preceding
asterisk).
1488 Pacey et al The American Journal of Sports Medicine
used previously in reviews of prognosis of musculoskeletal
conditions.
10,28
Two reviewers independently assessed the
quality of each study. Ambiguities were resolved through
discussion, with a third reviewer consulted when agree-
ment could not be reached.
Data Extraction and Analysis
Data extracted from each included study by 2 authors
included participant age, gender, and sporting activity;
number of participants; length of follow-up; measure and
definition of GJH; definition of injury used; injuries stud-
ied; and reported injury data. In line with the American
Academy of Pediatrics system, sporting activities were
classified into 3 groups—collision or contact sports, limited
contact sports, and noncontact sports.
38
Four studies
23-26
included sporting activities in more than 1 of these groups
and were therefore classified into a fourth group referred
to as mixed sports. Where studies reported insufficient
data to be included in a meta-analysis, an attempt was
made to contact the authors by e-mail. From this, 4 sets
of more detailed data were obtained.
16,23,26,33
Odds ratios and 95% confidence intervals (CIs) were
generated using StatsDirect software (version 2.7.2,
StatsDirect Ltd, Altrincham, Cheshire, United Kingdom).
Because an odds ratio of 1 demonstrates that an injury is
equally as likely to occur in the hypermobile group as
the nonhypermobile group, where the odds ratio is
greater than 1, and the 95% CI does not include 1,
39
then statistical significance was accepted as P\.05.
Forest plots were generated (Figures 1-4) and the area
of each square symbol (odds ratio) is proportional to the
study’s weight in the meta-analysis, reflecting the weight-
ing of each study’s results in the random effects analysis
model. The open diamond symbol represents the com-
bined odds ratio.
Because of differences among the included studies in
the participants, sports, and measures used, the DerSimo-
nian and Laird random effects method
12
was used for the
meta-analysis and I
2
was calculated to determine the level
of heterogeneity, with an I
2
value of 100% representing
Odds ratio meta-analysis plot [random effects]
0.2 0.5 1 2 5 10
Soderman et al - CONTACT 1.27 (0.59, 2.72
)
Krivickas & Feinberg - MIXED 0.65 (0.27, 1.45
)
Diaz, Estevez & Guijo - CONTACT 3.40 (1.66, 7.15
)
combined [random] 1.43 (0.56, 3.67
)
odds ratio (95% confidence interval)
Figure 1. Meta-analysis of risk of lower limb joint injury to
hypermobile sports participants utilizing the standardized
generalized joint hypermobility definition.
Odds ratio meta-analysis plot [random effects]
0.2 0.5 1 2 5 10 100
Ostenberg & Roos - CONTACT 5.32 (1.59, 18.26)
Nicholas - CONTACT 25.74 (8.82, 77.28
)
Krivickas & Feinberg - MIXED 0.72 (0.20, 2.15)
Diaz, Estevez & Guijo - CONTACT 2.52 (1.02, 6.31)
combined [random] 3.98 (0.95, 16.55)
odds ratio (95% confidence interval)
Figure 2. Meta-analysis of risk of knee joint injury to hyper-
mobile sports participants utilizing the standardized general-
ized joint hypermobility definition.
Odds ratio meta-analysis plot [random effects]
0.1 0.2 0.5 1 2 5 10 100
Ostenberg & Roos 5.32 (1.59, 18.26)
Nicholas 25.74 (8.82, 77.28
)
Krivickas & Feinberg 1.19 (0.18, 5.69)
Diaz, Estevez & Guijo 2.52 (1.02, 6.31)
combined [random] 4.69 (1.33, 16.52)
odds ratio (95% confidence interval)
Figure 3. Meta-analysis of risk of knee joint injury to hyper-
mobile participants of contact sports utilizing the standard-
ized generalized joint hypermobility definition.
Odds ratio meta-analysis plot [random effects]
0.10.2 0.5 1 2 5 10 100
McHugh et al - MIXED 0.82 (0.32, 2.02)
Krivickas & Feinberg - MIXED 0.48 (0.11, 1.51)
Hiller et al - NON-CONTACT 1.40 (0.57, 3.41)
Diaz, Estevez & Guijo - CONTACT 5.33 (1.51, 23.50
)
Beynnon et al - CONTACT 1.38 (0.22, 6.01)
combined [random] 1.28 (0.62, 2.63)
odds ratio (95% confidence interval)
Figure 4. Meta-analysis of risk of ankle joint injury to hyper-
mobile sports participants utilizing the standardized general-
ized joint hypermobility definition.
Vol. 38, No. 7, 2010 Hypermobility and Joint Injuries 1489
TABLE 2
Included Studies
a
Author(s)
(Year)
Participants’
Gender and Age
Participants’
Sporting Activity
Length
of Follow-up
Injuries
Studied
Injury
Definition
Author(s)’ Conclusion:
Does GJH Affect
the Risk of Injury?
