Recent trials have shown that recombinant tissue plasminogen activator (rt-PA) is an effective thrombolytic agent in patients with acute myocardial infarction. Because rt-PA converts plasminogen to plasmin, which is known to activate complement in vitro, we tested the hypothesis that rt-PA can induce in vivo activation of complement. Studies were performed in 12 patients with acute myocardial infarction. Six control patients had patent coronary arteries and did not receive rt-PA; these patients had normal values of the components of the complement system C4a (409 ± 111 ng/ml) and C5a (8.8 ± 1.8 ng/ml) with a slight elevation of C3a (204 ± 6.6 ng/ml) in samples collected before coronary arteriography (253 ± 25 minutes after onset of pain). After coronary arteriography, there was a slight decrease in the values of C4a (224 ± 37 ng/ml), C5a (7.3 ± 1.3 ng/ml) and C3a (164 ± 35 ng/ml). The remaining six patients had complete coronary occlusion and received rt-PA (80 to 150 mg intravenously). In this treated group, before coronary arteriography the values of C4a (406 ± 51.6 ng/ml) and C5a (8.1 ± 1.9 ng/ml) were normal, and those of C3a were slightly elevated (250 ± 76 ng/ml). All complement values obtained before rt-PA were similar to those in the untreated group. However, after administration of rt-PA (but before any angiographically detectable reperfusion), there was a striking ihcrease in C4a (2,265 ± 480 ng/ml; p < 0.01), C3a (600 ± 89 ng/ml; p < 0.05) and C5a (30.0 ± 4.5 ng/ml; p < 0.05). Thus, rt-PA causes an immediate, marked activation of complement in patients with acute myocardial infarction. Measurement of plasma anaphylatoxin levels may be a useful means of assessing the activity of rt-PA, and the degree of complement activation might reflect the extent of thrombolysis. The significance of rt-PA-induced complement activation with respect to the extent of ischemic injury remains to be determined, but it is conceivable that the anaphylatoxin release might limit the otherwise beneficial effects of rt-PA.