The pharmacokinetics of cytosine arabinoside in the plasma and cerebrospinal fluid during conventional and high-dose therapy.

Abstract

Subcutaneous bolus administration of cytosine arabinoside (Ara-C) has been used as an alternative to continuous intravenous infusion. However, the pharmacokinetics of subcutaneous bolus Ara-C have not been satisfactorily documented in man. The plasma concentrations of Ara-C were therefore compared following intravenous bolus injection, subcutaneous bolus injection, and intravenous infusion in five patients with acute myelogenous leukemia. A triphasic decline in plasma Ara-C levels was seen after intravenous bolus administration. Subcutaneous bolus Ara-C was rapidly absorbed (mean absorption half-life 3.4 ± minutes) and then declined biexponentially with initial and terminal half-lives similar to intravenous bolus injection. Ara-C concentrations were above the steadystate infusion levels for only 40 minutes following the intravenous bolus and for 100 minutes following subcutaneous bolus administration. The decline in plasma Ara-C concentrations continued rapidly following both these routes of administration and after five hours were approximately 10% of the steady-state infusion levels. This study demonstrates that it is not possible to achieve comparable steadystate levels of Ara-C with the same total dose given by twice or thrice daily subcutaneous bolus injection as by continuous intravenous infusion. High-dose Ara-C has been used to treat patients with relapsed acute leukemia resistant to Ara-C in conventional doses and may provide a means of overcoming such resistance. Significant concentrations of Ara-C are achieved in the cerebrospinal fluid (CSF) during conventional-dose intravenous infusions, and it seemed likely that much greater levels would be reached during these high-dose infusions. The pharmacokinetics of Ara-C in the plasma and CSF during high-dose intravenous infusions (1 g/m2) has therefore been investigated. Five patients were treated with a three-hour infusion preceded by a loading dose, and five patients received a one-hour infusion with no loading dose. The CSF Ara-C concentrations were similar during both infusions, and the concentrations in the CSF at the end of the infusions were higher than the plasma Ara-C levels obtained with intravenous infusion at conventional doses. In addition CSF Ara-C levels were measured in a patient with an Ommaya reservoir during intravenous infusion of 100 mg/m2 and 1 g/m2. It was demonstrated that Ara-C equilibrated rapidly betweenthe plasma and CSF and crossed the blood brain barrier as effectively at 1 g/m2 as at 100 mg/m2. Extrapolation of these data suggests that during therapy with high-dose Ara-C at a dose of 3 g/m2 12 hourly, Ara-C levels in the CSF in excess of 100 ng/ml (410 nmol/liter are likely to be maintained almost continuously, and should therefore be effective therapy for CNS leukemia.