Abstract
BACKGROUND: People with the metabolically obese normal weight (MONW) phenotype have been confirmed to significantly increase the risk of unfavorable health consequences. This study aimed to investigate the relationships between traditional and novel anthropometric indices with the MONW phenotype and compare the predictive ability of different anthropometric indices in identifying individuals with the MONW phenotype. METHODS: This cross-sectional study involved a total of 26,332 participants aged 18 years or older with a normal weight from Nanjing, China. Sociodemographic information, biochemical measurements, and anthropometric indices were collected. The novel body fat anthropometric indices included body shape index (ABSI), body roundness index (BRI), abdominal volume index (AVI), weight-adjusted-waist index (WWI), body adiposity index (BAI), conicity index (CI), waist-hip-height ratio (WHHR), as well as traditional indices such as waist circumference (WC), hip circumference (HC), body mass index (BMI), waist-to-hip ratio (WHR), and waist-to-height ratio (WHtR).The prevalence ratio (PR) from modified poisson regression and area under the receiver-operating characteristic curve (AUC) were conducted to compare the association and predictive capacity of different obesity indicators for the MONW phenotype. All analyses were stratified by sex. RESULTS: Modified poisson regression analyses revealed that weight, WC, HC, BMI, WHR, WHtR, ABSI, BRI, AVI, WWI, BAI, CI, and WHHR were independently associated with higher risk of the MONW phenotype, regardless of whether they were treated as a continuous or categorical variable (P < 0.05). Notably, BRI demonstrated the strongest association in both men (highest quartile VS lowest quartile; PR = 3.14, 95%CI, 2.49, 3.96; P < 0.001) and women (PR = 4.63, 95%CI, 3.81, 5.62; P < 0.001). Receiver operating characteristic analysis indicated that AUC for the different anthropometric indices ranged from 0.50 to 0.80. BRI and WHtR had the largest AUC in both males (both AUC = 0.733; 95% CI, 0.717, 0.750) and females (both AUC = 0.773; 95% CI, 0.761, 0.786). The optimal cut-off points for BRI, determined by maximizing the Youden's index, were 3.102 (sensitivity: 63.2%, specificity: 36.2%) in males and 3.136 (sensitivity: 68.9%, specificity: 44.2%) in females. Moreover, BRI and WHtR exhibited the highest diagnostic accuracy in younger age groups, specifically those aged 18-34 in both sexes. CONCLUSIONS: BRI emerged as the optimal predictor and independent determinant of the MONW phenotype, regardless of gender. This association was particularly pronounced in young individuals.