Clinical utility of novel anthropometric indices in identifying type 2 diabetes mellitus among South African adult females

Aim To assess the clinical utility of novel anthropometric indices and other traditional anthropometric indices in identifying the risk of type 2 diabetes mellitus (T2D) among South African adult females. Methods In the first South African National Health and Nutrition Examination Survey (SANHANES-1), traditional [body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR), and waist-to-height ratio (WHtR)] and novel [a-body shape index (ABSI), abdominal volume index (AVI), body adiposity index (BAI), body roundness index (BRI), conicity index (CI), and Clínica Universidad de Navarra-Body Adiposity Estimator (CUN-BAE)] anthropometric indices were assessed. T2D was diagnosed using glycated haemoglobin (HbA1c) ≥ 6.5% among participants without known T2D. Basic statistics and multiple regression analyses were explored the association between anthropometric indices and newly diagnosed T2D. Receiver operating characteristic (ROC) curve analysis was measured the predictive ability of both traditional and novel indices. Results Among 2 623 participants, 384 (14.6%) had newly diagnosed T2D. All anthropometric indices mean values were significantly higher among participants with T2D ( most p < 0.001 ). Higher mean values increased T2D odds e.g., in the model adjusted for age, employment, residence, and population group, odds ratio (OR) and 95% confidence interval (CI) for T2D with some of anthropometric indices were: 1.86 (1.60–2.15) for WC, 1.84 (1.59–2.13) for WHtR, 1.73 (1.51–1.99) for AVI, 1.71 (1.49–1.96) for BRI and 1.86 (1.57–2.20) for CUN-BAE. The top quartile for all indices had the highest T2D odds (p < 0.05). These outcomes were the highest for WC, AVI, and CUN-BAE and remained so even after removing the confounding effects of age, employment, population group, and residence. Based on the ROC analysis, none of the anthropometrical indices performed excellently (i.e., had an area under the curve [AUC] > 0.80). The WC, WHtR, AVI, BRI, and CUN-BAE, however, performed acceptably (AUCs 0.70–0.79), while also exhibiting corresponding cutoff values of 86.65 cm, 0.57, 15.52, 3.83, and 38.35, respectively. Conclusions The data shows that traditional and novel anthropometric indices similarly identifying newly diagnosed T2D among adult South African females. We recommend the continuing the use of traditional indices, as they are affordable and easy to use in our setting.