cis-Diammine(1,1-cyclobutanedicarboxylato)platinum (CBDCA) is a nonnephrotoxic but myelosuppressive analogue of cisplatin (DDP) with greatly reduced protein binding and greatly increased renal excretion. Thus, CBDCA might produce undue toxicity in patients with decreased renal function. Twenty-two patients [14 females and 8 males; median age, 66 (range, 35 to 83); median Karnofsky performance status, 70 (range, 40 to 90)] with refractory tumors and renal dysfunction [creatinine clearance (CCr) 6 to 83 ml/min] were treated with 31 courses of i.v. bolus CBDCA every 4 to 5 weeks. Dosages were determined by pretreatment CCr. Patients with CCr greater than or equal to 40 ml/min received 400 mg/sq m; patients with CCr 20 to 39 ml/min received 250 mg/sq m; and patients with CCr 0 to 19 ml/min received 150 mg/sq m. Toxicities were assessed by weekly clinical and laboratory assessment. Responses were assessed in patients with measurable disease. Plasma pharmacokinetics and urinary excretion of total and ultrafilterable platinum were measured with flameless atomic absorption spectrometry. Observed toxicities were similar to those in patients with normal renal function. Myelosuppression, especially thrombocytopenia, was the major toxicity. Nausea and vomiting were mild to moderate. There was no ototoxicity, neurotoxicity, or nephrotoxicity or reduction in CCr due to CBDCA. Total body clearance of ultrafilterable platinum correlated highly with CCr. The percentage of reduction in platelet count correlated highly and linearly with the area under the curve (AUC) of plasma-ultrafilterable platinum. However, for any AUC, there was 17% greater platelet reduction in patients who had previously received extensive myelosuppressive chemotherapy than in nonpretreated patients. Since total body clearance is proportional to CCr, platelet reduction is proportional to AUC, and total body clearance = dosage/AUC, we have derived an equation to calculate a dosage that will produce a desired reduction in platelet count. Calculations for theoretical patients (both pretreated and nonpretreated) with CCr of 100 ml/min produce dosages very close to maximum tolerated dosages derived in actual Phase I trials. The actual clinical utility of these predictive equations must await validation in prospective studies with larger numbers of patients.