¹Department of Medicine, Otago University, Dunedin, New Zealand

²Department of Human Nutrition, Otago University, Dunedin, New Zealand

³Department of Preventive and Social Medicine, Otago University, Dunedin, New Zealand

Correspondence to: Dr A Goulding, Department of Medicine, University of Otago, PO Box 913, Dunedin, New Zealand. ailsa.goulding@stonebow.otago.ac.nz

OBJECTIVES: To determine whether girls and boys categorized from body mass index (BMI) values as overweight or obese for their age have lower bone mineral content (BMC) or lower bone area in relation to total body weight than children of normal adiposity.

DESIGN: Cross-sectional study in a university bone research unit.

SUBJECTS: Two hundred girls and 136 boys aged 3-19 y recruited from the general population by advertisement.

MEASUREMENTS: Total body BMC (g) and bone area (cm²) measured by dual energy X-ray absorptiometry (DXA) in relation to body weight (kg), lean tissue mass (kg) and fat mass (kg) in boys and girls of three different BMI percentile groupings: normal weight (BMI<85th percentile); overweight (85 to 94th BMI percentile); obese (95th BMI percentile).

RESULTS: Obese children had higher BMC, bone area, and fat mass for chronological age than those of normal body weight (P<0.001). In spite of this the observed values for age-adjusted total body BMC and bone area relative to body weight were each lower than predicted values, in both overweight and obese children (2.5-10.1% less, P<0.05) than in children of lower adiposity.

CONCLUSION: In overweight and obese boys and girls there is a mismatch between body weight and bone development during growth: their bone mass and bone area are low for their body weight.

International Journal of Obesity (2000) 24, 627-632

obesity; bone development; body composition; DXA; future health risks

Introduction

The increasing prevalence of obesity in childhood and adolescence observed in many countries is currently a major public health concern. This rising adiposity may be due in part to decreased physical activity.^1,2,3 Since physical activity is strongly osteogenic during growth^4,5 diminished participation in weight-bearing exercise by overweight children could affect bone development adversely.

Although there is a general perception that overweight children are skeletally advanced,⁶ we recently observed that a high proportion of children with distal forearm fractures were overweight. In spite of this, our fracture cases had lower bone mass and bone area density throughout their skeletons than fracture-free controls, and we suggested that a low bone mineral content relative to body weight might make overweight children vulnerable to fragility fractures.^7,8 Others report low bone density in the spine and radius of overweight children^9,10,11 and orthopaedic problems such as slipped capital epiphyses of the femora,¹² scoliosis and tibia varu¹³ are recognized health risks of childhood obesity.

The present study was undertaken to test the hypothesis that overweight and obese fracture-free children and adolescents have low bone mass or low bone area for their body weight. We chose to examine relationships of bone mineral content and bone area to total body weight in the whole body rather than at regional bone sites because these represent the whole skeleton and are therefore less subject to variability in site-specific maturation rates. No previous studies appear to have systematically evaluated the relationships of total bone mass or bone area to total body weight in fracture-free children of these differing grades of adiposity

Methods

New Zealand girls (n=200) and boys (n=136) aged 3-19 y were recruited from the general population by advertisement for studies of bone health, anthropometry and nutrition. All were of Caucasian ethnicity. The protocol was approved by the Ethics Committee of our hospital and informed written consent was obtained from each participant and a parent or guardian. No children taking medication affecting either body weight or bone growth were enrolled. No child had broken any bone at any stage of life.

Subjects came to the study centre where they answered a short health questionnaire. Parents/caregivers helped younger children answer the questions. Pubertal development (Tanner stage) was assessed as described previously⁷ and was confirmed in the boys after a physical examination by an endocrinologist. No measurements of bone age were obtained.

Children were weighed and measured without shoes in light clothing and bone and body composition was measured by dual energy X-ray absorptiometry (DXA). Height was measured with a Harpenden stadiometer and weight with an electronic scale. Body mass index (BMI) was determined as body weight divided by height squared (kg/m²). To measure total body bone mass (total bone mineral content, BMC, (g)), total bone area (cm²) and body composition (lean mass (kg), fat mass (kg), and fat percentage) a rectilinear total body scan was performed on each child using the fast or medium mode as appropriate for body size, with a Lunar DPX-L scanner (Lunar Corporation, Wisconsin, USA).⁷

Subjects were allocated to three groups according to their calculated BMI centile values for age, using standard BMI reference data used in our hospital.¹⁴ Group 1 were controls (BMI<85th centile), group 2 were overweight (BMI>85th centile and <95th centile) and group 3 were obese (BMI95 centile). Sexes were grouped separately.

