British Journal of Rheumatology 1994;33:56-59

HIGH PREVALENCE OF JOINT LAXITY IN WEST AFRICANS

F. N. BIRRELL,* A. O. ADEBAJO,* B. L. HAZLEMAN* and A. J. SILMANt
*Rheumatotogy Research Unit, Addenbrooke's Hospital, Cambridge CB2 2QQ;f ARC Epidemiology
Research Unit, University of Manchester, Stop ford Building, Oxford Road, Manchester Ml 3 9PT

SUMMARY
Previous surveys have suggested marked ethnic and geographical variation in the occurrence of joint hypermobility. We
investigated the prevalence of joint hypermobility and the influences of age, sex, body mass and occupation in a rural
Yoruba population in Nigeria.
The study sample consisted of 204 individuals aged 6-66 yr from the townships of Igbo-ora and Eruwa in south western
Nigeria. Sixty-eight had reported joint pain as part of a population survey of arthritic disorders and each was age and sex
matched with one household and one neighbour control. Joint hypermobility was assessed, at four peripheral sites
bilaterally and forward flexion of the trunk, by a single observer using the Beighton score. Each subject had weight and
height recorded, answered a brief questionnaire about occupation and joint symptoms and was examined for peripheral

Downloaded from http://rheumatology.oxfordjournals.org/ at York University Libraries on October 23, 2014

joint disease.
Only 11 (5%) of the subjects were negative at all five sites whereas 111 (54%) were hypermobile at three or more sites
including 23 (11 %) positive at all five. Using a score of 4/9 or greater as a cutoff, 88 (43 %) were positive, including 35 % of
males and 57% of females. There was a linear decline with age in females but a more rapid decline only to age 35 yr in
males. There was no relation to body mass or occupation.
We conclude that joint hypermobility amongst this population is substantially greater than that recorded for other
groups but is not associated with joint pain.
KEY WORDS: Hypermobility, Beighton score, Joint mobility, Joint pain, Body mass index, Occupation.

SINCE the syndrome of generalized joint hypermobility
was first described [1] and simple clinical criteria
derived for its measurement [2, 3], there have been
considerable efforts to understand the importance of
this condition. This has largely centred around quanti-
fying population frequency [3-9] and exploring the
possible associations with both articular and extra-
articular disease [10-12].
This study sets out to address these questions with
respect to a West African population. We have assessed
the frequency of hypermobility in this population and
its relationship to the factors of age, sex and BMI. In
addition, we have investigated possible associations
between hypermobility and either symptoms or signs
of musculoskeletal disorders. Beighton's original study
of hypermobility was carried out on an African popula-
tion, but there have been no attempts to compare the
South African Tswanas studied with another distinct
African population. We sought to test the impression
of some Nigerian clinicians that there is a high pre-
valence of joint laxity in Yoruba people.
SUBJECTS AND METHODS
The subjects were 204 individuals (116 women; 88
men) from the townships of Igbo-ora and Eruwa in the
rural area covered by the Ibarapa Community Health
Programme in Oyo state, south-western Nigeria. The
area has a population of approximately 150 000, con-
sisting almost entirely of the Yoruba ethnic group.
Indeed, all individuals included for study were Yomba.
The climate of Ibarapa is that of the rain forest belt,
which it borders, though the terrain is largely savannah
Submitted 6 October 1992; revised version accepted 22
February 1993.
with residual patches of forest near to water courses.
The local Yoruba people are mostly peasant farmers
and traders.
The study population was based on the results of a
screening survey for joint disease of some 2000 resi-
dents [13]. From that survey 68 individuals who
answered positive at interview to the current presence
of peripheral joint pain were recruited for the current
investigation. Each was matched with two comparison
individuals: one member of the extended family in the
same household, 'related control' and one unrelated
individual in the same residential compound, 'neigh-
bour control'.
Data was collected on occupation and history of joint
pain. Height and weight were measured, to calculate
BMI. Clinical assessment was made of joint mobility
and the presence of signs of joint disease (swelling, ten-
derness, deformity) was noted. Joint mobility was
scored according to Beighton's modification of the
Carter-Wilkinson scoring system [3].
All data were collected by one trained observer,
assisted by an interpreter fluent in both English and the
Yoruba language.
Analysis
Joint mobility was analysed separately in the three
groups, the joint mobility data were then pooled, giving
a total population sample of 204.
The associations between both joint mobility and
joint pain with possible predictor variables were
assessed and adjusted for possible confounders using
unconditional logistic regression analysis. The results
are expressed as odds ratios with 95% CI.

