Full text of DOI:10.1359/jbmr.2002.17.7.1273

JOURNAL OF BONE AND MINERAL RESEARCH
Volume 17, Number 7, 2002
© 2002 American Society for Bone and Mineral Research
Genetic and Environmental Determinants of Peak Bone
Mass in Young Men and Women
FIONA E.A. McGUIGAN,1 LIAM MURRAY,2 ALISON GALLAGHER,3 GEORGE DAVEY-SMITH,4
CHARLOTTE E. NEVILLE,3 ROB VANЈT HOF,1 COLIN BOREHAM,3 and STUART H. RALSTON1
ABSTRACT
Peak bone mass is an important risk factor for the development of osteoporosis in later life. Previous work has
suggested that genetic, intrauterine, and environmental factors all contribute to the regulation of bone mass, but the
ways in which they interact with each other to do so remain poorly understood. In this study, we investigated the
relationship between peak bone mass and polymorphisms of the vitamin D receptor (VDR), estrogen receptor (ER)
␣, and collagen type I␣1 (COLIA1) genes in relation to other factors such as birth weight, lifestyle diet, and exercise
in a population-based cohort of 216 women and 244 men in their early 20s. Stepwise multiple regression analysis
showed that body weight was the strongest predictor of bone mineral density (BMD) in women, accounting for
16.4% of the variance in spine BMD and 8.4% of the variance in femoral neck BMD. Other significant predictors
were VDR genotype (3.8%) and carbohydrate intake (1.6%) at the spine and vitamin D intake (3.4%) and ER
genotype (3.4%) at the femoral neck. Physical activity was the strongest predictor of BMD in men, accounting for
6.7% of the variance at the spine and 5.1% at the hip. Other significant predictors were body weight (5%) and ER
PvuII genotype (2.8%) at the spine and weight (3.4%) and alcohol intake (2%) at the femoral neck. Birth weight
was not a significant predictor of BMD at either site but COLIA1 genotype significantly predicted birth weight in
women, accounting for 4.3% of the variance. We conclude that peak bone mass is regulated by an overlapping but
distinct set of environmental and genetic influences that differ in men and women. However, much of the variance
in BMD was unexplained by the variables studied here, which suggests that either most of the genes that regulate
BMD remain to be discovered or major environmental influences on BMD exist that have not yet been identified.
(J Bone Miner Res 2002;17:1273–1279)
Key words: polymorphism, COLIA1, vitamin D receptor, estrogen receptor, bone mineral density, birth
weight
contribute significantly to the regulation of peak bone mass
in healthy individuals. The importance of genetic factors is
INTRODUCTION
EAK BONE mass is an important determinant of risk for illustrated by the results of twin and family studies, which
P
the development of osteoporosis later in life. Current
evidence suggests that genetic and environmental factors
have indicated that heredity accounts for between 50% and
85% of the variance in bone mineral density (BMD), de-
pending on the site examined.(1–4) However, environmental
factors such as calcium intake and physical activity also
have been shown to influence peak bone mass by modulat-
ing bone gain during childhood and adolescence.(5–7) Inter-
Dr. Stuart Ralston holds a patent on the use of COLIA1 geno-
typing as a genetic marker for osteoporosis. All other authors have
no conflict of interest.
1Aberdeen University, Aberdeen, Scotland.
2Queens University Belfast, Belfast, Northern Ireland.
3University of Ulster, Ulster, Northern Ireland.
4Department of Social Medicine, University of Bristol, Bristol, United Kingdom.
1273

1274
MCGUIGAN ET AL.
