Abstract
AIMS: Poorly controlled diabetes is closely associated with obesity. This study aimed to investigate the associations and predictive values of obesity-related indicators including waist circumference (WC), body mass index (BMI), weight-adjusted waist index (WWI), and waist-to-height ratio (WHTR) with uncontrolled diabetes mellitus (DM) in an elderly Chinese community population. METHODS: A cross-sectional study was conducted among 15,673 elderly diabetic patients from health examinations in Shangyu District in 2024. Receiver operating characteristic (ROC) curve analysis was used to compare the predictive performance of four obesity indicators for uncontrolled DM. Multivariable logistic regression was performed to assess their associations, while the threshold effects were detected by two piecewise linear models. Subgroup analyses were also performed. RESULTS: ROC analysis revealed WC had the highest Area under curve (AUC) (0.53, 95%CI: 0.52-0.54), significantly outperforming BMI (p < 0.001). Multivariable analysis demonstrated that WC, WWI and WHTR exhibited piecewise linear relationships with uncontrolled DM. Notably, WC showed a nonlinear association only in women (threshold = 70.5 cm), beyond which the risk of uncontrolled DM significantly increased (OR = 1.02, 95%CI: 1.01-1.03). WWI displayed a nonlinear pattern exclusively in men (threshold = 9.60 cm/√kg), with higher values associated with elevated risk (OR = 1.37, 95%CI: 1.24-1.51). WHTR exhibited nonlinear relationships in both genders (thresholds: men = 0.46, women = 0.45), with high risk of uncontrolled DM observed at above thresholds. CONCLUSION: WC was the strongest predictor of uncontrolled DM in the elderly population. The nonlinear relationships between WC/WWI and the risk of uncontrolled DM exhibit gender-specific threshold effects.