The aim of this study was to investigate intrasubject variability in ifosfamide (IFO) pharmacokinetics and metabolism which may influence clinical effect, since the pharmacology of this drug is dependent on metabolism. A group of 11 patients (ages 1-16 years) were studied on at least two occasions. IFO, 9 gm-2 was administered as a continuous infusion over 72 h. Plasma and urine samples were collected and concentrations of IFO and its metabolites were determined. Comparisons were made between courses in the same subject, allowing for differences in age and prior IFO treatment. There was a wide variation in drug (twofold) and metabolite (up to tenfold) AUCs between courses in the same patient. Although some patients did show an increase in clearance between courses (up to threefold), there was no significant consistent change in overall pharmacokinetics among the different courses studied in the same patient. There was a significant decrease (up to 63%) in the AUC of the inactive metabolite 3-dechloroethylifosfamide (3-DCI) in later courses compared with the first course studied (P = 0.032, paired t-test). This was matched by an increase in the AUC of the total dechloroethylated metabolites with course (P = 0.015, paired t-test). None of the other metabolites measured showed any consistent change in plasma or urine levels between courses. Overall, the AUC of parent drug correlated with age (r2 = 0.86, P = 0.011), and postinfusion half-life correlated with plasma bilirubin (r2 = 0.89, P = 0.007). This study demonstrated large and seemingly unpredictable intrasubject variability in IFO pharmacokinetics and metabolism during repeated administrations. Investigations relating the clinical effects of IFO to pharmacokinetics and metabolism must take this variation into account.