est has focused also on the possibility that environmental venous blood samples using standard techniques. Genotyp-
influences during intrauterine life may influence peak bone ing for the BsmI polymorphism of the vitamin D receptor
mass in the light of recent studies, which documented an (VDR) and PvuII and XbaI polymorphisms of the ER was
association between body weight during infancy and body carried out using polymerase chain reaction—restriction
size and bone mineral content (BMC) in young adults.(8,9) In fragment length polymorphism (PCR-RFLP)–based meth-
this study, we investigated the relative contribution of ge- ods as previously described.(17,18) Genotyping for the Sp1
netic and environmental variables to the regulation of peak binding site polymorphism of the COLIA1 gene was deter-
bone mass in a population-based cohort of young healthy mined by allelic discrimination using the Taqman system
men and women, focusing on the BsmI polymorphism of the (Applied Biosystems, Foster City, CA, USA) as follows. A
vitamin D receptor gene,(10) the PvuII and XbaI polymor- 176-bp product was amplified using the primers (forward
phisms of the estrogen receptor (ER) ␣ gene,(11) and the Sp1 5Ј-GATGTCTAGGTGCTGGAGGTTAGGG-3Ј and re-
binding site polymorphism of the collagen type I␣1 (CO- verse 5Ј-CCTTCTCTCCTCCCTCGCC-3Ј) and the fluoro-
LIA1) gene.(12)
genic probes (5Ј TET-CCCGCCCACATTCCCTGG-3Ј and
5Ј-6-carboxy-fluorescein (FAM)-CCCGCCCCCATTCCC-
TGG-3Ј). The reaction mixture of 23 ␮l/well contained 12.5
␮l of Mastermix (Applied Biosystems), 9.25 ␮l of water, 50
nM of each primer, 50 nM of the FAM-labeled probe, 100
nM of the 6-carboxy-4,7,2Ј,7Ј-tetrachloro-fluoroscein (TET)-
PATIENTS AND METHODS
Study population
The study was conducted on participants of the Young labeled probe and 0.2 ␮g of genomic DNA. The cycling
Heart’s Project, a longitudinal cohort study of risk factors parameters were 1 cycle of 95°C for 10 minutes followed by
for coronary heart disease in school children from Northern 40 cycles of 95°C for 15 s and 60°C for 1 minute. Geno-
Ireland. The original 1015 subjects, aged 12–15 years of types were detected on a PRISM 7200 Sequence Detector
age, were enrolled by random sampling of schools from (Applied Biosystems). All samples were genotyped blind
throughout Northern Ireland in 1987 as described else- and 20% of all the samples were repeated.
where.(13)
All participants gave informed consent to being included
At the first study visit, each subject completed in the study, which was approved by the Research Ethics
interviewer-administered dietary and activity questionnaires Committee of the Queens University of Belfast, Northern
and underwent a medical examination and fitness testing. Ireland.
Information relating to parental height and weight, length of
gestation, and the number of children in the family was sion 10 (SPSS, Inc., Chicago, IL, USA). The ␹ test was
Statistical analysis was performed using SPSS, Inc. ver-
2
obtained also. Birth weight was obtained from computer- used to test for the Hardy-Weinberg equilibrium and to look
ized child health system records.
for differences between genotype distributions in men and
The participants were invited by letter to attend another women. The relationship between genotype and continuous
assessment in 1997 when the mean age of the cohort was 22 variables was assessed by one-way ANOVA. The general
years. Four hundred and eighty-nine (48.2%) subjects at- linear model (GLM)-ANOVA procedure was used to study
tended this additional assessment, but BMD scans and/or the association between genotype and BMD by entering
blood samples for DNA analysis were unavailable for 29 individual genotypes into the model as factors along with
subjects, resulting in a final study population of 460 (45.3% relevant covariates such as age, weight, height, nutrient
of the original cohort). Previously, we reported(14,15) that intake, smoking, alcohol, and exercise history. GLM-
socioeconomic position was the main determinant of non- ANOVA was similarly used to probe for gene-gene inter-
attendance at the 1997 visit, although other factors such as actions by including genotypes from more than one poly-
decreased height, increased skinfold thickness, and in- morphism in the analysis and building an interaction term
creased systolic blood pressure also were significantly as- into the model. Stepwise regression analysis was used to
sociated with failure to reattend.
