• bone mineral density;
  • physical activity;
  • body composition;
  • grip strength;
  • 16–20-year-old women


  1. Top of page
  2. Abstract
  3. Introduction
  4. Participants and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Abstract. Valdimarsson Ö, Kristinsson JÖ, Stefansson SÖ, Valdimarsson S, Sigurdsson G (Reykjavik Hospital, University of Iceland, Reykjavik, Iceland). Lean mass and physical activity as predictors of bone mineral density in 16–20-year old women. J Intern Med;245: 489–496.

Objective.  The aim of the study was to quantify the inter-relationship between bone mineral density and physical activity, muscle strength, and body mass composition in a group of healthy 16–20-year-old women.

Design. A cross-sectional study.

Setting.  Reykjavik area.

Subjects.  Two-hundred and fifty-four Icelandic Caucasian women aged 16, 18 and 20 years, randomly selected from the registry of Reykjavik.

Main outcome measures.  Bone mineral content (BMC) and density (BMD) in lumbar spine, hip, distal forearm and total skeleton and lean mass and fat mass were measured with dual energy X-ray absorptiometry (DEXA) and compared with grip strength measured with a dynamometer and physical activity as assessed by a questionnaire.

Results.  The lean mass had the strongest correlation with BMC and BMD, stronger than weight, height and fat mass, both in univariate analysis (r = 0.41–0.77; P < 0.001) and in linear regression analysis. The total skeletal BMD was logarithmically higher by hours of exercise per week (P < 0.001)). About 30% of variability in total skeletal BMD in this age group can be predicted by lean mass and physical exercise.

Conclusions.  Modifiable factors, such as exercise and adequate muscle seem to be significant predictors of the attainment of peak bone mass in women.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Participants and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The bone mass attained during adolescence is likely to be an important determinant of the risk of fracture later in life [1]. Although it is recognized that bone mass is controlled largely by genetic factors [2–5], environmental and lifestyle factors such as nutrition, physical activity, muscle strength and body composition may play roles of varying importance [6–8]. These factors are potentially modifiable, but their contribution to final peak bone mass is uncertain. Methods to optimize the potential for high peak bone mass must be applied during the years when rapid skeletal mineral acquisition occurs.

Although bodyweight has been identified as a determinant of bone mineral density (BMD), some controversy exists over the independent effects of its major components, lean mass and fat mass, as measured by dual-energy X-ray absorptiometry (DEXA) [29–13]. Faulkner et al. [10] and Reid et al. [12] concluded that bone-free, lean tissue was the most important predictor of total BMD in both boys and girls, more predictive than bodyweight alone. Physical activity is another factor that could affect both body composition and bone density, but the relationship between these variables and physical activity has not been well studied.

The main objective of our study was to investigate whether BMD is related to physical activity, muscle strength and basic constitutional parameters, such as lean mass, fat mass, weight and height in a group of healthy 16-, 18- and 20-year-old women. The inter-relationship between these factors has not been studied thoroughly in this age group in previous studies.

Participants and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Participants and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References


Three hundred and fifty-eight girls, aged 16, 18 and 20 (born in 1980, 1978 and 1976), were selected randomly from the registry of Reykjavik and invited to participate in the study. Fifty-one declined to participate for personal reasons, 41 could not be reached and seven were pregnant. Two hundred and fifty-nine participated (72.3%), and informed written consent was obtained from the girls or their parents. One girl was excluded because of a disease, and four had taken medication known to affect bone metabolism (i.e. steroids or anticonvulsants). The final study group included 254 healthy, Icelandic, Caucasian girls: 71 in the 16th year (78% participation rate), 64 in the 18th year (88.7%) and 119 in the 20th year (60.7%). The assessments were performed in February, March and April 1996. Ethical approval for the study was granted by the Ethical Committee of the Reykjavik Hospital. The assessment with regard to nutritional factors in the same study group has been published [13].

