Cerebrospinal Fluid Drug Concentrations and Clinical Outcome of Patients with Neoplastic Meningitis Treated with Liposomal Cytarabine.

BACKGROUND AND OBJECTIVE: Liposomal cytarabine is a slow-release formulation for intrathecal application in patients with neoplastic meningitis. Although standard dosing intervals range from 2 to 4 weeks, it is unclear whether sustained cytotoxic cerebrospinal fluid (CSF) concentrations can be achieved beyond 14 days from drug injection. The objective of this study was to assess CSF and plasma concentrations of liposomal cytarabine more than 2 weeks after lumbar drug administration and to correlate those findings with clinical outcome. METHODS: 66 matched CSF and plasma drug concentrations were analyzed by a validated liquid chromatography-tandem mass spectrometry method starting at day 13 from lumbar drug injection in 19 patients with neoplastic meningitis treated with liposomal cytarabine. CSF drug concentrations were correlated with clinical outcome. RESULTS: Overall response rate was 63.2% (12/19). 100% (9/9) of patients with positive CSF cytology at diagnosis achieved cytological complete remission, and none of the patients (0/19) experienced on-drug disease progression. In responding patients with controlled systemic disease, CNS-specific progression-free survival was 14 months (n = 4; range 5-25 months). The median CSF concentration of free cytarabine was 156 ng/ml (range 5-4581 ng/ml) and 146 ng/ml (range 5-353 ng/ml) in samples withdrawn at days 13-16 and at days 25-29 after intrathecal drug injection, respectively. Free cytarabine concentrations > 100 ng/ml were detected in 58.8% (20/34) and 53.3% (7/13) of the CSF samples obtained at days 13-16 and days 25-29, respectively. CSF drug concentrations did not differ significantly between responding and nonresponding patients. CONCLUSION: Liposomal cytarabine permits prolonged CSF drug exposure, with cytotoxic cytarabine concentrations that are detectable for 4 weeks in the majority of patients. The preserved clinical activity seen in patients with inferior CSF drug concentrations (< 100 ng/ml) suggests that maintaining lower cytarabine concentrations for a longer period of time may be similarly effective as using short peak concentrations.

[Cerebrospinal fluid examination after liposomal cytarabine intrathecal injection]. / Examen cytologique d'un liquide céphalorachidien après administration intrathécale de cytarabine liposomale.

The patient is a 36 year old female who presented breast cancer with leptomeningeal involvement. A systematic lumbar puncture was performed and sent to the laboratory for CSF analysis. CSF examination using wet mount preparation showed a large number of round spherules. After discussion with the ordering physician, we learnt that the patient had received intrathecal liposomal cytarabine injection 19 days earlier. Cytarabine liposomes are spherules with a granular interior and range in size from 10-30 µm. It can be confused with leukocytes and lead to spurious elevation of CSF leukocytes count. Care needs to be taken in interpreting CSF results in patients who have received intrathecal liposomal cytarabine.

Accidental overdose of intrathecal cytarabine in children.

OBJECTIVE: To report 2 cases of intrathecal cytarabine overdose in children with cancer, neither of whom underwent cerebrospinal fluid (CSF) exchange per current recommendations or developed apparent toxicity related to the event. CASE SUMMARY: A 17-year-old female with newly diagnosed acute myeloid leukemia received 177 mg of intrathecal cytarabine rather than the appropriate dose of 70 mg. She was monitored closely with no apparent toxicity from the event. A 4-year-old boy with newly diagnosed precursor B-cell acute lymphoblastic leukemia received 175 mg of intrathecal cytarabine rather than the appropriate dose of 70 mg. CSF was immediately withdrawn and intrathecal hydrocortisone was instilled for possible antiinflammatory effect. He developed no apparent toxicity from the event. DISCUSSION: Cytarabine is an important chemotherapeutic agent in the treatment of leukemia. One case report of intrathecal cytarabine overdose was identified in the literature, which recommended CSF exchange as management. Neither child in our report underwent CSF exchange or developed apparent toxicity related to the event. Institutional changes were made in both cases to prevent similar events. CONCLUSIONS: These cases demonstrate that measures such as CSF exchange are not uniformly required for cytarabine overdose.

Sequential arabinosylcytosin with or without fludarabine in paracmastic patients with acute myeloid leukemia.

