To determine the efficacy of amonafide in patients with advanced, measurable pancreatic adenocarcinoma, 15 patients previously untreated with chemotherapy were entered on a phase II trial. The starting dose was 400 mg/m2 administered daily over 1 hr for 5 consecutive days repeated every 3 weeks. Because of grade 4 myelosuppression observed in the initial 2 patients, the daily starting dose was decreased to 350 mg/m2. Of the 15 patients, 14 were evaluable. Amonafide failed to produce clinical responses in the doses and schedule employed. Grade 4 granulocytopenia was observed in 3 of 9 courses at 400 mg/m2, 3 of 12 courses at 350 mg/m2 and in 1 of 4 courses at 300 mg/m2. Grade 4 thrombocytopenia was observed in 3 courses at 400 mg/m2. Nonhematologic toxicities included mild nausea and vomiting and skin rashes.