Journal of the Royal Society of Medicine Volume 82 December 1989

Osteoporosis - an epidemiological perspective:

753

review

C Cooper MA MRCP The Rheumatology Unit, Bristol Royal Infirmary, Bristol BS2 8HW and the
MRC Environmental Epidemiology Unit, Southampton General Hospital
Keywords: epidemiology; osteoporosis; fracture; bone mass; risk factors

Osteoporosis, defined as a reduction in bone mass1,

constitutes a major public health problem through its
association with age-related fractures. The most
important sites at which these fractures occur are the
hip, vertebra and distal radius. However, other
fractures such as those of the proximal humerus and
pelvis, are increasingly being recognized as associated
with osteoporosis. This review outlines the descriptive
epidemiology of the three major age-related fractures,
the patterns of change in bone mass with age, the
relationship between bone mass and fracture risk, and
the individual risk factors for osteoporosis.
The epidemiology of age-related fractures
Hip fracture
Hip fracture is the most dramatic consequence of osteoporosis, both in terms of ill health and cost2. Among
Caucasian women, the risk of sustainin a hip fracture
before the age of 90 years is estimated as one in three3.
The hospital costs alone of the 40 000 such fractures
which occur each year in England and Wales are over
160 million, and a considerable financial burden is
added by the large number of individuals who become
functionally disabled following hip fracture2. The
incidence of hip fracture rises exponentially with age
(Figure 1). Above 50 years of age, there is a female
to male incidence ratio of approximately two to one.
Hip fracture incidence also varies markedly with
ethnic group. In Negroid populations, for example, the

South African Bantu4 and the North American black5,

hip fracture is infrequent and rates in men and
women are similar. Recent data from Hong Kong6
show incidence rates intermediate between those
found in Caucasians and in Negroes.
Hip fractures show marked seasonality in their incidence, tending to occur more frequently during winter
in temperate countries (Figure 2)7. The majority of
these fractures, however, occur following falls indoors.
Hypothermia could increase the likelihood ofthese falls
by impairing neuromuscular coordination. Vitamin
D deficiency during winter provides an alternative
explanation8.
It is now clear that the last three decades have witnessed an increase in age-specific hip fracture incidence,
above that expected merely from the increasing
number of elderly in the population9. Such increases
have been documented in Britain, Scandinavia and
the United States'0. The reason for this secular trend
in western countries is unknown. It may reflect a
current increase in one or more environmental risk
factors, for example physical inactivity. Alternatively,
it might result from a cohort effect: risk factors acting
at some point earlier this century, giving progressive
increases in rates in successive generations.

Paper read to
joint meeting of
Sections of
Epidemiology &
Community
Medicine and
Orthopaedics
7 February 1989

Vertebral fracture
Vertebral fractures occur as partial (or wedge)
deformities and complete (or crush) fractures". Few

WOMEN

MEN

300

300

---@-- Distal radial

250

250

200

200

150

150

100

100

Hip

U
0

;
a0
VI

c
C.

50

55

65

75

85

55

65

75

85

0141-0768/89/

Age (yers)

120753-05/$02.00/0
Age (years)

Figure 1. Incidence of hip and distal radial fractures in British population samples. Rates
and Miller et al.'8