Descriptive results are presented as means and standard deviations. Stagewise regression was used to analyse the data.¹⁵ In the first stage log-transformed BMC and total body area were regressed on age and age² and total body weight, which had also been log-transformed. This provided equations from which the expected value of the dependent variable was calculated. The residual or the difference between the observed value and the predicted value was also calculated for each observation. In the second stage of the regression analysis the residuals were used as the dependent variables and regressed on two dummy variables so that comparisons could be made between those with a BMI between 85-94th centile and those with a BMI95th centile with those with BMI<85th centile, the reference group. The analysis was repeated using total body lean mass (Table 4) and then total body fat mass (Table 5) instead of weight. It was necessary to use age² in the model to account for decreasing rates of bone growth among the older children. As the data were log transformed before analysis, the comparison between the observed values, and those predicted from the regression analysis was determined from ratios with 95% confidence intervals.

Results

Table 1 shows the main characteristics of our study population. Of the girls, 165 subjects (82.5%) had BMI values placing them in group 1, 22 (11%) in group 2, and 13 (6.5%) in Group 3. Of the boys, 111 subjects (81.6%) had BMI values placing them in group 1, 17 (12.5%) in group 2, and 8 (5.9%) in group 3. Thus the distribution of obese and overweight children in our study did not differ from that expected in a normal population. Pubertal status was appropriate for age in all subjects; all girls over 14 y had reached menarche and all boys over 14 y had Tanner Stage grades of 3 or more.

In the girls, standard deviation scores for height-for-age¹⁶ (means (s.d.)) were similar in our three groups: group 1, 0.70 (1.13); group 2, 0.51 (1.24); group 3, 0.49 (1.67), showing that increasing adiposity was not associated with increased height. By contrast, obese boys (group 3) were significantly taller (P<0.05) than control boys (group 1): standard deviation scores for height-for-age being: group 1, 0.70 (1.08); group 2, 1.17 (1.13); group 3, 1.59 (1.84). As expected, in both sexes the standard deviation scores for weight-for-age (girls 0.34 (0.89); 1.24 (0.81); and 3.12 (1.62), boys 0.70 (1.0); 2.53 (1.03); and 4.79 (2.15)) were significantly higher in both overweight and obese subjects vs those of group 1 (P<0.001).

Variables in relation to chronological age

In obese girls, but not in overweight girls, BMC, bone area and lean mass were each increased relative to chronological age (P<0.005). In both sexes fat mass was increased relative to chronological age in both overweight and obese subjects (P<0.001). In the boys, BMC and bone area values for chronological age were increased in both overweight (P<0.003) and in obese groups (P<0.001) but lean mass relative to chronological age was increased only in the overweight boys (P<0.005). Obese boys (group 3) and boys of normal weight (group 1) had similar lean mass for chronological age.

BMC in relation to body weight in children of different adiposity (age-adjusted data)

Figure 1 shows the relationship of bone mineral content to total body weight according to gender in the whole study sample. Table 2 shows both the observed BMC and BMC predicted from weight after adjusting for age and age-squared for the three BMI groups. The ratio of the observed to predicted values was significantly less in overweight and obese groups for both girls and boys.

Bone area in relation to body weight in children of different adiposity

In our age-adjusted data the relationship of area with BMC was very close in the three BMI groups in boys and girls with observed values ranging from 99.6-101.6% of predicted values (data not shown). In the age-adjusted data, bone area relative to body weight followed a similar pattern as BMC relative to body weight. Predicted area values for weight were lower than observed values in groups 2 and 3 in both sexes, suggesting inadequate enlargement of the bones to fully compensate for increased body weight in these groups (Table 3).

BMC in relation to lean tissue mass in children of differing adiposity

After age-adjustment, the obese groups of both sexes, but not the overweight groups, had higher BMC in relation to given lean tissue mass, suggesting adaptive increases in BMC had occurred in children with high adiposity but not in those with moderate adiposity (Table 4).

