The pharmacokinetic values of 1-beta-D-arabinofuranosylcytosine (ara-C) in plasma and its active metabolite 1-beta-D-arabinofuranosylcytosine 5'-triphosphate (ara-CTP) in circulating blast cells were studied in 11 patients with acute leukemia. ara-C was administered as a 2-h infusion (3 g/m2) followed in 12 to 24 h by a continuous infusion for 4 days in 10 patients and for 7 days in one. A steady-state concentration of ara-C in plasma (94 +/- 32 microM) was reached by the end of the 2-h infusion. Its elimination was biphasic with an initial and terminal t1/2 of 0.44 +/- 0.10 h and 2.8 +/- 0.9 h, respectively. The accumulation of ara-CTP in leukemic cells was linear and continued for up to 2 h after the bolus infusion. ara-CTP elimination was monophasic with a median t1/2 of 3.4 h (range, 1.25 to 18.9 h). The disposition of ara-C and 1-beta-D-arabinofuranosyluracil during continuous infusion was linear with dose rate over the dose range of 70 to 3000 mg/m2/day. The area under the concentration versus time curve for ara-CTP in leukemic cells was not related to the dose infused, but rather appeared to be intrinsic to the cells of each individual. As a general finding, the pharmacokinetic values of ara-CTP in circulating blasts were more heterogeneous than those of ara-C in plasma. There were marked differences in the absolute concentrations of ara-C in plasma and ara-CTP in leukemic cells at different times after the bolus infusion and also during continuous infusion. No correlation was evident between the determinants of ara-C pharmacokinetic values and those of ara-CTP. Thus, it is concluded that the pharmacokinetics of ara-C in plasma cannot predict for the metabolism of ara-CTP in leukemic cells.