Pharmacokinetics of intrathecal gemcitabine in nonhuman primates.

PURPOSE Gemcitabine is an excellent candidate for regional therapy. We quantified cerebrospinal fluid (CSF) and plasma concentrations of gemcitabine and its inactive metabolite, 2',2'-difluorodeoxyuridine (dFdU), in nonhuman primates given intrathecal gemcitabine. EXPERIMENTAL DESIGN Three nonhuman primates received 5 mg of gemcitabine via lateral ventricle. CSF was sampled from the fourth ventricle in all of the animals and the lumbar space in one, and one had plasma sampled. One animal had ventricular CSF sampled after receiving 5 mg intralumbar gemcitabine. Gemcitabine and dFdU were measured by high-performance liquid chromatography. Three additional animals had 5 mg intralumbar gemcitabine administered weekly for 4 weeks and were monitored for toxicity. RESULTS At 37 degrees C in vitro, gemcitabine was stable in CSF. Ventricular delivery of gemcitabine produced peak ventricular CSF gemcitabine concentrations of 297 +/- 105 microg/ml. After 6 h, the concentrations were <0.03 microg/ml. Intrathecal gemcitabine was rapidly and extensively converted to dFdU. CSF dFdU concentrations increased to 82 microg/ml at 1 h and then declined to very low values by 24 h. After intraventricular administration, CSF gemcitabine and dFdU area(s) under the curve (AUC) were 251 +/- 85 and 249 +/- 88 microg/ml x h. Intralumbar gemcitabine produced lower ventricular CSF gemcitabine and dFdU concentrations than did intraventricular gemcitabine. The plasma gemcitabine AUC associated with 5 mg of intraventricular gemcitabine was 2 mg/ml x h, which was >200-fold lower than the CSF gemcitabine AUC in the same animal. Transient CSF pleocytosis was the only toxicity observed. CONCLUSIONS Our results demonstrate a large pharmacokinetic advantage of intrathecal gemcitabine and support a planned Phase I clinical trial of this dosing strategy.