Osteoporos Int (2001) 12:588–594 ß 2001 International Osteoporosis Foundation and National Osteoporosis Foundation

Osteoporosis International

Original Article Weight and Body Mass Index at Menarche are Associated with Premenopausal Bone Mass M. Blum1, S. S. Harris1, A. Must2, S. M. Phillips2, W. M. Rand2 and B. Dawson-Hughes1 1 Calcium and Metabolic Bone Laboratory, Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University; and 2Department of Family Medicine and Community Health, Tufts University, Boston, Massachusetts, USA

Abstract. Adolescence is a critical time for skeletal growth and mineralization. Exposure to protective or detrimental factors during this period may influence peak bone mass attainment and subsequent development of osteoporosis. In order to evaluate the association of body size during adolescence with subsequent adult bone mass, we conducted a follow-up study of a communitybased cohort of girls who participated in a growth and sexual maturation study 30 years ago. Data from the original study included age at menarche, height at menarche and weight at menarche. Follow-up evaluation of 119 subjects, now premenopausal women ages 40–45 years, included bone mineral density (BMD) measurements of the total body, lumbar spine, femoral neck, total hip, and ultradistal radius by dual-energy X-ray absorptiometry. After adjustment for current adult weight and other factors related to bone mass, weight at menarche was found to be positively associated with subsequent adult BMD. Similarly, body mass index (BMI) at menarche was positively associated with adult BMD. In contrast, age at menarche was not found to predict adult BMD. When the subjects were divided into quartiles based on their BMI at menarche, subjects in the lowest quartile of BMI at menarche had adult mean BMD that was 8–15% lower at the measured sites compared with subjects in the highest quartile of BMI at menarche. In conclusion, low body weight and low BMI

Correspondence and offprint requests to: Miriam Blum, MD, Calcium and Metabolic Bone Laboratory, Jean Mayer United States Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, Massachusetts, USA. Tel: +1 (617) 556 3064. Fax: +1 (617) 556 3305. e-mail:[email protected]

at menarche appear to be significant predictors of reduced bone mass in healthy premenopausal women ages 40–45 years. Keywords: Adolescence; Body mass index; Bone mass; Menarche; Weight

Introduction Peak bone mass is an important determinant for the risk of osteoporosis later in life [1]. The age at which peak bone mass is achieved is not certain, and estimates range from age 14 years to the fourth decade of life [2–5]. Adolescence is a critical time for skeletal growth and mineralization [3,6–9], and exposure to protective or detrimental factors during this period may influence the attainment of peak bone mass and the subsequent development of osteoporosis. Body size is one of the strongest predictors of bone mass. Cross-sectional studies have demonstrated a positive association between body size and bone mass in subjects of all age groups [10–16]. The influence of body size at one age on subsequent bone mass has also been described. Childhood weight has been shown to be associated with adolescent total body bone mineral content (BMC) in boys and girls with a mean age of 15.1 years [17]. Similarly, early body weight at age 1, 5 and 10 years was found to be associated with BMC at the lumbar spine and femoral neck in young adult women at age 21 years [18]. Additionally, weight change during adulthood appears to affect bone mass in elderly women [19,20]. To our knowledge, the influence of adolescent

Weight and BMI at Menarche are Associated with Premenopausal Bone Mass

body size at menarche on later bone mass has not been previously reported. Prior studies have examined the association between the timing of puberty and subsequent bone mass. Some have described no association [4,21,22] and others have reported a negative association [23–26] between age at menarche and bone mineral density (BMD). These studies have primarily been cross-sectional studies that rely on recall for age at menarche. In order to evaluate the association of adolescent body size at menarche and age at menarche with subsequent peak bone mass, we conducted a follow-up study of a community-based cohort of girls who had participated in a growth and sexual maturation study 30 years ago.