The purpose of this study is to assess how fludarabine (Fa) influences arabinosylcytosin's (Ara-C) mode of action. Plasma, cerebrospinal and urine samples were withdrawn from two study groups at specific time points and analyzed by HPLC. Group A was treated with Ara-C only whereas Group B was treated with Fa+Ara-C. The two study groups are all undergoing complete remission (CR). The Ara-C dose for Group A was 3g/m2 x 2, and the AUC(0-4) was 5.131 +/- 0.936. The Ara-C dose for Group B was 2g/m2 x 2, and the AUC(0-4) was 12.245 +/- 3.863. The AUC(0-4) for Group B is more than twice the AUC(0-4) for Group A, and these results indicate that Fa conduces a synergistic increase in the concentration and AUC of Ara-C in plasma and in cerebrospinal fluid. The pharmacokinetics between the different dose treatments was statistically different (P = 0.016). The differences in the ratios of C(Ara-u) to C(Ara-C), and in the Tmax between Groups A and B could indicate whether or not Ara-C combined with Fa. Although Group B demonstrates a higher AUC(0-4) with lower doses of Ara-C (2 g/m2 x 2), the adverse drug reaction (ADR) and bone inhibition were not more pronounced in Group B compared to Group A. These results are based on a limited number of case studies, hence, additional studies are necessary to support and prove this hypothesis.

Flow cytometric detection of liposomal cytarabine in cerebrospinal fluid of patients treated with intrathecal chemotherapy.

We report the unusual flow cytometric detection of liposomes in the cerebrospinal fluid of patients with aggressive non-Hodgkin's lymphoma and acute myeloid leukemia, treated with intrathecal liposomal cytarabine.

Management of leptomeningeal malignancy.

Leptomeningeal carcinomatosis is defined as malignant infiltration of the pia matter and arachnoid membrane. Leukaemias and lymphomas, lung, breast cancer and melanoma are the primary tumours commonly associated with leptomeningeal carcinomatosis. Diagnosis is based on compatible symptoms and signs, cytological evidence of malignancy in the cerebrospinal fluid, and neuroimaging studies. Treatment is largely palliative (median survival 2-4 months). Patients with lympomatous or leukaemic meningitis, chemosensitive tumours such as breast cancer, low tumour burden, minimal neurological deficits, good performance status and controllable systemic disease survive longer with occasional long-term responses. Available treatment options include focal radiation therapy to CNS sites of bulky, symptomatic or obstructive meningeal deposits, intrathecal cytotoxic therapy and systemic chemotherapy. No evidence of superiority of intrathecal treatment compared with best palliative care (including radiation therapy and systemic treatment) is available from clinical trials. Novel treatment approaches include intrathecal liposomal Ara-C, the development of new cytotoxic compounds, signal transduction inhibitors and monoclonal antibodies for intrathecal or systemic use. Until data from multi-centre randomised trials are available, rationalisation of therapy should be done by stratifying patients to prognostic groups. High-risk patients will only survive for a few weeks and are better managed with supportive measures, whereas low-risk patients justify vigorous cerebrospinal fluid-directed treatment combined with radiation therapy and systemic chemotherapy.

Pharmacokinetics of nimustine, cytosine arabinoside, and methotrexate in cerebrospinal fluid during cerebrospinal fluid perfusion chemotherapy.

This report investigates the pharmacokinetics of nimustine (ACNU), cytosine arabinoside (Ara-C), and methotrexate (MTX) in cerebrospinal fluid (CSF) during CSF perfusion chemotherapy. A 47-year-old Japanese man with spinal cord, cerebellum and brain stem dissemination of oligo-astrocytoma received nine courses of CSF perfusion chemotherapy with ACNU, Ara-C, and MTX. A CSF perfusion chemotherapy solution was perfused via an Ommaya reservoir in the ventricle, and was discharged by drainage though another Ommaya reservoir in the lumbar spinal canal. CSF samples via Ommaya reservoirs in the lumbar spinal canal were obtained during the fifth and eighth courses of treatment. The concentrations of ACNU and Ara-C in CSF were measured by HPLC, and the MTX concentrations by fluorescence polarization immunoassay. In the fifth course of treatment, a CSF injection chemotherapy solution, consisting of 5 mg of ACNU dissolved in 20 ml of artificial CSF, was injected over a few minutes using the Ommaya reservoir. Next, a CSF perfusion chemotherapy solution, consisting of 10 mg of Ara-C and 5 mg of MTX dissolved in 100 ml of artificial CSF, was perfused over 2 h. In the eighth course of treatment, a CSF perfusion chemotherapy solution, consisting of 5 mg of ACNU, 10 mg of Ara-C and 5 mg of MTX dissolved in 100 ml of artificial CSF, was perfused over 2 h. In both treatments, the highest concentrations of Ara-C and MTX in CSF were observed 1 or 2 h after the end of perfusion, with the values of each drug being similar. The CSF AUCs of Ara-C and MTX in each treatment were of similar values. Although the highest concentration of ACNU in CSF was observed in the fifth treatment 1 h after injection (an injection chemotherapy of ACNU plus a perfusion chemotherapy of Ara-C and MTX), the concentration of ACNU in CSF was undetectable in the eighth treatment (a perfusion chemotherapy of ACNU, Ara-C and MTX). We were successful in administering all anticancer drugs, and reaching a level of over 1.0 microg/ml concentration in CSF of the lumbar spinal canal, using an injection chemotherapy of ACNU plus a perfusion chemotherapy of Ara-C and MTX; this was done even though the drugs, in particular ACNU, underwent some perfusion-period dependent decomposition.

