JOURNAL OF WOMENS HEALTH

Volume 15, Number 9, 2006

Mary Ann Liebert, Inc.

The Association between Body Mass Index and

Osteoporosis in Patients Referred for a
Bone Mineral Density Examination

KOFI ASOMANING, M.B.Ch.B., M.S.,1 ELIZABETH R. BERTONE-JOHNSON, Sc.D.,2

PHILIP C. NASCA, Ph.D.,2 FREDERICK HOOVEN, Ph.D.,3
and PENELOPE S. PEKOW, Ph.D.2

ABSTRACT

Purpose: Osteoporosis affects 46 million (13%18%) postmenopausal white women in the

United States. Most studies to date on risk factors for osteoporosis have considered body mass
index (BMI) only as a possible confounder. In this study, we assess the direct relationship
between BMI and osteoporosis.
Methods: We conducted a cross-sectional study among women aged 5084 years referred by
their physicians for a bone mineral density (BMD) examination at Baystate Medical Center
between October 1998 and September 2000. BMI was determined prior to the BMD exami-
nation in the clinic. Information on other risk factors was obtained through a mailed ques-
tionnaire. Ordinal logistic regression was used to model the association between BMI and os-
teoporosis, controlling for confounding factors.
Results: BMI was inversely associated with BMD status. After adjustment for age, prior
hormone replacement therapy (HRT) use, and other factors, odds ratios (OR) for low, high,
and obese compared with moderate BMI women were 1.8 (95% CI 1.2-2.7), 0.46 (95% CI 0.29-
0.71), and 0.22 (95% CI 0.14-0.36), respectively, with a significant linear trend (p  0.0001) across
BMI categories. Evaluating BMI as a continuous variable, the odds of bone loss decreased
12% for each unit increase in BMI (OR  0.88, 95% CI 0.85-0.91).
Conclusions: Women with low BMI are at increased risk of osteoporosis. The change in risk
associated with a 1 unit change in BMI (58 lb) is of greater magnitude than most other
modifiable risk factors. To help reduce the risk of osteoporosis, patients should be advised
to maintain a normal weight.

INTRODUCTION density at the hip,1 making it the most common

human bone disease.2 Osteoporosis is character-

I T IS ESTIMATED THAT between 13% and 18% (46

million) of postmenopausal white women in
the United States have osteoporosis, and an ad-
ditional 30%50% (1317 million) have low bone
ized by low bone density and microarchitectural
deterioration of bone tissue, leading to bone
fragility and an increased risk of fracture. Osteo-
porotic fractures are the most common and seri-

1Harvard School of Public Health, Boston, Massachusetts.

2Department of Public Health, University of Massachusetts, Amherst, Massachusetts.
3University of Massachusetts Medical School, Shrewsbury, Massachusetts.

