The clinical pharmacokinetics of trimetrexate were determined in 11 patients during the phase I trial. The plasma drug disappearance curve was triphasic, with a t1/2 alpha of 8 +/- 5 minutes, t1/2 beta of 102 +/- 48 minutes, and t1/2 gamma of 15.2 +/- 5.7 hours. The AUC was 373 +/- 336 (micrograms/ml) hr (normalized to a dose of 200 mg/m2), volume of distribution by the area method (Varea) was 25.2 +/- 16.1 L/m2, total clearance (CL) was 14 +/- 8 ml/min/m2, and renal clearance (CLR) was 8 +/- 6 ml/min/m2. Four patients who received 190 to 200 mg/m2 did not develop severe toxicity. However, three patients who received 120 to 210 mg/m2 developed severe myelosuppression, skin rash, and stomatitis. This latter group had significantly longer terminal half-lives, greater AUCs, smaller Vareas, and lower rates of CL and CLR. One of these patients received an unusually large total amount of trimetrexate (470 mg) because of his obesity. The remaining two patients had renal problems. One developed toxicity despite having received a reduced dose (120 mg/m2) because of impaired renal function. The other patient, with normal renal function, had ascites and had undergone a unilateral nephrectomy for renal carcinoma. These data suggest that prolonged exposure to high trimetrexate levels may lead to increased toxicity. Dosage adjustment may have to be considered for patients who have renal dysfunction.