Sex-specific optimal cut-off points for metabolic health indicators to predict incident type 2 diabetes mellitus

We aimed to determine the optimal cut-off points for metabolic health indicators, including insulin resistance (IR), glucose, insulin, BMI, and waist circumference, in middle-aged nondiabetic people to predict future type 2 diabetes mellitus (T2DM).

The data came from 12,543 Atherosclerosis Risk Communities Study participants, including 5758 men and 6785 women. They did not have diabetes at baseline and were followed for incident T2DM within 3, 6, and 9 years. IR was estimated using four IR metrics: HOMA-IR, METS-IR, TyG index, and TG/HDL-C. We used the Youden index to determine the optimal cut-off values. In females, the cut-off points for glucose to predict incident T2DM ranged from 96 to 102 mg/dL, with Area Under the Curve (AUC) values of 0.64-0.85. In males, the cut-off points ranged from 102 to 106 mg/dL, with AUC values of 0.60-0.83. For HOMA-IR, the cut-off points in females varied from 2.4 to 3.2, with AUC values of 0.69-0.78, while they ranged from 2.8 to 3.2 in males. The optimal cut-off values for METS-IR, TyG index, TG/HDL-C, insulin, BMI, and waist circumference were 40-43, 8.6-8.9, 2.0-3.2, 9-15 μU/mL, 28-29 kg/m², and 91-97 cm in women, and 44-45, 8.8-8.9, 2.9-3.2, 11-12 μU/mL, 27-29 kg/m², and 99-103 cm in men.

The optimal threshold for each predictor's prediction of incident T2DM varied by sex. The eight predictors' order of predictive performance were fasting glucose, HOMA-IR, METS-IR, insulin, BMI, waist circumference, TyG index, and TG/HDL-C.