Sex and race/ethnicity specific reference predictive equations for abdominal adiposity indices using anthropometry in US adults.

BACKGROUND AND AIMS: Abdominal adiposity indices have stronger associations with cardiometabolic risk factors compared to anthropometric measures but are rarely used in large scale studies due to the cost and efficiency. The aim of this study is to establish sex and race/ethnicity specific reference equations using anthropometric measures. METHODS AND RESULTS: A secondary data analysis (n = 6589) of healthy adults was conducted using data from National Health and Nutrition Examination Survey 2011-2018. Variables included in the analyses were anthropometric measures (height; weight; waist circumference, WC) and abdominal adiposity indices (android percent fat; android to gynoid ratio, A/G ratio; visceral adipose tissue area, VATA; visceral to subcutaneous adipose area ratio, VSR). Multivariable prediction models were developed using quantile regression. Bland-Altman was used for external validation of prediction models. Reference equations to estimate android percent fat, A/G ratio, VATA and VSR from anthropometric measurements were developed using a randomly selected subsample of 4613. These reference equations for four abdominal adiposity indices were then cross-validated in the remaining subsample of 1976. The measured and predicted android percent fat, A/G ratio, VATA and VSR were not statistically different (p > 0.05) except for the A/G ratio in Asian males and VSR in White females. The results of Bland-Altman further revealed that ≥93% of predicted abdominal adiposity indices fell within the limits of agreement (±1.96 standard deviation). CONCLUSION: The sex and race/ethnicity specific reference equations for abdominal adiposity indices established using anthropometrics in the present study have strong predictive ability in US healthy adults.