Tissue plasminogen activator (t-PA) causes fibrinogen proteolysis when a2-antiplasmin levels fall, and this may contribute to t-PA-induced hemorrhage. Because clot-bound plasmin is protected from a2-antiplasmin inhibition, we tested the possibility that a2-antiplasmin supplementation would block t-PA-induced fibrinogenolysis and bleeding without affecting thrombolysis. When added to human or rabbit plasma, a2-antiplasmin inhibits t-PA-induced fibrinogenolysis, but has little effect on the lysis of 1251-fibrin clots. To examine its effect in vivo, rabbits with preformed 125I-labeled-jugular vein thrombi were randomized to receive t-PA, t-PA and a2-antiplasmin, or saline. a2-Antiplasmin infusion produced a modest decrease in t-PAinduced thrombolysis (from 40.2% to 30.1%, P = 0.12), but reduced fibrinogen consumption from 87% to 27% (P = 0.0001), and decreased blood loss from standardized ear incisions from 5,594 to 656 ,ul (P < 0.0001). We hypothesize that a2-antiplasmin limits t-PA-induced hemorrhage by inhibiting fibrinogenolysis and subsequent fragment X formation because (a) SDS-PAGE and immunoblot analysis indicate less fragment X formation in a2-antiplasmin treated animals, and (b) when added to a solution of fibrinogen and plasminogen clotted with thrombin in the presence of t-PA, fragment X shortens the lysis time in a concentration-dependent fashion. These findings suggest that fragment X incorporation into hemostatic plugs contributes to t-PA-induced bleeding. By blocking t-PA-mediated fibrinogenolysis, a2-antiplasmin supplementation may improve the safety of fibrin-specific plasminogen activators. (J. Clin. Invest. 1993Invest. . 91:1343Invest. -1350