1-beta-D-Arabinofuranosylcytosine (ara-C) is used empirically at a low, conventional, or high dose. Ara-C therapy may be optimal if it is directed by the clinical pharmacokinetics of the intracellular active metabolite of ara-C, 1-beta-D-arabinofuranosylcytosine 5'-triphosphate (ara-CTP). However, ara-CTP has seldom been monitored during low- and conventional-dose ara-C therapies because detection methods were insufficiently sensitive. Here, with the use of our newly established method (Cancer Res., 56, 1800 -- 1804 (1996)), ara-CTP was monitored in leukemic cells from acute myelogenous leukemia patients receiving low- or conventional-dose ara-C [subcutaneous ara-C administration (10 mg / m(2) ) (3 patients), continuous ara-C infusion (20 or 70 mg / m(2) / 24 h) (7 patients), 2-h ara-C infusion (70 mg / m(2) ) (4 patients), and 2-h infusion of N(4)-behenoyl-1-beta-D-arabinofuranosylcytosine, a deaminase-resistant ara-C derivative (70 mg / m(2) ) (6 patients)]. Ara-CTP could be determined at levels under 1 microM. There was a close correlation between the elimination half-life values of the plasma ara-C and the intracellular ara-CTP. The presence of ara-C in the plasma was important to maintain ara-CTP. The continuous ara-C and the 2-h N(4)-behenoyl-1-beta-D-arabinofuranosylcytosine infusions maintained ara-CTP and the plasma ara-C longer than the subcutaneous ara-C or the 2-h ara-C infusion. They also afforded relatively higher ara-CTP concentrations, and consequently produced ara-CTP more efficiently than the 2-h ara-C infusion. Different administration methods produced different quantities of ara-CTP even at the same dose.