1028
ASSOCIATION OF BMI AND OSTEOPOROSIS
ous consequence of osteoporosis and are also an
important cause of disability and mortality in el-
derly people.
Body mass index (BMI), a measure of body
composition, may be associated with risk of os-
teoporosis.3 Individuals with low BMI typically
have low stores of body fat and lower circulating
estrogen levels, which help prevent loss of bone
tissue.46 Because body fat also acts as a cushion,
low BMI may increase the risk of fracture as a re-
sult of a fall.7,8
Limited epidemiological research has been
done on the direct relationship between BMI and
osteoporosis. Most previous studies have evalu-
ated other risk factors, including age, weight,
race, maternal history of fractures, use of calcium
supplements, and menopausal history,913 and
developed indices based on age and weight (in
addition to other risk factors) to help identify wo-
men at risk. The easiest index to calculate is the
Osteoporosis Self-assessment Tool (OST), which
is based on age and weight.14 Dargent-Molina et
al.15 showed that low weight is predictive of low
bone mineral density (BMD) (defined as a T-score
3.5 in that study). Compared with women
with weight 66 kg, women with weight 5966
kg had a 3-fold increase in the risk of low BMD
(crude OR  2.87, 95% CI 1.73-4.74). Other stud-
ies14,1620 have shown results of similar magni-
tude and direction for weight as a risk factor for
osteoporosis. BMI is likely to be more predictive
of osteoporosis than weight alone because it ad-
justs for differences in height and is more reflec-
tive of body composition. To examine the associ-
ation between BMI and osteoporosis, we
conducted a cross-sectional study among patients
who had been referred by a physician for a BMD
examination at Baystate Hospital, a large tertiary
care hospital in western Massachusetts.
MATERIALS AND METHODS
In order to obtain a population at high risk for
osteoporosis and who were capable of accurately
recalling relevant information, while also mini-
mizing the number of eligible women who might
be lost to follow-up because they had been insti-
tutionalized or died between the BMD examina-
tion and questionnaire mailing, subjects were
limited to women aged 5084. Study participants
were selected from among the 1769 patients aged
5084 who had a BMD examination between Oc-
1029
tober 1998 and September 2000. We first excluded
those with a missing address or invalid address
or for whom only single BMD scores were
recorded. We then selected 100% of the remain-
ing osteoporotic women and selected every other
patient within the normal and osteopenic cate-
gories, after sorting women alphabetically to en-
sure an adequate number of osteoporotic women
in our population.
Questionnaires were mailed to a total of 958 se-
lected patients. Respondents were divided into
BMD categories based on the T-score from their
BMD examination, using National Osteoporosis
Foundation and World Health Organization
(WHO) definitions.1,21,22 Two BMD measure-
ments, one at the hip (proximal femur) and one
at the lumbar spine, were performed at a single
visit. Osteoporosis was defined as one bone mass
measurement falling below 2.5 standard devia-
tions (SD) from the mean for adult women of the
same race. Osteopenia was defined as bone mass
between 1 and 2.5 SD below the same mean. Nor-
mal bone density was defined as bone mass mea-
surement greater than 1 SD below the mean.1,21,22
If a subject was osteoporotic at either the hip or
spine, we classified them as osteoporotic. If they
were not osteoporotic at either site but were os-
teopenic at least at one site, they were classified
as osteopenic. If they were normal at both sites,
they were classified as normal.
BMD was measured by dual x-ray absorp-
tiometry (DXA), a standard bone densitometry
study. Measurement of BMD using the DXA ma-
chine is the most commonly used and accepted
standard for measuring bone mass.20
BMI was assessed by measuring the weight
and the height of the patients during their first
referral visit. This was done by a single nurse be-
fore the BMD examination. Weight was measured
with a standardized weight column scale. Height
was measured with the patient standing and us-
ing a stadiometer. The exact measurement of
height in the elderly person is very difficult to de-
termine because of frailty and spinal deformity,
but Kirk and Hawke23 have demonstrated that
using the stadiometer has good agreement with
the more accurate assessments using knee height
or demispan (sternal notch to finger roots with
arms outstretched laterally). We then calculated
BMI as weight in kilograms divided by height in
meters squared.
Information on medication use and other risk
factors for osteoporosis at the time of the BMD
