• new fracture predictors;
  • adolescents;
  • children;
  • bone density;
  • weight


  1. Top of page
  2. Abstract
  7. Acknowledgements

Predictors of childhood fractures have not been investigated previously. This study was undertaken to determine whether a previous history of forearm fracture, low bone mineral density (BMD; both areal bone mineral density [aBMD, g/cm2] and volumetric bone mineral apparent density [BMAD, g/cm3]), or anthropometry, influence fracture risk in young girls. At baseline, two cohorts of girls, aged 3–15 years, were evaluated: 100 had recently broken a forearm (group 1) and 100 were fracture free (group 2). Four years later we restudied 170 of these girls (82 from group l and 88 from group 2). We now report the relationships of previous fracture history, baseline BMD (measured by dual-energy X-ray absorptiometry), baseline weight, and height to risk of new fracture. More new fractures occurred in group l (37 fractures in 24 girls) than in group 2 (8 fractures in 7 girls; p = 0.0007). The independent predictors for occurrence of a new fracture at any skeletal site in a multivariate model adjusting for age, weight, total body aBMD, and fracture history were previous fracture (hazard ratio [HR], 3.28; 95% CI, 1.41-7.64); age (HR per l-year increase, 0.91; 95% CI, 0.84-0.99); total body aBMD (HR per l SD decrease, 1.92; 95% CI, 1.31-2.81); and body weight (HR per l SD increase, 1.49; 95% CI, 1.06-2.08). Girls with two risk factors together had substantially greater fracture risk: previous fracture and low spinal BMAD (HR, 9.4; 95% CI, 2.8-32.0), previous fracture and high body weight (HR, 10.2; 95% CI, 2.8-37.6), or previous fracture and low total body aBMD (HR, 13.0; 95% CI, 3.9-43.1). We conclude that previous forearm fracture, low total body aBMD, low spinal BMAD, and high body weight each increase risk of new fractures within 4 years in young girls. Interventions to reduce the risk of fractures, particularly forearm fractures, in girls warrant further study.


  1. Top of page
  2. Abstract
  7. Acknowledgements

SOME INDIVIDUALS break more bones than others. Adults with low bone density(1–3) or a history of fractures(4–8) are prone to further fractures. In older people previous fragility fractures predict new fractures independently of density.(9) Anthropometry also influences fracture risk, with tall and light adults being susceptible to fracture.(10,11) However, the effects of low bone mineral density (BMD), previous fracture history, or anthropometry on fracture risk during childhood and adolescence are unknown. This is because of a paucity of BMD measurements in children with fractures and because of the absence of longitudinal studies in this group. Nevertheless, Landin(12) observed that repeat fracturing was more common among girls and boys with previous fractures than in the general population. A quarter of all fractures in childhood occur in the distal forearm.(12) In New Zealand we estimate that 7500 of these fractures annually affect children aged 3–15 years, an incidence of approximately 1% of this age group. In 1994/1995 we conducted a case-control study of young girls with distal forearm fractures and showed they had lower areal BMD (aBMD, g/cm2) than fracture-free controls.(13) In addition, more cases than controls in this study had high adiposity. The present study was undertaken to determine whether a history of distal forearm fractures, baseline aBMD, calculated measures of volumetric bone mineral apparent density (BMAD, g/cm3), or simple anthropometry (height or weight) influenced subsequent fracture risk over a 4-year period.


  1. Top of page
  2. Abstract
  7. Acknowledgements


At baseline (1994/1995) we recruited 100 girls with recent distal forearm fractures (group l) and 100 age-matched friend controls who had never broken any bones (group 2) and measured their bone density.(13) The 200 girls were white and between 3 and 15 years old. After 4 years, we obtained a further fracture history by reinterviewing 170 girls and their parents (96.6% of the subjects who continued to live locally). Of the 30 girls who were not reinterviewed, 6 declined invitations and 24 had left Dunedin. Our local ethics committee approved the study.

