Examining the Relationship Between Obesity and Predicted Height in Children

Analysis indicates that current measures may underestimate predicted adult height in obese girls, and overestimate predicted adult height in obese boys.

Ovidiu A. Galescu, MD, a pediatric endocrinologist at the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), National Institutes of Health, introduced his presentation at ENDO 2015 by describing how pediatric obesity is associated with advancement of skeletal maturation and increase in height during childhood.

However, Galescu said that the accuracy of skeletal age-based adult height predictions using the classical Bayley-Pinneau tables when applied to obese children has not been established. The system has also shown limitations in validity in inflammatory bowel disease, congenital adrenal hyperplasia, precocious puberty, and in therapy with GnRH analogs. Galescu and his colleagues hypothesized that skeletal age-based height predictions would overestimate final adult height among obese as compared to non-obese children.

Therefore, a sample of non-obese (BMI percentile <95th percentile) and obese (BMI percentile ≥95th percentile) children age 5.5 to 15 years were prospectively enrolled for longitudinal investigation. Age- and sex-specific BMI percentiles were calculated according to CDC 2000 standards. Skeletal age was assigned using the Greulich and Pyle radiographic atlas and predicted adult height (PAH) was calculated according to the Bayley-Pinneau tables. Participants were followed to near-final adult height (>97% of full growth) as determined by age/skeletal age (>14 years for girls; >15.5 years for boys).

The primary analysis was by ANCOVA, with the difference between predicted adult height (PAH) and near-final adult height (FAH) (PAH-FAH) as the dependent variable and age, sex, race, obesity status and the sex*obesity status interaction as independent variables.

A total of 191 patients (mean age 9.3±1.8 years; 52.9% female and 47.1% male; 60.2% white, 37.7% black, 2.1% Asian) were enrolled and followed to FAH. At baseline, the mean bone age was 9.9±1.9 years; 50% were prepubertal and 34% were obese. There were no statistically significant differences between the non-obese and obese groups with respect to sex, race, or mid-parental target height. At baseline, the obese children were younger than the non-obese (8.7±1.4 vs. 9.5±1.8 years, p=0.002) but the bone age relative to chronological age was significantly greater among the obese group (+1.5±1.1 vs. +0.3±0.9 years, p<0.0001) and the height SD Z-score was significantly greater in the obese group (+1.45±0.97 vs. +0.29±0.99, p<0.0001).

Final adult height was documented at mean age 19.9 ±3.2 years. Using ANCOVA, PAH-FAH was significantly predicted by sex, age, and the interaction between sex and obesity status. Among obese girls, FAH was significantly underestimated by PAH (-2.3±1.1 cm, CI -4.5 to -0.08 cm, p=0.035). However, among obese boys, FAH appeared to be overestimated by PAH (+2.4±1.2 cm, CI -0.01 to +4.7 cm). Among non-obese children, there were no significant differences found between FAH and PAH.

In this large pediatric cohort, enriched for obesity and followed prospectively to adult height, obese children were initially taller than the non-obese and also had advanced skeletal age, consistent with adipose tissue aromatase action increasing circulating estrogens. Surprisingly in obese girls, despite their tendency to have earlier onset of puberty, FAH was underestimated, while in obese boys, who generally show a delayed entry into gonadarche, FAH showed a trend towards being overestimated by the Bayley-Pinneau method. These data suggest the Bayley-Pinneau tables for predicting adult height using skeletal age may need to be revised to take into account obesity.

Galescu speculated that the unexpected results in girls suggest that sex hormones might play a role, mentioning evidence from animal studies regarding the relationships between estrogens, estrogen receptor alpha (ER-α) and sexual dimorphism but said that there were other possible explanations.

Regarding the sex differences, the speaker was asked if there might be a timing issue in predicting the height with respect to onset of puberty. Galescu said their experimental model allowed for the difference in onset of puberty between females and males.