Pharmacokinetics of cytosine arabinoside, methotrexate, nimustine and valproic acid in cerebrospinal fluid during cerebrospinal fluid perfusion chemotherapy.

This report investigates the pharmacokinetics of cytosine arabinoside (Ara-C), methotrexate (MTX), nimustine (ACNU) and valproic acid (VPA) in cerebrospinal fluid (CSF) during CSF perfusion chemotherapy. A 28-year-old Japanese woman with disseminated glioblastoma was, on admission, on a stable oral regimen of prolonged-release VPA tablets (Depakene-R), 400 mg twice a day, for seizure control. Twelve courses of CSF perfusion chemotherapy with Ara-C, MTX, and ACNU were administered. Plasma samples and CSF samples via Ommaya reservoirs were obtained during the eleventh course of treatment. The Ara-C and ACNU concentrations were measured by HPLC. The MTX and VPA concentrations were measured by fluorescence polarization immunoassay. During CSF perfusion chemotherapy, the highest CSF concentrations of Ara-C, MTX, and ACNU were observed at the end of the perfusion and decreased in a monoexponential pattern. The half-lives of Ara-C, MTX, and ACNU were 2.65, 3.52, and 0.71 h, respectively. No anticancer drugs were detectable in plasma during CSF perfusion chemotherapy. Before CSF perfusion chemotherapy, the free VPA concentration in plasma was 14.4% of the total VPA concentration. The mean total and free VPA concentrations in plasma were 78.0+/-0.8 and 10.9-0.3 microg/ml, respectively. The free VPA concentrations in plasma and in CSF were of similar values. At the end of perfusion, the lowest free VPA concentration in CSF was 30.3% of that at the initiation of perfusion. The free VPA concentrations in CSF at 3, 7, 23, and 47 h after the end of perfusion were 79.8, 94.5, 100.9, and 100.9% respectively of that at the initiation of perfusion. During CSF perfusion chemotherapy, the ratio of free VPA concentrations to the total VPA in CSF was 86.3+/-6.9%. The VPA concentrations in CSF rapidly decreased during the CSF perfusion but recovered to pre-treatment levels within 7 h.

Pharmacokinetics of anticancer drugs in cerebrospinal fluid.

OBJECTIVE: To examine the pharmacokinetics of anticancer drugs in the cerebrospinal fluid (CSF) during chemotherapy by the lumbar-ventricular (LV) and ventricular-lumbar (VL) routes. CASE SUMMARY: A 69-year-old Japanese woman with disseminated glioblastoma received two LV and four VL courses of CSF perfusion chemotherapy with methotrexate, nimustine, and cytarabine hydrochloride. Samples of CSF from the ventricles and lumbar spinal canal were obtained via Ommaya reservoirs during one LV and one VL course. Drug concentrations in the CSF were measured by fluorescence polarization immunoassay or HPLC. RESULTS: During LV CSF perfusion, the highest CSF drug concentrations in both the ventricles and the lumbar spinal canal were observed at the end of perfusion. During treatment, the concentrations of all three drugs in the lumbar spinal canal were higher than those in the ventricles. The CSF AUC of methotrexate in the ventricles was 16.1% of that in the lumbar spinal canal. During VL CSF perfusion, the highest drug concentrations were also observed at the end of perfusion. The drug concentrations in the lumbar spinal canal were initially lower than those in the ventricles. However, the concentrations of methotrexate and cytarabine in the lumbar spinal canal exceeded those in the ventricles 3 hours after perfusion. The AUC of methotrexate in the lumbar spinal canal was 174.9% of that in the ventricles. CONCLUSIONS: The pharmacokinetics of anticancer drugs in ventricular CSF differ from those in lumbar CSF during LV and VL perfusion chemotherapy.

