Thirty-nine patients with advanced solid tumors, including 28 with ovarian cancer, were entered in a phase I-II trial of a new platinum analog, JM9. Twenty-three patients had received prior chemotherapy which did not include cisplatin. Based on preliminary information from an ongoing study, our starting dose was 180 mg/m2. The total dose of JM9 was administered in 1 L of saline infused over 1 hour, with no additional hydration or electrolyte supplementation. Courses were repeated at 3-week intervals or after full recovery from thrombocytopenia. One hundred thirty-nine courses (range, one to six per patient) were administered at four dose levels: 180 mg/m2 (13 courses); 240 mg/m2 (64 courses); 300 mg/m2 (45 courses); and 350 mg/m2 (17 courses). The dose-limiting toxic effect was thrombocytopenia, which was dose-related and cumulative. Median platelet count nadirs were 50, 47, 25, and 28 X 10(9)/L for previously treated patients at dose levels of 180, 240, 300, and 350 mg/m2, respectively. For patients who had not received prior chemotherapy, the corresponding values were 403, 61, 44, and 36 X 10(9)/L. The nadir was predictable at Day 14 with recovery by Day 21 in earlier courses, but with delay of recovery to Days 28-42 in later courses and at higher dose levels. Twenty-five courses of chemotherapy in 15 patients were associated with a platelet count nadir of less than 20 X 10(9)/L, but despite this, serious hemorrhage was rare. Leukopenia was dose-related and mild; the median wbc count (X 10(9)/L) was 2.2 (range, 1.0-7.3) at the highest dose level of 350 mg/m2. The leukocyte count nadir was later than that for the platelet count (Days 21-28), and recovery was often not complete by the time of retreatment. All patients showed a progressive rise in mean corpuscular volume in successive courses, often accompanied by a fall in hemoglobin. Transfusions were required in 14 patients, 12 of whom had received prior chemotherapy. Nausea and vomiting, starting within 1 hour of drug administration, occurred in all patients, but appeared to be less severe and prolonged compared to that occurring with cisplatin. Diarrhea occurred in most patients at the two higher dose levels. There was no evidence of significant renal impairment, electrolyte disturbance, hearing loss, or peripheral neuropathy. Two patients had mild allergic reactions shortly after drug infusion and two others developed vasculitic rashes which were self-limiting.(ABSTRACT TRUNCATED AT 400 WORDS)