Subjects and Methods Subjects The Newton Girls Study began in 1965 as a prospective study of growth and sexual maturation in a communitybased cohort of 9- and 10-year-old girls. This cohort has been described previously in detail [27–29]. The girls were recruited from local public schools in a New England town (Newton, Massachusetts), and were followed at monthly intervals. Seven hundred and ninety-three girls were enrolled in the original study and 635 had body size measurements assessed at menarche. In 1998–1999, we attempted to recontact all of the subjects in the original study to invite them to participate in a follow-up study. Sixty-two percent of the original subjects responded to our inquiries. The present study population of 119 women is a subgroup of a larger cohort of follow-up participants. This subgroup consists of women who were eligible to participate in this study and were able to come to our center in Boston, MA for one visit. Ninety-seven percent of the participants were white and the remaining 3% were Hispanic, Asian or from the Pacific Islands. Exclusion criteria for this study included: incomplete menarcheal data, natural or surgical menopause, pregnancy or lactation within the previous 4 months, past use of oral glucocorticoids for more than 4 consecutive months, and history of a medical condition affecting bone metabolism. The protocol was approved by the Human Investigation Review Committee of Tufts University, and written informed consent was obtained from each participant.

Measurements The original study obtained data prospectively by a series of questionnaires sent to the participants’ homes each month and completed either by the girls’ mothers or by the girls themselves. The data recorded includes the precise day of appearance of first menstrual blood (age at menarche) as well as weight at menarche and height at menarche for each of the participants.

589

In the follow-up study, assessments of oral contraceptive use, cigarette use, alcohol use, reproductive history, and weight-bearing exercise during adolescence (defined as miles walked per day, ages 14–17 years) were made with a questionnaire designed for this study and completed by the subjects at home. Additional information about participants’ medical history, family history, medication use, milk consumption during adolescence (defined as cups of milk per day, ages 10– 17 years) and current use of calcium supplements was obtained in a medical interview. Assessment of current dietary calcium intake was made with the Fred Hutchinson Food Frequency Questionnaire [30], and total calcium intake was calculated by adding daily dietary calcium intake (mg) and calcium supplements (mg). Current body weight was measured with a conventional digital scale. Weight change since menarche was calculated by subtracting weight at menarche from current adult body weight. Current height was measured with a wall-mounted Harpenden stadiometer. Body mass index (BMI) was calculated by dividing the weight (in kilograms) by the square of the height (in meters).

Bone Densitometry BMD of the total body, lumbar spine (L2–L4) and hip was measured by dual-energy X-ray absorptiometry (DXA; model DPX-L scanner, Lunar Radiation, Madison, WI). The coefficient of variation (CV) of these scans in our laboratory is 0.62 for the total body, 1.0 for the spine and 2.1 for the femoral neck [31]. One spine scan was excluded from analysis due to marked vertebral scoliosis. BMD of the ultradistal radius was performed using DXA (model DTX 200 scanner, Osteometer Meditech, Copenhagen, Denmark). The CV of this measure in our laboratory is 2.0%.

Statistical Analysis Characteristics of the participants were reported as means or proportions. Associations between variables were examined using Pearson correlation coefficients and general linear models for multivariate adjustments. Covariates that are known to affect BMD were examined in the models; those that affected the parameter estimates were retained in the model. All adjusted models contained either current adult weight or weight change since menarche, current adult height, smoking history, adolescent milk consumption and adolescent weight-bearing exercise as covariates. When appropriate, either weight and height at menarche or BMI at menarche was included. For some of the analyses, the study population was divided into quartiles based on BMI at menarche. Characteristics across menarcheal BMI quartiles were compared by analysis of variance and proportions by chi-square tests. Analysis of covariance was used to control for differences between

590

M. Blum et al.

the groups in factors related to BMD. All analyses were done in SPSS (version 9.0, Chicago, IL).