Baumhauer
et al
5
(1995)
73 males, 72
females aged 18-
23 years
Lacrosse, soccer,
field hockey
(college)
Not stated Lateral
ankle
sprains
Trainer review Ankle: No significant
difference
Beynnon et al
6
(2001)
50 males, 68
females aged
18-23 years
Lacrosse, soccer,
field hockey
(college)
Not stated Lateral
ankle
sprains
Medical review Ankle: No significant
difference
Davies and
Gibson
9
(1978)
185 males aged
23 65 years
Rugby union (1st
class—British,
Welsh, and
English)
1 season Any injury,
any body
part
Self-report: Temporary
interruption to game OR
impaired ability to train
or play
Overall: No significant
difference
Decoster et al
11
(1999)
147 males, 163
females aged 20 6
4 years
Lacrosse (college) 1 season Any injury,
any body
part
Trainer report: Injury
resulting in missing at
least 1 practice or game
Overall: No significant
difference
Ankle: Increased risk
Diaz et al
13
(1993)
675 males aged 17
years
Soldier recruits 2 months Any injury,
any body
part
Medical review: Lesions
or alterations to the
locomotor system
requiring treatment
Overall: Increased
risk
Hiller et al
16
(2008)
21 males, 94
females aged
12-16 years
Ballet and dance
(high school)
Up to 13
months
(until 1st
lateral
ankle
sprain)
Lateral
ankle
sprains
Self-report: Swelling or
bruising from inversion
injury and limping
for .1 day
Ankle: No significant
difference
Hopper et al
17
(1995)
72 females aged 15-
36 years
Netball (A grade) 14 weeks Any injury,
any body
part
Medical review Overall: No significant
difference
Kalenak and
Morehouse
21
(1975)
72 males American football
(college)
1-4 seasons Knee
ligaments
Medical review: Medial,
lateral, or anterior
cruciate ligament
Knee: No increased
risk
Krivickas and
Feinberg
23
(1996)
131 males, 70
females aged
17-21 years
Football, baseball,
basketball, soccer,
cross-country,
golf, diving, and
others (college)
1 year Any type of
back or
lower
extremity
injury
Trainer or medical
review: Injury that
limited activity or
prevented play in
practice or game
Overall: Decreased
risk
Lompa et al
24
(1998)
71 males, 34
females aged
9-24 years
Amateur
gymnastics,
volleyball and
basketball
12 months Any injury,
any body
part
Medical review Overall: Increased
risk
Lysens et al
25
(1989)
118 males, 67
females aged
17-19 years
Physical education
(college)
1 year Any injury,
any body
part
Medical review: Injury
during sport session
resulting in 3 days
absence from sport
Overuse: Increase risk
McHugh et al
26
(2006)
101 males, 68
females, aged
14-18 years
Athletes (high
school)
2 years Lateral
ankle
sprains
Trainer review: Ankle
injury with inversion
mechanism resulting in
missing 1 game or
practice
Ankle: No significant
difference
Nicholas
32
(1970)
139 males American football
(professional)
At least 1
season
Knee
ligaments
Medical review: 3°
ligament rupture
requiring surgery within
2 weeks due to
instability
Knee: Increased risk
Ostenberg and
Roos
33
(2000)
123 females aged
14-39 years
Soccer (all levels) 1 season (7
months)
Any injury,
any body
part
Medical review: Absence
from at least 1 practice
or game
Overall: Increased
risk
Knee: Increased risk
(continued)
1490 Pacey et al The American Journal of Sports Medicine
complete heterogeneity. Potential publication bias was not
tested because of the substantial heterogeneity and low
number of studies included within the meta-analysis.
19, 44
Although differing definitions of GJH were identified
between studies, obtaining the original raw data sets
allowed us to use a standardized criterion of 4of9on
the Beighton scale to indicate GJH, as recommended by
the British Society of Rheumatology.
37
Where the scale
used to determine the extent of hypermobility was not
a 9-point scale, the point closest in percentage to 4 of 9
(44.44%) was used (eg, 2 of 5, 3 of 6).
Therefore, odds ratios for all lower limb, knee, and
ankle joint injuries related to GJH are reported in 2
ways: first, by the original authors’ definition; and second,
by the standardized definition using the British Society of
Rheumatology criterion for defining GJH (4 of 9 or equiv-
alent). Thus the odds ratios for each area (ankle, knee,
overall lower limb) and for each definition (authors’ and
standardized) are pooled from differing numbers of studies
and individual data sets, depending on the data that could
be extracted.
Participants in the 6 studies that provided hypermobil-
ity scores as continuous data
13,16,23,26,32,42
were then clas-
sified on the basis of extent of hypermobility into 1 of 3
categories—not hypermobile (0 of 9 to 3 of 9 or equiva-
lent), hypermobile (4 of 9 to 6 of 9 or equivalent), or
extremely hypermobile (7 of 9 to 9 of 9 or equivalent)—
withamethodusedpreviously.
13,23,43
The 95% CIs for
the injury rates of participants in each of these 3 catego-
ries were calculated, and the difference in injury propor-
tions between categories was tested with the zstatistic.
3
The number of individual data sets able to be used in
each calculation differs from the odds ratio calculations
depending on the number of data sets able to be catego-
rized into these 3 groups.
