Cerebrospinal fluid concentrations of vincristine after bolus intravenous dosing

A surrogate marker of brain penetration


  • Stewart J. Kellie M.B., B.S.,

    Corresponding author
    1. Oncology Unit and Department of Paediatrics and Child Health, the University of Sydney, The Children's Hospital at Westmead, Westmead, New South Wales, Australia
    • Oncology Unit, The Children's Hospital at Westmead, Sydney, Locked Bag 4001, Westmead, New South Wales 2145 Australia
    Search for more papers by this author
    • Fax +612-9845-2171

  • Draga Barbaric M.Med.,

    1. Oncology Unit, The Children's Hospital at Westmead, Westmead, New South Wales, Australia
    Search for more papers by this author
  • Pauline Koopmans B.Sc.,

    1. The Pharmacy Laboratory, University Hospital, Groningen, the Netherlands
    Search for more papers by this author
  • John Earl Ph.D.,

    1. Department of Biochemistry, The Children's Hospital at Westmead, Westmead, New South Wales, Australia
    Search for more papers by this author
  • Deborah J. Carr R.N.,

    1. Oncology Unit, The Children's Hospital at Westmead, Westmead, New South Wales, Australia
    Search for more papers by this author
  • Siebold S. N. de Graaf M.D.

    1. Department of Paediatrics, University Hospital, Groningen, the Netherlands
    Current affiliation:
    1. University Medical Center St. Radboud, Nijmegen, the Netherlands
    Search for more papers by this author



Vincristine (VCR) is used widely in oncology practice, and regular dosing is commonly associated with the development of sensorimotor or autonomic neuropathies. However, the incidence of VCR-related central nervous system (CNS) toxicity is comparatively low, suggesting that the blood-brain barrier may limit drug penetration into the brain parenchyma. This study determined whether measurable concentrations of VCR could be detected in the cerebrospinal fluid (CSF), as a surrogate marker of brain parenchyma penetration, after bolus intravenous injection in children without primary CNS pathology.


The authors studied 17 pediatric patients ages 2.5–14.1 years (median, 6.8 years) with acute lymphoblastic leukemia or non-Hodgkin lymphoma without evidence of leptomeningeal disease. Patients received VCR 1.5 mg/m2 by intravenous bolus injection followed at varying intervals by lumbar puncture for scheduled intrathecal methotrexate administration under general anesthesia. Paired VCR concentrations in both plasma and CSF were measured in each patient simultaneously at times ranging from 8 minutes to 146 minutes after the VCR injection. Three patients were studied twice. The paired samples were stored at −40 °C until analysis using a high performance liquid chromatography assay with a sensitivity of 0.1 μg/L in CSF and 0.4 μg/L in plasma.


Plasma VCR concentrations ranged from 2.2 μg/L to 91.2 μg/L. No measurable VCR concentrations were detected in the CSF samples.


Measurable concentrations of VCR in CSF are not achieved after the administration of standard intravenous bolus doses of VCR. The current observations are consistent with the relative rarity of VCR-related CNS neurotoxicity compared with the commonly observed sensorimotor and autonomic neuropathies. These findings suggest that the penetration of VCR into the brain parenchyma of patients with a relatively intact blood-brain barrier is low and that VCR may have a limited role in the CNS-directed therapy of these patients. Cancer 2002;94:1815–20. © 2002 American Cancer Society.