Activity questionnaire

Physical activity (PA) was assessed by the questionnaire developed by Slemenda et al. [5] which is a modified version of the reports of the National Children and Youth Fitness Study [14] and more suitable for clinical use. The subjects were asked questions regarding physical education classes (frequency, time and intensity) and regarding the time spent each week in 15 specific activities, nonweight-bearing exercises, biking and swimming, and weight-bearing exercises such as football, basketball, volleyball, handball, gymnastics, alpine skiing, cross country skiing, dancing, weight lifting, aerobics, running, badminton, golf, swimming, bicycling and one open-ended question on other physical activities. Walking was excluded from this analysis because of subjects’ difficulty in estimating time spent on this. The frequency of activity during the last 3 months was used.

Assessments of anthropometric findings

Measurements of height (cm) and weight (kg) were made on the same day as bone mineral measurements. Height was measured to the nearest 0.5 cm with subjects barefoot using a SECA stadiometer. Weight was measured to the nearest 0.5 kg with subjects barefoot and in light clothing using a SECA electronic scale. The menarcheal age was defined as the time since menarche. The girls were asked about menstruation history and irregularities.

Assessment of bone mineral density and body composition

Dual-energy X-ray absorptiometry (DEXA) (Hologic QDR-2000 plus, Hologic Inc., Waltham, MA) was used to measure bone mineral content (BMC, g) and bone area (BA, cm2) of the lumbar vertebrae (L2-L4), proximal end of the left femur (femoral neck, trochanteric and intertrochanteric regions, the totality referred to as total hip), dominant forearm (1/3 distal, mid-distal and ultra-distal regions, the totality referred to as total forearm) and total skeleton. From these two measurements, areal bone mineral density (BMD, g/cm2) was calculated. The reproducibility for replicate scans of the same individual with repositioning was 1.0% for the spine and 1.8% for the proximal femur (CV, coefficient of variation). Phantom calibration was made daily. Total body soft tissue composition, in terms of fat mass and lean mass (both in kg), were also measured by the same DEXA instrument (enhanced whole body array mode); the CV for lean mass was 1.7% and fat mass 2.3%. A single trained radiographer performed all the DEXA measurements.

Grip strength measurement

The isometric grip strength of the dominant arm was measured with a Jamar adjustable hand-held dynamometer. The girls sat with the elbow flexed to 90°, shoulder adducted and forearm in a neutral position. The adjustable handle was placed in the second position so the grip space was 4.7 cm. Three measurements were performed, where the instrument was pressed for a few seconds, and the resting period between the tests was at least 1 min. The values recorded were the average of the two highest scores.

Statistical analysis

Spearman’s correlation coefficients (r) were calculated for each age group and for the age groups combined.

Multivariate analysis was performed by multiple linear regression. Attributable percentage (r2) explained by each independent variable was calculated as the square of the partial correlation coefficients (r2).

As the distribution of physical activity was very skewed, a logarithmic transform (ln(PA + 1)) was used and proved to be a better predictor of BMC and BMD than PA itself.

The same applies to the variable years since menarche (YSM) where the transform ln(YSM + 1) was used and fat mass where ln(fat mass) was used.

Spearman’s correlation coefficients (r) were calculated between BMC and BMD on the one hand and the explanatory variables on the other hand.

Spearman’s correlation coefficients are invariant under logarithmic transformation of the variables.

P-values less than 0.05 were considered significant. The program package used was SPIDA [15].


  1. Top of page
  2. Abstract
  3. Introduction
  4. Participants and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The characteristics of the participants are shown in Table 1. The mean age of menarche was 13.1 years. None of the participants reported amenorrhoea for more than 3 months. Maximal height was achieved by the 18th year, but the girls’ weight was greater with increasing age. The fat mass was 30.1% of bodyweight for the 16-year-olds, but 35.1% for the 20-year-olds, whereas the lean mass was 64.7% and 60.3%, respectively. At the same time, the mean total bodyweight ranged from 57.8 kg to 65.0 kg, so the mean lean mass ranged from 37.4 kg to 39.2 kg, and mean fat mass from 17.4 kg to 22.8 kg.

Table 1.  Study group characteristics Thumbnail image of

The 16-year-olds were significantly (P < 0.01) more active physically than the 18-and 20-year-old girls. The grip strength was similar in all groups.

About 20% of the whole group exercised more than 7 h per week, whereas 20% trained less than 1/2 h per week.