0263-7103/94/010056 + 04 $08.00/0 © 1994 British Society for Rheumatology

56
 NAVASA ETAL.: JOINT LAXITY IN WEST AFRICANS
RESULTS
Relationship of joint mobility to pain
The first question addressed was whether the pres-
ence of joint pain was associated with a high Beighton
score. The unadjusted odds ratio suggested the oppo-
site—that joint mobility was negatively associated with
joint pain, with each increment in Beighton score asso-
ciated with a 10% reduction in joint pain. However this
was almost completely explained by age = Beighton
score decreased and pain increased with age. Thus,
after age adjustment, there was no influence of
Beighton score on the presence of pain. In view of this
finding, there is no reason to suppose that the articular
mobility of the group of joint pain affected individuals
differs from those of similar age and sex in the popula-
tion studied.
Prevalence of joint mobility
The overall estimated prevalence of hypermobility,
by the number of the five Beighton criteria sites
affected is presented in Table I. This approach to the
analysis of the Beighton score values emphasizes multi-
site laxity and does not distinguish between unilateral
and bilateral involvement at one of the four bilateral
sites, e.g. both knees hypermobile equals only one site
affected. Hypermobility at one or more joint sites was
remarkably common = only 11 (5%) of the subjects
were negative at all five sites, while 37 (18%) were lax
at one, 45 (22%) lax at two, 55 (27%) lax at three and 33
(16%) were lax at four of the five criterion sites (four
bilateral).
Overall 111 (54%) were hypermobile at three or
more sites including 23 (11%) positive at allfive.Using
the conventional score of 4/9 as a cutoff for hyper-
mobility, 88 (43%) were positive, including 35% of
males and 57% of females.
Forward flexion of the trunk was the most frequent
site for laxity, followed by the 5th MCPJ. The fre-
quencies for elbow and knee hypermobility were
similar whereas the 'wrist' was the least likely site to
meet the criterion for laxity (Fig. 1).
Other factors associated with joint mobility
The influences of age and sex on joint mobility are
shown in Fig. 2. The median score was higher in
females than males at all ages, except for the youngest
age group (6-15 yr) where very similar laxity was
observed. The score also declined with age in both
sexes, although the pattern of decline was not con-
TABLEI
Prevalence of hypermobility by number of affected sites
Percentage
No. of (number) of
hypermobile sites individuals affected
5(11)
18 (37)
22 (45)
27 (55)
16(33)
11(23)
57
sistent between the sexes with the suggestion that the
effect of age was greater in females. Small numbers in
each of the individual age groups (see Fig. 2) precludes
more definitive conclusions. The independent effects
of age and sex together with those of body mass and
occupation in statistically predicting a Beighton score
of 4 or greater were examined by logistic regression.
Hypermobility is, of course, a graded trait, but this cri-
terion score of 4 or greater was chosen for ease of com-
parison with existing literature. The results are shown
in Table II. This shows a threefold excess risk in
females and a 4% decline in risk for each year of age.
There was no effect of either body mass or occupation.
Further investigation showed that a proportion of
Downloaded from http://rheumatology.oxfordjournals.org/ at York University Libraries on October 23, 2014
individuals in the comparison groups were also
affected by joint pain. This was equal between the two
control groups, with half of each group (related, 49%;
neighbour, 53%) reporting any joint pain. This effec-
tive inclusion of a proportion of joint pain individuals
reduced the number of pain free controls, but analysis
comparing those individuals free from joint pain
against the pain group confirmed no excess of hyper-
mobility associated with joint pains.
DISCUSSION
Joint mobility
The prevalence of hypermobility in our study popu-
lation was higher than that found by any previous
investigators. This includes Tswana South Africans [3],
Iraqi students [6], Omanis [9], Indians [7] and Cau-
casians [7, 8,12], when matched for age and sex.
Indeed, if expressed in terms of 'pauci-articular
hypermobility' = a term Larsson [8] suggested to
embrace all with one or more lax joints—a remarkable
97% of females and 91% of our male sample show at
least one lax joint, casting doubt on the usefulness of
this assignation as distinct from normality in this
population.
100
80
S 60
I
£ 40
o
09
t 20
0
5thMCPS Writs Elbow Knee Spine
Site of hypermobility
FIG. 1.—Prevalence of hypermobility by site.
58 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 33 NO. 1

