The initial metabolism of the oxazaphosphorine cytostatic ifosfamide (IF) consists of two different pathways: ring oxidation at carbon-4 forms the cytostatically active metabolite 4-hydroxyifosfamide (4-OH-IF, "activated ifosfamide"), whereas side-chain oxidation with liberation of the presumably neurotoxic compound chloroacetaldehyde (CAA) that may also be responsible for IF-associated nephrotoxicity results in the formation of the cytostatically inactive metabolites 2-dechloroethylifosfamide (2-DCE-IF) and 3-dechloroethylifosfamide (3-DCE-IF). The pharmacokinetics of IF and its metabolites were investigated in 11 patients with bronchogenic carcinoma receiving IF on a 5-day divided-dose schedule (1.5 g/m2 daily). Blood samples were drawn on days 1 and 5 for up to 24 h after the start of the IF infusion. IF, 2-DCE-IF, and 3-DCE-IF were simultaneously quantified by gas chromatography (GC) with an NIP flame-ionization detector (NPFID), CAA was determined by GC with an electron-capture detector (ECD), and the highly unstable compound 4-OH-IF was measured using a high-performance liquid chromatography (HPLC) assay with fluorometric detection of 7-OH-quinoline, which is formed by the condensation of 4-OH-IF-derived acrolein with m-aminophenol. As compared with the values obtained on day 1, on day 5 the terminal half-life and AUC values determined for IF were reduced by 30% (6.36 vs 4.06 h and 1781 vs 1204 nmol h ml-1, respectively), whereas the maximal concentration (Cmax) values were not affected significantly (199.1 vs 181.1 nmol ml-1). This known phenomenon is explained by autoinduction of hepatic IF metabolism and was paralleled by increased metabolite levels. The mean Cmax values determined for 4-OH-IF, CAA, 3-DCE-IF, and 2-DCE-IF (on day 1/on day 5) were 1.51/2.59, 2.69/4.85, 12.9/26.5, and 8.6/16.7 nmol ml-1, respectively. The corresponding AUC values were 11.3/16.5, 30.3/34.3, 146/354, and 111/209 nmol h ml-1, respectively. As calculated by intraindividual comparison, the mean Cmax (day 5): Cmax (day 1) ratios for 4-OH-IF, CAA, 3-DCE-IF, and 2-DCE-IF were 1.94*, 2.05*, 2.52*, and 2.33*, respectively; the corresponding AUC (day 5): AUC (day 1) ratios were 1.51*, 1.29, 2.34*, and 2.23*, respectively (* P < 0.05). These data reveal that during fractionated-dose IF therapy the cancerotoxic effect of the drug increases. If the assumed role of CAA in IF-associated neurotoxicity and nephrotoxicity is a dose-dependent phenomenon, the probability of developing these side effects would also increase during prolonged IF application.