To evaluate the role of selective intra-arterial low-dose thrombolytic therapy (SILDT) as an alternative to the surgical management of acute arterial occlusion, the hospital records of 40 patients who underwent 43 SILDT treatments with either streptokinase (36) or urokinase (7) between December 1979 and March 1984 were reviewed. Twenty-eight patients underwent 30 treatments (group 1) for native arterial occlusion and 12 patients underwent 13 treatments (group 2) for prosthetic or autogenous graft occlusions. Therapy was deemed successful if subsequent surgical therapy was obviated. In group 1, SILDT was successful in 13 of 28 (45%) patients with 12 of 25 lower extremity occlusions and one of three upper extremity occlusions. Successful lysis in the native artery occlusion group fell into three categories: five patients were successfully treated for arterial thrombosis complicating percutaneous transluminal angioplasty (PTA); four patients required PTA after complete lysis revealed an underlying arterial stenosis; and only three required no further therapy after SILDT. SILDT failed in all three patients with the aortoiliac occlusions. Eleven patients with femoral artery occlusions and unsuccessful SILDT required six bypass procedures, three amputations, one embolectomy, and one PTA. In group 2 only 3 of 14 treatments (21%) were successful. Bypass revision was not possible in 11 patients and all required amputation. Systemic fibrinolysis was seen in 20 (59%) of 34 patients with available data. Neither fibrinogen levels nor fibrin degradation products predicted the occurrence of complications. Minor complications occurred in 18 of 43 (43%) treatments; small hematomas at the catheter entry site were most common. Minor complications occurred in 20 of 43 treatments (44%) and included severe local hemorrhage (four), distant bleeding (three), pulmonary embolism (four), myocardial infarction (three), unmasking of an aortoduodenal fistula (one), and clot migration requiring emergency thrombectomy (four). SILDT is most effective in acute arterial thrombosis complicating arteriography or percutaneous angioplasty. It may play a role in the patient in whom thrombolysis can reveal an underlying stenosis amenable to percutaneous angioplasty. This experience shows SILDT to be of limited value in the management of prosthetic autogenous graft occlusions. Finally, thrombolytic therapy is associated with significant morbidity and mortality rates and requires cautious monitoring to detect arterial thrombus migration, worsening tissue ischemia, venous thromboembolism, intracerebral hemorrhage, and local or systemic bleeding.