are

derived from Boyce et al.9

1989
The Royal
Society of
Medicine

754

Journal of the Royal Society of Medicine Volume 82 December 1989

Bone mass: measurement

and changes with aging
_
_
40
Early infomation about bone mass was obtained by
in vitro analysis of skeletal material, expressing
the results as the ash weight to volume ratio2'23.
ao -Although confined to autopsy specimens, these studies
showed that skeletal weight decreases with age above
____________
40 years in both sexes. A number of non-invasive
8 s
have subsequently
meods
of bone These
masmvary
easment
__
of validity
in their degrees
been developed.
M J J A S O N 0 J FM A
N D J F_M A __of_T
im
M985
Mand repeatability. In the 1960s a number of morphohvff of adssion. .
metric techniques were devised to estimate bone mass
from conventional radiographs24. These methods were
applied most frequently to the metacarpal, the spine
Figure 2. Month of admission ofpatients with hip fractures
and the proximal femur. Although some of these
in an English health district (Southampton) showing ?nrked
.techniques show moderately good correlations with
seasonal variation
in vitro assessment of bone mass at the measured site
epidemiological studies have been carried out of (the ash weight to volume ratio), they are limited by
technical considerations such as variable beam
vertebral fractures. This stems in part from the
quality, film response, radiation scattering conditions
the
diffiand
fractures
onset
of
such
asymptomatic
and variation in soft tissue thickness around the
culties inherent in their radiographic definition. In
measured skeletal site25. These problems limit the
a survey of women in Michigan, 2.7% aged from
repeatability of such measurements and hate confined
55-59 years had sustained one or more vertebral
fractures12. The prevalence rose to 20% at ages their use to epidemiological studies in which large
70-74 years. A study"in Denmark13 found that 4.5%,_ groups of subjects are investigated.
The problems inherent in these radiographic
of 70-year-old women randomly selected from the
general population had crush fractures and an techniques have largely bee,n overcome in the last
additional 18% had wedge fractures. Estimates' decade with the advent of 'i number of valid and'
repeatable measurement methods. The most widely
of crush fracture prevalence in the UK derive
used ofthese is photon absorptiometryo. The princple
from female hospital outpatients in Leeds14 where
of this technique is that an isotope emitting a photon
prevalences rise from 2.5% at age 60 years to 7.5%
beam is made to perform a rectilinear scan across a
at age 80 years. Wedge fractures were considerably
skeletal site. The change in beam intensity over the
more frequent with a prevalence of arounqd 60% in
bone is proportional to the bone mineral content ofthe
women aged over 75 years.
site. If soft tissue thickness around the measurement
The annual incidence of vertebral fractures in the
site can be made effectively onstant, only one energy
general population has not been directly measured.
level is required (single photon absorptiometry)27. If
Estimates have generally been derived from changes
not, two energy levels are required (dual photon
in prevalence with age. A radiological survey in
Finland"1 derived an annual incidence of thoracic absorptiometry)2. Photon absorptiometric measurespine crush fractures in women over age 65 years of ments are most frequently made in the forearm,
1.8/1000/year. This figure is comparable with a North, lumbar spine and femoral neck. They have high.
degrees of validity and are sufficiently,repeatable to
American estimate of 2.3/1000/year in,women aged
follow changes in the bone mass of individuals over
45 years and over who had densitometrically normal
spinal radiographs16. Rates derived for Caucasian periods of three to five years. The repeatability of
absorptiometric measurements in the axial skeleton
women aged 50 years and over in Rochester, Minnesta
has recently been enhanced by the substitution ofthe
are considerably higher, at 15.4/1000/year16. The
radioisotope by an X-ray source (dual energy X-ray
difference may stem, however, from inclusion in this
last study of both wedged and ballooned vertebrae as
absorptiometry)P. Other non-invasive techniques, for
example, quantitative computed tomography and
fractures.
neutron activation analysis, also provide valuable
Most studies report a greater prevalence of vertebral
information on skeletal status but their use is
fractures in women than in men17, although the
currently confined to research centres.
female to male ratio has varied from two to seven.
The widespread use of non-invasive measurements
There is little information on ethnic variation
ofbone mass has shed light on the patterns of change
in vertebral fracture occurrence, although rates in
in bone mass with aging. The adult skeleton is
Negroes and in inhabitants of Jerusalem have been
composed of two types of bone: cortical, which
reported to be lower than those in the United
predominates in the shafts of long bones, and
States16.
trabecular, which is the major constituent,of thepxial
skeleton. The peak levels of cortical' and trabecular
Distal radial fracture
bone mass are thought to be attained by the third or
Distal radial fractures are the most common fractures
in Caucasian women below 75 years of age. In the
fourth decade' of life and bone loss commences soon
UK18, Sweden19 and the USA20 the incidence in after3W. A biphasic pattern of loss has been identified
for both cortical and trabecular bone: a protracted
women has been shown to increase linearly from age
slow phase, which occurs in both sexes, and a
40 to 70 years. Thereafter rates stabilize. In men,
transient accelerated phase, which occurs in women
incidence remains constant between age 20 and 80
after the menopause. During the slow phase, bone is
years. As with hip fractures, there is a winter peak
lost at a rate between 0.3% and 1% of peak bone mass
in incidence21. Distal radial fractures, however- are
annually. During the accelerated postmenopausal
most frequently associated with outdoor falls, often
period, 108l rates may rise to between 3% and 5% of
during periods of snow and ice.

