Vincristine (VCR) is widely used to treat patients with malignant disease; among the patients treated with VCR are children with brain tumors. In vitro studies have demonstrated that the cytotoxic activity of VCR is related to both extracellular concentration and duration of exposure. The attainment of higher plasma concentrations by injecting larger bolus doses of VCR has been limited by concerns about neurotoxicity. One possible alternative strategy for enhancing the antitumor efficacy of VCR involves prolonging the duration of in vivo exposure. Therefore, the authors explored the neurotoxicity and pharmacokinetics of VCR administered via a 96-hour continuous infusion after administration of a conventional bolus dose in a pediatric population.
The current study included 16 patients, 11 of whom were males. The median age of the study population was 4.8 years (range, 1.7–15.8 years). The diagnoses included intrinsic pontine glioma (n = 4), ependymoma (n = 5), astrocytoma (n = 3), medulloblastoma/primitive neuroectodermal tumor (PNET; n = 2), ganglioglioma (n = 1), and choroid plexus carcinoma (n = 1). Of the 16 patients, 5 were newly diagnosed, and the remaining 11 had disease recurrences, 8 of which arose after radiotherapy. Treatment included cyclophosphamide 65 mg/kg administered intravenously over 1 hour on Day 1, a bolus of VCR 1.5 mg/m2 administered intravenously on Day 2, and VCR 0.5 mg/m2 per 24 hours administered via continuous intravenous infusion on Days 2–5. Thus, a total VCR dose of 3.5 mg/m2 was administered via infusion over 4 days. Fifteen patients received 2 courses of treatment at 21–28-day intervals, and a total of 31 treatment courses were administered. VCR concentrations in plasma samples were measured using high-performance liquid chromatography.
Jaw pain, constipation, mild abdominal pain, and depressed reflexes were common. However, only 1 of 31 courses was associated with Grade III toxicity, and no Grade IV toxicity (e.g., cranial nerve palsy, ileus, inappropriate antidiuretic hormone secretion, seizures, hallucinations, etc.) was noted. The steady-state plasma concentration of VCR during continuous infusion ranged from 1 to 3 μg/L in all patients. Responses after 2 courses were evaluated in 14 of 16 patients. A complete response was noted in one patient (astrocytoma), a partial response in three patients (one each with astrocytoma, ependymoma, and PNET), stable disease in seven patients, and disease progression in three patients.
Vincristine (VCR) has been used for more than 4 decades to treat children with acute lymphoblastic leukemia (ALL) and currently is included in many treatment programs for children with tumors arising within or outside the central nervous system (CNS). However, dose-limiting neurotoxicity, characterized principally by a mixed sensory-motor peripheral and autonomic neuropathy, has prevented further intravenous bolus dose escalation.1
Many factors, including inadequate drug dosing and suboptimal scheduling, may contribute to ineffective therapy.2, 3 Pharmacokinetic investigations have revealed rapid systemic clearance of VCR following bolus doses. However, in vitro studies in xenograft models have demonstrated that cell kill by vinca alkaloids increases progressively with exposure time.4–6 Continuous infusion of chemotherapy drugs may improve therapeutic efficacy, particularly when the drug half-life is short and drug activity is cell cycle dependent. VCR is excreted rapidly after bolus injection and exerts its cytotoxic effect by binding with tubulin, a subunit of microtubules, and arresting cell division at metaphase by blocking tubulin polymerization.7, 8
The goal of the current study was to investigate the pharmacokinetic characteristics of continuous-infusion VCR and the neurotoxicity and extraneural toxicity associated with two courses of VCR and cyclophosphamide in pediatric patients with newly diagnosed or recurrent CNS tumors; in addition, we assessed the antitumor efficacy of this drug pair. VCR and cyclophosphamide were selected because of the documented activity of each agent individually9–15 and in combination with each other16–19 against a variety of pediatric CNS tumors and because of the ethical limitations associated with using VCR monotherapy to treat a group of patients with either newly diagnosed disease or with a first disease recurrence after radiotherapy.
MATERIALS AND METHODS
All patients age < 18 years with recurrent primary CNS tumors and newly diagnosed patients age < 60 months with low-grade gliomas and intrinsic pontine tumors were eligible for study enrollment. Histologic verification of the diagnosis was mandatory for all patients except those with diffuse intrinsic pontine tumors or surgically unresectable optic pathway tumors. Patients who had previously experienced treatment failure after receiving VCR and cyclophosphamide were ineligible. Histologic verification of the diagnosis also was mandatory for all patients except those with neurofibromatosis and those with progressive optic pathway or intrinsic pontine tumors. Before study entry, patients were required to have radiologically measurable residual disease at the primary site or elsewhere within the neuraxis, with the malignancy having an area > 1.5 cm2 as calculated using the largest perpendicular cross-sectional diameters identified on magnetic resonance imaging (MRI) evaluation with and without gadolinium enhancement. Responses were assessed by measuring changes in the dimensions of enhancing tumors. A history of previous seizures was not a contraindication against enrollment.
