OBJECTIVE--To examine if the risk for CHD increases progressively with increases in the BMI of normoglycemic, hyperinsulinemic, morbidly obese women (BMI > or = 35 kg/m2). RESEARCH DESIGN AND METHODS--Insulin sensitivity was evaluated by calculating an ISI following an OGTT. There was a significant linear relationship between ISI and BMI fitted by two straight lines intersecting at a point corresponding to a BMI of 29.7 +/- 1.5 kg/m2. Significant linear relationships between insulin sensitivity and BMI were obtained below and above this breakpoint. Similarly, a breakpoint for the relation between dBP and BMI corresponding to a BMI > or = 33.7 +/- 3.4 kg/m2 was obtained. Significant linear relationships between BMI and plasma fasting glucose, triglyceride, cholesterol, HDL cholesterol, sBP, or dBP were not observed in the women with a BMI > 35 kg/m2. RESULTS--Compared with lean (BMI < 27) women of similar age, the morbidly obese patients appear to be at a higher risk for CHD. This is suggested by statistically significant increases in fasting insulin (mean +/- SD; 187 +/- 137 vs. 64.2 +/- 16.2 pM) and triglyceride levels (128 +/- 78.1 vs. 73 +/- 25 mg/dl), sBP (132 +/- 114 vs. 104 +/- 15.8) and dBP (84 +/- 72 vs. 67 +/- 2.1 mmHg), and decreases in HDL cholesterol (1.03 +/- 0.44 vs. 1.29 +/- 0.82 mM) and apo A-I (91 +/- 55 vs. 122 +/- 35 mg/dl) concentrations. CONCLUSIONS--It appears that there may be a threshold of body mass up to which insulin sensitivity is associated with CHD risk. Above this threshold, there does not appear to be a progressive increase in the risk factors for CHD with increases in BMI.