Nonlinear association between changes in fasting plasma glucose and the incidence of diabetes in a nondiabetic Chinese cohort

Abstract

Abstract Background Limited data show that changes in fasting plasma glucose (FPG changes) are related to the incidence of type 2 diabetes (T2D). We aimed to correlate FPG changes with incident diabetes and evaluate FPG changes as a marker to screen participants at high risk of T2D in China. Methods A total of 116,816 individuals were followed during a median follow-up of 3.10 years by secondary analysis in a nondiabetic Chinese cohort. The turning points were derived from a receiver operating characteristic curve. Hazard ratios (HRs) were evaluated by Cox proportional hazards models. Results A total of 2669 cases of T2D were identified (788 women and 1881 men). The age-standardized incidence of diabetes was 12.87 per 1000 person-years (women: 11.04; men: 14.69). A nonlinear relationship between FPG changes and incident diabetes is shown by the fitting curves. The curves were categorized into three stages by two turning points (-0.04 and 1.25 mmol/L) and conformed to the hook-like pattern: an initial decrease (stage-1), then a transient sharp elevation (stage-2), followed by a slow increase (stage-3). HRs per SD of FPG changes on incident diabetes varied with stage: stage-1: 0.16 (0.12, 0.23), stage-2: 0.20 (0.15, 0.28) and stage-3: 0.22 (0.16, 0.31). Compared with stage-1, the HR in stage-3 was significantly higher at 28.05 (23.99, 32.79), while the increase in stage-2 was slight at 2.16 (1.79, 2.61), and the HR in stage-3 rose to 30.09 (25.02, 36.19). Conclusions FPG changes had a strong correlation with the incidence of T2D and was a steady indicator that was used to distinguish the participants at high risk of diabetes.