The chemotherapy regimen consisted of cyclophosphamide 65 mg/kg administered intravenously over 1 hour with sodium mercaptoethanesulfonate (mesna) on Day 1, followed by double maintenance fluids for 24 hours with mesna. A bolus of VCR 1.5 mg/m2 (maximum bolus dose, 2.0 mg) was administered intravenously on Day 2, followed by a 96-hour continuous infusion of VCR 0.5 mg/m2 per 24 hours on Days 2–5. The total VCR dose delivered during each course was 3.5 mg/m2 for patients with body surface area < 1.3 m2 and 2.0 mg bolus plus 2.0 mg/m2 for patients with body surface area < 1.3 m2. The treatment course was repeated once after an interval of approximately 21 days, provided that the platelet count was > 100 × 109/L and the neutrophil count was > 1.0 × 109/L. Patients were evaluated frequently via physical examination, blood counts, and serum chemistry assays. After completion of the study protocol, patients without progressive disease were offered additional cyclophosphamide and VCR–containing chemotherapy according to institutional guidelines. Clinical follow-up assessment was performed by one of the authors (S.J.K.) and will continue indefinitely for surviving patients.
The protocol was approved by the institutional ethics committee of The Children's Hospital at Westmead (Sydney, Australia). Written informed consent was obtained from all participants' parents.
Blood samples for measurement of plasma VCR concentrations were collected at 0 (blank), 5, 30, 120, and 135 minutes and at 6, 12, 24, 72, and 96 hours from the administration of the bolus dose. Blood samples were collected from a site that was far from the infusion site in 4 mL heparinized polyethylene terephthalate tubes (Greiner Labortechnik, Poitiers, France) and centrifuged at 3000 rpm for 10 minutes. The plasma was separated, transferred into 10 mL polypropylene tubes (JHN CT32020; Selby Biolab, Melbourne, Australia), and stored at −40 °C. Batches of these plasma samples were transferred to University Hospital Groningen (Groningen, The Netherlands) on dry ice (−57 °C), with en route reicing between Sydney and Groningen. Questions concerning the stability of VCR during shipment were explored previously and are reported elsewhere.20
VCR concentration in plasma samples was measured using high-performance liquid chromatography with on-line column extraction and electrochemical detection; the detection sensitivity was 0.5 ng/mL.21 Pharmacokinetic data analysis was performed using ADAPT II software22 (Biomedical Simulations Resource, Los Angeles, CA) and a maximum a posteriori Bayesian algorithm. A two-compartment first-order model was fitted to the data. Parameter values (mean and variance) for the Bayesian ‘priors’ applied to the current study were determined in a previous study using weighted least-squares parameter estimation.5
Evaluation of Response
Response was determined at the completion of two courses of chemotherapy using MRI scans with and without gadolinium by comparing the products of the two largest cross-sectional diameters of the tumor before and after two courses of treatment. Patients were ineligible for assessment of response unless they were not receiving corticosteroids, or were on reducing doses of corticosteroids. A complete response (CR) was defined as the disappearance of all disease, and a partial response (PR) was defined as a reduction of ≥ 50% in tumor area. Patients with an increase or decrease of < 25% in the product of cross-sectional diameters were considered to have stable disease (SD), and patients with an increase of > 25% in tumor area were considered to have progressive disease (PD). Neurotoxicity was graded according to a modified version of the National Cancer Institute Common Toxicity Criteria (Table 1).
Table 1. Vincristine Neurotoxicity
SIADH: syndrome of inappropriate antidiuretic hormone secretion.
Achilles reflex abnormality; mild paresthesia
Constipation; paresthesia; jaw, throat, or bone pain; myalgia; mild weakness
Severe motor weakness; gait impairment; wrist/foot pain; sensory loss; severe abdominal pain
SIADH; motor paralysis; cranial nerve palsy; ileus; seizures; hallucinations
No. of courses affected/total courses
A total of 31 courses of chemotherapy were administered to 16 children enrolled in the study. Fifteen patients received two courses of treatment, and one patient received one course. Clinical characteristics are summarized in Table 2. Five patients with newly diagnosed disease and three previously untreated (except for surgery) patients were enrolled at the time of tumor progression. Eight patients were enrolled with disease recurrence or PD after receipt of previous craniospinal radiotherapy (n = 1), local radiotherapy (n = 6), or both (n = 1). Two of 16 patients had neurofibromatosis type 1, and 4 patients had a history of grand mal seizures related to their previous diagnosis of CNS tumor. Six of 16 patients had ≤ 2.5 times the upper normal level of gamma-glutamyl transpeptidase (GGT), 2 patients had 2.5–5 times the upper normal level of GGT, and 1 patient had > 5 times the upper normal level of GGT. One patient had mild pretreatment elevation of bilirubin levels (< 1.5 times the upper normal level). Pretreatment renal function, assessed by evaluation of serum urea and creatinine levels, was mildly abnormal (Grade 1) in two patients.