DOI 10.1002/cncr.10397

Vincristine (VCR) is an effective chemotherapy agent that is used widely in infants and children with hematologic malignancies and solid tumors. The dose limiting toxicity of VCR is peripheral neuropathy, characterized by progressive motor, sensory, and autonomic involvement in varying combinations. Neuropathies involving the cranial nerves, including bilateral ptosis, extraocular paresis, recurrent laryngeal and facial palsies, and optic neuropathy with visual loss, have been reported. VCR also may cause autonomic neuropathy, characterized by constipation, ileus, postural hypotension, urinary retention, and impotence. Central nervous system (CNS) neurotoxicity, characterized by inappropriate antidiuretic hormone (vasopressin) secretion, depression, insomnia, seizures, and visual hallucinations, is rare.1–5

The distribution of chemotherapy drugs into the CNS is influenced by a number of biologic and physicochemical factors, including protein binding and lipid solubility of the drug and the expression of P-glycoprotein in the blood-brain barrier.6–10 We postulated that measurable concentrations of VCR may not be achieved in the cerebrospinal fluid (CSF) in children with a relatively intact blood-brain barrier after conventional intravenous (IV) dosing. This hypothesis was tested in a cohort of children with lymphoproliferative malignancies who were receiving regular doses of IV VCR and scheduled, periodic lumbar punctures for intrathecal methotrexate.



A study sample of 17 patients was enrolled. There were 10 male patients and 7 female patients who ranged in age between 2.5 years and 14.1 years (median, 6.8 years). Pharmacokinetic studies were performed once in 14 patients and twice in 3 patients. Sixteen patients were receiving treatment for acute lymphoblastic leukemia at the time of the study: 14 patients who were in first remission and 2 patients who were in second remission. Three patients had previous evidence of CNS involvement with leukemia but were in complete remission at the time of this study. One patient with T-cell non-Hodgkin lymphoma was enrolled. Two patients were studied during induction, 4 patients were studied during CNS consolidation phase, and 11 patients were studied during remission maintenance therapy. Fourteen patients were enrolled on the Study VI of the Australian and New Zealand Children's Cancer Study Group,11 and 3 patients were on a modified version of the New York II protocol.12 Tests of hepatic and renal function were normal in all patients enrolled on study. The study was approved by the Institutional Ethics Committee at The Children's Hospital at Westmead, Sydney, and written informed consent was obtained from a parent of each patient.


Patients were studied at a time that coincided with the scheduling of intrathecal methotrexate by lumbar intrathecal injection under general anesthesia. VCR 1.5 mg/m2 (maximum dose, 2 mg) was administered at baseline by rapid IV bolus injection through a central venous catheter. The lumbar puncture was performed at variable time intervals, ranging from 4 minutes to 146 minutes after the IV VCR bolus. A sample of CSF was obtained prior to intrathecal methotrexate administration, including 1 mL of CSF collected and stored at −40 °C in glass tubes for this study. A simultaneous peripheral venipuncture was used to obtain 3–4 mL blood for a paired plasma sample for VCR assay. The central venous catheter was not used for blood sampling to eliminate any possibility of contamination by VCR adherence to the wall of the catheter. After centrifugation at 3000 rpm for 10 minutes, the plasma was separated and aliquoted into 10-mL polypropylene tubes (JHN CT32020; Selby Biolab) and stored at −40 °C. CSF and plasma samples were transferred to the University Hospital, Groningen, the Netherlands on dry ice (−80 °C) with en route icing between Sydney and Groningen.


The VCR concentration in plasma was measured using high-performance liquid chromatography (HPLC) with on-line column extraction and electrochemical detection with a sensitivity of 0. 4 μg/L.13 The sensitivity of the assay for VCR in CSF could be improved to 0.1 μg/L according to Good Laboratory Practice Rules. All analyses were performed in duplicate, and an average concentration was obtained.