Results Characteristics of the participants at adolescence obtained prospectively (age, weight, height and BMI at menarche) and retrospectively (walking, milk consumption) are shown in Table 1. At menarche, 10% of the girls were at or below the 15th percentile of BMI and 7% were at or above the 85th percentile for Caucasian girls according to the first United States National Health and Nutrition Survey (NHANES I) [32], which was conducted at about the time the girls were evaluated in the original study [33]. Neither weight at menarche nor BMI at menarche was significantly correlated with age at menarche (r = 0.095, NS; r = –0.157, NS, respectively), but height at menarche was positively correlated with age at menarche (r = 0.354, p50.01). Selected characteristics Table 1. Characteristics of 119 subjects at adolescence Age at menarchea (years) Weight at menarchea (kg) Height at menarchea (cm) BMI at menarchea (kg/m2) Milk servingsb (cups/day) 51 1–2 3 or more Walkingb (miles/day) 51 mile 1–2 miles 3–5 miles 45 miles

12.8
1.2 46.8
6.0 156.7
6.4 19.1
2.2

(10.0–16.0) (34.1–71.8) (136.3–172.5) (14.0–27.6)

14.3% (17) 46.2% (55) 39.5% (47) 37.0% 51.2% 7.6% 4.2%

(44) (61) (9) (5)

Values given are mean
SD (range) or percentage (n). a Assessed prospectively.

of the participants at follow-up are shown in Table 2. The group as a whole was educated, with 85% having completed college. As expected, measurements of body size at menarche including weight, height and BMI, were strongly correlated with adult body size measurements. Weight change since menarche was not significantly correlated with weight at menarche, but was correlated with current adult weight (Table 3). To examine the relationship between body size at menarche and adult BMD, we evaluated associations of weight at menarche and BMI at menarche with adult BMD. Both weight at menarche and BMI at menarche were significantly correlated with adult BMD at all skeletal sites measured. Height at menarche was not significantly associated with adult BMD at any site, and age at menarche was significantly correlated with adult BMD only at the ultradistal radius, where there was a small negative association (Table 4). Current adult weight was significantly correlated with adult BMD at all sites (r = 0.251 to 0.553, p50.01), and weight change since menarche was significantly correlated with adult BMD at all sites except the lumbar spine (r = 0.365 to 0.483, p50.01). In multivariable regression analyses, after adjustment for current adult weight, adult height, smoking history, height at menarche, age at menarche, adolescent walking and adolescent milk consumption, weight at menarche was significantly associated with adult BMD at the total body (p = 0.006), lumbar spine (p = 0.024) and total hip (p = 0.035). After similar adjustments (but substituting BMI at menarche for weight at menarche and height at menarche), BMI at menarche was also significantly associated with adult Table 3. Correlation coefficients between body size at menarche and body size at follow-up Adult weight

Adult height

Adult BMI

Weight changea

0.50** 0.18* 0.42** 0.91**

0.40** 0.89** –0.20* 0.14

0.38** –0.12 0.51** 0.90**

0.09 –0.02 0.12

Table 2. Characteristics of 119 subjects at follow-up Age (years) Weight (kg) Height (cm) BMI (kg/m2) Weight change since menarche (kg) Current total calcium intakea (mg/day) Current alcohol use 3 drinksb or fewer per month 1–3 drinks per week 4 drinks or more per week Current smoker Current oral contraceptive use Ever pregnant no. of pregnancies Bone mineral density (g/cm2) total body lumbar spine femoral neck total hip ultradistal radius

41.7
0.8 (40.0–45.0) 65.7
14.3 (44.2–113.2) 163.2
6.0 (145.5–180.0) 24.6
5.2 (27.3–44.8) 18.9
12.5 (1.7–64.3) 1101.7
578.5 (261.7–2737.4)

*Significant correlation (p50.05). **Significant correlation (p50.01). Weight change since menarche.

a

52.9% (63) 30.3% (36) 16.8% (20) 8.5% (10) 12.0% (14) 83.8% (100) 2.9
1.5 (1–9) 1.17
0.07 1.25
0.15 0.98
0.14 1.02
0.13 0.41
0.06

(1.03–1.39) (0.96–1.57) (0.71–1.40) (0.77–1.38) (0.30–0.66)

