Comparative Analysis of Anthropometric Indices for  Predicting Obstructive Sleep Apnea among U.S. Adults

Background: Obstructive sleep apnea syndrome (OSA) is a significant public health issue with adverse outcomes such as cardiovascular disease and metabolic disorders. Obesity is a primary risk factor for OSA, and traditional obesity indicators like BMI may not accurately reflect fat distribution, which is crucial in OSA development. Other anthropometric indicators like waist circumference (WC), body mass index (BMI),body roundness index(BRI), visceral adiposity index (VAI),lipid accumulation product(LAP),a body shape index(ABSI) and weight-to-waist index (WWI) have been proposed for better assessment of obesity-related health risks. Methods: A cross-sectional study was conducted using data from the 2005-2008 and 2015-2018 NHANES, which included a total of 8,966 participants. We used restricted cubic spline (RCS) analysis, logistic regression analysis, and subgroup analysis to assess the associations of anthropometric indices with OSA. We assessed the ability of the anthropometric indices to identify OSA using receiver operating curve (ROC) analysis and area under the curve (AUC)and compared them using the Delong test. Results: Based on NHANES data, this study revealed that waist circumference (WC) and body mass index (BMI) were the strongest predictors of obstructive sleep apnea (OSA) risk, with their associations remaining significant after adjusting for all covariates. Subgroup analysis revealed that this positive correlation was particularly strong in the 40-60 years age group, non-Hispanic whites, inactive individuals, and those with higher education. ROC analysis confirmed that WC and BMI had the highest AUC values of 0.637 and 0.633, respectively, making them the most effective in distinguishing OSA cases. Although the predictive power of the BRI, LAP, VAI, and WWI was weaker, the BRI still showed some predictive ability. Bonferroni correction and DeLong tests further validated WC and BMI as the most reliable indicators for predicting OSA. Conclusions: WC, BMI, BRI, LAP, VAI and WWI were significantly correlated with OSA. It is worth noting that WC and BMI, two traditional anthropometric indicators, show the best predictive ability compared with other new measurement indicators. Among them, WC has the characteristics of high efficiency and simple measurement in predicting OSA in clinical application, and is the best choice for early detection and treatment strategies.