Pharmacokinetics of intralumbar DTC-101 for the treatment of leptomeningeal metastases.

PURPOSE: To study the results of intralumbar administration of chemotherapeutic agents for the treatment of neoplastic meningitis and to determine how these results relate to the variable and often subtherapeutic drug concentrations in the ventricular cerebrospinal fluid (CSF) compartment. Ventricular and lumbar pharmacokinetic studies were done following intralumbar administration of DTC-101, an extended-release formulation of cytarabine. PATIENTS AND METHODS: Nine patients (age range, 23 to 67 years; median age, 42 years) with leptomeningeal metastases were treated with 18 courses of intralumbar DTC-101. Eight patients who were treated with 14 courses underwent pharmacokinetic CSF sampling from the lumbar sac and lateral ventricle. Cytarabine concentrations were determined by using high-pressure liquid chromatography. RESULTS: Following intralumbar administration of DTC-101, therapeutic free cytarabine concentrations were achieved rapidly in both of the ventricular and lumbar CSF compartments and maintained for 2 weeks. The mean pharmacokinetic parameters of free cytarabine in the lumbar and ventricular CSF compartments were as follows: maximum concentration, 226 and 6.06 mg/L; half-life, 277 and 130 hours; and area under the concentration vs time curve, 4120 and 598 micrograms/h per milliliter, respectively. CONCLUSIONS: Intralumbar administration of DTC-101 results in extended cytotoxic free cytarabine concentrations in both of the lumbar and ventricular regions of CSF and allows an every-other-week drug-dosing schedule.

Alkylsulphonic acid ion pairing with radial compression columns for determining plasma or cerebrospinal fluid 1-beta-D-arabinofuranosylcytosine in pediatric pharmacokinetic analysis.

In order to accurately and precisely measure plasma and cerebrospinal fluid (CSF) 1-beta-D-arabinofuranosylcytosine (Ara-C) in pediatric samples with adequate sensitivity and without interference, we have developed a reversed-phase ion-pairing technique utilizing the free amino group of the pyrimidine ring of Ara-C. Optimum resolution and separation was achieved utilizing a 4 microns C18 radial compression column. Ara-C and the internal standard 8-bromo-cyclic-AMP eluted at 6.5 and 4.6 min, respectively, with complete resolution. The minimum detectable amount is 2.5 pmol in a 50-microliters volume. The assay was linear in both plasma and CSF. Intra- and inter-day assay precision were less than 4% and 9%, respectively, for plasma with similar results obtained for CSF. Neither endogenous compounds nor commonly co-administered drugs interfere. Validity for our method was supported by the successful assay of over 400 pediatric plasma and 50 CSF samples for pharmacokinetic analysis. The method offers accuracy, precision, sensitivity and efficiency for plasma or CSF Ara-C determination.

Extended CSF cytarabine exposure following intrathecal administration of DTC 101.

PURPOSE: The aims of the present study were to determine ventricular and lumbar CSF pharmacokinetics and the maximum-tolerated dose (MTD) of DTC 101 (DepoFoam; DepoTech Corp, La Jolla, CA) following intraventricular administration. PATIENTS AND METHODS: Twelve patients with neoplastic meningitis were treated with escalating doses of DTC 101. CSF samples were obtained from the right lateral ventricle or from the lumbar subarachnoid space and cytarabine concentrations were determined by high-performance liquid chromatography. RESULTS: Therapeutic ventricular CSF concentrations were maintained for 9 +/- 2 days following administration of a single dose of DTC 101 into the lateral ventricle. Lumbar cytarabine concentrations became equal to those in the ventricle within the first 6 hours after intraventricular injection, and the subsequent decay in concentrations of free and total cytarabine were the same at both sites. Following intralumbar administration, the peak ventricular concentration of free cytarabine was reached within 1 day, and therapeutic ventricular CSF levels were maintained for several days. Therapeutic intralumbar concentration of free cytarabine was maintained for up to 14 days. The MTD was 75 mg of DTC 101, and seven of nine patients manifested cytologic responses. CONCLUSION: Extended CSF exposure to therapeutic cytarabine concentrations was achieved after a single intraventricular or intralumbar dose of DTC 101, permitting drug administration once every 2 weeks.

