Predicting BMI in young adults from childhood data using two approaches to modelling adiposity rebound

Abstract

OBJECTIVE: To identify the age of adiposity rebound and the value of its associated BMI and examine their association with BMI at ages 18 and 21 y for males and females. DESIGN: A longitudinal study of a large cohort of people born in Dunedin, New Zealand between 1972-1973. SUBJECTS: Four hundred and seventy-four males and 448 females aged between birth and 21 y. MEASUREMENTS: BMI was derived from measurements of weight and height made when the participants were born and at intervals from age 3-21 y. RESULTS: When a random coefficients model was fitted to the data for those who had five or more measures of BMI between age 3 and age 18 y, adiposity rebound occurred at 6.0 y of age for boys and 5.6 y for girls. The values of BMI associated with these were 15.7 kg/m2 for boys and 15.5 kg/m2 for girls. The correlations between age at adiposity rebound and BMI at ages 18 and 21 y were between -0.72 and -0.65 for boys and -0.59 and -0.47 for girls. These were higher than those derived from fitting individual curves or from deriving the adiposity rebound from data collected up to age 11 y. The correlation between BMI at age 7 y and BMI at ages 18 and 21 y were 0.70 and 0.61 for boys and 0.56 and 0.52 for girls. The correlations between measures of skeletal maturity at age 7 y and adiposity rebound were statistically significant for boys but not for girls. CONCLUSIONS: BMI in early adulthood was associated with both age of adiposity rebound and BMI at that age. As the correlations between BMI at age 7 y and BMI at ages 18 and 21 y were similar in magnitude, BMI at age 7 y may be a more practical way of predicting BMI in early adulthood.