Anthropometry and bone density measurements

At baseline children were weighed (electronic scale) and measured (Harpenden stadiometer) without shoes in light clothing. aBMD was measured in the forearm (at ultradistal and 1/3 distal radius), spine, and total body by dual-energy X-ray absorptiometry (Lunar DPX-L scanner, software packages 1.3z and l.5e; Lunar Corp., Madison, WI, U.S.A.) as described previously.(13) In our laboratory the CVs (%) based on l0 repeat scans of adults were 1.61 for ultradistal aBMD, 1.52 for 1/3 distal radius aBMD, 0.67 for L2-L4 aBMD, 0.70 for total body aBMD, 1.52 for total body bone mineral content (BMC), 2.52 for total body fat mass, and 1.11 for total body lean mass.

In addition, to take account of differences in bone size that contribute so much to gains in BMC during growth and that confound aBMD measurements in children,(14) we calculated BMAD values at different regional sites according to the following formulas: in the L2-L4 lumbar spine as BMC/Area3/2(14,15) and at the 1/3 distal radius as BMC/Area2.(16) We evaluated the ability of these volumetric measures to predict new fractures over 4 years in our l70 girls.

Tanner stage of pubertal development

This was assessed at baseline and at follow-up by self-assessment(17) Participants and their accompanying caregivers were shown pictures and written descriptions of Tanner stages of pubertal development(18) and asked to select the stage of breast development that most resembled their current appearance.

Fracture history

The 4-year follow-up fracture histories were recorded from 170 girls (82 in group 1 and 88 in group 2). Investigators confirmed reported fractures by checking radiology records. All fractures sustained during follow-up were treated at Dunedin hospital. Although one child broke her distal radius outside the district served by our hospital she did return for final treatment to Dunedin Hospital Fracture Clinic and had X-rays taken there.


The data were analyzed using Cox regression. The Anderson and Gill method was used to take into account the recurrent fractures in some individuals. Robust SEs were used to calculate 95% CIs. Values for aBMD, BMAD, and anthropometric measures were log-transformed and converted to Z scores using the coefficients derived by regressing values on age using the original 100 girls in our fracture-free cohort.(13)


  1. Top of page
  2. Abstract
  7. Acknowledgements

This study evaluates subsequent fracture risk in a high proportion of our original study population (85%). Table 1 shows the baseline characteristics of the l70 subjects who were followed up. Importantly, differences in baseline anthropometry and baseline aBMD between girls who could not participate and those who were followed up were not statistically significant. In accord with our reported findings in the total sample,(13) age-adjusted aBMD was reduced in group l in the total body, 1/3 distal radius, and lumbar spine. The ratios in Table 1 show that in group 1 aBMD values at these sites averaged only 97.1, 93.9, and 96.2%, respectively, of the values present in group 2. Reductions in aBMD were not explained by smaller bone size because in group l the age-adjusted BMAD values also were significantly lower, both for the spine and the 1/3 distal radius, volumetric measures at these sites averaging only 93.1% and 95.7%, respectively, of the values present in group 2. Although in the original 200 subjects more girls with previous fractures were overweight and overfat,(13) mean values for groups l and 2 at follow-up did not differ sufficiently to achieve statistical significance for body mass index (BMI), body fat, or lean mass in age-adjusted data. Twenty-nine girls of the 82 group 1 girls seen at follow-up had already reported a previous history of fracture before sustaining their baseline distal forearm fracture: 2 reporting 3 previous fractures, 11 reporting 2 previous fractures, and 16 reporting 1 previous fracture.

Table Table 1.. Descriptive Variables for Fractured and Fracture-Free Cohorts of Girls at Baseline
Thumbnail image of

During follow-up significantly more radiologically confirmed fractures occurred in girls from group 1 than from group 2 (Fig. 1). There were 45 fractures in 31 girls (29.3% of group l and 8.0% of group 2) during the 4 years of follow-up (Table 2). The distal radius was the most common site of fracture. The trauma associated with fracture was classified according to Landin(12) as slight (21 cases, 5 controls) or moderate (16 cases, 3 controls). As in our earlier report,(13) the majority of new fractures occurred on playground equipment, with minor falls incurred while running or during participation in team ball sports. There were no fractures from traffic accidents or severe trauma. Furthermore, at baseline the reported time spent in physical activity causing puffing or sweating (hours per week; mean ± SD) was similar in the 31 girls who sustained new fractures and the 139 girls who did not break any bones during follow-up (10.52 ± 4.31 and 10.43 ± 4.26, respectively). Of the 11 girls who suffered more than 1 new fracture during the 4 years of follow-up, 6 girls had already sustained 1 or more fractures before baseline. In spite of this, not 1 of these 11 girls had any diagnosed illness, nor was there any evidence of any of them participating in behavior placing them at particularly high risk of fracture. Like girls sustaining a single new fracture, their fractures occurred during sports and play of a general nature.