identify predictors of BMD and birth weight. For this anal-
All subjects underwent a physical examination and com- ysis, we created dummy variables for individual genotypes,
pleted a lifestyle questionnaire, including details of diet, such that there were two levels for each polymorphism
smoking history, alcohol intake, exercise history, and fam- (coded 0 or 1) comprising homozygotes for one allele and
ily history of osteoporosis. Nutrient intakes were calculated the combined group of heterozygotes and homozygotes for
from the dietary histories using a computerized food anal- the other allele. For VDR, the levels were bb (0) versus Bb
ysis program (IUNA WISP; Tinuviel Software, Warrington, and BB (1); for ER PvuII, the levels were pp (0) versus Pp
UK) and physical activity was assessed using a modification and PP (1); for ER XbaI, the levels were xx (0) versus Xx
of the Baecke questionnaire,(16) which records work-related and XX (1); and for COLIA1, the levels were SS (0) versus
activity, sports-related activity, and nonsports leisure activ- Ss and ss (1).
ity. A total activity score was obtained from the sum of
scores in these three domains to give a total score ranging
RESULTS
from 3 (low physical activity) to 15 (high physical activity).
BMD was measured at the femoral neck and lumbar spine
using a GE Lunar Expert-XL dual-energy X-ray absorpti-
All individuals were in good health at the time of study
ometer (GE Lunar, Madison, WI, USA). Genotyping was and virtually all of the women (97.8%) were menstruating
performed on leukocyte DNA extracted from peripheral normally. None of the individuals were taking medications

GENETIC AND ENVIRONMENTAL ON BMD
1275
TABLE 1. CLINICAL CHARACTERISTICS OF THE POPULATIONS
Variable
Clinical
Women (n ϭ 216)
Men (n ϭ 244)
p Value
Current smokers
18.6%
34%
0.016
Age (years)
Birth weight (g)
Height (m)
22.6 Ϯ 1.6
22.3 Ϯ 1.6
3459 Ϯ 587.8
1.77 Ϯ 6.74
75.4 Ϯ 11.55
1.25 Ϯ 0.14
1.14 Ϯ 0.17
0.025
0.015
0.0001
0.0001
0.0001
0.0001
3335 Ϯ 494
1.64 Ϯ 6.26
64.5 Ϯ 11.5
1.18 Ϯ 0.117
1.05 Ϯ 0.147
Weight (Kg)
Spine BMD (g/cm2)
Hip BMD (g/cm2)
Nutritional
Alcohol (U/week)
Energy (KC)
Calcium (mg)
Vitamin D (␮g)
Vitamin C (␮g)
Protein (g)
Fat (g)
Carbohydrate (g)
16.7 Ϯ 19.1
2035 Ϯ 604
750 Ϯ 287.7
1.85 Ϯ 2.1
94.9 Ϯ 69.2
70.3 Ϯ 22.2
74.3 Ϯ 25.1
259.2 Ϯ 92.6
7.4 Ϯ 1.2
46.7 Ϯ 41.4
3131 Ϯ 815
1166 Ϯ 470.4
2.32 Ϯ 1.66
96.7 Ϯ 62.3
101.2 Ϯ 27.6
114.1 Ϯ 35.9
367.6 Ϯ 106.6
7.9 Ϯ 1.4
0.001
0.001
0.001
0.001
0.79
0.001
0.001
0.001
0.001
Exercise score
Genetic
COLIA1
SS
133 (61.6%)
69 (31.9%)
13 (6%)
146 (61.9%)
79 (33.5%)
11 (4.7%)
Ss
ss
␹2 ϭ 4.71,
p ϭ 0.79
VDR
BB
Bb
45 (20.8%)
105 (48.6%)
54 (25%)
41 (17.4%)
112 (47.6%)
82 (34.8%)
␹2 ϭ 4.0,
p ϭ 0.14
bb
PvuII
PP
Pp
40 (18.5%)
95 (44%)
56 (25.9%)
49 (22.4%)
109 (49.8%)
61 (27.8%)
␹2 ϭ 0.17,
p ϭ 0.92
pp
XbaI
XX
Xx
31 (14.4%)
87 (40.3%)
73 (33.8%)
30 (13.7%)
106 (48.4%)
83 (37.9%)
␹2 ϭ 0.62,
p ϭ 0.73
xx
known to affect bone metabolism and only 8 subjects re- associations were observed between ER PvuII genotype and
ported having a positive family history of osteoporosis. hip BMD in women (Fig. 1). We also looked for evidence
Other relevant characteristics of the study population are of gene-gene interactions in regulating BMD by entering
summarized in Table 1. As expected, men were taller, two or more genotypes into the model simultaneously, but
heavier, and had a higher birth weight and higher BMD no significant interactions were observed (data not shown).