school [8], when no comment was made on the magni-

tude of the difference. This difference may be due to
either an extrinsic effect, such as may be caused by
training, or to an intrinsic effect, such as may be hor-
monally induced. There is some evidence for the
former effect, but only on forward flexion, from work
on ballet dancers [14]. This cannot account for our
observed difference = which is largely due to differ-
I ences at the elbow and knee. Moreover, in this study,
no particular occupation, or type of occupation
0)
m (manual/non-manual), was associated with an excess of
a
a 16 hypermobile individuals, which is consistent with no
•3 significant training effect accruing from daily tasks.
Without more detailed information on lifestyle it is dif-
ficult to attribute any of the excess hypermobility found

Downloaded from http://rheumatology.oxfordjournals.org/ at York University Libraries on October 23, 2014

6-15 16-25 26-35 36-45 46-55
Age (yr)
FIG. 2.—Median joint mobility score by age and sex.
Having found such a high prevalence of hypermobil-
ity in this West African population, two questions,
particularly, must be addressed (i) methodological = is
this representative? (ii) interpretation = is it impor-
tant, as regards population morbidity?
Representativeness
The sample was derived in part from positive respon-
ders to a joint pain survey, selected by stratified
random sample [13]. The sample is thus not random. In
interpreting the results the question should be
addressed as to whether the prevalence of joint laxity,
or the estimates of its association with possible pre-
dictors are biased because of population selection. This
seems unlikely for three reasons (i) the prevalence of
hypermobility was similar in all groups and, as the data
showed, uninfluenced by joint pain, (ii) The subjects
being examined were unaware of the hypothesis being
tested and were recruited independently of any prior
expectations to their joint mobility, (iii) The within-
group association between joint mobility and the other
variables were examined after adjustment for other
potential confounders.
As expected, joint mobility declined with age, but
the possible difference in age trends between the sexes
was of interest. This trend of later female decline was
also apparent in Larsson's study at a New York music
TABLE II
Factors associated with high joint mobility
in this Yoruba population compared to Beighton's
Tswana group [3] to training.
The evidence in favour of the latter hypothesis, of
intrinsic influence, is supported by a study examining
56+ hypermobility in pregnancy [15]. It was concluded that
there was increased joint laxity during pregnancy, as
measured by the hyperextensometer. Moreover,
hyperextension was more marked in multigravida than
primigravida, suggesting a cumulative effect. Given the
high prevalence of multiparous women in rural
Nigeria, our hypothesis is that factors related to preg-
nancy are responsible for the observed slower decline
in female joint laxity. A foremost candidate among
these factors must be the relaxins, which have much
indirect evidence linking them with public symphyseal
laxity [e.g. 16]. More recent biochemical evidence on
the effects of relaxin on collagen and procollagenase
activity, in vitro [17] suggests that its effect may be
ubiquitous on collagen throughout the body, poten-
tially affecting the ligamentous/capsular component to
joint stability.
Thus, a high joint laxity for age would be expected in
a population which has been affected by relaxins (par-
ous women), against controls (men and nulliparous
women). One could test whether more pregnancies are
associated with a sustained excess of hypermobility and
also establish if factors apart from pregnancy are
important in maintaining joint laxity, by a retrospective
study. A longitudinal study of joint mobility, using a
hyperextensometer and Beighton's criteria throughout
pregnancy could provide better evidence both of the
time scale of hypermobility changes and correlation to
relaxin levels and other potentially associated factors.
TABLE III
Influence of Beighton score on the presence of joint pain
Adjusted for Odds ratio 95% CI

