Identifying the best metric to assess adiposity change in children: A comparison between changes in dual-energy x-ray absorptiometry (DEXA) and body mass index

Background: Accurately measuring childhood adiposity change has important clinical management and public health surveillance implications. Dual-energy X-ray absorptiometry (DEXA) is the method of choice to estimate fat mass, but is cost-prohibitive and proxies such as body mass index (BMI) are oftentimes used. However, limited data exist on the validity of BMI metrics to measure adiposity change among youth. Objective: Compare correlations between different BMI change metrics with different DEXA change metrics. Methods: Data were from QUALITY, a prospective cohort of children 8-10 years-old at recruitment from Québec, Canada (n=557). Height and weight were measured by trained nurses at baseline (2008) and follow-up (2010). Metrics of BMI change were: raw unit, adjusted for median BMI (BMI%), and age-sex-adjusted with the US CDC growth curves expressed as Z-scores (BMIz) or percentiles (BMI%ile). Metrics of DEXA change were raw unit (total fat mass: (TFM), percent (TFM%), height-adjusted (fat mass index; FMI) and internally age-sex-adjusted (TFMz). Spearman rank correlations between BMI change metrics and DEXA change metrics were derived. Results: Correlations ranged from 0.59 to 0.86. TFM change correlated most highly with BMI change (r=0.86), TFM% change with BMI% and BMIz changes (r=0.65), TFMz change with BMI% change (r=0.80), and FMI change with BMI and BMI% changes (r=0.83). Correlations with BMI%ile change were consistently the lowest for all DEXA change metrics. Conclusions: In 8-10 year-old children followed-up over 2 years, changes in BMI (raw unit) or in BMI% are the best proxies for changes in TFM or in FMI. BMI%ile performs less well.