We correlated the incidence and degree of exercise induced ventricular arrhythmias (EIVA) with the angiographic severity of coronary artery disease (CAD) in 162 patients with a history of stable effort angina, all showing a positive exercise stress test for myocardial ischemia and a greater than or equal to 70% stenosis of a major coronary artery. Patients were grouped according to the following criteria: presence of electrocardiographic evidence of old transmural myocardial infarction (MI), number of significant coronary stenoses and number of left ventricular (LV) areas showing abnormal segmental wall motion (ASWM). The incidence of EIVA in patients with multivessel CAD was higher than in patients with single vessel CAD, but this difference was not statistically significant. The number of LV areas with ASWM was better correlated with the frequency of EIVA, which was 20.0% in patients with normal LV wall motion, 31.2% in patients with 1 area of ASWM, 54.0% in patients with 2 areas of ASWM (p less than 0.005 vs normal LV wall motion), 74.1% in patients with 3 or more areas of ASWM (p less than 0.001 vs normal LV wall motion and 1 area of ASWM), and 81.8% in patients with LV aneurysm (p less than 0.001 vs normal LV wall motion and 1 area of ASWM, p less than 0.005 vs 2 areas of ASWM). Patients with old MI showed a significantly higher incidence of EIVA than those without MI (p less than 0.001), but this difference was due to the more severe LV asynergy in the MI group. In conclusion, our results show that, in a selected population of patients with CAD, the incidence of EIVA correlates better with the extent of LV segmental wall motion abnormalities than with the number of diseased coronary arteries or the presence of an old transmural MI.