Treating Children with Acute Lymphoblastic Leukemia and Down Syndrome: Pharmacokinetics Provides Insight into Vincristine Therapy

Children with Down syndrome have an increased risk of developing leukemia, both myeloid and acute lymphoblastic leukemia (ALL). Down syndrome children with myeloid leukemia have a much better prognosis compared to those children without Down syndrome who develop acute myeloid leukemia [1]. This is compared with ALL, where historically children with Down syndrome treated with risk-stratified treatment protocols have had worse outcomes than children without Down syndrome. The poorer outcome in the Down syndrome patient population has been attributed to several factors including inherent differences in tumor biology, increased therapy-related toxicities, and higher rate of infectious complications [2].

To address the question of why children with ALL and Down syndrome have a poorer outcome than their non-Down syndrome peers, investigators have studied the clinical and biological characteristics of ALL in children with Down syndrome. In most studies, the decrease in early peripheral blood blast count after prednisolone or multi-agent chemotherapy in children with ALL and Down syndrome has not been different from children without Down syndrome [2]. The rate of remission induction also has generally similar between the two groups, if the Down syndrome patients receive the more aggressive induction therapy associated with high-risk disease [3]. Lastly, results of a recent analysis of the CCG experience with ALL and Down syndrome children showed decreased event free and overall survival in patients enrolled in CCG trials compared to children with ALL and without Down syndrome. However, in a subset analysis they found that those children with ALL and Down syndrome that were classified as high-risk and received intensive therapy did as well as their non-Down syndrome peers [3].

To evaluate the differences in leukemic biology, Whitlock et al. analyzed data from Down syndrome children with ALL enrolled on CCG-1952 [4]. Patients eligible for this study met National Cancer Institute criteria for standard risk, including age ≥1 and < 10 years as well as white blood cell count < 50 × 10⁹/L. In this trial, children with Down syndrome had a 4-year event free survival that was statistically worse than those children without Down syndrome. The investigators retrospectively classified patients using current COG criteria for low-risk ALL, and noted that only 7% of the Down syndrome children met the criteria compared with 30% of the children without Down syndrome. The authors noted that this was likely due to the children with Down syndrome not having many of the biological features associated with better outcome (e.g., high hyperdiploidy, triple trisomies of chromosomes 4, 10, and 17, or expression of TEL-AML1 fusion transcript). When the patients with the low-risk ALL criteria were removed from the life-table analysis, the event free survival in both groups was equivalent. They concluded that the poorer outcome in the children with ALL and Down syndrome has at least in part a biological basis, and that with appropriately defined (i.e., biological) subsets the outcomes might be similar.

Children with ALL and Down syndrome have been noted to have increased therapy-related toxicities when compared with children without Down syndrome. These toxicities include cardiotoxicity, infection, mucositis, and myelosuppression. The increased incidence of mucositis and myelosuppression in children with Down syndrome is most likely attributable to their increased sensitivity to methotrexate. Several factors have been proposed to account for this increased sensitivity including alterations in methotrexate polyglutamination, enhanced folate depletion due to increased gene dosage and expression of cystathionine β-synthase, as well as overexpression of reduced folate carrier whose gene is located on chromosome 21. Lastly, delayed methotrexate clearance has been reported in children with ALL and Down syndrome, even in the setting of normal renal function [5]. It is crucial that an optimal balance between antileukemic effect and treatment related toxicity must be established for children with ALL and Down syndrome, and questions exist regarding the optimal dosage and frequency of most anticancer drugs for these children.

What contribution can pharmacokinetics make to define the optimal dosage and frequency of anticancer drugs and to attain this balance between efficacy and toxicity for children with Down syndrome and ALL? The report by Lönnerholm et al. in this issue of Pediatric Blood & Cancer addresses this by reporting the vincristine pharmacokinetics in children with ALL and Down syndrome compared to 92 non-Down syndrome children [6]. The authors have previously reported the disposition of vincristine in 98 children with ALL [7], six of whom had Down syndrome. In the present analysis, they present the pharmacokinetic data for the six Down syndrome patients separately and compare the results with the 92 non-Down syndrome patients. The authors use a pharmacokinetic limited sampling model and MAP-Bayesian pharmacokinetics to analyze the vincristine concentration-time data. They found that the median values for the pharmacokinetic estimates were not statistically different between the two patient populations. Thus, they conclude no pharmacokinetic rationale exists for altering vincristine dosing in children with Down syndrome and ALL.

Although the study provided important data regarding vincristine dosing in children with Down syndrome, it is important to closely examine the study methodology. First, it was a retrospective comparison of previously analyzed data. However, with the limited number of children with Down syndrome and acute lymphoblastic leukemia, investigators should take every opportunity to present data that might provide insight into the pharmacokinetics of drugs in this patient population. Second, the authors established their conclusions on data from only four patients. However, as noted above, this was a very limited patient population and that must be considered when interpreting the results. Lastly, the use of population priors (mean and variance) from the non-Down syndrome group in a MAP-Bayesian analysis may lead to bias in the calculation of the final parameter estimates in the Down syndrome group. However, the authors acknowledge this limitation and indicate that if a difference existed, sufficient data were available to detect one. The availability of newer statistical algorithms in population pharmacokinetic software [8] may aid in reducing some of the bias and lack of robustness currently observed in such limited data problems. In support of their findings, the genetic basis for increased sensitivity to methotrexate and altered methotrexate clearance is clearly plausible because the candidate genes are on chromosome 21. Whereas, the genes for the primary drug metabolizing enzymes (e.g., CYP3A) and transporters (e.g., MDR1) for vincristine are found on chromosome 7.

Why would one be interested in the pharmacokinetics of vincristine, a drug that was approved by the US FDA in 1963 and has been used in pediatric oncology for over 45 years? Although vincristine has been a standard of chemotherapy regimens for many childhood cancers, including acute lymphoblastic leukemia, rhabdomyosarcoma, Wilms tumor, and various lymphomas, very few vincristine pharmacokinetic data in children have been published [7,9]. However, it is apparent from these early studies that vincristine has wide inter- and intra-patient variability. Given the general lack of pharmacokinetic data for vincristine, dosing is generally empirical and not based on pharmacokinetic data. For example, dosage reductions are recommended for infants, and dosage is calculated on a weight rather than surface area basis. This paucity of pharmacokinetic data for vincristine in infants is currently being addressed by a collaborative study in the Children’s Oncology Group (ADVL06B1).

The large variability in vincristine pharmacokinetics and the subsequent need to clearly define vincristine dosing became clinically relevant with the recent report of a relation between vincristine systemic exposure and its antileukemic effect [10]. The investigators showed that in 40 standard risk (e.g., B-cell precursor) ALL patients, the relative risk of relapse was significantly increased for patients with vincristine clearance values above the median or AUC values below the median. In a multivariate analysis including clearance and sex, they found a trend (p=0.057) for vincristine clearance (and AUC values) to have an impact on relative risk of relapse. These results linking vincristine pharmacokinetics and clinical outcome should encourage other investigators to perform studies to define the effects of age, disease state, concomitant medications, metabolic interactions, demographics, biochemical characteristics, pharmacogenetics, and clinical characteristics on vincristine disposition. The ultimate goal would be to design evidence-based vincristine dosing regimens to enhance the use of this important drug in pediatric oncology.