BMC in relation to fat mass in children of differing adiposity

In the age-adjusted data, only the boys (both overweight and obese subgroups) had higher observed than predicted values for BMC in relation to fat mass than males of normal weight (Table 5).

Discussion

Using a sample from the general population of fracture-free children, we demonstrate for the first time that important discrepancies between bone area and bone mineral content relative to body weight occur in overweight and obese children during growth. Our overweight and obese children had lower bone area and bone mass for their body weight than subjects with body weights in the healthy range. We consider that this mismatch between high body weight and bone development during growth may place considerable strains on the bones and joints of overweight and obese children. We obtained some evidence of adaptative increases in BMC relative to both lean mass and to fat mass in our youngsters with high adiposity, suggesting some skeletal compensations to elevate bone mass had taken place. Nevertheless, observed BMC values relative to weight remained substantially lower than predicted BMC values in all our overweight groups, these reductions averaging 8-10% in our obese groups of children.

All techniques used to estimate body composition have drawbacks, and concerns have been expressed regarding the ability of DXA scanners to discern BMC and bone area with equal accuracy at all depths of overlying tissue.^17,18,19 These problems are attributed to small variations in edge detection. However, given that BMC has been shown to increase slightly when the amount of overlying tissue has been deliberately increased artificially using lard,¹⁷ and that BMC generally decreases during weight loss^18,20 we are confident that the observed BMC values for our overweight and obese children are if anything slightly overestimated, rather than underestimated. Furthermore, the scanner we used is a pencil beam DXA and pencil beam scanners are superior to fanbeam DXA scanners for determining body composition. DXA scanners were originally designed to measure BMC and they do this extremely accurately. The overweight children we measured have weights which lie well within the normal range for adults. Using the same brand of scanner (Lunar) as was used in the present study, Van Loan et al ¹⁹ showed that no significant changes occurred in either BMC or bone area in fourteen overweight women who lost an average of 15.6 kg in fifteen weeks.

Our findings that both BMC and bone area are low relative to body weight in overweight children are perhaps surprising since there is good evidence that obese adolescents undergo early puberty and have advanced skeletal maturation (as shown by bone age) relative to chronological age, than leaner children.²¹ However, previous investigators do not appear to have examined the quantitative relationships between bone mass or bone area to body weight in overweight/obese groups of children vs children of normal body weight. In support of our findings that total body bone area and BMC were low for body weight in groups 2 and 3, are earlier reports of decreased spinal density for body weight in obese children.^10,11 A recent preliminary report, based on measurements in 32 obese children, also found that increased levels of obesity in youth were associated with increased fat mass but not increased bone mineral content.²²

Our study was cross-sectional in design so we cannot say whether the mismatch in skeletal development for body weight which we observed represents a temporary lag between bone growth and bone mineral accrual relative to body weight, which will later be corrected by compensatory increased skeletal growth, or whether there will be a lasting deficit in bone mass for body weight into adult life. Prospective studies are needed to show whether the mismatches in BMC and bone area relative to weight which we report are transient or lasting. Since overweight adults have higher bone mass than lighter individuals²³ it seems probable adaptation will occur over time. On the other hand inactive obese adults can have low bone density.⁵ It may simply be that in childhood and adolescence the pace of gain in adipose tissue outstrips possible compensatory increases in bone size or mineral accrual. Alternatively, our overweight subjects may be insufficiently active to optimize bone gain. Our results suggest that many overweight adolescents fail to adapt bone development adequately to cope with their excess weight. This may increase their likelihood of sustaining fractures when they fall as we have suggested elsewhere.⁸ The weight/bone mass imbalance will place high strains on growing bones and joints and may induce lasting joint damage which could contribute to the pathogenesis of osteoarthritis in adult life. Obesity in youth is already an established risk factor for adult osteoarthritis.²⁴

We recognise that at present there is no consensus regarding the definition of overweight or obesity in children. We chose to group subjects as overweight or obese by using the 85th and 95th BMI centile values for age because BMI is readily measured very accurately without the need for sophisticated technology. Furthermore, there is good agreement between BMI and adiposity measured by DXA so that children above these centiles tend to be overfat.^25,26 These cutoffs have been employed to examine the prevalence of overweight in nationally representative cross-sectional surveys.²⁷ They have also been recommended as useful in selecting youngsters for interventions to reduce adiposity.²⁸