Journal of the Royal Society of Medicine Volume 82 December 1989

peak mass per year. Rates of bone loss vary at

different skeletal sites, depending upon whether they
are axial or appendicular, and according to their
relative contents of cortical and trabecular bone.
Extrapolations of bone mass at any given site from
measurements made at a different site are therefore

uncertain31.
Ethnic variations in bone mass have been reported.
Negroes in the USA have higher peak bone mass than
Caucasians, whether measured by ashing studies23
or neutron activation analysis32. However, a South
African study33 using metacarpal morphometry failed
to find differences in cortical bone mass between these
two ethnic groups. Data from Japanese residents in
Hawaii34 have suggested lower forearm bone mineral
content than in US Caucasians of similar age, and
Polynesian women in New Zealand have been shown
to have greater appendicular bone mass than their
Pakeha counterparts35.
Bone mass and fracture risk
Bone mass is an important determinant of bone
strength. Positive correlations have been documented,
for both the vertebra"* and femoral neck37, between
the ash density of bone and its compressive strength.
However, variation in bone mass predicts only
40-50% of the variation in bone strength in such
studies, reflecting the importance of other aspects of
skeletal structure in determining its strength. These
qualitative abnormalities of bone which serve to
reduce its strength i'nclude: accumulated fatigue
damage, ineffective architectural arrangement of
bony elements, age-related&changes in the mechanical
properties of bone and osteoid accumulati6nw.
The relation, however, between bone lfiass and
fracture risk appears to differ for the three most
frequently encountered fracture sites. Reduced spinal
trabecular bone mass has consistently been found in
patients with vertebral fractures when compared with
non-fracture controls39.
The role of osteoporosis in the occurrence of hip
fracture is less marked. Stochastic models can wholly
explain the variation of hip fracture incidence with
age and sex on the basis of changes in bone mass
with aging4e. Yet, sx case-ntrol studies have showi
far greater overlap between the femoral neck bone
mass values of hip fracture cases and non-fracture
controls than was found for vertebral fracture41. As
most hip fractures follow a fall, these results have
fuelled the hypothesis that hip fracture.risk depends
more on propensity to trauma than on bone mass. A
recent study, however, in which allowance was made
for the risk of falling, clearly implicated osteoporosis
as a determinant of hip f
in the relatively young
elderly (50-74 years) but emphasized the importance
of other factors at older ages42.
The contribution of osteoporosis to the risk of distal
radial fractures is least well understood. Measurements of forearm bone mineral content in patients
with such fractures show a distribution which iS
similar to that of non-fracture controls"43 and it is
possible that osteoporosis has a relatively small part
to play in their pathogenesis as compared with factors
predisposing to trauma.
The usefulness of non-invasive measurement
methods of bone mass as screening tools to predict
future fracture risk is unknown. This question has
become all the more relevant with the realition that
postmenopausal hormone replacement therapy is able

Figure 3. Factors involed in the pathogenesis ofhip fractures

to retard bone loss in women, as well as reducing

fracture incidence". The balance, however, of risks
and benefits of hormone replacement therapy remain
unclear. Until further infortion becomes available
on the risk ofbreast cancer with such treatment, and
the effect of opposed therapy in preventing cardiovascular disease, it would seem prudent not to
recommend it for all menopausal women but to
confine it to a group at high risk f osteoporosis. Bone
mass measurements have been proposed as one
method of selecting such a high risk group45. Two
published analyses of this Sulect47 have questioned
the ability ofthese measurements to identify women
at increased risk of hip fractures (the major public
health problem), and concluded that the costs of
such bone mineral screening outweighs the potential
benefits. They must be placed, however, against
recent longitudinal data supporting a relationship
between bone mass and the risk of both spine and
non-spine fractures in Hawaiian Japanese women"8.
Although the numbers of fracures encountered in the
first six years of this study were necessarily small,
the study continues and points to the requirement for
a rigorous assessment ofthe costs and benefits of bone
mineral screening.