Table 2. Patient Characteristics
Age at time of VCR/cyclophosphamide chemotherapy (yrs)
Previous XRT or chemotherapy
Time from diagnosis to VCR/cyclophosphamide chemotherapy (mos)
Results of the pharmacokinetic studies are listed in Table 3 and Figure 1. Peak plasma concentrations of VCR 5 minutes after injection were similar to those previously observed after a 1.5-mg/m2 bolus injection.5 After approximately 1 hour, the concentration-time curve exhibited a plateau, indicating that the rates of infusion and elimination had reached equilibrium. Primary pharmacokinetic parameters, including the volume of the central compartment (Vc), first-order rate constants for transfer between the central and peripheral compartments (kcp,kpc), and the elimination rate constant (kel), were estimated as described in Materials and Methods. Secondary parameters, including distribution (t1/2,α) and elimination half-lives (t1/2,β), total body clearance (CL), and apparent volume of distribution at steady state (Vd,ss), were calculated from the model. Plasma concentration at steady state (Css) was calculated as the quotient of the rate of infusion (Ri) and CL (i.e., Css = Ri/CL). An example of a set of measured concentration-time data points obtained from an individual patient is shown in Figure 1. The calculated pharmacokinetic parameters for Patient 9 are within the range of average values for the study (Table 3).
SD: standard deviation; Vc: volume of central compartment; kcp: first-order rate constant for transfer from central to peripheral compartment; kpc: first-order rate constant for transfer from peripheral to central compartment; kel: elimination rate constant; t1/2,α: distribution half-life; t1/2,β: elimination half-life; CL: total-body clearance; Vd,ss: volume of distribution at steady state; Css: plasma concentration at steady state.
The side effects of treatment with continuous-infusion VCR with cyclophosphamide were mild and well tolerated in each patient. The most common toxicities were mild myelosuppression and neuropathy. Detailed analyses of neurologic and nonneurologic toxicity are provided in Tables 1 and 4. The median interval between courses of chemotherapy was 25 days (range, 20–30 days). Hemopoietic growth factors were not used. One patient with recurrent medulloblastoma associated with supratentorial metastases died after an intratumoral hemorrhage 10 days after commencing protocol treatment. His platelet count was 324 × 109/L at readmission, and there were no clinical features suggestive of recent seizure activity. We could not demonstrate any relation between the concentration (peak and/or steady-state) of VCR and neurotoxicity.
CR or PR was documented in four patients, including two with previously untreated progressive low-grade astrocytoma or ependymoma and another with a brainstem metastasis from a previously treated frontal primitive neuroectodermal tumor. Patients 4 and 14 had objective responses that did not meet the criteria for a PR. These patients were classified as having SD.
The current study has demonstrated that the dose of VCR can be escalated safely in children with active CNS tumors using a 96-hour continuous-infusion technique. Serious CNS neurotoxicities, particularly encephalopathy and coma, have been described in adults with a disrupted blood-brain barrier receiving VCR via continuous infusion or following accidental high-dose bolus.23–25 However, neither cranial nerve palsy (e.g., hearing loss, facial palsy, hoarseness/dysphagia, ophthalmoplegia, or ptosis) nor CNS neurotoxicity (e.g., syndrome of inappropriate antidiuretic hormone secretion [SIADH], seizures, hallucinations, depression, insomnia, or coma) complicated any of the 31 treatment courses in the current study. Peripheral neuropathy, including motor/sensory and autonomic neuropathy, was common but mild.