RESULTS
The search identified 4841 studies, of which 18 met all
inclusion criteria (Table 2). Of the 18 studies, 4 considered
only ankle injuries,
5,6,16,26
3 considered only knee inju-
ries,
21,32,46
1 study pooled all leg injuries,
42
and all others
considered both lower limb and injuries to other areas of
the body.
{
Attempts were made to contact authors of all
studies by e-mail with a request to provide de-identified
data of sufficient detail to be included in the meta-
analyses. Authors were not contacted if the complete
data were available within the publication, as was the
case for 2 studies. Three of the 10 authors who responded
were able to provide the required data. No specific injury
data relating to GJH could be obtained for 8 of the 18
articles and consequently these studies are not included
in the meta-analyses. Hip joint injury data were only
available from 1 study
23
that reported 3 hip joint injuries,
generating an odds ratio of 1.33 (95% CI, 0.12-14.94; P5
.82) utilizing the standardized definition of GJH.
TABLE 2 (continued)
Author(s)
(Year)
Participants’
Gender and Age
Participants’
Sporting Activity
Length
of Follow-up
Injuries
Studied
Injury
Definition
Author(s)’ Conclusion:
Does GJH Affect
the Risk of Injury?
Pasque and
Hewett
34
(2000)
418 males aged 14-
19 years
Wrestling (high
school)
1 season (3
months)
Any injury,
any body
part
Medical review: Any
condition reducing
function resulting in
seeking medical care OR
Self-report: Loss of
practice or match and
loss of 1 day from
athletic participation
Overall: No significant
difference
Soderman
et al
42
(2001)
146 females aged
20.6 64.7 years
Soccer (2nd and
3rd division)
1 season (7
months)
Leg injuries Self- or coach report:
Absence from 1 practice
or game
Traumatic : Increased
risk
Stewart and
Burden
43
(2004)
51 males aged 23.6
63.3 years
Rugby union (1st
class New
Zealand)
1 season Shoulder,
hip, knee,
ankle,
wrist, and
hand
injuries
Medical review: Any
condition reducing
function requiring
treatment
Overall: Increased
risk
Uhorchak
et al
46
(2003)
1021 males, 177
females aged 17-
23 years
Cadets 4 years ACL rupture Medical review: ACL
rupture on arthroscopy
Overall: Increased
risk
a
GJH, generalized joint hypermobility; ACL, anterior cruciate ligament.
{References 9, 13, 17, 23-25, 33, 34, 43, 46.
Vol. 38, No. 7, 2010 Hypermobility and Joint Injuries 1491
Methodological Quality
The 2 reviewers scored 108 quality criteria and initially
agreed on 94 (87%) of these (k50.74; 95% CI, 0.62-0.87).
Further discussion led to agreement on 107 (99%) (k5
0.98; 95% CI, 0.95-1.0), requiring only 1 quality criterion
to be decided by a third reviewer. Nine studies (50%)
defined the population from which the sample was drawn,
5 (28%) clearly described methods for assembling a repre-
sentative sample, and 10 (56%) reported follow-up of at
least 80%. All studies (100%) quantified prognosis, no
study reported blinded assessment of outcome measures,
and 9 studies (53%) reported statistical adjustments. Final
methodological quality scores for these studies ranged
from 1 of 6 to 5 of 6 (Table 3).
Objective Measures Used
Seven different measures of GJH were used by the 18 stud-
ies, incorporating varying cutoff points to indicate the pres-
ence of hypermobility (Table 4). Of the 8 studies that
reported using the modified 9-point Beighton scale, 4
described different criteria for positive identification.
Joint Injury
All Lower Limb Joint Injuries. Original authors’ definition:
Odds ratios and confidence intervals for all lower limb joint
injuries were calculated from 3 studies using the original
authors’ definition of GJH (Table 5) giving a combined odds
ratio of 1.71 (95% CI, 0.60-4.85; P5.31; I
2
580.9%).
Standardized definition: Calculation of odds ratios
using the standardized definition of GJH was possible for
the same 3 studies, enabling pooling of 1047 individual
data sets (Figure 1) and demonstrating that 14% of the
sporting participants suffered all lower limb joint injuries.
Odds ratios ranged from 0.65 to 3.40 with a combined odds
ratio of 1.43 (95% CI, 0.56-3.67; P5.46; I
2
581%). Figure
5 reveals that overlap between the 95% CIs of each mobil-
ity status group and the difference in the proportion of
participants with lower limb joint injuries between nonhy-
permobile participants and the combined hypermobile and
extremely hypermobile participants was not significant (z
51.69; P5.09).
Knee Joint Injuries. Original authors’ definition: From
the 5 studies that reported knee injury data, a significant
relationship between hypermobility status and risk of
knee joint injury was found with a combined odds ratio of
2.62 (95% CI, 1.04-6.58; P5.04; I
2
574.1) (Table 5).