Table Table 2.. Sites of the Radiologically Confirmed Fractures Sustained During 680 Patient Years of Follow-Up
Thumbnail image of
thumbnail image

Figure Figure 1. Probability of not experiencing a fracture (A) at any site and (B) in the distal radius alone in groups l (previous forearm fractures) and 2 (no previous fractures).

Download figure to PowerPoint

Pubertal development(18) was similar in cases and controls.(13) At baseline in the l70 girls studied prospectively there were in groups l and 2, respectively, 40 and 48 girls in Tanner stage l, 20 and 18 girls in Tanner stage 2, 12 and 8 girls in Tanner stage 3, and 10 and 14 girls in Tanner stage 4/5. During follow-up, pubertal progression was also similar in groups l and 2 and was appropriate for age. Thus at follow-up in groups 1 and 2, respectively, there were 9 and 12 girls in Tanner stage 1, 12 and 14 girls in Tanner stage 2, 16 and 21 girls in Tanner stage 3, and 45 and 41 girls in Tanner stage 4/5 (χ2 = 1.23 and p = 0.74 for this pubertal distribution). Moreover, when the average age at which menarche had been achieved was calculated at follow-up for girls who had progressed to this stage, it did not differ in the two groups, being (mean ± SD) 12.8 ± 1.1 years for 54 girls in group l and 12.8 ± 0.9 years for 52 girls in group 2. During follow-up, new fractures occurred in groups l and 2, respectively, in l5 and 3 girls with baseline pubertal gradings of Tanner stage l, in 2 and 3 girls of Tanner stage 2, in 5 and 1 girl of Tanner stage 3, and in 2 and 0 girls of Tanner stage 4/5.

Table 3 shows the hazard ratios (HRs) for previous fracture history, age, total body aBMD, and body weight on risk of new fracture (both at all sites and at the distal radius alone) in all l70 girls. In the unadjusted analysis, previous fracture and low total body aBMD each increased risk significantly. In the multivariate model, the presence of at least one previous fracture at baseline, a reduction in total body aBMD, and an increase in body weight increased the risk of further fractures. Fracture risk decreased with increasing age. We also considered height, as a covariate, but because it did not add to the explanatory power of the model, it was not included.

Table Table 3.. Hazard Ratios for the Risk of New Fractures at Any Site and for New Fractures in the Distal Radius Only in 170 Girls Studied for 4 Years
Thumbnail image of

Next, we tested the influence of low aBMD values (per l SD decrease) at specific skeletal sites on risk of further fractures. None were significant predictors of fracture. The HRs adjusted for previous fracture, age, and weight for fractures at any site were ultradistal radius, 1.08 (95% CI, 0.75-1.56); 1/3 distal radius, 1.31(95% CI, 0.98-1.76); and lumbar spine L2-L4, 1.33 (95% CI, 0.97-1.82). HRs adjusted for previous fracture, age, and weight for new distal radius fractures only were ultradistal radius aBMD, 1.19 (95% CI, 0.75-1.87); aBMD 1/3 distal radius, 1.22 (95% CI, 0.80-1.84); and aBMD lumbar spine L2-L4 1.39 (95% CI, 0.91-2.11).

By contrast, we found some support for site-specific BMAD measures as predictors of new fracture. The univariate HRs for new fractures at any site (per l SD decrease) were 1.38 (95% CI, 1.09-1.75) for low spinal BMAD and 1.29 (95% CI, 0.95-1.75) for low 1/3 distal radius BMAD. Neither of the unadjusted BMAD measures were predictive of new fractures of the forearm alone, the HRs being 1.30 (95% CI, 0.94-1.79) for low spinal BMAD and 1.25 (95% CI, 0.81-1.92) for low 1/3 distal radius BMAD. Multivariate analyses showed that HR for fractures at any site, adjusted for previous fracture, and age and weight, were significant for spinal BMAD 1.34 (95% CI, 1.02-1.75) but not for 1/3 distal radius BMAD 1.14 (95% CI, 0.89-1.46). HRs for new distal radius fractures only, adjusted for previous fracture, age, and weight, were not significant: spinal BMAD, 1.28 (95% CI, 0.88-1.85), and 1/3 distal radius BMAD, 1.06 (95% CI, 0.76-1.50).

