Monitoring children and adolescents with severe obesity: body mass index (BMI), BMI z-score or percentage above the International Obesity Task Force overweight cut-off?

Abstract

Body mass index (BMI) metrics are widely used as a proxy for adiposity in children with severe obesity. The BMI expressed as the percentage of a cut-off percentile for overweight or obesity has been proposed as a better alternative than BMI z-scores when monitoring children and adolescents with severe obesity. Annual changes in BMI, BMI z-score and the percentage above the International Obesity Task Force overweight cut-off (%IOTF-25) were compared with dual-energy X-ray absorptiometry (DXA) derived body fat (%BF-DXA) in 59 children and adolescents with severe obesity. The change in %BF-DXA was correlated with the change in %IOTF-25 (r=0.68) and BMI (r=0.70), and somewhat less with the BMI z-score (r=0.57). Cohen's Kappa statistic to detect an increase or decrease in %BF-DXA was fair for %IOTF-25 (κ=0.25; p=0.04) and BMI (κ=0.33; p=0.01), but not for the BMI z-score (κ=0.08; p=0.5). The change in BMI was positively biased due to a natural increase with age. Changes in the BMI metrics included in the study are associated differently with changes in %BF-DXA. The BMI z-score is widely used to monitor changes in adiposity in children and adolescents with severe obesity, but the %IOTF-25 might be a better alternative.