Individual risk factors for osteoporosis

The bone mass of any individual over the age of
around 50 years depends upon the peak bone mass
which they attained at maturity, and its subsequent
rate of loss. Relatively little is known of the factors
which influence peak bone mass. Twin studiesO9'0
have demonstrated that monozygotic twin pairs have
a greater concordance for bone density than dizygotic
twin pairs, suggesting agenetically determined limit
to the attainment of bone mass. Other studies have
pointed to the roles of hormonal, mechanical and
nutritional factors in modulating skeletal growth5l.
More is known of the risk factors for bone loss.
Foremost among these is a premature (before age 45
years) menopause through oophorectomy, which is,
asoiated with accelerated bone loss as well as an
increased risk of hip, vertebral and distal radial
fractures52. The importance of other reproductive
variables, such as delayed menarche, parity, lactation
and use of oral contraceptive preparations is les clear.
Low body weight is also an important determinant
of osteoporosis and is associated-nwith increased
fracture risku. There are two mechanisms through
which thin build might influence bone mass. Adipose
tissue in postmenopausal women is a major source of
circulating oestrogeu (peripherally produced from
androstenedione in the adipocyte). High body weight

755

756

Journal of the Royal Society of Medicine Volume 82 December 1989

might also increase skeletal loading and thereby

protect against bone loss.
Other lifestyle risk factors include cigarette smoking,
heavy alcohol consumption, physical inactivity, certain
diseases (thyrotoxicosis, hyperparathyroidism, rheumatoid arthritis, partial gastrectomy) and the use of
certain medications (especially corticosteroids)3.
Considerable controversy surrounds the relatioship
between a low dietary calcium intake and osteoporosis54. Although most epidemiological studies have
failed to detect an appreciable relationship between
customary calcium intake and rate of bone loss5,
metabolic balance studies have suggested a negative
effect of the menopause on calcium balance performance in women56 and trials of calcium supplementation in postmenopausal women have in some
cases demonstrated a reduction in rate of bone lossa.
Three recent studies have addressed the issue as to
whether reduced calcium intake is a risk factor for
hip fracture. Two of these57'5, one in California and
the other in Hong Kong, reported a statistically
significant increase in fracture risk associated with
declining intake even after adjustment for potential
confounding variables. The third5 from the UK
failed to document such an effect. It is noteworthy that
in the two positive studies, the calcium consumption
of the control groups was low and a threshold level
below which calcium nutrition becomes critical cannot
be excluded. However, even if future studies do
confirm a beneficial effect of calcium supplementation
in later life, the magnitude of that effect is likely to
be considerably smaller than that of hormone
replacement therapy.

Conclusion
Age-related fractures pose a major public health
problem. Their incidence varies with age, sex and
ethnic group. This variation is explained in part
by changes in bone mass with aging. The precise
interaction, however, between bone mass and
propensity to trauma in the causation of these
fractures remains unclear. A number of individual
risk factors for osteoporosis have been delineated.
Further areas for research include a more detailed
evaluation of methods to identify high risk individuals
before the occurrence of significant bone loss and
the applications of population-based preventive
strategies.
References
1 Consensus Conference: Osteoporosis. JAMA 1984;252:
799-802
2 Fractured neck of femur: prevention and management.
The Royal College of Physicians of London, 1989
3 Melton LJ, Riggs BL. Epidemiology of age-related
fractures. In: Avioli LV, ed. The osteoporotic syndrome.
New York: Grune and Stratton, 1983:45-72
4 Solomon L. Osteoporosis and frcure ofthe femoral neck
in the South African Bantu. J Bone Joint Surg 1969;

5OB:2-13
5 Bollett AJ, Engh G, Parson W. Epidbmiology of osteo-

porosis: sex and race incidence of hipi*factures. Arch

Intern Med 1965;116:191-4
6 Lau EMC. Personl communication
7 Bastow MD, Rawlings J, Allison SP. Undernutrition,
hypothermia and iinjury in elderly with fractured femur:
an injury response to altered metabolism? Lancet 1983;i:
143-6
8 Aaron JE, Gallagher JC, Anderson JC, et al. Frequency
of osteomalacia and osteoporosis in fractures of the
proximal femur. Lancet 1974;i:229-33

9 Boyce WJ, Vessey MP. Rising incidence of fracture of

the proximal femur. Lancet 1985;i:150-1
10 Melton LJ, O'Fallon WM, Riggs BL. Secular trends in
the incidence of hip fractures. Calcif Tissue Int
1987;41:57-64
11 Harma M, Heliovaava M, Aromaa A, Kneckt P. Thoracic
spine compression fractures in Finland. Clin Orthop

