Abstract
BACKGROUND: Although cellular and animal studies have reported that resolvin D1 (RvD1) and resolvin D2 (RvD2) are mechanisms involved in the development of type 2 diabetes mellitus (T2DM), the impact of RvD1 and RvD2 on the risk of T2DM at a population level remains unclear. METHODS: We included 2755 non-diabetic adults from a community-based cohort in China and followed them for seven years. Cox proportional hazards model was used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the association of RvD1 and RvD2 with T2DM probability. Time-dependent receiver operator characteristics (ROC) curve was used to evaluate the predictive performance of RvD1 and RvD2 for the risk of T2DM based on the Chinese CDC T2DM prediction model (CDRS). RESULTS: A total of 172 incident T2DM cases were identified. Multivariate-adjusted HRs (95% CI) for T2DM across quartiles of RvD1 levels (Q1, Q2, Q3 and Q4) were 1.00, 1.64 (1.03-2.63), 1.80 (1.13-2.86) and 1.61 (1.01-2.57), respectively. Additionally, body mass index (BMI) showed a significant effect modification in the association of RvD1 with incident T2DM (P (interaction) = 0.026). After multivariate adjustment, the HR (95% CI) for T2DM in the fourth compared with the first quartile of RvD2 was 1.94 (95% CI: 1.24-3.03). Time-dependent ROC analysis showed that the area under time-dependent ROC curves of the "CDRS+RvD1+RvD2" model for the 3-, 5- and 7-year risk of T2DM were 0.842, 0.835 and 0.828, respectively. CONCLUSIONS: Higher RvD1 and RvD2 levels are associated with a higher risk of T2DM at the population level.