Values given are mean
SD (range) or percentage (n). Includes supplements. b 1 drink = 12 oz of beer, 4 oz of wine or 1 oz of liquor. a

Weight at menarche Height at menarche BMI at menarche Weight changea

Table 4. Correlation coefficients of age at menarche and body size at menarche with adult BMD Site/BMD

Age at menarche

Height at menarche

Weight at menarche

BMI at menarche

Total body Lumbar spine Femoral neck Total hip Ultradistal radius

–0.11 –0.11 –0.20 –0.05 –0.19*

0.09 0.09 0.15 0.09 0.02

0.38** 0.24** 0.34** 0.36** 0.32**

0.36** 0.21* 0.27** 0.34** 0.34**

*Significant correlation (p50.05). **Significant correlation (p50.01).

Weight and BMI at Menarche are Associated with Premenopausal Bone Mass

591

Table 5. Characteristics of 119 subjects by quartiles of BMI at menarche BMI quartile at menarche I (n = 29) BMI at menarche (kg/m2) Age at menarche (years) Adolescent milk servings (cups/day) 51 1–2 3 or more Adolescent walking(miles/day) 51 mile 1–2 miles 3–5 miles 45 miles Weight change since menarche (kg) Current BMI (kg/m2) Current age (years) Current smokers Current use of OCP Ever pregnant no. of pregnancies Current total calcium intake (mg/day)

16.7
0.2* 13.1
0.2

II (n = 30) 18.2
0.1* 13.0
0.2

III (n = 30) 19.4
0.1* 12.5
0.2

IV (n = 30) 21.9
0.3 12.7
0.2

10.3% 51.7% 37.9%

20.0% 23.3% 56.7%

20.0% 56.7% 23.3%

6.7% 53.3% 40.0%

34.5% 55.2% 6.9% 3.4% 16.9 21.7 41.6 10.3% 13.8% 72.4% 3.0 1158.3

36.7% 53.3% 3.3% 6.7% 19.0 23.6 41.9 13.3% 10.3% 89.7% 2.9 1139.6

36.7% 53.3% 6.7% 3.3% 15.2 23.9 41.7 0 6.9% 80.0% 2.9 1093.3

40.0% 36.7% 3.3% 10.0% 24.7 29.3 41.7 10.0% 16.7% 93.1% 2.7 1014.6


1.6*
0.6*
0.1


0.4
136.8


1.8
0.6*
0.2


0.4
88.8


2.1*
0.8*
0.2


0.3
121.3


3.0
1.0
0.2


0.3
69.7

Values given are mean
SE or percentage of respective group. OCP, oral contraceptive pill. Ranges of menarcheal BMI quartiles (kg/m2): I [14.0–17.5]; II [17.6–18.7]; III [18.7–20.4]; IV [20.5–27.6]. *Significantly differs from highest menarcheal BMI quartile (p50.05).

BMD at the total body (p = 0.007), lumbar spine (p = 0.034) and total hip (p = 0.044). In these adjusted models, neither age at menarche nor height at menarche was found to predict adult BMD at any of the measured sites. Since weight at menarche and BMI at menarche were both significantly correlated with current adult weight, we next substituted weight change since menarche (which was not significantly correlated with weight at menarche or BMI at menarche) for current weight in the models. After adjusting for weight change since menarche and additional variables as above, both weight at menarche and BMI at menarche were significantly associated with adult BMD at all measured sites (p50.01). For every additional 10 lb (4.5 kg) of weight at menarche, adult BMD was approximately 2% higher at the total body and lumbar spine and 4% higher at the femoral neck, total hip and ultradistal radius. We next divided the participants into quartiles based on their BMI at menarche. Characteristics of the women by quartiles of menarcheal BMI are given in Table 5. There were no significant differences in indices of reproductive history, oral contraceptive use, smoking, adolescent walking or milk consumption across the quartiles. Age at menarche and current adult age were also not significantly different across the quartiles. The quartiles did differ in BMI at menarche, weight change since menarche or current adult BMI, as indicated in Table 5. After adjusting for weight change since menarche and additional variables as before, there was a significant difference across the quartiles in adult BMD

