ABSTRACTObjectiveThis study investigates the association between estimated glucose disposal rate (eGDR), a measurement of insulin resistance, and microvascular and macrovascular complications in patients with type 2 diabetes (T2D).MethodsThis cross‐sectional study enrolled 7471 patients with T2D from 2010 to 2023. The eGDR was calculated using waist circumference, HbA1C levels, and hypertension status. Logistic regression analysis and restricted cubic splines were utilised to examine the relationship between eGDR and vascular complications, including nephropathy, retinopathy, and coronary artery disease (CAD). The robustness of the results and between‐group interactions were examined by sensitivity and subgroup analysis. Furthermore, receiver operating characteristic (ROC) curve analysis was employed to assess the discriminatory value of the adjusted model for T2D vascular complications.ResultsAmong participants, 56.5% were female, with a mean age of 57.04 ± 11.05 years and a median of 8 years of diabetes duration. In the final adjusted model, each unit increase in the standard deviation of eGDR was significantly associated with a 23.6%, 24.8% and 29.6% decrease in the odds of nephropathy, retinopathy, and CAD, respectively. There was a significant association between higher eGDR quartiles compared to Q1 for all complications (p < 0.05). The Q4 group had the lowest adjusted odds ratios (ORs) compared to the Q1 group for all complications; the OR of Q4 was 0.549 for nephropathy, 0.360 for retinopathy, and 0.396 for CAD (p < 0.001). The restricted cubic spline for nephropathy followed a negative nonlinear association with eGDR, whereas for retinopathy and CAD, it followed a negative linear pattern. The effect of eGDR was consistent among different subgroups. The ROC curve analysis of the adjusted model showed good discriminatory power for all complications.ConclusionIn patients with type 2 diabetes, a higher eGDR was significantly associated with a lower risk of microvascular and macrovascular complications, regardless of well‐known confounders.