In recent years, there has been some controversy regarding the use of BMI as an indicator of good or poor health. BMI is obtained using measurable objectives, weight and height. In some persons, BMI may be an inadequate measure of health as it does not always accurately measure body composition. However, BMI can be used as a measure of excess weight, and it has been noted that children with an elevated BMI are at risk of obesity in adulthood and various metabolic diseases. Unlike adult BMI, children’s BMI is conveyed over wide percentile ranges for each age and sex because a child’s height and weight are not changing proportionally during the various growth stages13. Oh et al. studied various factors affecting bone age and found that children with increased waist circumference also had advanced skeletal age14. Another large-scale study reported that increased BMI (greater than or equal to 85th percentile) had a high association with increased body fat percentage rather than lean mass in the pediatric population compared to those with a lower BMI, and therefore, BMI could be a useful screening tool for obesity, which in turn can affect skeletal age15.
From a current consecutive cohort of orthopaedic patients at our facility we report a significantly greater discrepancy (average 13 months) between CA and SA in overweight and obese children when compared to non-overweight children of the same age and sex (no discrepancy) when evaluating hand radiographs with the Greulich and Pyle Atlas. Skeletal age discrepancy decreased as chronological age increased for overweight females, as noted in a previous study16.
In a retrospective study, Mack et al. examined skeletal maturation in orthodontic patients via analysis of cervical radiographs and dental age as a function of BMI percentile. Interestingly, they found both cervical vertebral stage and dental age to be more advanced (relative to chronological age) in subjects with increased BMI percentile9. In a similarly designed study, Akridge et al. reported a trend in accelerated SA in obese children as compared to age-matched cohorts. They found no overall statistically significant difference16. Of note, Akridge et al. utilized a single observer with Fishman’s hand-wrist analysis, contrary to our use of two blinded observers using the Greulich and Pyle method. There have been other case reports of advanced skeletal age in obese children, one in particular by Caffey in 1978, of a 2-year-old obese boy with a SA approximately three times his CA17.
A discrepancy between CA and SA has been seen among particular ethnicities. In a design similar to this current study, Ontell et al. found SA to be statistically significantly more advanced than CA in African American girls, Hispanic girls, Asian boys, and Hispanic boys. Of interest, this relationship was the most profound in late childhood to adolescence18. The authors however did not control for BMI, a potentially confounding variable in the study.
Another factor that may be taken into consideration when studying the variations between CA and SA is growth velocity. In a study done by De Simone et al, results showed that growth velocity varies between obese and non-obese children during the various pubertal stages. Obese children presented with increased SA earlier in life19.
A defined biochemical pathway establishing between obesity and advanced SA has not yet been identified, however a generalized pathway has been proposed. Klein et al. found leptin levels in obese children with advanced SA to be correlated. Further, this level of leptin was significantly higher in the obese children than in the non-obese children. The authors go on to suggest that leptin may be responsible for the earlier onset of puberty and advanced skeletal age in obese children. They propose that leptin may affect the hypothalamic-pituitary axis by increasing LH and FSH directly at the pituitary, or possibly influencing the axis upstream at the hypothalamus10. Both in vitro and in vivo studies indicated that leptin can also have direct effect on osteoblast proliferation and differentiation and inhibit osteoclast formation which can account for increased bone growth and growth plate maturation20–21. Elevated levels of insulin are also thought to be implicated with advanced SA in obese children. In a cohort of 74 overweight and obese children 4 to 13 years evaluated by Pinhas-Hamiel et al., hyperinsulinemia was associated with a 6.8-fold increased risk for advanced SA, independent of the degree of obesity22. Oh et al. also found that obese children with metabolic syndrome were more likely to have advanced skeletal age compared to obese children without metabolic syndrome further indicating hormonal implications on bone maturation14.
Despite present literature, the exact relationship between a child’s advanced SA and overweight or obese status is thoroughly complex and involves many factors that are still not completely understood.
In this study, both right- and left-hand radiographs were used. Considering that the Greulich and Pyle method of bone age evaluation is performed using left-handed radiographs, it is possible that our use of both right- and left-hand radiographs may have decreased the reliability of our findings. This is unlikely however, considering that differences between right- and left-hand maturation levels (using the Greulich and Pyle method) have been found to be insignificant in relation to the estimation of the maturational stages of the skeleton as a whole12.
Other methods of calculating bone age include the Tanner-Whitehouse methods, which was not utilized in this study. Previous studies have suggested that the Greulich and Pyle method of bone evaluation may show greater intra-observer variability than the Tanner-Whitehouse method but is a faster and preferred method for assessing bone age in pediatric radiologists and endocrinologists11, 23–24. Also, the study done by Chaimotre et al., which included a large multi-ethnic sample, found that the Greulich and Pyle method produces excellent correlation between skeletal age and chronological age and is still a reliable source for assessing bone age accurately25. Alshamrani and Offiah also demonstrated the accuracy of the GP method in a mutli-ethnic sample by comparing the method to BoneXpert and found no statistically significant difference between the two methods26. Future studies may attempt to utilize other forms of bone age evaluation and the effects of obesity or elevated BMI on skeletal age. The precision of our results using the GP method to compare relative differences in skeletal age between elevated and normal body mass index is evidence of the effects of elevated BMI on accelerated bone age.
Although our sample size was small, the ICC indicated that we had sufficient subject variability reflective of the average pediatric population and was still able to achieve excellent rater reliability. Comparing our results to similar studies that looked at other factors, such as waist circumference, we can conclude that being overweight or obese does impact skeletal maturation; therefore, the pediatric population cannot be treated for orthopaedic conditions based solely on chronological age.
Based on our cohort, children with weight above normal must be approached with the expectation of significantly earlier skeletal maturity when providing orthopaedic, or other age relative care. Although BMI used for this study was obtained at only one point in time, the increased SA observed is indicative that increased weight can have its implications on orthopaedic care in children because increased SA may not be a transient or acute issue but rather one that has been occurring over time due to physiological responses of chronic increased weight rather than acute weight changes. Expected fracture patterns and treatment options can vary based on age. Pogorelic et al described surgical indications for flexible intramedullary nails for proximal humerus fractures to be based on varying degrees of intolerable translation and angulation, and/or open physis depending on age. Generally, children < 10 years that do not meet criteria for surgery are treated conservatively given the bone has good remodeling potential27. Nguyen et al demonstrated differing scaphoid fracture patterns depending on age with distal scaphoid fractures occurring more often in a younger cohort compared to proximal fractures in an older cohort28. Casting duration for lower leg fractures is generally 8 weeks for patients 11–17 years and 6 weeks for patients 5–10 years29. Since many orthopaedic treatment recommendations are based on chronological age, one should carefully determine the skeletal age of overweight or obese patients before deciding definitive treatment.
While the conclusions of the study are highly translational for this population, there still exists limited orthopedic literature on skeletal age effects in children with an elevated BMI. Further studies should ideally include a multicenter cohort, with an emphasis on obtaining a racially diverse patient population to further explore ethnicity as an effect modifier on skeletal age in the children with elevated BMI.