Finally, we examined the effects of having two risk factors in combination on fracture risk. We divided our sample into groups using cutoffs of spinal BMAD values below 1 SD to define low spinal BMAD, body weight above 1 SD to define high body weight, and total body aBMD below 1 SD to define low total body aBMD. Although the interaction effect was not significant, the HR for previous fracture and low spinal BMAD together was 9.4 (95% CI, 2.8-32.0). The HR for previous fracture and high body weight together was 10.2 (95% CI, 2.8-37.6). The HR for previous fracture and low total body aBMD together was 13.0 (95% CI, 3.9-43.1). None of the girls had high body weight and low spinal BMAD or high body weight and low total body aBMD.


  1. Top of page
  2. Abstract
  7. Acknowledgements

This is the first study to examine prospectively the influence of fracture history, BMD, and anthropometry as risk factors for further fracture in girls. We show that previous fractures, low total body aBMD, and high body weight each independently raises the risk of new fractures in growing children. Spinal BMAD was the only site-specific bone measure to predict new fractures and we showed that this increased risk occurred independently of previous fracture history, age, and body weight. Our study also shows that girls with previous fractures and low total body aBMD, or previous fractures and high body weight, or previous fractures and low spinal BMAD, have substantially higher fracture risks than girls with a single risk factor. Because BMD and body weight are both modifiable by diet and exercise,(19–25) our results strongly suggest that raising the mineral density of the bones and reducing high body weight each have the potential to reduce fractures. This seems especially important in girls with forearm fractures, a group our study identifies as being at particular risk of sustaining new fractures. In all children, achieving good BMD and maintaining healthy weight should help to prevent fractures.

After adjustment for low total body aBMD, age, and body weight, we show that girls with a recent fracture of the forearm at baseline had a 3.3-fold greater risk of sustaining new fractures during follow-up than girls who had never broken any bone at baseline. This result shows that the increased risk of fracture we have shown in girls with a previous history of fracture is not explained fully by low aBMD in the total skeleton or high weight and is in agreement with the view that adults who have suffered one or more fragility fractures at any site are at increased risk of having subsequent fractures for any given value of bone density.(26) Increased fragility may involve differences in balance or behavior (more falls and greater risk-taking), rapid bone turnover, or differences in bone shape or microarchitecture. Adults with forearm fractures have twice the risk of sustaining vertebral and hip fractures(5) and a greater risk of fractures at other sites.(6) Thus, current guidelines use a history of previous fracture in adult life to select high-risk individuals for osteoporosis intervention.(11,27)

In girls, we found that each reduction of 1 SD of total body aBMD, equivalent to a 6.4% change, approximately doubled the risk of new fractures at any site and in the distal radius/ulna alone. Similarly, each 1 SD decrease in total body aBMD approximately doubles fracture risk in older people.(2,3) Our findings are commensurate with the view that dense bone resists fracture more effectively than less dense bone.

In older people reductions in aBMD values at the sites where bones break generally predict fracture better than low aBMD measurements elsewhere.(28,29) However, this is not always the case.(30,31) We tested the ability of aBMD density measurements in the radius and spine to predict fractures because most fractures occurred in the forearm and people with forearm fractures often have low spinal aBMD.(13,32) However, in our children these regional aBMD measurements did not predict new fractures as well as total body aBMD. There continues to be some uncertainty regarding the relative merits of regional aBMD and total body aBMD in predicting fractures and it is of interest to note that Nordin et al.,(31) in a large prospective study of adults, also found total body aBMD to be superior to regional aBMD measurements, in discriminating between postmenopausal women with and without spinal fractures. Possibly, the lower variability of total body aBMD measurements compared with regional aBMD measurements gave total skeletal aBMD greater predictive power in our study. Low total body aBMD may in consequence provide a better global measure of skeletal fragility than reduced aBMD values at specific skeletal sites in children. Alternatively, the failure of regional aBMD values to predict fracture may be due to the rapid growth of our children. Although aBMD values track in childhood,(33) bones at different sites do not mature synchronously(34–36) and aBMD values at different sites may change rapidly during maturation.(37) Thus, site-specific aBMD values that were low at baseline could have improved during the 4-year follow-up of our study without significantly improving total body aBMD. For example, in some children radial aBMD values may have been low at baseline because the periosteal envelope had briefly expanded more than bone mineral accrual(15,35); subsequently, the low aBMD at this site could correct itself as catch-up of mineral accrual occurred and the forearm strengthened. Furthermore, changes in physical activity,(22) effects of hormones on different bone surfaces during maturation,(38,39) or altered nutrition(19,21,24) in individuals could have acted to modify bone mineral accrual and affected mineralization at some sites more than others.

