The 2 most commonly used oxazaphosphorines are cyclophosphamide and ifosfamide, although other bifunctional mustard analogues continue to be investigated. The pharmacology of these agents is determined by their metabolism, since the parent drug is relatively inactive. For cyclophosphamide, elimination of the parent compound is by activation to the 4-hydroxy metabolite, although other minor pathways of inactivation also play a role. Ifosfamide is inactivated to a greater degree by dechloroethylation reactions. More robust assay methods for the 4-hydroxy metabolites may reveal more about the clinical pharmacology of these drugs, but at present the best pharmacodynamic data indicate an inverse relationship between plasma concentration of parent drug and either toxicity or antitumour effect. The metabolism of cyclophosphamide is of particular relevance in the application of high dose chemotherapy. The activation pathway of metabolism is saturable, such that at higher doses (greater than 2 to 4 g/m2) a greater proportion of the drug is eliminated as inactive metabolites. However, both cyclophosphamide and ifosfamide also act to induce their own metabolism. Since most high dose regimens require a continuous infusion or divided doses over several days, saturation of metabolism may be compensated for, in part, by auto-induction. Although a quantitative distinction may be made between the cytochrome P450 isoforms responsible for the activating 4-hydroxylation reaction and those which mediate the dechloroethylation reactions, selective induction of the activation pathway, or inhibition of the inactivating pathway, has not been demonstrated clinically. Mathematical models to describe and predict the relative contributions of saturation and autoinduction to the net activation of cyclophosphamide have been developed. However, these require careful validation and may not be applicable outside the exact regimen in which they were derived. A further complication is the chiral nature of these 2 drugs, with some suggestion that one enantiomer may have a favourable profile of metabolism over the other. That the oxazaphosphorines continue to be the subject of intensive investigation over 30 years after their introduction into clinical practice is partly because of their antitumour activity. Further advances in analytical and molecular pharmacological techniques may further optimise their use and allow rational design of more selective analogues.