Our study has some limitations. The participants were volunteers and so may not be a representative sample of the current pediatric population in New Zealand. However every fracture-free child aged 3-19 y enrolled in Dunedin bone studies was included in our present analysis and since none were excluded on BMI classification we had no bias for degree of adiposity. Reduced physical activity is considered an important contributor to the genesis of obesity and adversely influences bone mineral accrual²⁹ whereas exercise has dual benefits of augmenting BMC and reducing body fat. Unfortunately we do not have any information on the physical activity of our subjects, the duration of their obesity or any information on hormone levels which undoubtedly influence bone growth and mineral accrual.^30,31 The strengths of our study include the use of large samples of both genders, a wide age range, precise measurements of body composition supplied by DXA and the use of well-accepted BMI classifications of overweight and obesity. Moreover, our predictions were based on regressions from the total samples of each gender, which provide conservative estimates of expected bone mineral content and area in groups of different adiposity.

Conclusion

In conclusion our study confirms the hypothesis that overweight and obese children have lower bone area and bone mass relative to body weight than their leaner peers. Further longitudinal investigations will be needed to define the reasons for this and to determine whether the decrease is transient or persistent. Because the mismatch between bone mass and body weight which we have documented during growth could increase the current and/or future vulnerability of these individuals to bone fractures and to later osteoarthritis, studies examining the influence of childhood obesity upon development of fractures and arthritis and should also be undertaken. These topics have important public health implications.

Acknowledgements

This study was supported by the Health Research Council of New Zealand. We thank the participants and their parents for their willing co-operation.

1 Prentice A, Jebb S. Obesity in Britain: gluttony or sloth? Br Med J 1995; 311: 437-439.

2 Gortmaker S, Must A, Sobol A, Peterson K, Colditz G, Dietz W. Television viewing as a cause of increasing obesity among children in the US, 1986-90. Arch Pediatr Adolesc Med 1996; 150: 356-362. MEDLINE

3 Rossner S. Childhood obesity and adulthood consequences. Acta Paediatrica 1998; 87: 1-5. MEDLINE

4 Slemenda CW, Miller JZ, Hui SL, Reister TK, Johnston CC. Role of physical activity in the development of skeletal mass in children. J Bone Min Res 1991; 11: (6) 1227-1233.

5 Frost H. Obesity, and bone strength and 'mass': a tutorial based on insights from a new paradigm. Bone 1997; 21: 211-214. MEDLINE

6 Klein KO, Larmore KA, de Lancey E, Brown JM, Considine RV, Hassink SG. Effect of obesity on estradiol level, and its relationship to leptin, bone maturation, and bone mineral density in children. J Clin Endocrinol Metab 1998; 83: (10) 3469-3475. MEDLINE

7 Goulding A, Cannan R, Williams S, Gold E, Taylor R, Lewis-Barned N. Bone mineral density in girls with forearm fractures. J Bone Min Res 1998; 13: 143-148.

8 Goulding A, Taylor R, Jones I, McAuley K, Manning P. Fractures: are overweight children at increased risk? Conference Proc NZ Dietetic Assoc 1998; 3: 68-70.

9 Zamboni G, Soffiati M, Giavarina D, Tato L. Mineral metabolism in obese children. Acta Paediat Scandinav 1988; 77: 741-746.

10 McCormick DP, Ponder SW, Fawcett D, Palmer JL. Spinal bone mineral density in 335 normal and obese children and adolescents: evidence for ethnic and sex differences. J Bone Min Res 1991; 6: 507-513.

11 De Schepper J, Van den Broeck M, Jonckheer MH. Study of lumbar spine bone mineral density in obese children. Acta Paediatr 1995; 84: 313-315. MEDLINE

12 Loder R, Aronson D, Greenfield M. The epidemiology of bilateral slipped capital femoral epiphysis. J Bone Joint Surg 1993; 75A: 1141-1147.

13 Davids J, Huskamp M, Bagley A. A dynamic biomechanical analysis of the etiology of adolescent tibia vara. J Pediat Orthop 1996; 16: 461-468.

14 Must A, Dallal G, Dietz W. Reference data for obesity: 85th and 95th percentiles of body mass index (Wt/ht²)¾a correction. Am J Clin Nutr 1991; 54: 773.