1030
examination was collected through question-
naires mailed to study subjects in December 2002.
We asked participants to report information on
their age, race, education, menopausal status, cig-
arette smoking, use of calcium and vitamin D
supplements, family history of osteoporosis,
medication use, and history of oophrectomy or
hysterectomy. The survey did not include any
questions on the secondary risk factors for osteo-
porosis, such as such as prior treatment with
steroids, chronic kidney disease, and parathyroid
and thyroid disease. The survey was designed
specifically for this study and included a number
of questions derived in whole or in part from the
Behavioral Risk Factor Surveillance System
(BRFSS), National Health and Nutritional Exam-
ination Survey (NHANES), and National Health
Interview Survey.
The Statistical Analysis System (SAS Institute,
Cary, NC), version 8.2, was used to complete all
data analyses. We divided women into four cat-
egories of BMI: low (19 kg/m2), moderate
(1925), high (2530), and obese (30). Wo-
men were also divided into categories based on
levels of other factors potentially associated with
both BMI and BMD, including age, current health
status, and prior use of hormone replacement
therapy (HRT) (Table 1).
We evaluated the relationship between risk fac-
tors and BMD categories using the Pearsons chi-
square test. Ordinal logistic regression analysis
was used to model the relationship between BMI
and three categories of BMD. The proportional
odds logistic model with maximum likelihood es-
timate was fit to the ordinal response of BMD cat-
egory (normal, osteopenia, and osteoporosis),
and 95% confidence intervals (CI) for the odd ra-
tios (OR) were calculated. Two cumulative logits
were obtained for the three response levels of
normal, osteopenia, and osteoporosis. Hence, the
ORs presented are those comparing osteoporosis
vs. osteopenia/normal and osteoporosis/osteo-
penia vs. normal status of bone density. The score
test was used to evaluate the assumption of pro-
portional odds, and the models were evaluated
for goodness of fit. We calculated age-adjusted
ORs using BMI as a categorical variable and also
as a continuous variable.
We then evaluated the relationship between
BMI and BMD categories, adjusting for the effects
of other factors, using backwards stepwise elim-
ination. We used a cutoff for the Wald statistic of
0.15 to retain a variable in our multivariable
ASOMANING ET AL.
model. Factors included in the final model were
age, use of calcium and vitamin D supplements,
use of HRT and bone active medications (e.g.,
alendronate, risedronate, calcitonin, and ralox-
ifene), and current health status. Other factors
evaluated but not included in the final model
because they were not associated with either
BMI or BMD included race, marital status,
smoking, and maternal history of osteoporosis
or fracture.
A test for trend over increasing BMI categories
was performed where each category median was
modeled as a continuous variable in the regres-
sion. We also assessed whether the relationship
between BMI and BMD differed in subgroups of
participants by stratifying by age group (60 vs.
60 years old), use of calcium/vitamin D (never
vs. current or past users), use of HRT (never vs.
current or past users), use of bone active drugs
(never vs. current or past users), and current
health status (excellent/very good/good vs.
fair/poor/very poor). Interactions between vari-
ables and BMI were considered statistically sig-
nificant if the Wald 2-sided p value for the inter-
action term in the multivariable model was 0.05.
RESULTS
Completed questionnaire responses were re-
ceived from 155 women classified as having os-
teoporosis (60% of eligible participants), 182
(58%) with osteopenia, and 169 (56%) women
with normal BMD. Responders and nonrespon-
ders did not differ by age group, osteoporosis sta-
tus, or BMI category. Age, calcium/vitamin D,
HRT, and health status categories appeared sig-
nificantly different and related to BMD status
(Table 1). For women in the oldest age category
(7584), a lower proportion had normal (7%) or
osteopenic (16%) rather than osteoporotic (30%)
bone density, whereas a reverse trend of more
normal (64%) and fewer osteopenic (33%) and os-
teoporotic (36%) was observed for women in the
youngest (5064) category (Table 1). A similar
trend across BMD categories was observed for
calcium/vitamin D when nonusers were com-
pared with users. Among users of HRT, 57% of
women had normal BMD status, 48% were os-
teopenic, and 31% were osteoporotic. For non-
HRT users, the percentages were 43%, 52%, and
68% for the same BMD categories.
Age-adjusted and multivariable ORs showing
ASSOCIATION OF BMI AND OSTEOPOROSIS 1031

