The pharmacokinetics of isophosphamide (IP) in man has been studied using 14C-labeled drug and differential extraction of unchanged drug from metabolites by CHCl3 extraction of plasma and urine. The plasma decay of IP is biphasic with a terminal half-life of 15.2 hours which is twice that reported for cyclophosphamide (CP). The fraction of drug metabolized is 49% for IP compared to 88% reported for CP. This is consistent with the six-times larger pseudometabolic pharmacokinetic rate constant reported for CP (0.26 hr-1) than found by us for IP (0.04 hr-1). The renal clearance for IP (21.3 ml/min) is two times that of CP (10.7 ml/min). In addition to an analysis of the kinetics of disposition and elimination of unchanged IP, an analysis of the pharmacokinetics of total metabolite of IP in plasma and urine was made. A nonlinear pharmacokinetic model which includes a term for biotransformation according to Michaelis and Menten revealed that the total metabolites do not diffuse from a small plasma space of 2.1 liters but are excreted directly via the urine. In contrast to the biexponential plasma decay of single, high-dose IP (5 g/m2), multiple-dose IP (2.4 g/m2/day X 3) shows a monoexponential decay with a half-life of 6.9 hours, a pseudometabolic rate constant of 0.08 hour-1, and a renal clearance of 18.7 ml/minute. The fraction of drug metabolized (79.7%) calculated from the pharmacokinetic parameters agrees with the fraction of drug recovered in urine as total metabolites (72.8%). This is distinctly different from high-dose IP (48.6%) but similar to the fraction of drug metabolized as reported for CP (88.0%). This suggests that the pharmacokinetic transfer constant for IP metabolism is dose dependent, an observation that may be useful in developing new IP regimens.