Optimal Cutoff Values for Anthropometric Adiposity Measures of Sri Lankan Adult Women

Abstract

Anthropometric adiposity measures (AAMs) such as body mass index (BMI), waist circumference (WC), and waist-to-hip ratio (WHR) are used to evaluate obesity status. Country-specific cutoff values of AAMs would provide more accurate estimation of obesity prevalence. This cross-sectional study was designed to determine the optimal cutoff values for AAMs, BMI, WC, hip circumference (HC), and WHR, of Sri Lankan adult women. The study was conducted in Galle, Sri Lanka, with 350 healthy community-dwelling middle-aged women aged 30–60 years, divided into two groups (Group A, n = 175 and Group B, n = 175). Total body fat percentage (TBFP) (kg) was measured with DXA. Body weight (kg), height (m), and WC and HC (cm) were measured. BMI (kg/m2) and WHR were calculated. Optimal cutoff values were determined by area under curve (AUC) in Receiver-Operating Characteristic (ROC) curve analysis using TBFP as the criterion at the TBFP level of 33% and 35% using the women in Group A. Then, the prevalence of obesity was determined in Group B while comparing the prevalence based on the cutoff values recommended by the World Health Organization (WHO) for Asians and the newly developed cutoff values for Sri Lankan women. Optimal cutoff values of AAMs which correspond to TBFP 33% are BMI, 24.5 kg/m2; WC, 80 cm; HC, 95 cm; and WHR, 0.83. TBFP 35% corresponds to the optimal cutoff values of BMI, 25.0 kg/m2; WC, 85 cm; HC, 100 cm; and WHR, 0.83. Prevalence of obesity (number, %) according to the WHO and newly defined cutoff values that correspond to TBFP 33% and 35% were as follows: BMI = 83 (47.4%), 98 (56.0%), 83 (47.4%); WC = 106 (60.6%), 106 (60.6%), 72 (41.1%); and WHR = 140 (80.0%), 106 (60.6%), 106 (60.6%). The observed cutoff values of BMI and WC in this study were within the ranges of those described by the WHO for Asian populations which correspond to the 33% and 35% TBFP levels, respectively. However, the WHR cutoff value of WHO (Asians) is lower when compared to the newly determined value for Sri Lankan females while overestimating the prevalence. More studies are needed to confirm these values before clinical use.