values than women. Dietary intake of all nutrients was
Stepwise linear regression analysis was performed to
significantly greater in men as was exercise score. Alco- define the relative contribution of genes and environmental
hol intake was also higher and more men were current factors to the regulation of BMD. All genotype information
smokers.
was entered into the model, along with dietary, exercise, and
One-way ANOVA analysis showed no significant differ- lifestyle variables; anthropometric variables; and birth
ences between the various polymorphisms studied and an- weight. The results of this analysis are shown in Table 2.
thropometric characteristics such as birth weight, BMD Weight was the most important predictor of spine BMD in
values, dietary nutrient intakes, alcohol intake, and age once women, although other significant predictors included car-
the number of comparisons had been taken into account bohydrate intake and VDR BsmI genotype. Weight also was
(data not shown). We similarly found no significant asso- the most important predictor of hip BMD in women but
ciation between genotype and categorical variables such as other significant predictors included vitamin D intake and
social class or smoking history (data not shown). On GLM- ER PvuII alleles. Exercise history was the most important
ANOVA analysis, we found significant associations be- predictor of spine BMD in men, along with ER PvuII
tween VDR BsmI genotype and spine BMD in women and alleles, weight, and energy intake. Physical activity also was
ER PvuII genotype and spine BMD in men. Significant the strongest predictor of hip BMD in men and other sig-

1276
MCGUIGAN ET AL.
FIG. 1. Genotype associations with BMD. BMD measurements were adjusted for age, birth weight, height, weight, smoking, exercise score,
vitamin C, alcohol, energy, calcium, protein, total fat, and carbohydrate intakes. Significant differences between genotypes as assessed by
GLM-ANOVA are indicated.
nificant predictors included weight and alcohol intake. The
results were almost identical when we repeated the regres-
sion analysis using dietary, exercise, and lifestyle data col-
lected at the age of 12–15 years (data not shown).
Overall, the regression model explained only 21.8% of
the total variance in spine BMD and 15.2% of the variance
in femoral neck BMD in women and 17.5% and 10.6%,
respectively, of the variance in BMD in men. These findings
are in close concordance with the observations made by
Rubin et al.(22) in a young female Canadian cohort and
suggests that most predictors of peak bone mass remain to
be defined.
In agreement with previous work, we found that body
weight was one of the most important determinants of BMD
at both the spine and the hip.(7,22) In women, weight ac-
counted for 60.2% of the explained variance in BMD at the
hip and 75% at the spine.
Although weight also was a significant predictor of BMD
at both sites in men, physical activity was by far the most
important determinant of BMD in men, accounting for
38.2% of the explained variance at the spine and 48.1% at
the hip. The positive association between physical activity
and peak BMD in men is consistent with previous data(7)
whereas lack of effect in women differs from that reported
previously.(22,23) The reason for this discrepancy is unclear
but may have been caused by the fact that few women in
this cohort actively participated in sports that involved high
peak strain (C. E. Neville, M. McGuinness, and C.A.G.
Boreham, unpublished data, 2000). In contrast, almost all
(91.2%) of the young women studied by Rubin et al.(22)
exercised for at least 30 minutes a week.
We also investigated predictors of birth weight, entering
COLIA1, VDR, and ER genotypes into the model, along
with maternal and paternal age, height, and weight; length
of gestation; parity; and multiple births. The results of this
analysis are summarized in Table 3. Significant predictors
of birth weight in women were length of gestation (8.6%),
mothers’ weight (4.4%), multiplicity (3.7%), COLIA1 ge-
notype (4.3%), and maternal age (2.6%). Significant predic-
tors in men were length of gestation (32%), fathers weight
(5.3%), multiplicity (3.9%), and mothers’ age (3.1%) and
height (2.2%). We also divided subjects into tertiles to look
for an interaction between birth weight and genotype in
predicting adult BMD as reported by Dennison et al.,(9) but
no significant interactions were observed (data not shown).
