The frequency of hemodialysis-associated hemorrhage was studied prospectively in two successive, parallel, heparin-controlled studies using epoprostenol (PGI2; average dose, 4.1 ng/kg.min) as the sole antithrombotic agent. Sixty-three patients with active or recently active bleeding underwent 163 hemodialysis treatments in each of which prospective bleeding risk was assessed. PGI2 was associated with up to 50% overall reduction in the frequency of bleeding, particularly in the highest risk circumstances. PGI2 also allowed successful completion of the full, prospectively prescribed hemodialysis time in the most treatments (82% versus 93% with heparin). Furthermore, the efficiency of hemodialysis using PGI2, as indicated by the reduction in concentration of blood urea nitrogen and serum creatinine, was equal to that using heparin, even though there was a tendency toward modest reduction in residual volume of the hollow fiber dialyzer and slightly more frequent early termination of treatment from dialyzer clotting with PGI2. No severe vasodilatory side effects of PGI2 were observed during these studies. Hypotension was equally frequent during hemodialysis with heparin as with PGI2. The current results suggest that PGI2 should be considered as a substitute for heparin during high-risk hemodialysis because PGI2 may reduce the incidence of dialysis-associated bleeding without severe adverse side effects.