The mean values for BMC and BMD are shown in Table 2. There was a significant difference in total skeletal BMC, mean 12.1%, and BMD, mean 5.2%, from 16 to 20 years. Thus, the 20-year-olds had about 2% greater total skeletal BMD than the 18-year-olds (P < 0.01), whereas there was no difference in height. Between 16 and 20 years there was, however, less difference in the total hip, 7.4% for BMC and 5.2% for BMD and in the spine the difference was 7.9% for BMC and 5.6% for BMD with P < 0.01 for all comparisons. In the total forearm the difference was similar as for total skeletal measurement, 11.3% for BMC and 7.1% for BMD. Concomitantly, total skeletal area was 6.6% higher and spinal and total hip area 2.2% higher (P < 0.01).

Table 2.  Mean bone mineral content (BMC-g) and bone mineral density (BMD-g/cm2) for different age groups Thumbnail image of

The correlation between BMC at different measurement sites was greatest between total skeleton and other sites (r = 0.75–0.83). The correlation between the spine and total hip was r = 0.71–0.76 and between the spine and forearm, r = 0.59–0.70 (lowest in the older girls). The correlation coefficients for BMD were almost the same as for BMC between total skeleton and other measurement sites, but slightly less elsewhere (data not shown).

Univariate analysis

In Table 3 the correlation coefficients (r) between BMC/area/BMD and the physical characteristics in the total group are shown, but the results were fairly similar in all age groups. Of the significant variables, lean mass had the greatest correlation with BMC/area/BMD at all sites, exceeding the correlations for weight, height and fat mass. The correlation coefficients between weight and lean mass was 0.67 (P < 0.01) and height 0.40 (P < 0.01), whereas the coefficient between lean mass and fat mass was 0.33 (P < 0.01) and height 0.69 (P < 0.01).

Table 3.  Spearman’s correlation coefficients (r) between bone mineral content (BMC) and bone mineral density (BMD) and subjects’ characteristics in the total group Thumbnail image of

Physical activity correlated with BMC and BMD for total skeleton (0.14/0.21; P < 0.05/0.01) and total hip (r = 0.20/0.26; P < 0.001) but not for area size. The weight-bearing activity correlated significantly, especially with the BMD of weight-bearing sites, such as the total hip (r = 0.23; P < 0.001), but nonweight-bearing activity had a significant correlation only with the BMC and BMD of total hip (r = 0.15; P < 0.05). The grip strength correlated significantly with BMC/BMD at all sites, but most with the forearm and more with BMC than BMD. The number of years since menarche correlated with BMC and BMD for the whole group, most for the 16-year-olds, but a trend was also seen for the 18- and 20-year-old girls. Height and weight were used instead of body mass index (BMI), as they had a stronger effect in the calculations.

Physical activity correlated positively with lean mass, the highest correlation being for the 16-year-olds (r = 0.46; P < 0.001).

Multivariate analysis

To further examine the independent contributions of the modifiable lifestyle factors of muscle strength, physical activity and lean mass to BMC and BMD, the inter-relations were corrected, using a multiple linear regression analysis performed for each age group as well as for the total group ( Table 4). The results for each age group were comparable.

Table 4.  Linear regression analysis for the total group Thumbnail image of

The results for the total group (total skeleton) are shown in Table 4. Five factors explained 70.3% (Rsq: 0.703) of the variation in BMC of total skeleton. The lean mass explained 58.5%, fat mass 2.4%, height 4.2%, number of years since menarche 5.5% and physical activity 0.2%. The lean mass was the most important factor at all measurement sites. Height correlated also with all measurement sites except the femoral neck. The number of years since menarche correlated with all measurement sites except total hip and femoral neck and this variable was a stronger predictor than age, which was not a significant additional predictor. The only correlation found with grip strength was the total forearm, which explained 4%. Weight-bearing activity correlated with femoral neck and total hip and explained 5.2% and 4.7%, respectively, of the variation of BMC at these sites.


