A number of reports have described the frequency of coronary arterial narrowing in patients with valvular aortic stenosis. No published reports have examined the structure of the stenotic aortic valve in adults and related the valve structure to variables, including coronary arterial narrowing, useful in predicting that structure. One hundred eighty-eight patients having aortic valve replacement for isolated valvular aortic stenosis were studied. All patients were > 40 years of age at the time of aortic valve replacement, all had coronary angiograms preoperatively, and of 182 patients (97%) measurements of serum total cholesterol had been obtained and 184 (98%) had body mass index calculated. The structure of the operatively excised valve was classified as unicuspid or bicuspid (congenitally malformed), or tricuspid aortic valve. A logistic regression model was developed that found 4 factors (age, serum total cholesterol, angiographic coronary artery disease and body mass index) to be predictive of aortic valve structure: (1) Patients with at least 3 or all 4 factors high or present (i.e., age > 65 years, serum total cholesterol > 200 mg/dl, body mass index > 29 kg/m2 and coronary artery disease) had a low probability (10 to 29%) of having a congenitally malformed valve; (2) patients with at least 3 or all 4 factors low or absent (i.e., age < or = 65 years, serum total cholesterol < or = 200 mg/dl, body mass index < or = 29 kg/m2, and no coronary artery disease) had a high probability (72 to 90%) of having a congenitally malformed valve.(ABSTRACT TRUNCATED AT 250 WORDS)