at the total body (p50.001), lumbar spine (p = 0.007), femoral neck (p = 0.001) and ultradistal radius (p50.001) (Fig. 1). Similar findings were observed at the total hip (p50.001, data not shown). At each measured site, women who were in BMI quartiles I and II at menarche had significantly lower BMD than women in the highest quartile. Compared with women in the highest quartile, the adjusted mean BMD of women in the lowest quartile group of BMI at menarche was 8% lower at the total body, 11% lower at the lumbar spine, and 15% lower at the total hip, femoral neck and ultradistal radius.

Discussion This prospective study identifies a positive association between body size at menarche and subsequent adult bone mass in a cohort of premenopausal women ages 40–45 years. We found both weight and BMI at menarche to be significant predictors of adult BMD at the total body, lumbar spine and total hip. Since bone mass generally remains stable in young healthy adult women once peak bone mass is achieved [6,34,35], our data suggest that body size at menarche influences peak bone mass at these sites. Several longitudinal studies examining body size measurements in children ages 10 years or younger have reported an association of childhood body weight and subsequent bone mass. Duppe et al. [17] reviewed childhood health records of ninth-grade students with a mean age of 15.1 years. They reported a positive

592

M. Blum et al.

Fig. 1. Adult mean bone mineral density (BMD, g/cm2) measured at follow-up grouped by quartiles of body mass index (BMI, kg/m2) at menarche. BMD is adjusted for weight change since menarche, adult height, smoking history, age at menarche, adolescent walking and adolescent milk consumption. Ranges of BMI at menarche: quartile I [14.0–17.5]; quartile II [17.6–18.7]; quartile III [18.7–20.4]; quartile IV [20.5–27.6]. *Significantly differs from quartile IV (p50.01). +Significantly differs from quartile III (p40.01).

correlation between body weight at ages 4 and 6 years and adolescent total body BMC but not total body BMD for both boys and girls. No significant association was found between childhood body weight and adolescent femoral neck BMC or BMD [17]. Similarly, in a study of 153 women, Cooper, et al. [18] described a positive association between weight at 1 year, 5 years and 10 years and subsequent lumbar spine and femoral neck BMC but not BMD in subjects at age 21 years. Our study found more consistent associations between early body size and later adult BMD than these previous studies, perhaps because we evaluated body size measurements at puberty when a substantial proportion of total adult bone mass is accumulated. With the rapid accrual of bone mass during puberty, the growing adolescent skeleton may be particularly sensitive to the influence of body size. BMI in childhood and adolescence can be predictive of adult BMI [36] and in our study population there was a significant correlation between body size at menarche and adult body size. While it is possible that weight at menarche is a surrogate for adult body weight and not an independent predictor of adult BMD, we think this is doubtful because weight at menarche was significantly

associated with adult BMD after controlling for adult body size. In addition, when we substituted weight change since menarche (which was not significantly correlated with weight at menarche) for current adult weight in the models, weight at menarche remained significantly associated with BMD at all measured sites. Previous studies have established that in elderly women, low body weight [37–39] as well as weight loss between early or middle adulthood and old age [40,41] increases the risk of hip fracture. Our data suggest that lower body weight and lower BMI in adolescence may also be a risk factor for osteoporosis and fracture later in life. Women who were in the lowest quartile of BMI at menarche had subsequent adult BMD that was 11–15% lower at the lumbar spine and hip than women who were in the highest quartile of BMI at menarche. A 10% difference in bone mass is equivalent to almost 1 SD for bone density in the general population. A decrease of 1 SD in lumbar spine or hip density would be expected to result in an age-adjusted increase in hip fracture risk of 1.6 or 2.7, respectively [42]. Thus, a lower peak bone mass due to low body weight during adolescence could substantially increase the risk of osteoporotic fractures in later life.