During growth, increases in aBMD with advancing age are to a large extent due to increases in bone size, whereas BMAD measures are less influenced by increased size of the bones. In children three-dimensional estimates of BMD (BMAD) are considered to provide a better estimate of bone mineral accrual within an apparent volume of bone than the two-dimensional aBMD values.(14,40) Thus, in the present study low BMAD values might have been expected to predict new fractures more efficiently than low aBMD values. Therefore, it was interesting to note that low baseline spinal BMAD did predict new fractures over 4 years, whereas low baseline spinal aBMD values did not. Our findings support the value of BMAD measurements and suggest that girls with low spinal BMAD are more prone to sustain new fractures than those girls with higher spinal BMAD values.

Volumetric measures of BMD seem to be established in early childhood.(36) These variables track strongly, even in prepubertal children, and some researchers have suggested that low BMAD values may assist early identification of fragile skeletons.(33) However, volumetric BMAD values can alter transiently during growth in some individuals with changes in hormones or the environment.(33,34,36) These transient changes could have limited the ability of low BMAD values to predict new fractures in our study because a large proportion of our study population was experiencing rapid growth and pubertal change. In our study low radial BMAD values at baseline did not significantly predict new fractures, though the univariate HRs were suggestive of some adverse effect of low volumetric density increasing the risk of new fractures. Our study sample was relatively small and had insufficient power for low radial BMAD to predict fractures.

In adults, lighter individuals have a greater risk of fracture because their BMD tends to be low.(11) By contrast we found that high body weight increased the risk of new fracture, at any skeletal site (1.5-fold) and in the forearm (1.7-fold). We attribute elevated fracture risk among overweight children to biomechanical effects. During growth, children do not increase bone mineral and weight synchronously. Rather weight gain precedes gain in bone mineral. This dissociation between gains in body weight and bone mineral during growth may make bones especially fragile, particularly in obese children. We have shown previously that children with broken forearms are often overweight and overfat.(13,41) Although they may have appropriate total body aBMD for their age, this often is insufficient for their increased body weight. In fracture-free girls and boys we have shown recently that bone mass is significantly lower relative to body weight in overweight and obese children than in children of healthy body weight.(42) Accordingly, heavy children may fall with greater force and perhaps more awkwardly than lighter children. The association of increased fracture risk with increased body weight is of concern, given rising levels of obesity in children today.(43) Reduced levels of physical activity are considered to play an important role in the pathogenesis of this rising obesity.(44) It seems likely that insufficient exercise could contribute concomitantly to excessive gain of fat and to generation of too little bone for body weight among inactive children and adolescents. Load-bearing is acknowledged to be a critically important stimulant to osteogenesis in youth.(22,23,45–48)

Although some consider that tall children may be especially prone to fracture,(49) height did not influence fracture risk in our children, as occurs in adults.(10) Our observation that risk of new fractures decreases as girls get older also differs from the situation in adults, in which fracture risk rises with advancing age. It seems plausible that these differences are in part a consequence of changes in total bone mass and strength, which rise as children grow but decline as adults age.

Our study is novel and has a number of strengths because of its prospective design. These include careful evaluation of radiologically confirmed fractures over 4 years in two healthy groups of young girls chosen at baseline to be with or without any history of fracture. Although our study did not randomly select subjects, we recruited at baseline 84% of a consecutive series of girls treated for the same fracture type at the only hospital serving our district. Moreover, we followed up a high proportion of both cohorts 4 years later. We have comprehensive information concerning baseline bone density throughout the skeleton. We showed that pubertal development was similar in our two cohorts and appropriate for age. Thus, the low baseline measurements for aBMD and BMAD in group 1 were not explained by delayed pubertal maturation. A large number of new fractures occurred during the 4 years of follow-up and none were a consequence of major trauma.

