The early effects of tissue-type plasminogen activator (t-PA) on the "culprit" coronary lesion in patients presenting with unstable angina or non-Q wave myocardial infarction were determined by quantitative arteriography. Of 391 such patients, 306 satisfied clinical and arteriographic requirements for eligibility and received a 90-minute front-loaded infusion of t-PA (0.8 mg/kg i.v.; maximum, 80 mg) or placebo plus conventional antianginal therapy. All patients received full heparinization and a follow-up arteriogram 18-48 hours after treatment. A non-Q wave myocardial infarction (MI) was diagnosed in 97 patients (32%) after entry. In the entire patient population, among t-PA- and placebo-treated patients, respectively, 25% versus 19% (p = 0.25) of all culprit lesions achieved the primary study end point, measurable improvement (by > or = 10% reduction of stenosis or two Thrombolysis in Myocardial Infarction TIMI flow grades) at follow-up. Substantial improvement (by > or = 20% reduction of stenosis or two TIMI grades) was seen with t-PA in 15% of all culprit lesions versus 5% with placebo (p < 0.003). Arteriographically apparent thrombus was present at baseline in the culprit lesion of 107 patients (35%). Substantial improvement was more frequent with t-PA among lesions containing apparent thrombus (in 36% with t-PA versus 15% with placebo; p < 0.01), as it was among patients evolving a non-Q wave MI (33% versus 8%; p < 0.005). By multivariate analysis, the significant, independent predictors of substantial improvement include apparent thrombus (p = 0.0001), non-Q wave MI (p = 0.003), and t-PA use (p = 0.01). Both non-Q wave MI status and thrombus had been specified a priori as important variables. Arteriographically apparent intraluminal thrombus and improvement of the culprit lesion with either of these regimens were only moderately frequent in patients with unstable angina or non-Q wave MI. Substantial improvement of culprit lesions was more frequent with t-PA than with placebo overall and in two prospectively defined subgroups. The clinical relevance of these observations is being tested in the larger, ongoing clinical TIMI IIIB study.