Weight and BMI at Menarche are Associated with Premenopausal Bone Mass

Age at menarche is an important marker of pubertal development. By utilizing measurements made at the age of menarche rather than at a particular chronological age, we were able to take into account maturational differences that occur in adolescent girls of the same age. In examining age at menarche with regard to adult BMD, after adjustment for covariates, we did not find age at menarche to be associated with adult BMD at any site. While some prior reports have found age at menarche to be negatively associated with adult BMD at the lumbar spine [23–25] and at the wrist [26,43], in agreement with us, other investigators have not detected an association between age at menarche and adult BMD [21,22]. This study has several limitations. Our measurement of BMD with DXA is not a volumetric density, but is based on a projected two-dimensional area of bone. Since these measurements may be influenced by variations in bone size [44], we adjusted for current adult height and weight in all analyses. Secondly, the menarcheal measurements collected prospectively in the original study were obtained either by the girls selfreporting their height and weight, or by their mothers reporting the measurements. Although there is bound to be some bias inherent in these measurements, selfreported weight and height [45,46] as well as mothers’ reports of their children’s weight [47], have been demonstrated to be accurate measures. Finally, since this study followed a group of subjects with relatively homogeneous ethnic and socioeconomic backgrounds, our findings are not necessarily representative of the general population. In terms of body size at menarche, however, fewer of our subjects were very thin (BMI measurements of 15th percentile or lower), and fewer were obese (BMI measurements of 85th percentile or higher), compared with age- and race-specific BMI percentiles for girls in the United States [32]. In conclusion, low body weight and low BMI at menarche appear to be significant predictors of reduced bone mass in healthy premenopausal women. In contrast, we did not find age at menarche to be a predictor of adult bone mass. While obesity at any age can lead to multiple health problems, low adolescent body size may also be associated with subsequent health risks. Because excessive thinness is promoted as an ideal among many adolescent girls in Western countries, our finding that low adolescent body size may adversely influence adult bone mass is of concern. Acknowledgements. This material is based on work supported by the US Department of Agriculture, under agreement no. 58-1950-9-001. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors, and do not necessarily reflect the view of the US Department of Agriculture. This work was presented in part as an abstract at the ASBMR meeting in St Louis, Missouri in 1999.

References 1. Hui SL, Slemenda CW, Johnston CCJ. The contribution of bone loss to postmenopausal osteoporosis. Osteoporos Int 1990;1:30–4.

