We designed this study to identify patients with coronary artery disease (CAD), employing the exercise tolerance test, and to develop further criteria for ordering the exercise tolerance test in the preventive medicine population of the Cleveland Clinic Foundation. During 1987-1988 1,930 patients not known to have CAD were referred from the Department of Preventive Medicine for exercise tolerance tests as part of their periodic physical exams. We hypothesized that age was a major risk factor and ordered most (86.4%) tests on this basis: at age 40 and every two years after age 50. Twenty-five cases of CAD (25/1,930 or 1.3%) were found. One of 297 women was found to have CAD (0.3%). Seventeen patients were treated surgically and eight medically. Using age as an indication for testing detected 23 of 25 cases (92%). We compared the group with normal or nondiagnostic exercise tolerance tests and presumed not to have CAD (1,905 patients, median age 48) with the group with CAD (25 patients, median age 59). Age greater than 40, a total cholesterol level over 240 mg/dL, triglyceride level over 250 mg/dL, a total cholesterol to high-density lipoprotein ratio greater than 4.5, and a history of chest pain of any type were all significantly related to the presence of CAD. Testing men older than forty with two or more CAD risk factors, as has been recommended, would have resulted in finding five of the 25 cases (20%). Testing only patients who complained of any type of chest discomfort would have resulted in detecting 14 of the 25 cases (56%).(ABSTRACT TRUNCATED AT 250 WORDS)