DISCUSSION
The risk of developing symptomatic osteoporosis later in
life is influenced to a large extent by the levels of peak bone
mass attained in early adulthood. Peak bone mass is thought
to be determined by an interaction between lifestyle and
genetic factors,(19–21) but the relative contribution that these
factors make in regulating BMD is incompletely under-
stood.(22) In this study we have assessed the contribution of
anthropometric, lifestyle, and genetic factors to the regula-
tion of peak bone mass in a population-based cohort study
of young white men and women.
Of the other environmental variables assessed, vitamin
D intake had a positive effect on hip BMD in women
(3.4% of total variance; 22.3% of explained variance). In
men, energy intake had a positive effect on spine BMD

GENETIC AND ENVIRONMENTAL ON BMD
1277
TABLE 2. PREDICTORS OF BMD IN MEN AND WOMEN
Standarized
coefficient
Model summary
R2 (adj)
SE
␤
t
p
Spine BMD
Women (constant)
Weight (kg)
14.79
5.70
2.80
2.01
8.152
2.81
3.74
2.59
2.48
Ͻ0.0001
Ͻ0.0001
0.005
0.046
Ͻ0.0001
0.006
Ͻ0.0001
0.011
0.014
0.164
0.202
0.218
0.104
0.01
0.10
0.408
0.205
0.143
VDR bb genotype
Carbohydrate intake
Men (constant)
Physical activity
Weight (Kg)
0.067
0.117
0.145
0.175
0.138
0.134
0.133
0.131
0.198
0.261
0.180
0.174
ER pp genotype
Energy intake
Hip BMD
Women (constant)
Weight (Kg)
11.34
4.13
3.16
2.67
6.98
3.11
2.64
2.27
Ͻ0.0001
Ͻ0.0001
0.002
0.008
Ͻ0.0001
0.002
0.084
0.118
0.152
0.1353
0.1327
0.1302
0.306
0.238
0.201
Vitamin D intake
ER pp genotype
Men (constant)
Physical activity
Weight
0.051
0.085
0.106
0.164
0.161
0.159
0.221
0.187
0.161
0.009
0.024
Alcohol intake
The regression model included age, birth weight, height, weight, smoking, vitamin C, vitamin D, alcohol, energy, carbohydrate, protein,
and exercise scores.
TABLE 3. PREDICTORS OF BIRTH WEIGHT IN WOMEN AND MEN
Standardized
Model summary
R2 (adj)
coefficient
SE
␤
t
p
Females (constant)
Length of gestation (weeks)
Weight Mother (kg)
*Multiple birth
COLIA1 SS genotype
Age mother (years)
Males (constant)
Length of gestation (weeks)
Weight father (kg)
Multiple birtha
Age mother (years)
Height mother (m)
Ϫ2.098
3.78
2.424
3.584
3.282
2.62
Ϫ5.508
9.76
3.472
Ϫ3.958
3.143
2.978
0.037
0.000
0.011
0.000
0.001
0.01
0.000
0.000
0.001
0.000
0.002
0.003
0.086
0.130
0.167
0.210
0.236
490.34
478.5
468.36
456.12
448.31
0.257
0.167
Ϫ0.246
0.221
0.182
0.325
0.378
0.417
0.448
0.470
480.54
461.48
446.82
434.8
0.524
0.191
Ϫ0.211
0.167
425.93
0.163
The regression model included the variables gestation, multiplicity, maternal age, height and weight, paternal age, height, and weight.
a Multiple birth was a dichotomous variable (yes or no).
(3% of total variance; 17.1% of explained variance) and The significant association between VDR BsmI and
alcohol intake had a positive effect on hip BMD (2.1% of BMD, which we observed here, is in broad agreement
total variance; 19.8% of explained variance). The lack of with the results of previous studies, which have indicated
a significant association between COLIA1 genotype and that allelic variation in VDR contributes to the regulation
BMD is consistent with previous studies, which have of peak bone mass.(22,28) In this study, the bb genotype
indicated that this polymorphism may be associated with was associated with high BMD, which is consistent with
osteoporotic fractures by affecting age-related bone loss that originally reported by Morrison(29) and other
and bone fragility, rather than peak bone mass.(19,24–27) groups(28,30) but differs from that reported in young Ca-

1278
MCGUIGAN ET AL.
nadian women(22) and older women from the Nether-
lands(31) and Scotland.(17) This supports the view that the
BsmI polymorphism in the 3Ј region of VDR probably is
acting as a marker for a functional polymorphism else-
where in VDR or in another gene nearby.