Twenty paired plasma samples and CSF samples for VCR assay were obtained from 17 patients, including 3 patients who were studied twice at intervals ranging from 10 days to 6 months. Plasma VCR concentrations ranged from 2.2 μg/L to 91.2 μg/L, (median, 8.9 μg/L). The lumbar CSF concentration of VCR was < 0.1 μg/L, the sensitivity limit of the assay, in all 20 samples (Table 1). Even high peak concentrations in plasma of 91.2 μg/L and 83.1 μg/L at time (t) = 8 minutes and t = 9 minutes in Patients 1 and 2, respectively, were not associated with a CSF concentration above the lower limit of detection of 0.1 μg/L. This indicates that the ratio of plasma and CSF concentrations was > 800–900. Paired plasma and CSF samples also were obtained at later times after IV bolus injection to determine whether VCR may enter the CSF after some delay. No VCR was detectable in CSF samples that were obtained approximately 1 hour after IV injection, whereas the plasma concentration at the same time varied between 2.9 μg/L and 4.4 μg/L. Similarly, no VCR was detected in a CSF sample that was obtained at t = 146 minutes.

Table 1. Plasma and Cerebrospinal Fluid Vincristine Concentrations
SampleaGenderAge (yrs)Time after VCR bolus (minutes)CSF VCR (μg/L)Serum VCR (μg/L)
  • VCR: vincristine; CSF: cerebrospinal fluid.

  • a

    Two samples were obtained 10 days to 6 months apart from three patients: samples 5 and 9, samples 11 and 16, and samples 13 and 18.

  • b

    This patient had T-cell non-Hodgkin lymphoma.

1F6.18< 0.191.2
2F7.09< 0.183.1
3M7.812< 0.13.3
4F2.612< 0.114.9
5M3.613< 0.115.0
6M7.814< 0.114.7
7F7.615< 0.116.1
8M13.616< 0.15.8
9M3.717< 0.17.6
10bM4.018< 0.110.9
11M4.919< 0.18.2
12F8.319< 0.112.3
13M2.621< 0.19.7
14F14.126< 0.19.6
15F7.029< 0.16.7
16M4.462< 0.12.9
17M7.864< 0.14.1
18M2.667< 0.14.4
19M9.281< 0.14.1
20M3.3146< 0.12.2

Figure 1 demonstrates the HPLC chromatogram of a CSF sample that spiked in the laboratory with 2.14 μg of VCR. The chromatogram shows a significant peak at the retention time of VCR (Peak A). Peak B represents vinblastine used as an internal standard in the assay. Figure 2 shows a representative chromatogram of CSF (sample 20) that was obtained at t = 146 minutes after the injection of VCR. No VCR could be detected in the CSF sample. A representative chromatogram of plasma (sample 12) showing a VCR peak of 11.7 μg/L is demonstrated in Figure 3 (these data differ slightly from sample 12 in Table 1, because the tabulated results recorded the average of duplicate assays).

Figure 1.

Chromatogram of cerebrospinal fluid (CSF) spiked with a known concentration of vincristine (Peak A) and an internal standard, vinblastine (Peak B).

Figure 2.

Typical chromatogram of cerebrospinal fluid (CSF) (sample 20). Note the absence of a peak corresponding to the retention time of vincristine.

Figure 3.

Typical chromatogram of plasma (sample 12).


In this study, plasma VCR concentrations ranged from 91.1 μg/L at t = 8 minutes to 2.2 μg/L at t = 146 minutes. These concentrations are consistent with earlier studies of VCR disposition after conventional IV bolus dosing in pediatric patients with malignant disease.1, 14, 15 No measurable CSF concentrations of VCR were detected in any patient samples, ranging from samples that were drawn shortly after VCR injection, which were associated with high plasma levels, to delayed CSF samples that were obtained more than 2 hours after IV injection. This is further evidence that our negative results were not due simply due to slow transmembrane transport of VCR into CSF.