However, our study has some limitations. We have no information concerning the role of family history, frequency of falling, or changing patterns in nutrition or physical activity. We cannot say whether risk-taking behavior or balance differed in girls who sustained new fractures. Our study sample may not have been large enough to discern the ability of site-specific aBMD values to predict new fractures given the wide intra- and intersubject variations in growing children. We acknowledge that we have sought predictors of new fracture in 170 girls who were selected initially for a case-control study designed to compare BMD and anthropometry in girls with forearm fractures and girls who had never broken any bones. In the general population, the only available figures, which were calculated in Sweden,(12) suggest the cumulative risk of sustaining at least one fracture from birth to l6 years of age in girls is 27%. In our follow-up sample, 82 girls (48% of our sample, which had an average age of 10 years at baseline) had at least one previous fracture at baseline, while 88 girls had no history of fracture. However, the nature of the sample should not affect the estimate of the HR or relative risk for a previous fracture. Bias may have been introduced when assessing the bone measurements, because they were related to previous fractures. It is important to consider adjustments for confounders but it is possible that including previous fracture in the multivariate model will to some extent overadjust confounders. Nevertheless, using a sample with a high proportion of girls having a history of fracture had the advantage of increasing the power of our study to identify important predictors of new fracture in young girls.

In conclusion, we have shown that young girls who present with a broken forearm have a substantially increased risk of sustaining new fractures during later childhood or adolescence. Thus, occurrence of a distal forearm fracture in a young girl is a clinically important warning signal of future fracture risk. In addition, low total body aBMD, high body weight, and low spinal BMAD values also independently increase the risk of new fractures. Efforts should be made to encourage all children, but particularly those with a history of fracture, to adopt a lifestyle that optimizes bone mineral gain and maintains healthy body weight with the aim of reducing their risk of future fracture.


  1. Top of page
  2. Abstract
  7. Acknowledgements

We are grateful to all participating children and their families and staff of Fracture Clinic for their help. This research was funded by the Health Research Council of New Zealand.