2. Slemenda CW, Christian JC, Williams CJ, Norton JA,
Johnston CC 1991 Genetic determinants of bone mass in adult
women: A re-evaluation of the twin model and the potential
importance of gene interaction on heritability estimates.
J Bone Miner Res 6:561–567.
3. Lutz J, Tesar R 1990 Mother-daughter pairs: Spinal and fem-
oral bone densities and dietary intakes. Am J Clin Nutr 52:
872–877.
The positive association that we observed between the ER
PvuII pp genotype and BMD also is consistent with the
findings reported by other workers who studied postmeno-
pausal women from Japan(11) and Scotland.(18) However, it
should be noted that several other investigators have failed
to detect an association between the PvuII and XbaI poly-
morphisms of the ER and BMD.(32–34) Currently, the mech-
anisms by which these intronic polymorphisms of the ER
gene contribute to regulation of BMD remain unclear.
There is some evidence to suggest that birth weight acts
as a predictor of bone mass in early(35) and later life.(36,37)
The availability of data on birth weight from individuals
who took part in this study allowed us to explore the
hypothesis that intrauterine environment influences suscep-
tibility to osteoporosis. In this study we found no significant
association between birth weight and BMD, although this is
in agreement with the study of Cooper and colleagues who
similarly failed to detect such an association in young
women from the United Kingdom.(8) Although Cooper re-
ported a positive association between birth weight and
BMC, we found no such association once parental height
and weight had been taken into account. Other workers have
reported an allelic association between VDR genotype and
birth weight. We did not detect an association between VDR
genotype and birth weight in this study, but we did detect an
association between COLIA1 genotype and birth weight.
In summary, our study has revealed that environmental
factors, together with polymorphisms in the VDR and ER
candidate genes explain ϳ18% of the variance in peak bone
mass in women and 14% in men. Although it remains
possible that environmental variables not assessed in this
study also may contribute to the variance in peak bone mass,
it seems much more likely that the bulk of the unexplained
variance is caused by allelic variation in candidate genes,
which have yet to be defined. Further studies also are
required to assess more fully the relationship between ge-
netic factors and the determinants of growth and body size.
If the pathogenesis of osteoporosis can be traced back to
childhood or earlier, then identification of the genes respon-
sible for these effects will have a major impact on the
diagnosis, prediction, and prognosis of the disease.
4. Gueguen R, Jouanny P, Guillemin F, Kuntz C, Pourel J, Siest
G 1995 Segregation analysis and variance components analy-
sis of bone mineral density in healthy families. J Bone Miner
Res 12:2017–2022.
5. Bonjour JP, Carrie AL, Ferrari S, Clavien H, Slosman D,
Theintz G, Rizzoli R 1997 Calcium-enriched foods and bone
mass growth in prepubertal girls: A randomized, double-blind,
placebo-controlled trial. J Clin Invest 99:1287–1294.
6. Johnston CC Jr, Miller JZ, Slemenda CW, Reister TK, Hui S,
Christian JC, Peacock M 1992 Calcium supplementation and
increases in bone mineral density in children. N Engl J Med
327:82–87.
7. Welten DC, Kemper HC, Post GB, van Mechelen W, Twisk J,
Lips P, Teule GJ 1994 Weight-bearing activity during youth is
a more important factor for peak bone mass than calcium
intake. J Bone Miner Res 9:1089–1096.
8. Cooper C, Cawley M, Bhalla A, Egger P, Ring F, Morton L,
Barker D 1995 Childhood growth, physical activity, and peak
bone mass in women. J Bone Miner Res 10:940–947.