The current results are partly in agreement with the observations by Jackson et al.,16 who studied the plasma and CSF pharmacokinetics of VCR using a radioimmunoassay methodology in six adults with leukemia or lymphoma, all of whom had a history of meningeal involvement, including two patients who were treated with cranial irradiation. Those investigators reported concentrations of VCR equivalents in CSF ranging from 0.8 × 10−9 M to 1.1 × 10−9 M (0.7–1.0 μg/L) after a 2 mg VCR IV bolus dose in two patients and a spot CSF concentration of VCR of 2.6 × 10−9 M (2.4 μg/L) in a third patient with active meningeal leukemia who received concurrent cranial irradiation. The sensitivity of the radioimmunoassay in that study reportedly was 5 × 10−10 M to 1 × 10−9 M (0.45–0.9 μg/L) VCR equivalents.17 No detectable CSF concentrations of VCR were detectable in three patients, and the authors speculated that active meningeal disease or concurrent irradiation may have accounted for the highest of the CSF VCR concentrations observed. Those authors concluded that the penetration of VCR into CSF in humans is poor, and the low or negligible concentrations of VCR are unlikely to be lethal to tumor cells.16, 18 The extent to which CSF VCR concentrations reflect peritumoral drug concentrations is not known. The authors used the term VCR equivalents to indicate a lack of specificity, because their radioimmunoassay measured total parent drug, metabolites, and degradation products. A VCR concentration of this magnitude would have been detected by our HPLC assay, which measures only parent drug to a detection limit of 0.1 μg/L. Although cytotoxicity is determined by the retention of intracellular VCR above a threshold concentration until the cell cycle proceeds to mitosis,19 our results suggest that inadequate CSF penetration may limit the efficacy of VCR in CNS-directed therapy in children with relatively intact blood-brain barriers.

The correlations between plasma drug concentration, CSF penetration, and CNS parenchymal penetration require further study. The delivery of chemotherapy drugs into the CNS and CSF is influenced by a number of biologic and physicochemical factors, including protein binding and lipid solubility of the drug; the expression of P-glycoprotein in the blood-brain barrier; and, in patients with CNS neoplasms, local tumor blood flow, interstitial pressure, and the pathophysiology of the microvasculature adjacent to the tumor.6–10 In patients with an intact blood-brain barrier and blood-CSF barrier, CSF:plasma ratios of < 0.05 have been reported earlier in chemotherapy agents used commonly in pediatric oncology (Table 2).

Table 2. Cerebrospinal Fluid to Plasma Ratios of Chemotherapy Agents Commonly Used in Pediatric Oncology
DrugCSF: plasma ratioReference
  1. CSF: cerebrospinal fluid; ND: not detected.

Methotrexate0.01–0.03Thyss et al.32
Cytosine arabinoside0.06–0.25Slevin et al.33
Etoposide0.018Handke et al.34
L-asparaginaseNDBalis et al.35
DaunorubicinNDBalis et al.35
Thiotepa1.0Heideman et al.36
Cisplatin0.029DeGregorio et al.37
Vincristine0.05Balis et al.35
NDCurrent study

The current results may have significance for the use of VCR in selected indications in pediatric neuro-oncology. VCR is used commonly as adjuvant therapy (or in prolonged maintenance) in pediatric neuro-oncology protocols enrolling patients with completely resected tumors or in children with low-grade gliomas.20–23 The extent to which the blood-brain barrier is disrupted in patients with previously resected CNS tumors is not known. We speculate that the blood-brain barrier may be relatively intact in patients with previously resected CNS tumors who are receiving adjuvant or maintenance chemotherapy. Although the duration of exposure of tumor cells to low concentrations of drug may contribute to cytotoxicity,18, 24 the absence of detectable VCR in the CSF of patients in whom the paired plasma concentrations are 800–900 times higher suggests that cytotoxic VCR levels are unlikely to be achieved in the CSF or peritumoral tissues in patients if the blood-brain barrier is not disrupted significantly.

The clinical data supporting the efficacy of single-agent VCR in the treatment of pediatric patients with CNS tumors is scarce and is based mostly on subjective clinical or computed tomography scan assessments of response.25–31 The current results question the extent of CSF penetration by VCR in patients in whom it is believed that the blood-brain barrier is relatively intact and also raise the possibility that the therapeutic contribution of VCR to these patients may be less than anticipated.