TABLE 1. DISTRIBUTION OF RISK FACTORS FOR OSTEOPOROSIS BY BMD CATEGORY:

THE BAYSTATE OSTEOPOROSIS STUDY, 19982000

Normal Osteopenic Osteoporotic

(n  155)b (n  181)b (n  168)b
BMD category n (%) n (%) n (%) p valuea

Age (years) 0.0001

5064 93 (64) 91 (33) 55 (36)
6574 42 (29) 52 (31) 52 (34)
7584 11 (7) 27 (16) 44 (30)
Race 0.11
White/non-Hispanic 136 (91) 154 (88) 135 (83)
Other 13 (9) 22 (12) 27 (17)
Education 0.03
 High school graduate 18 (12) 29 (17) 35 (22)
High school graduate 71 (48) 90 (51) 87 (54)
College graduate 60 (40) 56 (32) 39 (24)
Marital status 0.0001
Married 89 (60) 106 (61) 70 (43)
Divorced, separated, single 39 (26) 36 (20) 34 (21)
Widowed 21 (14) 33 (19) 58 (36)
Maternal history of osteoporosis 0.15
No 61 (39) 68 (46) 53 (43)
Yes 17 (11) 34 (23) 25 (20)
Dont know 77 (40) 79 (31) 90 (37)
Maternal history of fracture 0.02
No 83 (54) 112 (62) 87 (52)
Yes 10 (6) 23 (13) 18 (11)
Dont know 62 (40) 46 (25) 63 (37)
Smoking 0.41
Never 74 (51) 74 (43) 83 (52)
Former 60 (41) 82 (48) 61 (38)
Current 11 (8) 16 (9) 16 (10)
Calcium/vitamin D 0.0001
No 136 (88) 141 (78) 109 (65)
Yes 19 (12) 40 (22) 59 (35)
Bone active drugs 0.01
No 65 (42) 77 (43) 95 (57)
Yes 90 (58) 104 (57) 73 (43)
HRT 0.0001
No 66 (43) 94 (52) 115 (68)
Yes 89 (57) 87 (48) 53 (31)
Lives alone 0.001
No 109 (72) 129 (72) 90 (54)
Yes 41 (28) 52 (28) 76 (46)
Current health status 0.003
Excellent/very good/good 119 (82) 133 (78) 104 (65)
Fair/poor/very poor 27 (18) 38 (22) 55 (35)
ap value derived from chi-square statistics.
bNumbers may not sum to 155 normal, 181 osteopenic, and 168 osteoporotic women because of missing data.

the relationship between BMI and BMD cate-
gories are presented in Table 2. ORs were deter-
mined for BMI in categories and as a continuous
measure. After adjustment for age, use of cal-
cium/vitamin D, HRT, bone medications, and
other factors, BMI was inversely associated with
BMD categories, with evidence of a significant
linear trend (p  0.0001). The ORs presented com-
pare osteoporosis vs. osteopenia/normal and os-
teoporosis/osteopenia vs. normal. Adjusted ORs
for low, high, and obese BMI compared with
women of moderate BMI were 1.76 (95% CI 1.2-
2.7), 0.46 (95% CI 0.29-0.71), and 0.22 (95% CI 0.14-
0.36), respectively. The odds of osteopenia/os-
teoporosis decreased 12% for each unit increase
in BMI (OR  0.88, 95% CI 0.85-0.91).
1032 ASOMANING ET AL.

TABLE 2. AGE-ADJUSTED AND MULTIVARIATE ODDS RATIO

Age-adjusted Multivariate
BMI category n ORa ORb 95% CI

BMI
Low (19 kg/m2) 22 3.06 1.76 1.16, 2.68
Moderate (1925) 192 1.00 1.00
High (2530) 166 0.54 0.46 0.29, 0.71
Obese (30) 123 0.28 0.22 0.14, 0.36
0.0001*
BMI (continuous, kg/m2) 505 0.90 0.88 0.85, 0.91
aSingle odds ratios estimate for two cumulative logits of osteoporosis vs. osteopenia/

normal and osteoporosis/osteopenia vs. normal for BMI.

bAdjusted for age (5064, 6574, 7584), prior use of calcium/vitamin D (yes, no), prior

use of HRT (yes, no), prior use of bone active drug (yes, no), and current health status
(excellent/very good/good, fair/poor/very poor).
*p value for trend using category medians as a continuous variable in the regression
model.