9. Dennison EM, Arden NK, Keen RW, Syddall H, Day IN,
Spector TD, Cooper C 2001 Birthweight, vitamin D receptor
genotype and the programming of osteoporosis. Paediatr Peri-
nat Epidemiol 15:211–219.
10. Morrison NA, Yeoman R, Kelly PJ, Eisman JA 1992 Contri-
bution of trans-acting factor alleles to normal physiological
variability: Vitamin D receptor gene polymorphism and circu-
lating osteocalcin. Proc Natl Acad Sci USA 89:6665–6669.
11. Kobayashi S, Inoue S, Hosoi T, Ouchi Y, Shiraki M, Orimo H
1996 Association of bone mineral density with polymorphism
of the estrogen receptor gene. J Bone Miner Res 11:306–311.
12. Grant SF, Reid DM, Blake G, Herd R, Fogelman I, Ralston SH
1996 Reduced bone density and osteoporosis associated with a
polymorphic Sp1 binding site in the collagen type I alpha 1
gene. Nat Genet 14:203–205.
13. Boreham CA, Twisk J, Savage MJ, Cran GW, Strain JJ 1997
Physical activity, sports participation, and risk factors in ado-
lescents. Med Sci Sports Exerc 29:788–793.
14. van Lenthe FJ, Boreham CA, Twisk JW, Savage MJ, Murray
L, Smith GD 2001 What determines drop out in prospective
studies of coronary heart disease risk factors between youth
and young adulthood: The Young Hearts Study. J Epidemiol
Commun Health 55:681–682.
15. van Lenthe FJ, Boreham CA, Twisk JW, Strain JJ, Savage JM,
Smith GD 2001 Socio-economic position and coronary heart
disease risk factors in youth. Findings from the Young Hearts
Project in Northern Ireland. Eur J Public Health 11:43–50.
16. Baecke JA, Burema J, Frijters JE 1982 A short questionnaire
for the measurement of habitual physical activity in epidemi-
ological studies. Am J Clin Nutr 36:936–942.
ACKNOWLEDGMENT
17. Houston LA, Grant SFA, Reid DM, Ralston SH 1996 Vitamin
D receptor polymorphism, bone mineral density and osteopo-
rotic vertebral fracture: Studies in a UK population. Bone
18:249–252.
18. Albagha OM, McGuigan FE, Reid DM, Ralston SH 2001
Estrogen receptor alpha gene polymorphisms and bone min-
eral density: Haplotype analysis in women from the United
Kingdom. J Bone Miner Res 16:128–134.
19. Harris SS, Patel MS, Cole DE, Dawson-Hughes B 2000 As-
sociations of the collagen type Ialpha1 Sp1 polymorphism
with five-year rates of bone loss in older adults. Calcif Tissue
Int 66:268–271.
This study was supported by a grants from the Wellcome
Trust (to C.B., L.M., and S.H.R.), a grant from the British
Heart Foundation (to C.B. and L.H.), and a cooperative
group grant from the MRC (to S.H.R.).
REFERENCES
1. Pocock NA, Eisman JA, Hopper JL, Yeates MG, Sambrook
PN, Eberl S 1987 Genetic determinants of bone mass in adults.
A twin study. J Clin Invest 80:706–710.

GENETIC AND ENVIRONMENTAL ON BMD
1279
20. Kelly PJ, Eisman JA, Sambrook PN 1990 Interaction of ge- 31. Uitterlinden AG, Pols HAP, Burger H, Huang Q, van Daele
netic and environmental influences on peak bone density.
Osteoporos Int 1:56–60.
21. Krall EA, Dawson-Hughes B 1993 Heritable and life-style
determinants of bone mineral density. J Bone Miner Res
8:1–9.
22. Rubin LA, Hawker GA, Peltekova VD, Fielding LJ, Ridout R,
Cole DE 1999 Determinants of peak bone mass: Clinical and
genetic analyses in a young female Canadian cohort. J Bone
Miner Res 14:633–643.
23. Taaffe DR, Robinson TL, Snow CM, Marcus R 1997 High-
impact exercise promotes bone gain in well-trained female
athletes. J Bone Miner Res 12:255–260.
