Glycemic Cluster Analysis of Non-Diabetic Japanese Individuals Using the Triglyceride-Glucose Index Identifies an At-Risk Group for Incident Cardiovascular Disease Independent of Impaired Glucose Tolerance

Background/Objective: Type 2 diabetes mellitus (T2DM) is a known risk factor for cardiovascular disease (CVD); however, not all individuals with T2DM have the same CVD risk. Traditional glycemic cluster analyses of individuals with diabetes using the homeostatic model assessment estimate of insulin resistance (HOMA-R) as an indicator of insulin resistance have identified severe insulin-resistant diabetes as a high-risk factor for incident CVD. To extend this observation, we further classified non-diabetic individuals to identify risk clusters for incident CVD. Methods: We performed hierarchical cluster analysis of 577 non-diabetic Japanese individuals using the glycated hemoglobin level, body mass index, homeostasis model assessment estimate of β-cell function, and the triglyceride-glucose (TyG) index instead of the HOMA-R to assess incident CVD risk over up to 188 months. Results: Analysis using a multiple risk factor-adjusted Cox proportional-hazard regression model showed that in addition to impaired glucose tolerance (IGT), among the four clusters observed, a cluster labeled “low insulin secretion (Low-IS (TyG))” was a risk factor for incident CVD (hazard ratio, [95% confidence interval]: 2.77 [1.11-6.91]). Furthermore, when the participants were stratified on the basis of the presence of IGT, the Low-IS (TyG) cluster was associated with an increased risk of CVD only in the non-IGT group (3.29 [1.32-8.18]), but not in the IGT group (1.66 [0.34-8.15]). Conclusions: Glycemic clustering incorporating the TyG index identified a novel at-risk group for incident CVD among non-diabetic individuals, offering a useful tool for early CVD risk stratification beyond traditional measures of glucose tolerance.