Standardized definition: Four studies assessed injury
risk in male army recruits, female soccer players, profes-
sional male American football players, and athletes
engaged in mixed college level sports allowed pooling of
1167 individual data sets (Figure 2) and calculation of
odds ratios using the standardized definition of GJH. The
overall knee injury rate in these studies was 8.65%, with
odds ratios for knee joint injury in those with GJH ranging
from 0.72 to 25.74, giving a combined odds ratio of 3.98
(95% CI, 0.95-16.55; P5.06; I
2
588.4%).
TABLE 3
Methodological Rating
Author(s) (Year) Defined
Sample
a
Representative
Sample
b
Complete
Follow-up
c
Prognosis
d
Blinded
Outcome
e
Statistical
Adjustment
f
Methodological
Score
Baumhauer et al
5
(1995) Yes Yes No Yes No No 3
Beynnon et al
6
(2001) Yes No Yes Yes No Yes 4
Davies and Gibson
9
(1978) No No Yes Yes No No 2
Decoster et al
11
(1999) No Yes No Yes No No 2
Diaz et al
13
(1993) Yes No No Yes No N/A 2
Hiller et al
16
(2008) Yes Yes Yes Yes No Yes 5
Hopper et al
17
(1995) No No Yes Yes No Yes 3
Kalenak and Morehouse
21
(1975) No No No Yes No No 1
Krivickas and Feinberg
23
(1996) Yes No Yes Yes No Yes 4
Lompa et al
24
(1998) Yes No Yes Yes No No 3
Lysens et al
25
(1989) No No Yes Yes No Yes 3
McHugh et al
26
(2006) No No Yes Yes No Yes 3
Nicholas
32
(1970) Yes Yes No Yes No No 3
Ostenberg and Roos
33
(2000) No No Yes Yes No Yes 3
Pasque and Hewett
34
(2000) No No Yes Yes No No 2
Soderman et al
42
(2001) Yes No No Yes No Yes 3
Stewart and Burden
43
(2004) No No No Yes No No 1
Uhorchak et al
46
(2003) Yes Yes No Yes No Yes 4
a
Description of source of participants with defined inclusion and exclusion criteria.
b
Participants randomly selected or consecutive cases.
c
At least 1 prognostic factor available from 80% of study population at 3-month follow-up or later.
d
Studies must provide raw data, percentages, survival rates, or continuous outcomes.
e
Assessor unaware of at least 1 prognostic factor, used to predict prognostic outcome, at time prognostic outcome was measured.
f
For at least 2 prognostic factors with adjustment factor reported.
1492 Pacey et al The American Journal of Sports Medicine
Data were extracted from 4 studies comprising 1043 par-
ticipants in the contact activities of soccer, gridiron, basket-
ball, diving,
38
and Army recruit training (Figure 3). Army
recruit training was included as a contact activity because
of the nature of combatant activities and the high-impact
body contact with the ground likely to be involved. Using
the standardized definition of GJH, a combined odds ratio
of 4.69 (95% CI, 1.33-16.52; P5.02; I
2
582.81%) was calcu-
lated, indicating a significant increased risk of knee joint
injury for hypermobile participants playing contact sports.
Three studies of participants in contact sports
13,23,32
provided knee injury data that could be categorized into
the 3 mobility-status groups (not hypermobile, hypermo-
bile, and extremely hypermobile, using the standardized
definition of GJH). When the proportion of nonhypermobile
participants who sustained knee injuries was compared
with those who were hypermobile or extremely hypermo-
bile (combined), the difference was statistically significant
(z54.08; P\.001), with the hypermobile groups injured
more often (Figure 5).
TABLE 4
Objective Measures Used
a
Objective Measure of Hypermobility Description of Measure (Including Variations) Cutoff Point Indicating Hypermobility
Modified 9-point Beighton
5,6,11,16,17,23,33,43
5th finger MCP dorsiflexion .90°
(passively extend fingers parallel to forearm
11
)
Passive thumb to forearm
Hyperextension of elbows .10°
Hyperextension of knees .10°(.15°
43
)
Palms flat on the floor
3
17
4
5,6,16,23,33,43
5
11
6-point Beighton and Horan
34
Passive 5th finger MCP joint extension .90°
Thumb to forearm passively
Elbow hyperextension .10°
Knee hyperextension .10°
Ankle hyperextension .45°
Palms to floor with knees straight
Cutoff not used, treated
as continuous scale
Modified 5-point Carter and Wilkinson
13,24
Passive dorsiflexion of 5th finger .90°
(passive hyperextension of the fingers
parallel to the forearm
24
)
Thumb to forearm
Hyperextension of elbow (.5°
13
)
Hyperextension of knee (.5°
13
)
Palms on floor (femoral anteversion
24
)
2
13
3
24
Modified 10-point Carter and Wilkinson
42
Passive hyperextension of fingers to
parallel with forearm
Passive thumb to forearm flexor aspect
Elbow hyperextension .10°
Knee hyperextension .10°
Dorsiflexion of ankle .30°
5
42
5-point Nicholas
21,26,32
Upper extremity laxity—hypothenar eminence
inclines cephalad in vertical plane with elbows
extended and forearms supinated (hyperextension
of the elbow with the wrist in supination and
the shoulder flexed to 90°
26
)
Hyperextension of the knee 20°(.10°
26
)
Over pivot ability—feet .180°degrees heel to
heel and toes out (hip, knee, and ankle max
external rotation with knees flexed 15°-30°)
Knees or ankles parallel to floor when lying or
sitting in either external rotation (lotus position)
or internal rotation (thumb to forearm
with wrist flexed
26
)
Flex spine so palms touch floor with knees
fully extended
1
21,32
3
26
8-point Wynne and Davies
46
Thumb to volar aspect of forearm
5th MCP hyperextension .90°
Elbow hyperextension
Knee hyperextension
5
46
a
Original references cited within text with no description of test provided for 5-point Beighton and Horan
9
or combination of joint laxity,
looseness, and mobility test.