Three factors were significantly related to total BMD and explained 37.1% of total skeletal BMD for the total group. Lean mass explained 28.3%, years since menarche 5.9% and physical activity 2.9%. When physical activity was separated into weight-bearing and nonweight-bearing activity, most of the association with BMD was limited to weight-bearing activity. As for BMC, lean mass was the most significant factor and correlated with BMD at all measurement sites. The number of years since menarche correlated most with spine and total forearm. Physical activity correlated most with total hip and femoral neck, explaining 5%, and fat mass, explaining 10% at these sites. Fat mass, however, was not significant for total skeletal BMD. Height was not significantly correlated with BMD.

Figure 1 shows that there was a significant, positive relationship between training hs per week and total skeletal BMD in all age groups. The difference in BMD was, however, more marked from 0 to 3 h per week and most marked amongst the 16-year-olds. Thus, girls training more than 7 hs per week had 3.6–5.8% greater BMD than those training less than half an hour per week. There was no upper threshold of this relationship.


Figure 1. Results from linear regression of total skeletal BMD on logarithmic transform of physical activity (PA), adjusted for time since menarche. The whole group: BMD = 1.003 + 0.020xln(PA + 1), P < 0.001; 16-year-olds: BMD = 0.939 + 0.034xln(PA + 1), P = 0.002; 18-year-olds: BMD = 0.989 + 0.035xln(PA + 1), P < 0.001; 20-year-olds: BMD = 1.027 + 0.018xln(PA + 1), P = 0.03.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Introduction
  4. Participants and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This cross-sectional study of 16-, 18-and 20-year-old girls consistently showed a significant relationship between lean mass and bone mineral density, in both univariate and multivariate analyses. The lean mass (nonfat, soft tissue) has been shown to be 45–60% of muscle mass [10]. We therefore conclude that this relationship between lean mass and BMD is most likely related to muscle mass. The positive association of total skeletal BMD with grip strength (in a univariate analysis), a rough measure of muscle strength [16] further supports this notion. This relationship was, however, weaker than for muscle mass which could partly be explained by the effect of different hand sizes, whereas the same dynamometer position was used for all participants.

We also found a significant relationship between physical activity, mostly weight-bearing activity, and total BMD, especially axial BMD. The physical activity had a stronger correlation with hip BMD than with that of the forearm and spine. We found a significant, positive relationship between lean mass and physical activity, greatest in the youngest age group, who were also the most physically active. However, the possibility of self-selection by physique, i.e. the more muscular girls are more likely to engage in physical exercise, cannot be ruled out, but the positive relationship between physical activity and total BMD (independent of lean mass) makes this possibility less likely. Also, a common genetic effect on muscle mass and bone size cannot be excluded. This inter-relationship between physical activity, lean mass and grip strength supports, however, the hypothesis that physical activity plays an important role in achieving peak bone mass. The magnitude of the association with physical activity varies between 5 and 15% in other cross-sectional studies [17–22], whereas prospective studies have shown a smaller difference [2324]. In our study, weight-bearing activity had a greater association with BMD than nonweight-bearing activity (P < 0.01) (e.g. swimming) as some other studies have suggested [52425]. Some other studies [26] have not found a significant association with exercise as we have, however, the authors point out that this component had limited statistical power in their study. Parsons et al. however, did not identify any lifestyle and anthropometric factors influencing bone mineral in 18- to 21-year-olds [2].

Our results also suggest that muscle mass has a greater correlation with bone mass than adipose tissue, at least at this age, a finding that is consistent with some recent studies [1026]. The lean mass and physical activity predicted about 30% of the variance in BMD in our study, which is a very similar finding to that of Henderson et al. for 18-year-old girls [17]. Altogether 70.3% of total skeletal BMC variability and 37.1% of BMD could be explained by factors measured in our study ( Table 4). It is of interest in this respect that genetic factors have been estimated to contribute 60–80% of the peak bone mass [672327] which might indicate common genetic influence on bone mass and some factors measured in our study such as muscle mass.