Prolongation of drug exposure in cerebrospinal fluid by encapsulation into DepoFoam.

Prolonged maintenance of a therapeutic drug concentration in the cerebrospinal fluid is required for optimal treatment of leptomeningeal leukemia or carcinomatosis with cell cycle-specific antimetabolites. The pharmacokinetics of 1-beta-D-arabinofuranosylcytosine (ara-C) encapsulated into DepoFoam (Depo/Ara-C) was studied in six rhesus monkeys after intrathecal injection into the lumbar sac. Following a single 2-mg dose, the Depo/Ara-C concentration decreased biexponentially with initial and terminal half-lives of 14.6 and 156 h, respectively. The free drug concentration remained above the reported minimal cytotoxic level of 0.1 micrograms/ml (0.4 microM) for more than 672 h (28 days). In contrast, the half-life of ara-C following an intralumbar bolus dose of unencapsulated drug in a single animal was 0.74 h. A single intrathecal injection of Depo/Ara-C can maintain a therapeutic drug concentration in the cerebrospinal fluid for a very prolonged period.

Treatment of leptomeningeal metastasis with intraventricular administration of depot cytarabine (DTC 101). A phase I study.

Depo cytarabine (DTC 101 [formerly identified as Depo/Ara-C]) is a slow-releasing, depot formulation in which cytarabine is encapsulated within the aqueous compartments of microscopic (DepoFoam) particles. A phase I trial of DTC 101, given intraventricularly, was conducted in patients with leptomeningeal metastasis. Nine patients were given 1 to 7 cycles of DTC 101 in doses ranging from 25 to 125 mg that were administered via an Ommaya reservoir into the lateral ventricle. The dose-limiting toxic reaction was encephalopathy that occurred at the 125-mg dose level. All toxic episodes but one were transient and reversible, with the total duration of toxicity lasting from 1 to 7 days. The ventricular concentration of free cytarabine released from DTC 101 into cerebrospinal fluid decreased biexponentially with an initial half-life of 7.2 +/- 1.7 (+/- SEM) hours and a terminal half-life of 140 +/- 49 hours. The cerebrospinal fluid was cleared of malignant cells within 3 weeks of initial therapy in five of six cytologically evaluable patients. The duration of response ranged from 2 to more than 14 weeks, with a median of over 11 weeks. In conclusion, DTC 101 appears to be a pharmacologically attractive agent for use against leptomeningeal metastasis. The toxic episodes that occur with this therapy are well tolerated by patients.

Plasma and cerebrospinal fluid pharmacokinetics of 1-beta-D-arabinofuranosylcytosine and 1-beta-D-arabinofuranosyluracil following the repeated intravenous administration of high- and intermediate-dose 1-beta-D-arabinofuranosylcytosine.

We examined the plasma and cerebrospinal fluid (CSF) pharmacokinetics of 1-beta-D-arabinofuranosylcytosine (ara-C) and 1-beta-D-arabinofuranosyluracil (ara-U) in 19 patients with acute leukemia in order to determine whether ara-C or ara-U accumulate in these fluid compartments over time. Plasma and CSF samples were obtained just prior to the conclusion of the first and seventh, and immediately before the second and eighth, 2-h, twice-daily i.v. infusions of 3 g/m2/dose of ara-C (n = 10), 2 g/m2/dose of ara-C (n = 3), and 0.75 g/m2/dose of ara-C (n = 6). There was no accumulation of ara-C in the plasma or CSF, or of ara-U in the plasma following repeated ara-C infusions, ara-U did accumulate in the CSF; the end-dose 1/end-dose 7 CSF ara-U ratio was 0.35 +/- 0.12 and significantly different from this ratio for CSF ara-C (2.10 +/- 3.01; P = 0.004). The end-dose 7 CSF ara-U level was greater than the end-dose 1 CSF ara-U level in all paired specimens. There was a significant correlation between the dose of ara-C administered and the end-dose plasma ara-C and the end-dose CSF ara-U levels (P less than 0.02). One patient who received 3 g/m2/dose of ara-C developed neurotoxicity; his plasma and CSF ara-C and ara-U levels were not extraordinary during the period of ara-C administration, but he had persistent CSF ara-U demonstrable 75 h after his final ara-C dose. CSF ara-U accumulation might underlie the pathophysiology of ara-C-induced neurotoxicity. Intermediate doses of ara-C given i.v. (0.75 g/m2/dose over 2 h) appeared to generate therapeutic CSF ara-C levels and cleared CSF leukemia in one patient.