Analyses stratified by age group, current
health status, and prior use of calcium/vitamin
D, HRT, or bone active drugs did not suggest that
the BMI-BMD association varied over levels of
these factors, and no interaction term was signif-
icant (p  0.05) when added to the final model.
DISCUSSION
In this cross-sectional study of women referred
for a BMD examination, we observed that women
with a lower BMI have a significantly higher risk
of osteoporosis compared with normal weight
women. This association persisted after control-
ling for age, current health status, and prior use
of calcium/vitamin D, HRT, and bone active
drugs. BMI appears to have the strongest associ-
ation with BMD in this population.
Two mechanisms have been proposed to ex-
plain how body mass may affect osteoporosis and
risk of fracture. It has been suggested that body fat
provides an indirect protection from bone loss by
providing a source and depot for the peripheral
conversion of androstenedione to the metaboli-
cally active estrogen, estrone.5,6,24 When this es-
trogen depot is deficient, there is an increase in the
rate of bone turnover, which results in a faster rates
of bone loss.58 The second potential mechanism
suggests that in early adulthood, heavy individu-
als tend to reach a higher peak BMD than thinner
individuals and exert a greater load on their
weight-bearing joints, leading to the development
of higher BMD.24 Such heavy individuals are less
likely to be osteoporotic in their old age.
In our study, each one unit increase in BMI was
associated with a significant 12% decrease in risk.
This is consistent with results obtained by Nitzan-
Kaluski et al.,25 who reported a significant 8% de-
crease with each one unit increase in BMI. Simi-
larly, Munasinghe et al.26 observed that women
with osteoporosis had lower BMI compared with
patients with normal BMD (23.7 vs. 28.5 kg/m2,
p  0.001).
As with other studies,3,10,13,14,1628 results from
our analysis suggest that the proportion of wo-
men with bone loss increases with age, with a 4-
fold higher risk of bone loss in women age 75
compared with those 5064 years (data not
shown). In elderly women, this is explained by a
gradual loss of BMD, often referred to as age-re-
lated or senile bone loss, which starts in midlife29
and continues over time. Bone loss occurs more
rapidly in women in the early years after meno-
pause. Women who have been postmenopausal
for more than 5 years lose BMD at a rate of 1%2%
per year. The relative amount lost by women dur-
ing their 70s is greater than during their 60s, al-
though the rate of loss decreases.30 Maintaining
a healthy body mass and not becoming under-
weight may help reduce the effects of aging on
BMD loss.
BMI is likely to be a better predictor of osteo-
porosis than weight alone, as it more accurately
measures body composition. In our study, the ini-
tial medical record review did not provide the in-
dividual heights and weights (only the calculated
BMI), and we are unable to compare results for
BMI and weight in our population. We used a sin-
gle measurement of BMI assessed at the same
ASSOCIATION OF BMI AND OSTEOPOROSIS
time as BMD to classify subjects in terms of body
composition. A single measurement may not be
representative of long-term body composition,
does not allow us to evaluate the effects of BMI
in early adulthood or weight loss or gain over
time, and may contribute to misclassification.
Any such misclassification is likely to be nondif-
ferential, and the effect of such a bias will be to
underestimate the true relationship between BMI
and bone density. We were unable to assess the
reason women were referred for a BMD exami-
nation. Referred populations may include a large
proportion of patients with previously recog-
nized risk factors for osteoporosis or other dis-
eases associated with osteoporosis. They could
have been referred for screening for low bone