593

2. Matkovic V, Fontana D, Tominac C, Goel P, Chestnut CH III. Factors that influence peak bone mass formation: a study of calcium balance and the inheritance of bone mass in adolescent females. Am J Clin Nutr 1990;52:878–88. 3. Theintz G, Buchs B, Rizzoli R, Slosman D, Clavien H, Sizonenko PC, et al. Longitudinal monitoring of bone mass accumulation in healthy adolescents: evidence for a marked reduction after 16 years of age at the levels of lumbar spine and femoral neck in female subjects. J Clin Endocrinol Metab 1992;75:1060–5. 4. Lu PW, Briody JN, Ogle GD, Morley K, Humphries IRJ, Allen J, et al. Bone mineral density of total body, spine and femoral neck in children and young adults: a cross-sectional and longitudinal study. J Bone Miner Res 1994;9:1451–8. 5. Recker RR, Davies KM, Hinders SM, Heaney RP, Stegman MR, Kimmel DB. Bone gain in young adult women. JAMA 1992;268:2403–8. 6. Sabatier JP, Guaydier-Souquieres G, Laroche D, Benmalek A, Fournier L, Guillon-Metz F, et al. Bone mineral acquisition during adolescence and early adulthood: a study in 574 healthy females 10–24 years of age. Osteoporos Int 1996;6:141–8. 7. Kroger H, Kotaniemi A, Kroger L, Alhava E. Development of bone mass and bone density of the spine and femoral neck: a prospective study of 65 children and adolescents. Bone Miner 1993;23:171–82. 8. Katzman DK, Bachrach LK, Carter DR, Marcus R. Clinical and anthropometric correlates of bone mineral acquisition in healthy adolescent girls. J Clin Endocrinol Metab 1991;73:1332–9. 9. Slemenda CW, Reister TK, Hui SL, Miller JZ, Christian JC, Johnston CCJ. Influences on skeletal mineralization in children and adolescents: evidence for varying effects of sexual maturation and physical activity. J Pediatr 1994;125:201–7. 10. Glastre C, Braillon P, David L, Cochat P, Meunier PJ, Delmas PD. Measurement of bone mineral content of the lumbar spine by dual energy X-ray absorptiometry in normal children: correlations with growth parameters. J Clin Endocrinol Metab 1990;70:1330– 3. 11. Rubin K, Schirduan V, Gendreau P, Sarfarazi M, Mendola R, Dalsky G. Predictors of axial and peripheral bone mineral density in healthy children and adolescents, with special attention to the role of puberty. J Pediatr 1993;123:863–70. 12. Henderson NK, Price RI, Cole JH, Gutteridge DH, Bhagat CI. Bone density in young women is associated with body weight and muscle strength but not dietary intakes. J Bone Miner Res 1995;10:384–92. 13. Davee AM, Rosen CJ, Adler RA. Exercise patterns and trabecular bone density in college women. J Bone Miner Res 1990;5:245– 50. 14. Fehily AM, Coles RJ, Evans WD, Elwood PC. Factors affecting bone density in young adults. Am J Clin Nutr 1992;56:579–86. 15. Pocock N, Eisman J, Gwinn T, Sambrook P, Kelly P, Freund J, et al. Muscle strength, physical fitness, and weight but not age predict femoral neck bone mass. J Bone Miner Res 1989;4:441–8. 16. Lindsay R, Cosman F, Herrington B, Himmelstein S. Bone mass and body composition in normal women. J Bone Miner Res 1992;7:55–63. 17. Duppe H, Cooper C, Gardsell P, Johnell O. The relationship between childhood growth, bone mass, and muscle strength in male and female adolescents. Calcif Tissue Int 1997;60:405–9. 18. Cooper C, Cawley M, Bhalla A, Egger P, Ring F, Morton L, et al. Childhood growth, physical activity and peak bone mass in women. J Bone Miner Res 1995;10:940–7. 19. Holbrook TL, Barrett-Conner E. The association of lifetime weight and weight control patterns with bone mineral density in an adult community. Bone Miner 1993;20:141–9. 20. Felson DT, Zhang Y, Hannan MT, Anderson JJ. Effects of weight and body mass index on bone mineral density in men and women: The Framingham Study. J Bone Miner Res 1993;8:567–73. 21. Sowers MR, Clark MK, Hollis B, Wallace RB, Jannausch M. Radial bone mineral density in pre- and perimenopausal women: a prospective study of rates and risk factors for loss. J Bone Miner Res 1992;7:647–57. 22. Melton LJ, Bryant SC, Wahner HW, O’Fallon WM, Malkasian

594

23. 24.