The pathophysiology and cytotoxicity of VCR can, in part, be understood in terms of the disruption of cytoplasmic microtubules. Except in rare cases (e.g., in human erythrocytes), microtubules are found in the cytoplasm of eukaryotic cells and form a diverse collection of subcellular structures. The microtubular network disappears during cytokinesis, and the microtubular bundles are believed to form the spindle apparatus that effects the movement of daughter chromosomes and the division of cytoplasm, organelles, and cytoskeletal proteins.26 The cytotoxicity of VCR is cell cycle specific, resulting in the inhibition of mitosis.8, 9, 27 However, in vitro and animal model studies suggest additional mechanisms of action, including cytolytic effects on nondividing cells, induction of apoptosis, interference with the tumor blood supply, and interference with the synthesis of DNA, RNA, and intracellular proteins.28–30
Microfilaments play a key role in axoplasmic transport in neurons and are located throughout the length of the axon. They act as ‘tracts’ for axoplasmic transport of proteins or vesicles. VCR exposure results in inhibition of the microtubule assembly, defects in axonal transport, and axonal degeneration. Microtubular disruption occurs at very low VCR concentrations and appears to be significantly influenced by the duration of exposure in vitro.8, 31 Clinically, however, the incidence of toxicity, particularly neurotoxicity, is related to a complex interplay among patient age, body weight and composition, VCR dose, schedule, and pharmacokinetics.
Previous reports of the use of high-dose VCR administered via continuous infusion in children have involved multiagent bone marrow transplant conditioning regimens, preventing a careful assessment of toxicity attributable to VCR.32, 33 Pinkerton et al.32 reported significant muscle pain at a dose level of 4 mg/m2 but minimal gastrointestinal or other features of neurotoxicity. These investigators, however, observed that the identification of neurotoxicity that was specifically related to VCR within the broad spectrum of toxicity associated with multidrug myeloablative regimens is difficult. The Associazione Italiana Ematologia Oncologia Pediatrica Bone Marrow Transplant Group reported on a study of high-dose VCR (4 mg/m2) without mention of neurotoxicity.34 It is noteworthy that this Italian group reported a trend toward improved event-free survival at 4 years among children receiving VCR. Studies of continuous-infusion VCR in adults have not involved randomized trials, and comparative observations of the neurotoxicity and efficacy associated with continuous-infusion and bolus doses are not available.33, 35
Heightened awareness of the importance of dose intensity and logical drug scheduling based on pharmacokinetic and in vitro investigations has contributed substantially to patient outcomes. Virtually all drug-sensitive tumors exhibit steep dose-response curves.36 The relation between suboptimal dosing and adverse clinical outcome was noted as early as the 1960s, in a randomized trial involving children with acute leukemia.37 Although intrinsic or acquired drug resistance may adversely affect cure rates, ineffective drug dosing or scheduling represents an important factor that may contribute to ineffective therapy. The positive relation between drug dose and host toxicity is readily apparent. However, the dose-response effect is less clear, although higher doses, in general, have yielded superior response rates in drug-sensitive tumors.2, 3
In a Phase I clinical study of VCR administered via continuous infusion in adults, Jackson et al.38 recommended a dose schedule of 0.5 mg/m2 per day × 5 days preceded by a bolus of 0.5 mg/m2. Although neurotoxicity was common at higher dose levels (0.75 mg/m2 and 1.0 mg/m2 daily for 5 days), significant toxicity at the 0.5 mg/m2 level was mild, comprising ileus (1 of 19 courses), hyporeflexia (5 of 19 courses), myalgia (3 of 19 courses), and hyponatremia (not SIADH; 1 of 19 courses). In vitro and animal studies have demonstrated the positive relation among tumor cell sensitivity to VCR, intracellular drug retention, and cytotoxic activity.39–41
The CL observed in the current study (203 mL/m2 per minute) was similar to the corresponding value obtained in a study of VCR monotherapy for patients with ALL (228 mL/m2 per minute) but was substantially lower than the CL for patients with ALL who also received prednisone (431 mL/m2 per minute) at the time of VCR administration. This observation may be explained by the hypothesis of Groninger et al.,42 who suggested that a steroid-induced increase in cytochrome P450 3A4 enzyme activity may result in a higher VCR elimination rate in children receiving prednisone. Our inability to demonstrate any effect of concomitant use of steroids on clearance (data not shown) may be attributable to the relatively low (and sometimes tapering) doses of steroids used in the current study population compared with the children who were treated for ALL.
Continuous infusion of chemotherapeutic agents is more resource intensive and expensive than bolus dosing. Although outpatient continuous-infusion techniques are available, to date, studies involving VCR have required inpatient care. Theoretically, continuous infusion offers advantages, including 1) a greater probability of having the chemotherapeutic agent present in the cellular environment at the time of cell replication within tumors containing heterogeneous cell populations, 2) enhanced intracellular availability of drugs as a result of diffusion or active transport dependent on both extracellular concentration and the length of time for which the drug is in contact with the cell membrane, and 3) the ability to prolong the duration of administration, which may result in an improved therapeutic index for clinically useful agents, such as VCR, that have short half-lives. Despite these theoretic advantages, trials of VCR administered via continuous infusion have been unable, as a result of their experimental design, to address the issues of comparative efficacy and toxicity.