density prior to treatment or for follow-up of
treatment. This could affect the impact and in-
terpretation of use of therapies, such as cal-
cium/vitamin D, HRT, or bone active medica-
tions, on bone density status. Information on
other exposures, such as maternal history of frac-
ture or osteoporosis, smoking status, and use of
bone active drugs, was collected by self-report
without corroboration from medical records or
other sources. Response to questions that re-
quired remote recall may have been inaccurate
because of the 24 year gap between the BMD ex-
amination and our study, but this is not likely to
differentially affect participants according to
BMD status and would also tend to underesti-
mate the association.
In addition, because we assessed BMI and
BMD at the same time, we are unable to assess
the temporal relationship of these two factors. It
is possible, although unlikely, that osteoporosis
preceded the development of a low BMI. Osteo-
porosis occurs in elderly people, who are most
likely to have attained a relatively stable BMI be-
fore the onset of osteoporosis.14,6 The physio-
logical mechanism also suggests that body mass
has an effect on osteoporosis, as opposed to os-
teoporosis affecting body mass. Caution is still
needed, however, making causal interpretations
of our results.
Although this study evaluated the association
between BMI and BMD at a population level, if
a causal effect is assumed at the individual level,
our results suggest that a 1 unit increase in BMI
is associated with a 12% reduction in risk of lower
bone density.
Results of our study suggest that healthcare
providers should consider increased screening for
1033
osteoporosis in postmenopausal women who are
underweight. Interventions are also suggested that
will enable these patients to build and maintain
bone density. Additional advice on changes to
make to reduce the risk of fracture include exer-
cise to improve balance and lower body strength
and the removal of home hazards to make the liv-
ing area safer. Although obesity or excessive
weight gain should not be recommended as a strat-
egy to maintain bone density, the strong relation
of low bone density with low BMI underlies the
importance of elderly patients not becoming un-
derweight as far as bone density is concerned.
ACKNOWLEDGMENTS
We are indebted to Stephen Gehlbach, M.D.,
M.P.H., for his support of this research. We
would also like to acknowledge the assistance of
the Department of Healthcare Quality, Baystate
Health Systems, in completing this study.
REFERENCES
1. Dawson-Hughes B, Gold DT, Rodbard HW, et al. The
physicians guide to prevention and treatment of os-
teoporosis. National Osteoporosis Foundation, 2000,
Washington, DC.
2. Siris ES, Miller PD, Barrett-Connor E, et al. Identifi-
cation and fracture outcomes of undiagnosed low
bone mineral density in postmenopausal women
Results from the National Osteoporosis Risk Assess-
ment. JAMA 2001;286:2815.
3. Ensrud KE, Lipschutz RC, Cauley JA, et al. Body size
and hip fracture risk in older women: A prospective
study. Am J Med 1997;103:274.
4. Wooton R, Breyson E, Elsasser U. Risk factors for frac-
tured neck of the femur in the elderly. Age Aging
1982;11:160.
5. Heaney RP. Bone mass and osteoporotic fractures.
Calcif Tissue Int 1990;47:63.
6. Kreiger N, Kelsey JL, Holford TR, OConnor T. An
epidemiologic study of hip fracture in post-meno-
pausal women. Am J Epidemiol 1982;116:141.
7. Nevitt MC, Cummings SR. Type of fall and risk of hip
and wrist fracturesThe Study of Osteoporotic Frac-
tures. J Am Geriatr Soc 1994;42:909.
8. Cummings SR, Nevitt MC. A hypothesisThe causes
of hip fractures. J Gerontol 1989;44:M107.
9. Bernstein J, Grisso JA, Kaplan FS. Body mass and frac-
ture riskA study of 330 patients. Clin Orthopaed
Rel Res 1999;364:227.
10. Hoover PA, Webber CE, Beaumont LF, Blake JM.
Postmenopausal bone mineral density: Relationship