15 Draper N, Smith H. Applied regression analysis, Second edition. Wiley: New York, 1981. ,

16 Tanner JM, Whitehouse RH. Clinical longitudinal standards for height, weight, height velocity, weight velocity, and stages of puberty. Arch Dis Child 1976; 51: 170-179. MEDLINE

17 Svendsen O, Haarbo J, Christiansen C. Accuracy of measurements of body composition by dual-energy X-ray absorptiometry in vivo. Am J Clin Nutr 1993; 57: 605-608. MEDLINE

18 Jensen L, Quaade F, Sorensen O. Bone loss accompanying voluntary weight loss in obese humans. J Bone Min Res 1994; 9: 459-463.

19 Van Loan M, Johnson H, Barbieri T. Effect of weight loss on bone mineral content and bone mineral density in obese women. Am J Clin Nutr 1998; 67: 734-738. MEDLINE

20 Pritchard J, Nowson C, Wark J. Bone loss accompanying diet-induced or exercise-induced weight loss: a randomised controlled study. Int J Obes 1996; 20: 513-520.

21 De Simone M, Farello G, Palumbo M, Gentile T, Ciuffreda M, Olioso P, Cinque M, De Matheis F. Growth charts, growth velocity and bone development in childhood obesity. Int J Obes 1995; 19: 851-857.

22 Southern M, Loftin M, Suskind R, Udall J, Wilson J, Hensel L, Hargis J, Blecker U. The impact of mild, moderate and severe obesity on upper and lower bone mineral content, lean and fat body mass. Int J Obes 1998; 22: (Suppl 3) P383.

23 Reid I, Plank L, Evans M. Fat mass is an important determinant of whole body bone density in premenopausal women but not in men. J Clin Endocrinol Metab 1992; 75: 779-782. MEDLINE

24 Cicuttini F, Spector T. Obesity, arthritis, and gout. Handbook of Obesity. Marcel Dekker: New York, 1998, pp, 741-752.

25 Goulding A, Gold E, Cannan R, Taylor R, Williams S, Lewis-Barned N. DEXA supports the use of BMI as a measure of fatness in young girls. Int J Obes 1996; 20: 1014-1021.

26 Pietrobelli A, Faith M, Allison D, Gallagher D, Chiumello G, Heymsfield S. Body mass index as a measure of adiposity among children and adolescents: A validation study. J Ped 1998; 132: 204-210.

27 Troiano R, Flegal K, Kuczmarski R, Campbell S, Johnson C. Overweight prevalence and trends for children and adolescents. Arch Pediatr Adolesc Med 1995; 149: 1085-1091. MEDLINE

28 Dietz W. How to tackle the problem early? The role of education in the prevention of obesity. Int J Obes 1999; 23: (Suppl 4) S7-S9.

29 Morris F, Naughton G, Gibbs J, Carlson J, Wark J. Prospective ten-month exercise intervention in premenarcheal girls: positive effect on bone and lean mass. J Bone Min Res 1997; 12: 1453-1462.

30 Rizzoli R, Bonjour J-P. Hormones and bones. Lancet 1997; 349: (Suppl 1) 20-23. MEDLINE

31 Ferretti J, Capozza R, Cointry G, Garcia S, Plotkin H, Alvarez Filgueira M, Zanchetta J. Gender-related differences in the relationship between densitometric values of whole-body bone mineral content and lean body mass in humans between 2 and 87 years. Bone 1998; 22: 683-690. MEDLINE

Figure 1 Total body bone mineral content in relation to body weight in (a) girls (n=200) upper graph (log(BMC)= 3.02+1.15*log(weight), SEE=0.13, R²=0.94) and (b) boys (n=136) lower graph (log(BMC)=2.95+1.18*log(weight), SEE=0.13, R²=0.95). Group 1=BMI for age percentile<85; Group 2=BMI 85-94; Group 3=BMI95.

Table 4 Total body BMC (g) in relation to total body lean tissue mass (kg)^a

Table 1 Characteristics of our two study populations

Table 2 Total body BMC (g) in relation to total body weight^a

Table 3 Total body bone area (cm²) in relation to total body weight^a

Table 5 Total body BMC (g) in relation to total fat mass (kg)^a