The purpose of this study was to evaluate the clinical outcome of a large cohort of patients who suffered an acute myocardial infarction with absolutely normal epicardial coronary arteries at the post-myocardial infarction coronary angiogram. The aetiological and prognostic factors in this population were also analysed. Few data exist concerning the outcome, and aetiological and prognostic factors, of patients with myocardial infarction and angiographically absolutely normal coronary arteries. Ninety-one patients (34 females/57 males; mean age 50+/-13 years, range 24--78 years) admitted with an acute myocardial infarction had absolutely normal coronary arteries at the angiogram performed 6.2+/-4 days (range 1--15 days) after the myocardial infarction, defined by smooth contours and no focal reduction (NC). Of the 91 NC patients, 71 were evaluated prospectively, alongside a systematic search of all aetiological factors reported in the literature. The NC patients were matched for age, sex, and the same period of myocardial infarction onset with a group of 91 patients with coronary artery stenosis (>50% diameter stenosis) at the angiogram performed 7.3+/-4 days (range 1--15 days) after the myocardial infarction (SC). The percent of smokers was similar between the two groups; higher prevalence rates of coronary heart disease family history, obesity, hypertension, hypercholesterolaemia and diabetes mellitus were found in SC (P=0.043 to 0.0001). In NC, coronary spasm was found in 15.5%, congenital coagulation disorders in 12.8%, collagen tissue disorders in 2.2%, embolization in 2.2%, and oral contraceptive use in 1.1%. Left ventricular ejection fraction at hospital discharge was higher in NC (60%+/-13%) than in SC (55%+/-13%, P=0.04). The mean follow-up was 35 months (range 1--100 months). Kaplan-Meier event-free survival, with the combined end-point defined as death, reinfarction, heart failure and stroke was 75% in NC vs 50% in SC (P<0.0001). Survival rate was 94.5% in NC compared to 92% in SC (ns). Univariate predictors of events in NC were left ventricular ejection fraction (P=0.03), age (P=0.02), diabetes (P=0.01), and smoking (P=0.03). Using Cox multivariate analysis, independent predictors of long-term outcome in NC patients were left ventricular ejection fraction (P=0.003) and diabetes (P=0.004). Aetiological factors, predominantly coronary spasm and inherited coagulation disorder, can be detected in only one third of the patients with myocardial infarction and absolutely normal coronary angiograms despite a systematic search in a prospective population. Mortality rates are similar but morbidity is lower in myocardial infarction patients with absolutely normal coronary angiography compared with those with coronary artery stenosis. The only two independent factors predictive of poor outcome in myocardial infarction patients with normal coronary arteries are left ventricular function and diabetes. Copyright 2001 The European Society of Cardiology.