Predictor Odds ratio 95% CI Crude 0.89 0.81-0.99

Age 0.99 0.89-1.11
Age 0.96 0.93-0.98 Sex 0.87 0.78-0.97
Sex 2.94 1.66-5.23 Occupation (manual/non-manual) 0.94 0.84-1.18
Occupation (manual/non-manual) 1.00 0.51-1.97 BMI 0.90 0.81-1.00
BMI 0.99 0.91-1.06 All 0.97 0.86-1.10

•Odds ratios represent the increased risk of joint pain for each *Odds ratios represent the increased risk of joint pain for each
Beighton score increment (0-9). Beighton score increment (0-9).
 NAVASA ETAL: JOINT LAXITY IN WEST AFRICANS
Joint pain
Levels of reported joint pain were equivalent to
those found in other rural populations [9,18].
In contrast to other studies, however, which have
demonstrated a positive association of hypermobility
and joint pain [1,3,6,12], no significant association was
found in this population, despite the inclusion of a
group of pain positive individuals. Analysis of sub-
groups confirmed no excess of joint pain in those with
generalized joint laxity. This is in contrast to the excess
morbidity in the 16-24-yr-old group for those with high
joint mobility scores (7-9) in Pountain's larger (920
adults) Omani study [9]. Nor were we able to demon-
strate the association between obesity and joint pain
found by Pountain.
Thus, although we cannot exclude a small role for
joint pain in a young 'healthy' population, this is rather
insignificant in comparison with other causes of mus-
culoskeletal symptoms and signs in the population as a
whole. In Nigeria, these include septic arthritis, para-
sitic infections, OA and, less commonly, inflammatory
joint disease [13,19,20]. Although X-ray investigation
would have been a more sensitive indicator of OA,
clinical examination for swelling, tenderness, defor-
mity) did not reveal an excess of OA in those satisfying
hypermobility syndrome criteria. However, follow-up
of the same individuals would be interesting, to docu-
ment both changes in joint laxity and possible joint
pathology.
In conclusion, from the sample of 204 studied we
have observed that joint hypermobility in this popula-
tion is substantially greater than that recorded for
other population groups. Although a larger sample size
would be preferable, it appears that some degree of lax-
ity is the norm in this population, making it a good area
to investigate the significance of joint hypermobility
further. Whether further work should rely on the
Beighton criteria is open to debate; Bird [21] has pro-
posed a set of criteria for joint hypermobility, which
includes Beighton's criteria. As the most widely used
and, hence, comparable measure of generalized hyper-
mobility the Beighton criteria will continue to be the
standard for the foreseeable future, although in certain
circumstances the sensitivity of the hyperextensometer
is preferable.
Excessive joint mobility does not cause significant
morbidity in this population as a whole, either in terms
of joint symptoms or signs, nor was it related to occupa-
tion or BMI. Further work is clearly necessary to
demonstrate precisely the role of hypermobility in
disease causation, preferably using case control and
longitudinal studies.
Future studies might also delineate the factors we
have suggested as responsible for the excess of female