Our results indicate a temporary relationship with pubertal, hormonal changes reflected in the association of the number of years since menarche and total BMD. This association was most apparent in the 16-year-olds, but still present, although weaker, in the 18- and 20-year-olds. This is contrary to the persistent effect found in Japanese women [28]. It is of interest that BMC was 4% higher and BMD 2% higher in 20 years than 18-year-old girls after linear growth has stopped. The mechanisms by which BMC and BMD increase after linear growth has stopped remain undefined and may be due to increases in bone size or mineralization or decreases in bone remodelling. It is of interest that measured total skeletal area was 2.2% greater in 20-year-olds compared to 18-year-olds in our study group (P < 0.05), partly explaining the difference in BMC. However, a bias introduced by different participation rates amongst the study groups cannot be excluded. Faulkner et al. did not find any increase in Canadian girls between 17 and 21 years of age [29]. The difference in total skeletal BMD between the 16- and 20 year-old groups in our study was about 5% (P < 0.01).

This study indicates that lean mass (presumably half of it muscle mass) and physical activity predict about 30% of total skeletal BMD in girls aged 16–20. Although muscle mass is certainly genetically determined to some extent, it is also modified by physical activity, which in our study also was independently associated with BMD. It is therefore likely that the calculated percentage of variability in BMD attributed to physical activity in our results is a minimal figure. Five hs of training per week in this age group was associated with at least 2% higher bone mass, which is a greater association than is achieved by heavy exercise in adults [30]. If this association with physical activity is maintained until old age, the risk of osteoporotic fractures is likely to be substantially lower in active than in sedentary subjects. It is of interest that recent studies on twins have indicated that the association between BMD and lean mass is mediated mainly via environmental influences [31] and substantial role for environmental and lifestyle factors on BMD was evidenced in early adulthood [32].