25
MCP, metacarpophalangeal joint.
Vol. 38, No. 7, 2010 Hypermobility and Joint Injuries 1493
Ankle Joint Injuries. Original authors’ definition: Odds
ratios and CIs for ankle joint injury were calculated from 6
studies using the original authors’ definition of GJH (Table
5). The combined odds ratio was 1.34 (95% CI, 0.69-2.60;
P5.39; I
2
555.2%).
Standardized definition: Pooling 1361 individual data
sets from the studies providing data able to use the stan-
dardized definition of GJH, 8.74% of participants suffered
ankle joint injuries. The combined odds ratio of 1.28 (95%
CI, 0.62-2.63; P5.51; I
2
559.4%) was not significantly
different from 1 (Figure 4). Combined data from 4 stud-
ies
13,16,23,26
providing the continuous scores from 1244
sports participants demonstrated the 95% CIs between
the nonhypermobile, hypermobile, and extremely hyper-
mobile groups all overlapping (Figure 5). There was,
therefore, no significant difference in ankle injury rates
between these groups or the nonhypermobile and
combined hypermobile and extremely hypermobile group
(z51.22; P5.22).
Joint Injuries Incurred During Participation in
Different Sports: Contact, Limited Contact,
Noncontact, or Mixed
Relating lower limb, knee, and ankle joint injuries to the
type of sporting activity undertaken (contact, limited con-
tact, noncontact, or mixed), differences became apparent.
Studies involving mixed sporting activities had lower
odds ratios compared with those obtained for the contact
or collision sports, with no odds ratio equal to or above
1.00 for any of the mixed sporting activity studies, whereas
all collision or contact sporting activity studies generated
odds ratios greater than 1 when using the standardized
definition of GJH (Figures 1, 2, and 4).
DISCUSSION
The findings of this review and meta-analyses indicate
that for those people who have GJH, there is a signifi-
cantly increased risk of injury to the knee joint during
participation in contact sports; however, the ankle joint
is not at an increased risk of injury during any sporting
participation.
Much of the current literature is consistent with the
findings of the present review. Several prospective
studies of ankle injuries incurred during a range of
sporting activities have been undertaken, with most find-
ing no significantly increased risk of injury associated
with GJH.
5,6,16,23,26
A recent systematic review of the pre-
dictors of ankle injury reported that it was reduced ankle
dorsiflexion range of motion (hypomobility at the ankle)
that was a strong predictor of ankle sprains.
10
With
respect to the knee joint, our findings support studies
suggesting that individuals with increased knee hyperex-
tension are at increased risk of anterior cruciate
ligament injury,
31,36
one of the most common knee joint
injuries.
TABLE 5
Odds Ratios and 95% Confidence Intervals for Studies Using the Authors’ Definition of GJH
a
Author(s) (Year) Type of Sport Area of Injury Odds Ratio 95% Confidence Interval
Beynnon et al
6
(2001) Contact Ankle 1.38 0.22-6.01
Diaz et al
13
(1993) Contact Lower limb
Knee
Ankle
3.40
2.52
5.33
1.66-7.15
1.02-6.31
1.51-23.50
Decoster et al
11
(1999) Contact Ankle 1.99 0.57-6.30
Hiller et al
16
(2008) Noncontact Ankle 1.40 0.57-3.41
Kalenak and Morehouse
21
(1975) Contact Knee 1.08 0.36-3.24
Krivickas and Feinberg
23
(1996) Mixed Lower limb
Knee
Ankle
0.65
0.72
0.48
0.27-1.45
0.20-2.15
0.11-1.51
McHugh et al
26
(2006) Mixed Ankle 0.61 0.15-1.93
Nicholas
32
(1970) Contact Knee 15.56 3.61-138.32
Ostenberg and Roos
33
(2000) Contact Knee 5.32 1.59-18.26
Soderman et al
42
(2001) Contact Lower limb 2.24 0.79-6.71
a
GJH, generalized joint hypermobility.
Percent with injury
0
5
10
15
20
25
30
35
Knee Ankle
Injury location
**
** Not hypermobile
Hypermobile
Extremely hypermobile
All lower limb
Figure 5. Proportion of injuries by hypermobility status. Bars
represent 95% confidence intervals. **Difference in propor-
tions of injuries significant at P\.001 level.
