Because the initial evaluation of N,N',N''-triethylenethiophosphoramide (thioTEPA) preceded the standardized approach to the Phase I trials, uncertainty surrounds the recommended dose. Since it has recently been demonstrated that an almost 100-fold increase in dose can be administered in bone marrow transplant regimens, we conducted a Phase I reevaluation of thioTEPA. ThioTEPA was administered i.v. in 50 ml 5% dextrose in water over 10 min. Twenty-seven patients were entered at doses ranging from 30 to 75 mg/m2. The major toxic effect was myelosuppression; thrombocytopenia greater than or equal to grade 3 occurred in four of seven patients, and leukopenia greater than or equal to grade 3 in two of seven patients at 75 mg/m2. Among eight patients at 65 mg/m2 only two had greater than or equal to grade 3 myelosuppression making this the recommended new phase II dose for the majority of patients. Moderate (grade 2) easily controlled nausea and vomiting was the only other major side effect. There was no alopecia or mucosal or neurological toxicity. Three partial remissions were observed among nine previously treated ovarian cancer patients. Plasma concentrations of thioTEPA and its major active metabolite triethylenephosphoramide (TEPA) were measured by gas chromatography. The half-life of thioTEPA ranged from 51.6 to 211.8 min, and its pharmacokinetics was dose dependent; total body thioTEPA clearance decreased with increasing dose. The half-life of TEPA was considerably longer than that of the parent compound (3.0 to 21.1 h); as a result, the area under the plasma concentration-time curve (AUC) of TEPA was severalfold greater than that of the parent compound. The ratio of TEPA AUC to thioTEPA AUC decreased with increasing dose, suggesting that formation of TEPA is a saturable step in elimination. The AUC and total body clearance of thioTEPA, but not of TEPA, were closely correlated with neutrophil but not platelet toxicity.