The relationship between METS-IR and the risk of diabetes incidence in rural adults in China: A retrospective cohort study based on dynamic population.

Abstract

To evaluate the longitudinal association between the Metabolic Score for Insulin Resistance (METS-IR) and the risk of diabetes mellitus in rural Chinese adults. This retrospective cohort study included 53,120 participants aged ≥18 years from 2018 to 2023. Participants were stratified by quartiles of the METS-IR metrics. Cox proportional hazards models assessed the association between METS-IR and incident diabetes. Restricted cubic spline (RCS) models examined nonlinear trends. Subgroup analysis, interaction tests, and multiple sensitivity analyses were performed. Predictive ability was evaluated using time-dependent receiver operating characteristic (ROC) curves. During 176,413.4 person-years of follow-up (median 3.83 years), 14,397 participants developed diabetes. After multifactorial adjustment, METS-IR was significantly and positively associated with diabetes onset (hazard ratio (HR)=1.094,95% confidence interval (CI): 1.076-1.112, P < 0.001); those in the highest quartile group had a 1.435-fold higher risk compared to the lowest. RCS analysis revealed a nonlinear dose-response relationship. Kaplan-Meier curves confirmed increasing cumulative risk with higher METS-IR. Results remained robust across subgroups and sensitivity analyses. The area under the curve (AUC) for METS-IR predicting diabetes was 0.601 (1 year), 0.586 (3 years), and 0.599 (5 years). METS-IR is significantly correlated with the onset of diabetes, and the relationship is nonlinear. While it demonstrates limited discriminatory performance as a standalone screening tool, it remains suitable for initial risk stratification in primary health care institutions with limited resources.