1494 Pacey et al The American Journal of Sports Medicine
The increased risk of knee injury but not ankle injury in
sports participants with GJH has not been reported previ-
ously; however, it is consistent with ankle and knee joint
anatomy and biomechanics. Ankle stability relies on both
active (musculotendinous) and passive (ligamentous)
restraints to prevent injury, whereas the knee relies to
a greater extent on passive restraints. The most common
ankle injury, lateral ankle sprain, occurs as a result of
unrestrained inversion and plantar flexion.
8
Along with
the passive restraints to inversion provided by the lateral
ligament complex, the peroneus longus and brevis also pro-
vide active restraint. Both active and passive tissues, in
conjunction with the bony congruency of the talocrural
joint, provide restraint to all planes of the movement at
the ankle joint.
2
However, the knee has less bony congru-
ency and the alignment and action of surrounding muscu-
lature offer little active joint restraint. Most knee joint
injuries occur either at the end of extension range, or
involve unrestrained rotation, varus, or valgus force.
8
While the large hamstring muscle group, sartorius and
gracilis, provide some active restraint to the end range of
knee extension, their moment arms are small, rendering
them unlikely to be able to provide sufficient torque to
restrain extension. During rotatory, valgus, and varus
motion, passive restraint is provided by tension in the cru-
ciate, medial, and lateral collateral ligaments, respectively.
There is minimal active control of rotation and no active
control of varus or valgus movement at the knee joint,
27
which relies almost completely on the passive ligamentous
and capsular restraints to prevent injury. When the foot
remains in contact with the ground as forces are applied
to the joints (such as in the contact sport of football, where
the player wears studded or cleated shoes), greater inter-
nal torques are created at the knee joint than at the ankle
joint because of the increased distance from the ground,
again further predisposing the knee joint to injury. The
sporting participant with GJH may rely more on their
dynamic muscular control to maintain joint stability of
their lax lower limb joints than their nonhypermobile
peers, placing them at greater risk of musculotendinous
as well as capsuloligamentous injury.
The intention of the current study was to combine the
results from many available studies. However, of the 18
included studies, the results from 10 studies were ineligi-
ble for meta-analysis because of the method of reporting
data. Meta-analysis performed with access to individual
patient data sets is highly desirable as it allows consis-
tency of data analysis.
4
Variation in the definition and assessment of GJH is
evident within this review. Seven different objective meas-
ures of GJH, involving 10 differing measurement methods,
were used in the 18 studies. These methods of determining
GJH vary in terms of the particular joints assessed, the
range considered hypermobile, and the application of an
injury allowance point (Table 4). While all the tests aim
to be measures of GJH, upper limb mobility measures
are highly represented in the Beighton scale, while lower
limb measures dominate the Nicholas scale, raising ques-
tions about the face validity of either scale in relation to
generalized or whole body mobility.
Several differing cutoff points to indicate the presence of
GJH were used for the same tests of GJH reported within
this review. Every effort was made to establish the risk of
injury related to GJH utilizing all available data across the
18 studies by using the standardized criterion; however,
conversion of the 5-point scale to a 9-point scale may be
somewhat ambitious. The odds ratios calculated using
the authors’ original definition of GJH differed from those
calculated using the standardized definition of GJH, yet
the overall results were similar.
Definitional inconsistencies encountered during this
review were not limited to those concerning GJH. Different
definitions of ‘‘injury’’ were also used among the studies
reviewed. These varied from diagnostic definitions, for
example ligament rupture confirmed under arthroscopy,
32
to functional definitions such as absence from training or
a game because of injury.
33
Increased uniformity in the
reporting of injury incidence in prospective cohort trials
is recommended.
One of the confounders to determining the risk of lower
limb joint injury for those with GJH is that the studies
available cover a wide range of sports (from ballet to grid-
iron) as well as a range of levels of participation (from rec-
reational to occupational and elite). A previous review was
inconclusive as to whether professional competition or
amateur participation resulted in a greater risk of injury.
29
Sporting activities included in the present systematic
review varied extensively in both the demands of the sport-
ing tasks undertaken and the level at which participation
occurred (Table 2). The injury rate varied within the
included studies, which is likely to be attributable to
multiple factors, including not only sporting activity
undertaken but also the definition of ‘‘injury’’ used. Inves-
tigating risk of injury related to type of sport by categoriz-
ing those sports according to the American Academy of
Pediatrics classification
38
revealed a marked difference
between contact and other sporting activities. Generalized
joint hypermobility may indeed be protective against
injury in some limited contact and noncontact sports; how-
ever, the results of this study indicate that there is an
increased risk of injury, particularly to the knee, for partic-
ipants with GJH during contact activities.
In conclusion, the risk of ankle injury while participat-
ing in sporting activities is not altered by the presence of
GJH, yet individuals with GJH do have an increased risk
of knee injury during sporting activities, particularly dur-
ing contact sporting activities. Improved consistency in the
measurement of GJH and definitions of injury used within
research studies may assist in providing further evidence
as to which sports are associated with the least risk to
hypermobile individuals.
