Homeostasis model assessment of insulin resistance (HOMA-IR) is a less invasive, inexpensive, and less labor-intensive method to measure insulin resistance (IR) as compared with the glucose clamp test. The aim of this study was to evaluate the validity of HOMA-IR by comparing it with the euglycemic clamp test in determining IR. We assessed the validity of HOMA-IR by comparing it with the total glucose disposal rate measured by the 3-hour euglycemic-hyperinsulinemic clamp in subjects with type 2 diabetes (n = 47), impaired glucose tolerance (n = 21), and normal glucose tolerance (n = 22). There was a strong inverse correlation (r = -0.558; P < .001) between the log-transformed HOMA-IR and the total glucose disposal rate. There was moderate agreement between the 2 methods in the categorization according to the IR (weighted kappa = 0.294). The magnitude of the correlation coefficients was smaller in the subjects with a lower body mass index (BMI <25.0 kg/m2 , r = -0.441 vs BMI > or =25.0 kg/m2 , r = -0.615; P = .032), a lower HOMA-beta cell function (HOMA- beta <60.0, r = -0.527 vs HOMA- beta > or =60.0, r = -0.686; P = .016), and higher fasting glucose levels (fasting glucose < or =5.66 mmol/L, r = -0.556 vs fasting glucose >5.66 mmol/L, r = -0.520; P = .039). The limitation of the validity of the HOMA-IR should be carefully considered in subjects with a lower BMI, a lower beta cell function, and high fasting glucose levels such as lean type 2 diabetes mellitus with insulin secretory defects.