25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

M. Blum et al. GD, Judd HL, Riggs BL. Influence of breastfeeding and other reproductive factors on bone mass later in life. Osteoporos Int 1993;3:76–83. Ito M, Yamada M, Hayashi K, Ohki M, Uetani M, Nakamura T. Relation of early menarche to high bone mineral density. Calcif Tissue Int 1995;57:11–4. Boot AM, de Ridder MA, Pols HA, Krenning EP, de Muinck Keizer-Schrama SM. Bone mineral density in children and adolescents: relation to puberty, calcium intake and physical activity. J Clin Endocrinol Metab 1997;82:57–62. Rosenthal DI, Mayo-Smith W, Hayes CW, Khurana JS, Biller BMK, Neer RM, et al. Age and bone mass in premenopausal women. J Bone Miner Res 1989;4:533–8. Fox KM, Magaziner J, Sherwin R, Scott JC, Plato CC, Nevitt M, et al. Reproductive correlates of bone mass in elderly women. J Bone Miner Res 1993;8:901–8. Zacharias L, Rand WM, Wurtman RJ. A prospective study of sexual development and growth in American girls: the statistics of menarche. Obstet Gynecol Surv 1976;31:325–37. Zacharias L, Rand WM. Adolescent growth in height and its relation to menarche in contemporary American girls. Ann Hum Biol 1983;10:209–22. Zacharias L, Rand WM. Adolescent growth in weight and its relationship to menarche in contemporary American girls. Ann Hum Biol 1986;13:369–86. Block G, Woods M, Potosky A, Clifford C. Validation of a selfadministered diet history questionnaire using multiple diet records. J Clin Epidemiol 1990;43:1327–35. Johnson J, Dawson-Hughes B. Precision and stability of dualenergy x-ray absorptiometry measurements. Calcif Tissue Int 1991;49:174–8. Must A, Dallal GE, Dietz WH. Reference data for obesity:85th and 95th percentiles of body mass index (wt/ht2) and triceps skinfold thickness. Am J Clin Nutr 1991;53:839–46. National Center for Health Statistics. Plan and operation of health and nutrition examination survey, United States 1971–1973. Vital Health Stat [1] 1973;10a,10b:1–46, 1–77. Mazess RB, Barden HS. Bone density in premenopausal women: effects of age, dietary intake, physical activity, smoking and birth control pills. Am J Clin Nutr 1991;53:132–42. Matkovic V, Jelic T, Wardlaw GM, llich JZ, Goel PK, Wright JK,

36. 37.

38. 39. 40.

41.

42. 43. 44. 45. 46. 47.

et al. Timing of peak bone mass in Caucasian females and its implication for the prevention of osteoporosis. J Clin Invest 1994;93:799–808. Guo SS, Roche AF, Chumlea WC, Gardner JD, Siervogel RM. The predictive value of childhood body mass index values for overweight at age 35y. Am J Clin Nutr 1994;59:810–9. Ensrud KE, Lipschutz RC, Cauley JA, Seeley D, Nevitt MC, Scott J, et al. Body size and hip fracture risk in older women: a prospective study. Study of Osteoporotic Fractures Research Group. Am J Med 1997;103:274–80. Kiel DP, Felson DT, Anderson JJ, Wilson PW, Moskowitz MA. Hip fracture and the use of estrogens in postmenopausal women: the Framingham study. N Engl J Med 1987;317:1169–74. Paganini-Hill A, Chao A, Ross RK, Henderson BE. Exercise and other factors in the prevention of hip fracture: The Leisure World Study. Epidemiology 1991;2:16–25. Cummings SR, Nevitt MC, Browner WS, Stone K, Fox KM, Ensrud KE, et al. Risk factors for hip fracture in white women. Study of Osteoporotic Fractures Research Group. N Engl J Med 1995;332:767–73. Langlois JA, Harris T, Looker AC, Madans J. Weight change between age 50 years and old age is associated with risk of hip fracture in white women aged 67 and older. Arch Intern Med 1996;156:989–94. Cummings SR, Black DM, Nevitt MC, Browner W, Cauley J, Ensrud K, et al. Bone density at various sites for prediction of hip fractures. Lancet 1993;341:72–5. Johnell O, Nilsson BE. Life-style and bone mineral mass in perimenopausal women. Calcif Tissue Int 1984;36:354–6. Carter DR, Bouxsein ML, Marcus R. New approaches for interpreting projected bone densitometry data. J Bone Miner Res 1992;7:137–45. Stunkard AJ, Albaum JM. The accuracy of self -reported weights. Am J Clin Nutr 1981;34:1593–9. Palta M, Prineas RJ, Berman R, Hannan P. Comparison of selfreported and measured height and weight. Am J Epidemiol 1982;115:223–30. Coates TJ, Jeffery RW, Wing RR. The relationship between persons’ relative body weights and the quality and quantity of food stored in their homes. Addict Behav 1978;3:179–84. Received for publication 15 August 2000 Accepted in revised form 2 January 2001