1986;205:188-94
12 Smith RW, Rizek J. Epidemiologic studies of osteoporosis
in women of Puerto Rico and Southeastern Michigan
with special reference to age, race, national origin and
other related or associated findings. Clin Orthop
1966;45:31-48
13 Jensen GF, Christiansen C, Boesen J, Hegedus V,
Transbol I. Epidemiology of postmenqpausal spinal Lnd
long bone fractures. Clin Orthop 1982;106:75-81
14 Nordin BEC, Peacock M, Aaron J, Crilly RG, Heyburn
PJ, Horsman A, Marshall D. Osteoporosis and osteomalacia. Clin Endocrinol Metab 1980;9:177-204
15 Iskrant AP, Smith RW. Osteoporosis in women aged 45
years and over related to subsequent fractures. Public
Health Rep 1969;84:33-8
16 Melton LJ. Epidemiology of fractures. In: Riggs BL,
Melton LJ, eds. Osteoporosis: etiology, diagnosis and
management. New York: Raven Press, 1988: 133-54
17 Pogrund H, Makin M, Robin G, Meuczel J, Steinberg R.
Osteoporosis in patients with fractured femoral neck in
Jerusalem. Clin Orthop 1977;124:165-72
18 Miller SWM, Evans JG. Fractures ofthe distal forearm
in Newcastle: an epidemiological study. Age Aging
1985;14:155-8
19 Alffram PA, Bauer GCH. Epidemiology of fractures of
the forearm. J Bone Joint Surg 1962;42A:105144
20 Owen RA, Melton LJ, Johnson KA, Ustrup DM, Rigga
BL. Incidence of Colles fracture in a North American
community. Am J Public Health 1982;72:605-7
21 Ralis ZA. Epidemic of fractures during periods of snow
and ice. Br Med J 1982;282:603-5
22 Trtter MI, Broman GE, Peterson RR. Densities of bones
of white and negro skeletons. J Bone Joint Surg
1960;42A:50-8
23 Arnold JS, Bartley MH, Tout SA, Jenkins DP. Skeletal
changes in aging and disease. Clin Orthop 1966;
49:17-38
24 Barnett E, Nordin BEC. The radiological diagnosis of
osteoporosis. Clin Radiol 1960;11:166-74
25 Murby M, Fogelman I. Bone mineral measurements in
clinical practice.- Br J Hosp Med 1987;37:453-8
26 Mazess RB. Estimation of bone and skeletal weight by
direct photon absorptiometry. Invest Radiol 1971;
6:52-60
27 Christiansen C, Rodbro P, Jensen H. Bone mineral
content in the forearm measured by photon absorptiometry. Scand J Clin Lab Invest 1975;35:323-30
28 Nielsen SP, Krolner B. Photonbeam absorptiometry. In:
Dixon A St J, et aL eds. Osteoporosis. A muidisciplinary
problem. London: Academic Press, 1983:105-8
29 Mazess R, Collick B, Trempe J, Barden H, Hanson J.
Performance evaluation of a dual energy X-ray bone
densitometer. Calcif Tissue Int 1989;44:228-32
30 Riggs BL, Melton LJ. Involutional osteoporosis. NEngl
J Med 1986;314:1676-86
31 Riggs BL, Wahner HW, Dunn WL, Mazess RB, Offord
K?, Melton UJ. Differential changes in bone mineral
density 'of the appendiculae and axial skeleton with
aging. J Clin Invest 1982;67:328-35
32 Cohn SH, Abosamis C, Yasamura , A^loia JF, Zanzi I,
Elli KJ. Comparative skeletal mass and radial bone
mieral content inl blac and white women. Metabolism

1977;26:171-8
33 Bone density in aging Cacsan and Africa popuiations.
Lancet 1979;ii:1326-9
34 Yano K, Wasnich RD, Vogel JM, Heilbrun LK. Bone
mineral measurements among middle-aged and elderly
Japanese residents in H-awaii. Am J Epidemiol 1984;
119:751-64