PLA, van Duijn CM, Hofman A, Birkenhager JC, van Leeu-
wen JPTM 1996 A large-scale population-based study of the
association of vitamin D receptor gene polymorphisms with
bone mineral density. J Bone Miner Res 11:1241–1248.
32. Bagger YZ, Jorgensen HL, Heegaard AM, Bayer L, Hansen L,
Hassager C 2000 No major effect of estrogen receptor gene
polymorphisms on bone mineral density or bone loss in post-
menopausal Danish women. Bone 26:111–116.
33. Gennari L, Becherini L, Masi L, Mansani R, Gonnelli S,
Cepollaro C, Martini S, Montagnani A, Lentini G, Becorpi
AM, Brandi ML 1998 Vitamin D and estrogen receptor allelic
variants in Italian postmenopausal women: Evidence of mul-
tiple gene contribution to bone mineral density. J Clin Endo-
crinol Metab 83:939–944.
24. Mann V, Hobson EE, Aspden RM, Grant SFA, Li B, Robins
SP, Ralston SH 2001 COL1A1 Sp1 binding site polymorphism
predisposes to osteoporotic fracture by affecting bone density 34. Han KO, Moon IG, Kang YS, Chung HY, Min HK, Han IK
and quality. J Clin Invest 107:899–907.
1997 Non association of estrogen receptor genotypes with
bone mineral density and estrogen responsiveness to hormone
replacement therapy in Korean postmenopausal women. J Clin
Endocrinol Metab 82:991–995.
25. MacDonald HM, McGuigan FA, New SA, Campbell MK,
Golden MH, Ralston SH, Reid DM 2001 COL1A1 Sp1 poly-
morphism predicts perimenopausal and early postmenopausal
spinal bone loss. J Bone Miner Res 16:1634–1641.
26. McGuigan FE, Armbrecht G, Smith R, Felsenberg D, Reid
DM, Ralston SH 2001 Prediction of osteoporotic fractures by
bone densitometry and COLIA1 genotyping: A prospective,
35. Fewtrell MS, Prentice A, Jones SC, Bishop NJ, Stirling D,
Buffenstein R, Lunt M, Cole TJ, Lucas A 1999 Bone miner-
alisation and turnover in pre-term infants at 8–12 years of age:
The effect of early diet. J Bone Miner Res 14:810–820.
population-based study in men and women. Osteoporos Int 36. Yarbrough DE, Barrett-Connor E, Morton DJ 2000 Birth
12:91–96.
weight as a predictor of adult bone mass in postmenopausal
women: The Rancho Bernardo Study. Osteoporos Int 11:626–
630.
27. Uitterlinden AG, Burger H, Huang Q, Yue F, McGuigan FE,
Grant SF, Hofman A, van Leeuwen JP, Pols HA, Ralston SH
1998 Relation of alleles of the collagen type Ialpha1 gene to 37. Gale CR, Martyn CN, Kellingray S, Eastell R, Cooper C 2001
bone density and the risk of osteoporotic fractures in post-
menopausal women. N Engl J Med 338:1016–1021.
Intrauterine programming of adult body composition. J Clin
Endocrinol Metab 86:267–272.
28. Viitanen A, Karkkainen M, Laitinen K, Lamberg-Allardt C,
Kainulainen K, Rasanen L, Viikari J, Valimaki MJ, Kontula K
1996 Common polymorphism of the vitamin D receptor gene
is associated with variation of peak bone mass in young Finns.
Calcif Tissue Int 59:231–234.
29. Morrison NA, Qi JC, Tokita A, Kelly PJ, Crofts L, Nguyen
TV, Sambrook PN, Eisman JA 1994 Prediction of bone den-
sity from vitamin D receptor alleles. Nature 367:284–287.
30. Cooper GS, Umbach DM 1996 Are vitamin D receptor poly-
Address reprint requests to:
Fiona McGuigan, Ph.D.
Department of Medicine and Therapeutics
Foresterhill, Aberdeen AB25 2ZD, UK
morphisms associated with bone mineral density? A meta- Received in original form April 27, 2001; in revised form February
analysis. J Bone Miner Res 11:1841–1849. 12, 2002; accepted March 7, 2002.