BACKGROUND: Smoking is known to be a strong risk factor for premature atherosclerosis, myocardial infarction, and sudden cardiac death. Unexpectedly, in the reperfusion era, investigators have reported that patients who smoke have a more favorable prognosis after thrombolysis compared with non-smokers. Since smoking is associated with a relatively hyper-coagulable state, we hypothesized that the coronary occlusion responsible for infarction may be primarily thrombotic, with improved outcome relating to enhanced patency or the absence of a residual stenosis after thrombolytic therapy. METHODS AND RESULTS: To examine this issue, we evaluated 1619 patients treated with TPA, urokinase, or both in six consecutive myocardial infarction trials, of whom 878 (54%) were currently smoking. Patients underwent 90-minute and predischarge catheterizations, which were quantified blinded to the patients' smoking status. As expected, baseline fibrinogen (2.8 [2.5,3.6] versus 2.7 [2.4,3.5] g/dL, P = .003) and hematocrit (44% [41%, 47%] versus 43% [40%, 45%], P = .0001) levels were greater in smokers. Although there were no differences between smokers and nonsmokers with regard to 90-minute patency (73% versus 74%), smokers were more likely to have TIMI-3 flow (41.1% versus 34.6%, P = .034), with a larger minimum lumen diameter of the infarct stenosis both acutely (0.82 [0.51, 1.11] versus 0.72 [0.43, 1.04] mm, P = .0432) and at follow-up (1.2 [0.8, 1.74] versus 1.0 [0.7, 1.5], P = .002). Although smokers tended to have reduced in-hospital mortality compared with nonsmokers in univariate analysis (4.0% versus 8.9%, P = .0001), after adjustment for baseline differences between smokers and nonsmokers in age (54 [47, 62] versus 60 [54, 68] years, P < .0001), inferior infarct location (60% versus 53%, P < .0001), three-vessel disease (16% versus 22%, P < .001), and baseline ejection fraction (53% [44%, 60%] versus 50% [42%, 58%], P = .0069), smoking history was of no independent prognostic significance. CONCLUSIONS: Therefore, smokers have a relatively hypercoagulable state, documented by increased hematocrit and fibrinogen levels. Quantitative coronary angiographic analysis suggests that the mechanism of infarction in smokers is more often thrombosis of a less critical atherosclerotic lesion compared with nonsmokers. Enhanced perfusion status, as well as favorable baseline clinical and angiographic characteristics, may be responsible for the more benign prognosis of current smokers.