Purpose: Triglyceride-glucose (TyG) index is an emerging surrogate predictor of incident type 2 diabetes mellitus (T2DM). The study aimed to examine the association between TyG index and incident T2DM in a prospective Chinese cohort.
Methods: The data were collected in 1992 and recollected in 2007 from the same group of 687 participants. The association between TyG index and T2DM was analysed.
Results: During follow-up, 74 participants developed T2DM and the risk of T2DM increased with TyG index. The adjusted hazard ratio (HR) was 3.36 (95% CI: 1.52-7.39, P < 0.001) comparing the top TyG quartile to the bottom quartile. Smooth curve fitting revealed a nonlinear association and threshold effect between TyG index and incident T2DM with a nadir of risk when TyG index was around 8.51. For TyG ≤ 8.51, the risk of incident T2DM tended to decrease with per SD increase in TyG but no statistical significance was achieved (adjusted HR: 0.69, 95% CI: 0.43-1.12, P = 0.133). For TyG > 8.51, the risk of incident T2DM significantly increased by 38% with per SD increase in TyG (adjusted HR: 1.38, 95% CI: 1.14-1.67, P = 0.001). Time-dependent receiver operating curve suggested helpful discriminative power of TyG index for T2DM. It also significantly promoted the reclassification ability beyond the baseline risk model with net reclassification index of 0.159 (P = 0.020). Sensitivity analysis excluding participants with prediabetes demonstrated similar results.
Conclusions: The TyG index was a significant and independent predictor for future T2DM development. The shape of relationship will require further studies.
Keywords: Nonlinear association; Prediction; Triglyceride–glucose index; Type 2 diabetes mellitus.