Background: Some studies have found an association of obesity with type 2 diabetes only among individuals with high normal serum γ- glutamyltransferase (GGT) activity, not in those with low serum GGT. If this interaction reflected pathophysiology, it would have scientific and clinical importance. The findings failed to reach statistical significance, however, and no articles have focused on the topic. We investigated possible interactions between serum GGT and body mass index (BMI) and their effects on the risk of prevalent type 2 diabetes and homeostasis model assessment (HOMA) insulin resistance. Methods: We analyzed 4011 adults ≥40 years old who participated in the 3rd US National Health and Nutrition Examination Survey. Results: BMI was associated with prevalent diabetes only among persons with high normal serum GGT activity (P for interaction = 0.002). In the highest serum GGT quartile, adjusted odds ratios for BMI 25-29.9, 30-34.5, and ≥35 kg/m2 compared with BMI<25 kg/m2 were 3.1, 5.1, and 6.2, respectively (P for trend <0.001). In the lowest serum GGT quartile, BMI was not associated with diabetes; corresponding adjusted odds ratios were 1.0, 0.9, 1.8, and 0.8 (P for trend = 0.551). After prevalent diabetes was excluded, there was a parallel interaction with HOMA levels (P for interaction <0.001). Conclusions: BMI was not associated with prevalent type 2 diabetes when GGT was low normal, suggesting that obesity itself may not be a sufficient risk factor for type 2 diabetes. Practically, this interaction can be useful in clinical settings to identify individuals at high risk for type 2 diabetes.