The active 5'-triphosphate of arabinosyl-2-fluoroadenine (F-ara-ATP) increases the anabolism of arabinosylcytosine (ara-C), whereas ara-C 5'-triphosphate inhibits the phosphorylation of arabinosyl-2-fluoroadenine (F-ara-A) in human leukemia cells in vitro. These interactions have a potential impact on drug scheduling. Clinical trials of relapsed leukemia in which fludarabine (F-ara-A 5'-monophosphate) and ara-C were given in sequence provided the opportunity to evaluate the effects of ara-C infusion on two sequelae: the pharmacokinetics of F-ara-A in plasma and that of F-ara-ATP in leukemia cells. First, F-ara-A pharmacokinetics were altered by ara-C infusion. This was visualized as a transient increase in F-ara-A plasma levels during the ara-C infusion that was given 4 h after fludarabine. The perturbation in F-ara-A plasma levels was dependent on the dose ara-C. Second, peak F-ara-ATP concentrations were lower in leukemia cells of patients who received ara-C in addition to fludarabine as compared with those who received fludarabine alone. The terminal half-life of F-ara-A in plasma and the half-life of intracellular F-ara-ATP were reduced after the ara-C infusion in a concentration-dependent manner. Studies using purified deoxycytidine kinase support the conclusion that the increase in plasma levels of F-ara-A is in part the result of an effective competition by ara-C for phosphorylation by this enzyme, leading to a perturbation of the pharmacokinetics of intracellular F-ara-ATP.