ACKNOWLEDGMENT
The authors acknowledge Dr Louise Tofts for her assistance
in reviewing original data supplied by authors, Dr Paulo
Ferreira for the English translation of a Portuguese article,
and Dr Colleen Canning for her assistance with graphical
representation of data.
Vol. 38, No. 7, 2010 Hypermobility and Joint Injuries 1495
An online CME course associated with this article is
available for 1 AMA PRA Category 1 Credit
TM
at
http://ajsm-cme.sagepub.com. In accordance with the
standards of the Accreditation Council for Continuing
Medical Education (ACCME), it is the policy of The
American Orthopaedic Society for Sports Medicine
that authors, editors, and planners disclose to the learn-
ers all financial relationships during the past 12 months
with any commercial interest (A ‘commercial interest’ is
any entity producing, marketing, re-selling, or distrib-
uting health care goods or services consumed by, or
used on, patients). Any and all disclosures are provided
in the online journal CME area which is provided to all
participants before they actually take the CME activity.
In accordance with AOSSM policy, authors, editors, and
planners’ participation in this educational activity will
be predicated upon timely submission and review of
AOSSM disclosure. Noncompliance will result in an
author/editor or planner to be stricken from participat-
ing in this CME activity.
REFERENCES
1. Adib N, Davies K, Grahame R, Woo P, Murray KJ. Joint hypermobility
syndrome in childhood: a not so benign multisystem disorder? Rheu-
matology (Oxford). 2005;44:744-750.
2. Alter MJ. Anatomy and flexibility of the lower extremity and pelvic gir-
dle. In: Alter MJ, ed. Science of Flexibility. Champaign, Illinois:
Human Kinetics; 1996:239-240.
3. Altman DG. Comparing groups—categorical data. In: Altman DG, ed.
Practical Statistics for Medical Research. London: Chapman & Hall;
1991:234.
4. Altman DG. Systematic reviews of evaluation of prognostic variables.
BMJ. 2001;323:224-228.
5. Baumhauer JF, Alosa DM, Renstrom PA, Trevino S, Beynnon B. A
prospective study of ankle injury risk factors. Am J Sports Med.
1995;23:564-570.
6. Beynnon BD, Renstrom PA, Alosa DM, Baumhauer JF, Vacek PM.
Ankle ligament injury risk factors: a prospective study of college ath-
letes. J Orthop Res. 2001;19:213-220.
7. Birrell FN, Adebajo A, Hazleman BL, Silman AJ. High prevalence of
joint laxity in West Africans. Br J Rheumatol. 1994;33:56-59.
8. Brukner P, Khan K. Clinical Sports Medicine. Sydney: McGraw-Hill
Book Co; 1993.
9. Davies JE, Gibson T. Injuries in Rugby Union football. Br Med J.
1978;2:1759-1761.
10. de Noronha MA, Refshauge KM, Herbert RD, Kilbreath SL, Hertel J.
Do voluntary strength, proprioception, range of motion or postural
sway predict occurrence of lateral ankle sprain? Br J Sports Med.
2006;40:824-828.
11. Decoster LC, Bernier JN, Lindsay RH, Vailas JC. Generalized joint
hypermobility and its relationship to injury patterns among NCAA
lacrosse players. J Athl Train. 1999;34:99-105.
12. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin
Trials. 1986;7:177-188.
13. Diaz MA, Estevez EC, Sanchez Guijo P. Joint hyperlaxity and
musculoligamentous lesions: Study of a population of homoge-
neous age, sex and physical exertion. Br J Rheumatol. 1993;32:
120-122.
14. Grahame R. Hypermobility and hypermobility syndrome. In: Keer R,
Grahame R, eds. Hypermobility Syndrome—Recognition and Man-
agement for Physiotherapists. London: Butterworth-Heinemann;
2003:1-14.
15. Hakim AJ, Cherkas LF, Grahame R, Spector TD, MacGregor AJ. The
genetic epidemiology of joint hypermobility: a population study of
female twins. Arthritis Rheum. 2004;50:2640-2644.
16. Hiller CE, Refshauge KM, Herbert RD, Kilbreath SL. Intrinsic predic-
tors of lateral ankle sprain in adolescent dancers: a prospective
cohort study. Clin J Sport Med. 2008;18:44-48.
17. Hopper DM, Hopper JL, Elliott BC. Do selected kinanthropometric
and performance variables predict injuries in female netball players?
J Sports Sci. 1995;13:213-222.
18. Huston LJ, Greenfield ML, Wojtys EM. Anterior cruciate ligament inju-
ries in the female athlete: potential risk factors. Clin Orthop Relat Res.
2000;372:50-63.
19. Ioannidis JP, Trikalinos TA. The appropriateness of asymmetry tests
for publication bias in meta-analyses: a large survey. CMAJ.
2007;176:1091-1096.
20. Jessee EF, Owen DS, Sagar KB. The benign hypermobility syn-
drome. Arthritis Rheum. 1980;23:1053-1056.
21. Kalenak A, Morehouse CA. Knee stability and knee ligament injuries.
JAMA. 1975;234:1143-1145.
22. Klemp P, Stevens JE, Isaacs S. A hypermobility study in ballet
dancers. J Rheumatol. 1984;11:692-696.