Journal of the Royal Society of Medicine Volume 82 December 1989

35 Reid IR, Mackie M, Ibbertson HK. Bone mineral content
in Polynesian and white New Zealand women. Br Med
J 1986;292:1547-8
36 Weaver JK, Chalmers J. Cancellous bone: its strength
and changes with aging and an evaluation of some
methods for measuring its mineral content. JBone Joint
Surg 1966;48A:289-308
37 Leichter I, Margulies JY, Weinreb A, Mizrahi J, Robin
GC, Conforty B, Makin M, Bloch B. The relationship
between bone density, mineral content and mechanical
strength in the femoral neck. Clin Orthop 1982;163:
272-81
38 Heaney RP. Qualitative factors in osteoporotic fracture:
the state ofthe question. In: Christiansen C, et al., eds.
Osteoporosis 1987, 1. Proceedings of the International
Symposium on Osteoporosi Denmark, 1987. Copenhagen,
Osteopress ApS, 1987:281-7
39 Cann CE, Genant HK, Kolb FO, Etinger B. Quantitative
computed tomography for prediction of vertebral fracture
risk. Bone 1985;6:1-7
40 Horsman A, Marshall DH, Peacock M. A stochastic
model of age-related bone loss and fractures. Clin Orthop
1985;195:207-15
41 Cummings SR. Are patients with hip fractures more
osteoporotic? A review of the evidence. Am J Med
1985;78:487-94
42 Cooper C, Barker DJP, Morris J, Briggs RSJ. Osteoporosis, falls and age in fracture of the proximal femur.
Br Med J 1987;295:13-15
43 Horowitz M, Wishart JM, Bochner M, Need AG,
Chatterton BE, Nordin BEC. Mineral density of bone
in premenopausal women with fractured wrists. Br Med
J 1988;297:1314-15
44 Hunt K, Vessey MP. Longterm effects of post
menopausal hormone therapy. Br JHosp Med 1987;38:
450-60
45 Ross PD, Wasnich RD, Vogel JM. Detection of prefracture spinal osteoporosis using bone mineral absorptiometry. J Bone Mineral Res 1988;3:1-11
46 Cummings SR, Black D. Should perimen6pausal women

47

48

49

50
51
52

53
54
55

56
57

58

be screened for osteoporosis? Ann Intern Med 1986;

104:817-23
Hall FM, Davis MA, Baran DT. Bone mineral screening
for osteoporosis. N Engl J Med 1987;316:212-14
Wasnich RD, Ross PD, Heilbrun LK, Vogel JM.
Prediction of postmenopausal fracture risk with use of
bone mineral measurements. Am J Obstet Gynecol
1985;153:745-51
Smith DM, Nance WE, Kang KW, Christiansen JC,
Johnston CC. Genetic factors in determining bone mass.
J Clin Invest 1973;52:2800-8
Dequeker J, Nijs J, Verstraeten A, Geusens P, Gevers
G. Genetic determinants ofbone mineral content at the
spine and radius: a twin study. Bone 1987;8:207-9
Kanders B, Dempster DW, Lindsay R. Interaction of
calcium nutrition and physical activity on bone mass
in young men. J Bone Mineral Res 1988;3:145-9
Cann CE, Genant HK, Ettinger B, Gordan GS. Spinal
mineral loss of oopherectomised women: determination
by quantitative computed tomography. JAMA 1980;
244:2056-9
Cooper C, Barker DJP, Wickham C. Physical activity,
muscle strength and calcium intake in fracture of the
proximal femur in Britain. Br Med J 1988;297:1443-6
Kanis JA, Passmore R. Calcium supplementation of the
diet - II. Br Med J 1989;298:205-8
Riggs BL, Wahner HW, Melton LJ, Richelson LS, Judd
HL, O'Fallon WM. Dietary calcium intake and rates of
bone loss in women. J Clin Invest 1987;80:979-82
Heaney RP, Recker MR, Saville PD. Menopausal changes
in calcium balance performance. J Lab Clin Med
1978;92:953-63
Holbrook TL, Barrett-Connor E, Wingard DL. Dietary
calcium and risk of hip fracture: 14-year prospective
population study. Lancet 1988;ii:1046-9
Lau E, Donnan S, Barker DJP, Cooper C. Physical
activity and calcium intake in fracture of the proximal
femur in Hong Kong. Br Med J 1988;297:1441-3

(Accepted 6 July 1989)

757