DXA-Derived vs Standard Anthropometric Measures for Predicting Cardiometabolic Risk in Middle-Aged Australian Men and Women

Abstract

It is not clear if dual-energy X-ray absorptiometry (DXA) adiposity measures are superior to standard anthropometric measures for predicting cardiometabolic (CM) risk factors in a middle-aged general population. In the Busselton Healthy Ageing Study, we assessed a range of standard anthropometric and DXA-derived adiposity measures to predict metabolic syndrome (MetS) and CM risk factors in 4831 “baby boomers” aged 45–69 yr. Anthropometric and whole body DXA (GE Lunar Prodigy) measures were collected. Cross-sectional relationships of overall adiposity (BMI; DXA fat mass index, body fat %), central adiposity (waist circumference (WC); DXA trunk fat, android fat, abdominal visceral adipose tissue (VAT)) and ratio index (waist-to-hip ratio; DXA trunk/legs fat, android/gynoid ratio, VAT/total fat) with MetS and its components (as both continuous and binary outcomes) were evaluated using linear and logistic regression adjusting for age and lifestyle factors. Youden's Index was used to determine the optimal cut-points for predicting MetS. In linear regression analyses, central adiposity measures showed stronger associations with MetS score and CM risk factors than overall adiposity measures and fat ratio index, and DXA-VAT provided stronger associations than WC. Logistic regression models showed similar findings. For MetS diagnosis present in 35.9% of males and 24.4% of females, the highest odds ratio (95% CI) per SD change was observed for DXA-VAT (males: 5.02 [4.28, 5.88]; females: 3.91 [3.40, 4.49]), which remained significant (all p < 0.001) after further adjustment for BMI (males: 3.27 [2.65, 4.02]; females: 3.37 [2.79, 4.06]) or WC (males: 2.46 [1.95, 3.10]; females: 2.75 [2.21, 3.43]). The optimal DXA-VAT mass cut-point for predicting MetS was 1608 grams in males and 893 grams in females. DXA-VAT was superior to standard anthropometric and other DXA-derived adiposity measures for prediction of cardiometabolic risk factors, and has clinical utility for identifying middle-aged individuals at increased risk of MetS.