hypermobility observed and its slower decline with age
than in the male.
ACKNOWLEDGEMENTS
This work was supported by grants from The Common-
wealth Foundation and Emmanuel College, Cambridge.
Thanks also go to Aziz, the able interpreter; Bill
Brieger for help compiling questionnaires, Sara Holligan,
59
for advice on field work and Dr H. Bird for advice on the
study design.
REFERENCES
1. Kirk JH, Ansell BM, Bywaters EGL. The hypermobil-
ity syndrome—musculoskeletal complaints associated
with generalised joint laxity. Ann Rheum Dis 1967,26:
419-25.
2. Carter CO, Wilkinson J. Persistent joint laxity and
congenital dislocation of the hip. J Bone Joint Surg
[Br] 1964;46:40-5.
3. Beighton P, Solomon L, Soskolne CL. Articular
hypermobility in an African population. Ann Rheum
Dis 1973;32:413-48.
4. Silman AJ, Day SJ, Haskard DO. Factors associated
with joint mobility in an adolescent population. Ann
Downloaded from http://rheumatology.oxfordjournals.org/ at York University Libraries on October 23, 2014
Rheum Dis 1987;46:209-12.
5. Silverman S, Constantin L, Harvey W, Grahame R.
Survey of joint mobility and in-vivo skin elasticity in
London school children. Ann Rheum Dis 1975;34:
177-80.
6. Al-Rawi ZS, Al-Aszawi AJ, Al-Chalabi T. Joint mobil-
ity among university students in Iraq. Br J Rheumatol
1985^4:326-31.
7. Wordsworth P, Ogilvie D, Smith R, Sykes B. Joint
mobility with particular reference to racial variation
and inherited connective tissue disease. Br J Rheu-
matol 1987;26:9-12.
8. Larsson L-G, Baum J, Mudholkar GS. Hypermobility:
features and differential incidence between the sexes.
Arthritis Rheum 1987;30:1426-30.
9. Pountain G. Musculoskeletal pain in Omanis, and the
relationship to joint mobility and body mass index. Br
J Rheumatol 1991;3(h81-5.
10. Beighton PB, Grahame R, Bird HA. Hypermobility of
joints, 2nd edn. Berlin: Springer-Verlag; 1989.
11. Gedalia A, Person GA, Brewer EJ, Giannini EH. Juv-
enile episodic arthralgia and hypermobility. J Pediatr
1985;107:873-6.
12. Grahame R. The hypermobility syndrome. Ann
Rheum Dis 1990,49:190-200.
13. Holligan S, Birrell FN, Adebajo AO et al. Low preva-
lence of rheumatoid arthritis in a rural Nigerian popu-
lation. BrJ Rheumatol 1991,30:1.
14. Klemp P, Stevens JE, Isaacs S. A hypermobility study
in ballet dancers. J Rheumatol 1984;ll:692-6.
15. Bird HA, Calguneri M, Wright V. Changes in joint lax-
ity occurring during pregnancy. Ann Rheum Dis 40:
209.
16. Zarrow M, Holstrom EG, Salhanick HA. The concen-
tration of relaxin in the blood serum and other tissues
of women during pregnancy. J Clin Endocrinol 1955;
15:22-7.
17. Unemori EN. Relaxin modulates synthesis and secre-
tion of procollagenase and collagen by human dermal
fibroblasts. J Biochem Physiol 1990,265:10681-5.
18. Lawrence JS. The influence of climate on rheumatic
complaints. In: Rheumatism in populations. London:
Heinemann Medical; 1977:505-17.
19. Adebajo AO, Birrell F, Hazleman BL. The pattern of
rheumatic disorders seen amongst patients attending
urban and rural clinics in West Africa. Clin Rheum
1992;11:512-15.
20. Adebajo AO. Rheumatology in the Third World. Ann
Rheum Dis 1990,49:813-6.
21. Bird HA. Joint Hypermobility. Br J Rheumatol 1992;
31:205-6.