Plasma and cerebrospinal fluid pharmacokinetics of cytosine arabinoside in dogs.

Cytosine arabinoside (ara-C) is a component of many protocols for the treatment of CNS (central nervous system) leukemia and lymphoma in humans and dogs. It is also used for the prophylaxis of CNS metastasis in acute lymphoblastic leukemia. Although ara-C enters the cerebrospinal fluid (CSF) of human cancer patients after i.v. administration, it is unclear whether a similar CNS distribution occurs in humans whose blood-brain barrier has not been compromised by invasive disease. No information on the penetration of ara-C into the CSF in dogs is available. We studied the plasma and CSF pharmacokinetics of 600 mg/m2 ara-C in ten healthy male dogs after its administration as a rapid i.v. bolus (six dogs) or as a 12-h i.v. infusion (four dogs). Ara-C concentration in blood and CSF samples was determined by high-performance liquid chromatography (HPLC). After an i.v. bolus of ara-C, the mean plasma distribution half-life was 7.1 +/- 4.5 min and the mean elimination half-life was 69 +/- 28 min. The mean plasma clearance was 227 +/- 125 ml min-1 m-2. The peak concentration of ara-C in the CSF was 29 +/- 11 microM, which occurred at 57 +/- 13 min after the ara-C bolus. The CSF elimination half-life was 113 +/- 26 min. During a 12-h infusion of ara-C (50 mg m-2 h-1), the plasma steady-state concentration was 14.1 +/- 4.2 microM, the CSF steady-state concentration was 8.3 +/- 1.1 microM, and the CSF: plasma ratio was 0.62 +/- 0.14. The plasma elimination half-life was 64 +/- 19 min and the plasma clearance was 214 +/- 69 ml min-1 m-2. The CSF elimination half-life was 165 +/- 28 min. No clinically significant toxicity was observed over a 21-day period following drug administration in either of the treatment groups. Our data indicate that ara-C crosses the blood-brain barrier in normal dogs and that i.v. administration of this drug has potential as a treatment modality for neoplasia involving the CNS.

[Clinical effect and pharmacokinetics of intermediate dose Ara-C therapy in a patient with acute non-lymphocytic leukemia with two CNS recurrences].

A case of two repeated CNS recurrences of acute non-lymphocytic leukemia (M2) was treated with intermediate dose Ara-C therapy and achieved 2 complete remissions. The clinical effect and pharmacokinetics of intermediate dose Ara-C therapy in this patient were discussed. A 55-year-old male with acute non-lymphocytic leukemia (M2) achieved complete remission by combination chemotherapy of Behenoyl-ara-C, Daunorubicin, 6-Mercaptopurine and Prednisolone in July, 1985. He subsequently received consolidation and intensification therapy with periodical intrathecal injection of Methotrexate (MTX), but 13 months later he developed his first CNS recurrence which was resistant to the intrathecal administration of Ara-C and MTX. As he also relapsed systemically, Ara-C was administered in intermediate dose (1 g/m2 every 12 hrs for 5 days) and he achieved complete remission both in the CNS and systemic manifestations. Six months later he was diagnosed as having a second CNS recurrence and another systemic relapse. Intermediate dose Ara-C was administered again, and he achieved complete remission in the CNS and partial remission in systemic manifestations. Pharmacokinetic study revealed high peaks of Ara-C concentration in plasma (6.2 microM immediately after the end of the infusion) and high degree of its penetration into the CNS (5.6 microM at 3 hr after the end of the infusion) suggesting the effective and perhaps a uniform level of Ara-C is achieved throughout the CNS by this therapy. In 3 other patients without CNS involvement 0.88 +/- 0.44 microM of Ara-C, which is enough concentrations for its cytostatic effect, was detected at 3 hr after the end of infusion, suggesting the efficacy of the therapy for CNS prophylaxis. In this case the relapse occurred after repeated administration of antileukemic drugs, including Behenoyl-ara-C, an analog of Ara-C, and was resistant to the intrathecal administration of Ara-C. These findings suggest that intermediate dose Ara-C therapy was effective to overcome a resistance to antileukemic drugs, including Ara-C, and also, in some cases, more effective than intrathecal injection of antileukemic drugs for the treatment of CNS leukemia.