  1. Top of page
  2. Abstract
  7. Acknowledgements
  • 1
    Hui SL, Slemenda CW, Johnston CC 1989 Baseline measurement of bone mass predicts fracture in white women. Ann Intern Med 111:355361.
  • 2
    Marshall D, Johnell O, Wedel H 1996 Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ 312:12541259.
  • 3
    Schott AM, Cormier C, Hans D, Favier F, Hausherr E, Dargent-Molina P, Delmas PD, Ribot C, Sebert JL, Breart G, Meunier PJ 1998 How hip and whole-body bone mineral density predict hip fracture in elderly women: The EPIDOS prospective study. Osteoporos Int 8:247254.
  • 4
    Ross PK, Davis JW, Epstein RS, Wasnich RD 1991 Pre-existing fractures and bone mass predict vertebral fracture incidence in women. Ann Intern Med 114:919923.
  • 5
    Eastell R 1996 Forearm fracture. Bone 18:203S207S.
  • 6
    Cuddihy M-T, Gabriel SE, Crowson CS, O'Fallon WM, Melton LJ 1999 Forearm fractures as predictors of subsequent osteoporotic fractures. Osteoporos Int 9:469475.
  • 7
    Ettinger B, Black DM, Mitlak BH, Knickerbocker RK, Nickelsen T, Genant HK, Christiansen C, Delmas PD, Zanchetta JR, Stakkestad J, Gluer CC, Krueger K, Cohen FJ, Eckert S, Ensrud KE, Avioli LV, Lips P, Cummings SR 1999 Reduction of vertebral fracture risk in postmenopausal women with osteoporosis treated with raloxifene. JAMA 282:637645.
  • 8
    Nevitt MC, Ross PD, Palermo RL, Musliner T, Genant HK, Thomson DE 1999 Association of prevalent vertebral fractures, bone density, and alendronate treatment with incident vertebral fractures: Effect of number and spinal location of fractures. Bone 25:613619.
  • 9
    Wasnich R 1996 Vertebral fracture epidemiology. Bone 18:179S183S.
  • 10
    Hemenway D, Feskanich D, Colditz GA 1995 Body height and hip fracture: A cohort study of 90,000 women. Int J Epidemiol 24:783786.
  • 11
    National Osteoporosis Foundation 1998 Osteoporosis: Review of the evidence for prevention, diagnosis, and treatment and cost-effectiveness analysis. Osteoporos Int 4(Suppl):S1S88.
  • 12
    Landin LA 1983 Fracture patterns in children. Acta Orthop Scand Suppl 202:1109.
  • 13
    Goulding A, Cannan R, Williams SM, Gold EJ, Taylor RW, Lewis-Barned NJ 1998 Bone mineral density in girls with forearm fractures. J Bone Miner Res 13:143148.
  • 14
    Carter DR, Bouxsein ML, Marcus R 1992 New approaches for interpreting projected bone densitometry data. J Bone Miner Res 7:137145.
  • 15
    Bachrach LK, Hastie T, Wang M-C, Narasimhan B, Marcus R 1999 Bone mineral acquisition in healthy Asian, Hispanic, Black, and Caucasian youth: A longitudinal study. J Clin Endocrinol Metab 84:47024712.
  • 16
    Khosla S, Atkinson EJ, Riggs BL, Melton LJ 1996 Relationship between body composition and bone mass in women. J Bone Miner Res 11:857863.
  • 17
    Duke PM, Litt IF, Gross RT 1980 Adolescents' self-assessment of sexual maturation. Pediatrics 66:918920.
  • 18
    Tanner JM 1962 Growth at Adolescence, 2nd Ed. Blackwell Scientific Publications, Oxford, U.K.
  • 19
    Matkovic V 1996 Nutrition, genetics and skeletal development. J Am Coll Nutr 15:556569.
  • 20
    Bonjour J-P, Carrie A-L, 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 99:12871294.
  • 21
    Cadogan J, Eastell R, Jones N, Barker ME 1997 Milk intake and bone mineral acquisition in adolescent girls: Randomised, controlled intervention trial. BMJ 315:12551260.
  • 22
    Morris FL, Naughton GA, Gibbs JL, Carlson JS, Wark JD 1997 Prospective ten-month exercise intervention in premenarcheal girls: Positive effect on bone and lean mass. J Bone Miner Res 12:14531462.
  • 23
    Bradney M, Pearce G, Naughton G, Sullivan C, Bass S, Beck T, Carlson J, Seeman E 1998 Moderate exercise during growth in prepubertal boys: Changes in bone mass, size, volumetric density, and bone strength: A controlled prospective study. J Bone Miner Res 13:18141821.
  • 24
    Dibba B, Prentice A, Ceesay M, Stirling DM, Cole TJ, Poskitt EME 2000 Effect of calcium supplementation on bone mineral accretion in Gambian children accustomed to a low-calcium diet. Am J Clin Nutr 71:544549.
  • 25
    McKay HA, Petit MA, Schutz RW, Prior JC, Barr SI, Khan KM 2000 Augmented trochanteric bone mineral density after modified physical education classes: A randomized school-based exercise intervention study in prepubescent and early pubescent children. J Pediatr 136:156162.
  • 26