1034
to calcium intake, calcium absorption, residual estro-
gen, body composition, and physical activity. Can J
Physiol Pharmacol 1996;74:911.
11. Ross PD. OsteoporosisFrequency, consequences,
and risk factors. Arch Intern Med 1996;156:1399.
12. Lotz JC, Hayes WC. The use of quantitative computed
tomography to estimate risk of fracture of the hip
from falls. J Bone Joint Surg 1990;72A:689.
13. Joakimsen RM, Fonnebo V, Magnus JH, Tollan A, So-
gaard AJ. The Tromso Study: Body height, body mass
index and fractures. Osteoporos Int 1998;8:436.
14. Koh LK, Ben Sedrine W, Torralba TP, et al. Osteo-
porosis Self-Assessment Tool for Asians (OSTA) re-
search group. A simple tool to identify Asian women
at increased risk of osteoporosis. Osteoporos Int
2001;12:699.
15. Dargent-Molina P, Poitiers F, Breart G, EPIDOS
Group. In elderly women, weight is the best predic-
tor of a very low bone mineral density: Evidence from
the EPIDOS study. Osteoporos Int 2000;11:881.
16. Margolis KL, Ensrud KE, Schreiner PJ, Tabor HK.
Body size and risk for clinical fractures in older wo-
men. Study of Osteoporotic Fractures Research
Group. Ann Intern Med 2000;133:123.
17. van der Voort DJ, Dinant GJ, Rinkens PE, van der
Voort-Duindam CJ,van Wersch JW, Geusens PP. Con-
struction of an algorithm for quick detection of pa-
tients with low bone mineral density and its applica-
bility in daily general practice. Clin Epidemiol
2000;53:1095.
18. Lydick E, Cook K, Turpin J, Melton M, Stine R, Byrnes
C. Development and validation of a simple question-
naire to facilitate identification of women likely to
have low bone density. Am J Managed Care 1998;4:37.
19. Bauer DC, Browner WS, Cauley JA, et al. Factors as-
sociated with appendicular bone mass in older wo-
men. The Study of Osteoporotic Fractures Research
Group. Ann Intern Med 1993;118:657.
20. Hickson M, Frost G. A comparison of three methods
for estimating height in the acutely ill elderly popu-
lation. J Hum Nutr Dietetics 2003;16:13.
21. Kanis JA, Gluer CC. An update on the diagnosis and
assessment of osteoporosis with densitometry. Os-
teoporos Int 2000;11:192.
ASOMANING ET AL.
22. World Health Organization. Assessment of fracture
risk and application to screening for postmenopausal
osteoporosis. WHO technical report series 840.
Geneva, Switzerland: WHO, 1994.
23. Kirk S, Hawke T. Are the measures used to calculate
BMI accurate and valid for the use of older people? J
Hum Nutr Dietetics 2003;16:366.
24. Felson D, Zhang Y, Hannan M, Anderson J. Effects of
weight and body mass index on bone mineral density
in men and women: The Framingham Study. J Bone
Miner Res 1993;8:567.
25. Nitzan-Kaluski D, Chinieh A, Ifrah A, Merom D,
Green M. Correlates of osteoporosis among Jewish
and Arab women aged 4574 in Israel: National Wo-
mens Health Interview Survey. J Gender-Specific
Med 2003;6:17.
26. Munasinghe R, Botea V, Edelson G. Association
among age, height, weight, and body mass index with
discordant regional bone mineral density. J Clin Den-
sitom 2002;5:369.
27. Nguyen T, Center J, Eisman J. Osteoporosis in elderly
men and women: Effects of dietary calcium, physical
activity, and body mass index. J Bone Miner Res
2000;15:322.
28. Burger H, de Laet C, van Daele P, et al. Risk factors
for increased bone loss in an elderly population: The
Rotterdam Study. Am J Epidemiol 1998;147:871.
29. NIH Consensus Development Panel on Osteoporosis
Prevention, Diagnosis, and Therapy. Osteoporosis pre-
vention, diagnosis, and therapy. JAMA 2001;285:785.
30. Hannan M, Felson D, Dawson-Hughes B, et al. Risk
factors for longitudinal bone loss in elderly men and
women: The Framingham Osteoporosis Study. J Bone
Miner Res 2000;15:710.
Address reprint requests to:
Kofi Asomaning, M.B.Ch.B., M.S.
Harvard School of Public Health
665 Huntington Avenue
Building 1, Room 1404
Boston, MA 02115
E-mail: kasomani@hsph.harvard.edu