Comparing anthropometric measures for metabolic risk prediction: a stratified analysis

Abstract Background Body mass index (BMI) is widely used to assess adiposity, but alternative measures such as waist circumference (WC) and waist-to-height ratio (WHtR) may better predict metabolic risk. Given the heterogeneity of the evidence depending on the populations studied, this study proposes to compare the predictive power of these measures for different risk factors in different subgroups of different sexes, age groups and educational levels. Methods Data collected as part of the Belgian Health Examination Survey in 2018 were used. It consists of a cross-sectional representative sample of 1,183 Belgian adults who were examined using standardised anthropometric measurements, blood pressure and blood biomarkers. Weighted logistic regression was used to estimate risk prediction in the general population and in specific subgroups. Area under the curve (AUC) was calculated and compared using DeLong's test for correlated ROC curves. Results For hypertension, no overall differences were found in the predictive power of the measures. However, BMI performed significantly better in women (AUC = 0.78, 95% CI: 0.75-0.82) and in adults aged 65 + (AUC = 0.74, 95% CI: 0.67-0.81). For diabetes and prediabetes, WHtR outperformed BMI and WC in the overall model (AUC = 0.80, 95% CI: 0.77-0.84), with significant differences mainly in women and adults aged 40-65 years. For dyslipidaemia, no overall difference was observed, but WHtR showed better performance in subgroups, especially in women (AUC = 0.84, 95% CI: 0.80-0.88) and in those aged 40-65 years. No differences were found between education groups for any metabolic risk. Conclusions WHtR is a stronger predictor of diabetes and performs better in key demographic subgroups, while BMI remains a useful indicator in women and older adults. These findings support the use of population-specific measures of adiposity to improve the identification of individuals at metabolic risk. Key messages • Anthropometric performance varies by sex and age; using uniform measures may overlook individuals at high risk. • Waist-to-height ratio outperforms body mass index in predicting diabetes, while body mass index remains a stronger predictor of hypertension in women and older adults.