    Kanis JA, Delmas P, Burckhardt P, Cooper C, Torgerson D 1997 Guidelines for diagnosis and management of osteoporosis. Osteoporos Int 7:390406.
  • 27
    Kanis JA, McCloskey EV 1996 Evaluation of the risk of hip fracture. Bone 18:127S132S.
  • 28
    Eastell R, Wahner HW, O'Fallon WM, Amadio PC, Melton LJ, Riggs BL 1989 Unequal decrease in bone density of lumbar spine and ultradistal radius in Colles' and vertebral fracture syndromes. J Clin Invest 83:168174.
  • 29
    Cummings SR, Black DM, Nevitt MC, Browner W, Cauley J, Ensrud K, Genant HK, Palermo L, Scott J, Vogt TM 1993 Bone density at various sites for prediction of hip fractures. Lancet 341:7275.
  • 30
    Black DM, Cummings SR, Genant HK, Nevitt MC, Palermo L, Browner W 1992 Axial and appendicular bone density predict fractures in older women. J Bone Miner Res 7:633638.
  • 31
    Nordin BEC, Chatterton BE, Schultz GCG, Need AG, Horowitz M 1996 Regional bone mineral density interrelationships in normal and osteoporotic postmenopausal women. J Bone Miner Res 11:849856.
  • 32
    Peel NGA, Barrington NA, Smith TWD, Eastell R 1994 Distal forearm fracture as risk factor for vertebral osteoporosis. BMJ 308:15431544.
  • 33
    Ferrari S, Rizzoli R, Slosman D, Bonjour J-P 1998 Familial resemblance for bone mineral mass is expressed before puberty. J Clin Endocrinol Metab 83:358361.
  • 34
    Seeman E 1999 The structural basis of bone fragility in men. Bone 25:143147.
  • 35
    Magarey AM, Boulton TJC, Chatterton BE, Schultz C, Nordin BEC, Cockington RA 1999 Bone growth from 11 to 17 years: Relationship to growth, gender and changes with pubertal status including timing of menarche. Acta Paediatr 88:139146.
  • 36
    Bass S, Delmas PD, Pearce G, Hendrich E, Tabensky A, Seeman A 1999 The differing tempo of growth in bone size, mass, and density in girls is region-specific. J Clin Invest 104:795804.
  • 37
    Bonofiglio D, Maggiolini M, Marsico S, Giorno A, Catalano S, Aquila S, Ando S 1999 Critical years and stages of puberty for radial bone mass apposition during adolescence. Horm Metab Res 31:478782.
  • 38
    Seeman E 1998 Editorial: Growth in bone mass and size—are racial and gender differences in bone mineral density more apparent than real? J Clin Endocrinol Metab 83:14141419.
  • 39
    Libanati C, Baylink DJ, Lois-Wenzel E, Srinivasan N, Mohan S 1999 Studies on the potential mediators of skeletal changes occurring during puberty in girls. J Clin Endocrinol Metab 84:28072814.
  • 40
    Katzman DK, Bachrach LK, Carter DR, Marcus R 1991 Clinical and anthropometric correlates of bone mineral acquisition in healthy adolescent girls. J Clin Endocrinol Metab 73:13321339.
  • 41
    Goulding A, Taylor RW, Jones IE, McAuley K, Manning PJ 1998 Fractures: Are overweight children at increased risk? Conference Proc. N Z Dietetic Assoc 3:6870.
  • 42
    Goulding A, Taylor RW, Jones IE, McAuley KA, Manning PJ, Williams SM 2000 Overweight and obese children have low bone mass and area for their weight. Int J Obes 24:627632.
  • 43
    Rossner S 1998 Childhood obesity and adulthood consequences. Acta Paediatr 87:15.
  • 44
    Prentice AM, Jebb SA 1995 Obesity in Britain: Gluttony or sloth? BMJ 311:437439.
  • 45
    Kannus P, Haapasalo H, Sankelo M, Sievanen H, Pasanen M, Heinonen A, Oja P, Vuori I 1995 Effect of starting age of physical activity on bone mass in the dominant arm of tennis and squash players. Ann Intern Med 123:2731.
  • 46
    Khan NM, Bennell KL, Hopper JL, Flicker L, Nowson CA, Sherwin AJ, Crichton KJ, Harcourt PR, Wark JD 1998 Self-reported ballet classes undertaken at age 10–12 years and hip bone mineral density in later life. Osteoporos Int 8:165173.
  • 47
    Bass S, Pearce G, Bradney M, Hendrich E, Delmas PD, Harding A, Seeman E 1998 Exercise before puberty may confer residual benefits in bone density in adulthood: Studies in active prepubertal and retired female gymnasts. J Bone Miner Res 13:500507.
  • 48
    Bailey DA, McKay HA, Mirwald RL, Crocker PRE, Faulkner RA 1999 A six-year longitudinal study of the relationship of physical activity to bone mineral accrual in growing children: The University of Saskatchewan bone mineral accrual study. J Bone Miner Res 14:16721679.
  • 49
    Fournier P-E, Rizzoli R, Slosman D-O, Theintz G, Bonjour J-P 1997 Asynchrony between the rates of standing height gain and bone mass accumulation during puberty. Osteoporos Int 7:525532.