23. Krivickas LS, Feinberg JH. Lower extremity injuries in college ath-
letes: relation between ligamentous laxity and lower extremity muscle
tightness. Arch Phys Med Rehabil. 1996;77:1139-1143.
24. Lompa PA, Schio CL, Muller LM, Mallmann LF. Incidence of sports
lesions in athletes with and without familiar articular hypermobility
syndrome. Rev Bras Ortop. 1998;33(12):933-938.
25. Lysens RJ, Ostyn MS, Auweele YV, Lefevre J, Vuylsteke M, Renson
L. The accident-prone and overuse-prone profiles of the young ath-
lete. Am J Sports Med. 1989;17:612-619.
26. McHugh MP, Tyler TF, Tetro DT, Mullaney MJ, Nicholas SJ. Risk fac-
tors for noncontact ankle sprains in high school athletes: the role of
hip strength and balance ability. Am J Sports Med. 2006;34:464-470.
27. McMinn RMH, Hutchings RT, Pegington J, Abrahams PH. A Colour
Atlas of Human Anatomy. 3rd ed. London: Mosby-Year Book Europe
Ltd; 1993.
28. Muaidi QI, Nicholson LL, Refshauge KM, Herbert RD, Maher CG.
Prognosis of conservatively managed anterior cruciate ligament
injury: a systematic review. Sports Med. 2007;37:703-716.
29. Murphy DF, Connolly DAJ, Beynnon BD. Risk factors for lower extrem-
ity injury: a review of the literature. Br J Sports Med. 2003;37:13-29.
30. Murray KJ. Hypermobility disorders in children and adolescents. Best
Prac Res Clin Rheumatol. 2006;20:329-351.
31. Myer GD, Ford KR, Paterno MV, Nick TG, Hewett TE. The effects of
generalized joint laxity on risk of anterior cruciate ligament injury in
young female athletes. Am J Sports Med. 2008;36:1073-1080.
32. Nicholas JA. Injuries to knee ligaments: relationship to looseness and
tightness in football players. JAMA. 1970;212:2236-2239.
33. Ostenberg A, Roos H. Injury risk factors in female European football:
a prospective study of 123 players during one season. Scand J Med
Sci Sports. 2000;10:279-285.
34. Pasque CB, Hewett TE. A prospective study of high school wrestling
injuries. Am J Sports Med. 2000;28:509-515.
35. Pengel LH, Herbert RD, Maher CG, Refshauge KM. Acute low back
pain: systematic review of its prognosis. Br Med J. 2003;327:323-
327.
36. Ramesh R, Von Arx O, Azzopardi T, Schranz PJ. The risk of anterior
cruciate ligament rupture with generalised joint laxity. J Bone Joint
Surg Br. 2005;87:800-803.
37. Remvig L, Jensen DV, Ward RC. Are diagnostic criteria for general
joint hypermobility and benign joint hypermobility syndrome based
on reproducible and valid tests? A review of the literature. J Rheuma-
tol. 2007;34:798-803.
38. Rice SG, American Academy of Pediatrics Council on Sports Medi-
cine and Fitness. Medical conditions affecting sports participation.
Pediatrics. 2008;121:841-848.
39. Sackett DL, Straus SE, Richardson WS, Rosenberg W, Haynes RB.
Harm. In: Parkinson M, ed. Evidence-Based Medicine: How to Prac-
tice and Teach EBM. 2nd ed. Edinburgh: Harcourt; 2000:162-164.
1496 Pacey et al The American Journal of Sports Medicine
40. Simmonds JV, Keer RJ. Hypermobility and the hypermobility syn-
drome. Man Ther. 2007;12:298-309.
41. Smith R, Damodaran AK, Swaminathan S, Campbell R, Barnsley L.
Hypermobility and sports injuries in junior netball players. Br J Sports
Med. 2005;39:628-631.
42. Soderman K, Alfredson H, Pietila T, Werner S. Risk factors for leg injuries
in female soccer players: a prospective investigation during one out-
door season. Knee Surg Sports Traumatol Arthrosc. 2001;9:313-321.
43. Stewart DR, Burden SB. Does generalised ligamentous laxity
increase seasonal incidence of injuries in male first division club
rugby players? Br J Sports Med. 2004;38:457-460.
44. Terrin N, Schmid CH, Lau J. In an empirical evaluation of the funnel
plot, researchers could not visually identify publication bias. J Clin
Epidemiol. 2005;58:894-901.
45. Tofts LJ, Elliott EJ, Munns C, Pacey V, Sillence DO. The differential
diagnosis of children with joint hypermobility: a review of the litera-
ture. Pediatr Rheumatol Online J. 2009;7:1.
46. Uhorchak JM, Scoville CR, Williams GN, Arciero RA, St. Pierre P,
Taylor DC. Risk factors associated with noncontact injury of the ante-
rior cruciate ligament: a prospective four-year evaluation of 859 West
Point cadets. Am J Sports Med. 2003;31:831-842.
For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav
Vol. 38, No. 7, 2010 Hypermobility and Joint Injuries 1497