Our study is cross-sectional with all its limitations, reflecting associations but not revealing causes and effects. The results need therefore to be confirmed in a prospective intervention study with randomized groups of different physical exercise.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Participants and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study was supported by a grant from the Icelandic Research Council and Reykjavik Hospital Science Fund. We thank D. Oskarsdottir for performing all DEXA-measurements and H. Sigvaldason for statistical analysis.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Participants and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Ott SM. Bone density in adolescents. N Engl J Med 1991; 325: 1646 47.
  • 2
    Parsons TJ, Prentice A, Smith EA, Cole TJ, Compston JE. Bone mineral mass consolidation in young British adults. J Bone Mineral Res 1996; 11: 264 74.
  • 3
    Pocock NA, Eisman JA, Gopper JL, Yeates MG, Sambrook PN, Eberl S. Genetic determinants of bone mass in adults: a twin study. J Clin Invest 1987; 80: 706 10.
  • 4
    Seeman E, Hopper JL, Bach LA, Cooper ME, Parkinson E, McKay J. Reduced bone mass in daughters of women with osteoporosis. N Engl J Med 1989; 320: 554 58.
  • 5
    Slemenda CW, Miller JZ, Hui SL, Reister TK, Johnston Cc Jr Genetic determinants of bone mass in adult women: a reevaluation of the twin model and the potential importance of gene interaction on heritability estimates. J Bone Mineral Res 1991; 6: 1227 33.
  • 6
    Fehily AM, Coles RJ, Evans WD, Elwood PC. Factors affecting bone density in young adults. Am J Clin Nutr 1992; 56: 579 86.
  • 7
    Hirota T, Nara M, Ohguri M, Manago E, Hirota K. Effect of diet and lifestyle on bone mass in Asian young women. Am J Clin Nutr 1992; 55: 1168 73.
  • 8
    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.
  • 9
    Aloia JF, Vaswani A, Ma R, Flaster E. To what extent is bone mass determined by fat-free or fat mass? Am J Clin Nutr 1995; 61: 1110 14.
  • 10
    Faulkner RA, Bailey DA, Drinkwater DT, Wilkinson AA, Houston CS, McKay HA. Regional and total body bone mineral content, bone mineral density, and total body tissue composition in children 8–16 years of age. Calcif Tissue Int 1993; 53: 7 12.
  • 11
    Lohman T, Going S, Pamenter R et al. Effects of resistance training on regional and total bone mineral density in premenopausal women: a randomized prospective study. J Bone Mineral Res 1995; 10: 1015 24.
  • 12
    Reid IR, Legge M, Stapleton JP, Evans MC, Grey AB. Regular exercise dissociates fat mass and bone density in premenopausal women. J Clin Endocrinol Metab 1995; 6: 1764 68.
  • 13
    Kristinsson JÖ, Valdimarsson Ö, Sigurdsson G, Franzson L, Olafsson I, Steingrimsdottir L. Serum 25-hydroxyvitamin D levels and bone mineral density in 16–20-year-old girls: lack of association. J Int Med 1998; 243: 381 88.
  • 14
    Public Health Service Office of Disease Prevention and Health Promotion 1985. The National Children and Youth Fitness Study: A summary of findings. U.S. Department of Health and Human Services. JOPERD, January, pp. 13 93.
  • 15
    Gebski V, Leung O, McNeil D, Lunn D. SPIDA User’s Manual, Version 6. Statistical Computing Lab, Macquarie University of NSW, Australia, 1992.
  • 16
    Snow-Harter C, Bouxsein M, Lewis B, Charette S, Weinstein P, Marcus R. Muscle strength as a predictor of bone mineral density in young women. J Bone Mineral Res 1990; 5: 589 95.
  • 17
    Hendersson 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 Mineral Res 1995; 10: 384 93.
  • 18
    Kanders B, Dempster DW, Lindsay R. Interaction of calcium nutrition and physical activity on bone mass in young women. J Bone Mineral Res 1988; 3: 145 49.
  • 19
    Kröger H, Kotaniemi A, Vainio P, Alhava E. Bone densitometry of the spine and femur in children by dual-energy x-ray absorptiometry. Bone and Mineral 1992; 17: 75 85.
  • 20
    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 Paediatrics 1993; 123: 863 70.
  • 21
    Ruiz JC, Mandel C, Garabedian M. Influence of spontaneous calcium intake and physical exercise on the vertebral and femoral bone mineral density of children and adolescents. J Bone Mineral Res 1995; 10: 675 82.
  • 22
    Slemenda CW, Reister TK, Miller JZ, Christian JC, Johnston CC. Influences on skeletal mineralization in children and adolescents: Evidence for varying effects of sexual maturation and physical activity. J Paediatrics 1994; 12: 201 7.
  • 23
    Välimäki MJ, Kärkkäinen M, Lamberg-Allardt C et al. Exercise, smoking, and calcium intake during adolescence and early adulthood as determinants of peak bone mass. Br Med J 1994; 309: 230 35.
  • 24
    Welten DC, Kemper Hcg, Post GB et al. Weight-bearing activity during youth is more important factor for peak bone mass than calcium intake. J Bone Mineral Res 1994; 9: 1089 96.
  • 25
    Slemenda CW & Johnston CC. High intensity activities in young women: site specific bone mass effects among female figure skaters. Bone Mineral 1993; 20: 125 32.
  • 26
    Young D, Hopper JL, Nowson CA et al. Determinants of bone mass in 10- to 26- year old females: a twin study. J Bone Mineral Res 1995; 10: 558 67.
  • 27
    Metz JA, Anderson Jjb, Gallagher Pn Jr. Intakes of calcium, phosphorus, and protein, and physical activity level are related to radial bone mass in young adult women. Am J Clin Nutr 1993; 58: 537 42.
  • 28
    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 14.
  • 29
    Faulkner RA, Bailey DA, Drinkwater DT, McKay HA, Arnold C, Wilkinsson AA. Bone densitometry in Canadian children 8–17 years of age. Calcif Tissue Int 1996; 59: 344 51.
  • 30
    Nichols DL, Sanborn CF, Bonnick SL, Ven-Ezra V, Gench B, DiMarco NM. The effects of gymnastics training on bone mineral density. Med Sci Sports Exerc 1994; 26: 1220 25.
  • 31
    Nguyen TV, Howard GM, Kelly PJ, Eisman JA. Bone mass, lean mass, and fat mass: same genes or same environments? Am J Epidemiol 1998; 147: 3 16.
  • 32
    Hopper JL, Green RM, Nowson CA et al. Genetic, common environment, and individual specific components of variance for bone mineral density in 10- to 26-year-old females: a twin study. Am J Epidemiol 1998; 147: 17 29.

Received 12 March 1998; accepted 5 October 1998.