In order to determine the reasons for the low mortality after myocardial infarction in smokers compared with non-smokers (the smoker's paradox), the authors analysed the initial clinical data, the therapeutic interventions and hospital mortality in 790 consecutive patients (555 smokers, 235 non-smokers) admitted to hospital within 6 hours of the first symptoms of acute myocardial infarction and treated by intravenous thrombolytic agents and/or coronary angioplasty. Multivariate analysis with linear regression was used to identify the predictive factors of hospital mortality. The main differences between smokers and non-smokers were age (56 vs 67 years, p < 0.0001), gender (male, 90 vs 60%, p < 0.01), cardiogenic shock on admission (3 vs 8%, p < 0.01). TIMI 3 flow was obtained in the culprit artery in 84% of smokers and 79% of non-smokers (NS). Hospital mortality was 5% in the smoking population and 16% in non-smokers (p < 0.0001). In multivariate analysis, the variables of cardiogenic shock, age, gender and hypertension provided most of the prognostic information and tobacco consumption did not appear to have a protective effect. In patients admitted to hospital with acute myocardial infarction, identical incidences of early reperfusion are obtained in smokers and in non-smokers. However, mortality is higher in the non-smoking group due to more severe clinical characteristics on admission. Tobacco consumption is not a protective factor in the immediate period after acute myocardial infarction.