Longitudinal changes in BMI z-scores among 45 414 2–4-year olds with severe obesity

Abstract

Background: BMI z-scores (BMIz) based on the Centers for Disease Control and Prevention (CDC) growth charts among children do not accurately characterise BMI levels among children with very high BMIs. These limitations may be particularly relevant in longitudinal and intervention studies, as the large changes in the L (normality) and S (dispersion) parameters with age can influence BMIz.Aim: To compare longitudinal changes in BMIz with BMI expressed as a percentage of the 95th percentile (%BMIp95) and a modified z-score calculated as log(BMI/M)/S.Subjects and methods: A total of 45 414 2–4-year-olds with severe obesity (%BMIp95 ≥ 120).Results: Changes in very high BMIz levels differed from the other metrics. Among severely obese 2-year-old girls, for example, the mean BMIz decreased by 0.6 SD between examinations, but there were only small changes in BMIp95 and modified BMIz. Some 2-year-old girls had BMIz decreases of >1 SD, even though they had large increases in BMI, %BMIp95 and modified BMIz.Conclusions: Among children with severe obesity, BMIz changes may be due to differences in the transformations used to estimate levels of BMIz rather than to changes in body size. The BMIs of these children could be expressed relative to the 95th percentile or as modified z-scores.