Although thrombolysis with streptokinase has been compared with heparin anticoagulation for treating acute proximal deep venous thrombosis in several randomized trials, no individual study has had a sample of sufficient size to determine with adequate power both efficacy and safety. Therefore, results were pooled from six randomized studies in which phlebography was used to confirm the diagnosis and to assess therapy. Thrombolysis was achieved 3.7 times more often among patients treated with streptokinase than among patients treated with heparin (95 percent confidence limits 2.5, 5.7; p less than 0.0001). Only three studies allowed comparison of these drugs for major bleeding complications, which were 2.9 times greater with streptokinase than with heparin (95 percent confidence limits 1.1, 8.1; p = 0.04). Thus, in aggregate, streptokinase-treated patients achieved thrombolysis but also seemed to experience major bleeding complications more frequently than those assigned at random to receive heparin. Future trials of sufficient sample size should be undertaken to evaluate efficacy and safety. Such trials, which should include newer fibrinolytic agents, are necessary to determine optimal therapy for acute proximal deep venous thrombosis.