OBJECTIVE:To investigate the relation of the obesity and body-fat distribution with angiographically de®ned coronary atherosclerosis. DESIGN: Cross-sectional study in a clinical setting. SUBJECTS: Three hundred and twenty men (median age, 59 y) and 212 women (median age, 67 y) who underwent coronary angiography for suspected or known coronary heart disease at 5 cardiology departments between September 1996 and August 1997. Patients with disease duration b1 y were excluded. MEASUREMENTS: The body mass index (BMI) and the waist to hip circumference ratio (WHR) were used as main exposure variables, and either the presence of signi®cant coronary stenosis or the Gensini's score ( 10 vs`10) as an outcome variable, in a sex-speci®c multiple logistic regression analysis controlling for age, hospital, and other coronary risk factors. RESULTS: Among male patients, BMI was progressively higher with an increasing number of vessels involved (P trend 0.05); the adjusted odds ratios for the presence of signi®cant stenosis across quartiles of BMI were 1.0 (reference), 1.1, 1.9 and 2.5 (P trend 0.02), and the positive association was more pronounced for younger patients. Among females, however, such associations were not evident. Employing the Gensini's score as an outcome gave similar results. WHR was not signi®cantly associated with either outcome regardless of sex. CONCLUSION: These results suggested that BMI was predictive of coronary stenosis among male patients, but not among female patients. Unlike most previous studies, this study failed to detect a positive association with WHR.