|Home | About | Journals | Submit | Contact Us | Français|
For the majority of children with acute lymphoblastic leukemia (ALL), CNS prophylaxis consists of either intrathecal (IT) methotrexate or triple IT therapy (ie, methotrexate with both cytarabine and hydrocortisone). The long-term neurotoxicities of these two IT strategies have not yet been directly compared.
In this multisite study, 171 children with standard-risk ALL, age 1 to 9.99 years at diagnosis, previously randomly assigned to IT methotrexate (n = 82) or to triple IT therapy (n = 89) on CCG 1952, underwent neurocognitive evaluation by a licensed psychologist at a mean of 5.9 years after random assignment.
Patients who received IT methotrexate had a mean Processing Speed Index that was 3.6 points lower, about one fourth of a standard deviation, than those who received triple IT therapy (P = .04) after analysis was adjusted for age, sex, and time since diagnosis. Likewise, 19.5% of children in the IT methotrexate group had a Processing Speed Index score in the below-average range compared with 6.9% in the triple IT therapy group (P = .02). Otherwise, the groups performed similarly on tests of full-scale intelligence quotient, academic achievement, attention/concentration, memory, and visual motor integration. The association of treatment with measures of cognitive functioning was not modified by sex or age at diagnosis. In the post-therapy period, there were no group differences in special education services, neurologic events, or use of psychotropic medications.
This study did not show any clinically meaningful differences in neurocognitive functioning between patients previously randomly assigned to IT methotrexate or triple IT therapy except for a small difference in processing speed in the IT methotrexate group.
The introduction of presymptomatic CNS therapy during the 1970s was a major factor in the transformation of childhood acute lymphoblastic leukemia (ALL) to a highly treatable disease.1 CNS therapy was originally delivered with cranial radiation at 18 to 24 Gy, which had the undesired effects of diminished intelligence quotient (IQ), poor academic functioning, subsequent cancers, and endocrinopathies.2–4 Through a series of trials, it was determined that intrathecal (IT) chemotherapy could replace cranial radiation for the majority of children with ALL.5–7 Two primary IT medication strategies emerged: IT methotrexate and triple IT therapy (ie, methotrexate with both cytarabine and hydrocortisone). Assignment to one of these preparations varied by treatment protocol, cooperative group, and treatment center.
Only recently have the therapeutic efficacies of these two strategies been compared directly. In a Children's Cancer Group (CCG) study of 1,018 patients with childhood ALL, Matloub et al8 concluded that triple IT therapy was associated with a 6-year cumulative incidence of isolated CNS relapse of 3.4% ± 1.0% compared with 5.9% ± 1.2% for IT methotrexate (P = .004), but the therapies had similar event-free survival rates. Interestingly, triple IT therapy had worse overall survival rates, because the salvage rate after bone marrow relapse was worse than after CNS relapse. These results suggest that triple IT therapy may be particularly useful for patients with a higher risk of CNS relapse (eg, T-cell phenotype) if more effective systemic therapy is given as well.
With advances such as effective CNS prophylaxis, the vast majority of children with ALL are expected to become long-term survivors. The recently completed US cooperative trial CCG 1991, in which all patients received IT methotrexate, reported 4-year event-free survival rates approaching 90% and overall survival rates of approximately 95% for patients with standard-risk ALL.9 Therefore, the effect of different therapies on future quality of life has increasingly been considered as a critical factor in the selection of optimal treatment.
Unfortunately, multiple long-term studies suggest that survivors of childhood ALL who did not receive cranial radiation still may develop meaningful deficits in neurocognitive function.10–15 Investigators have consistently identified difficulties in attention, working memory, processing speed, mathematics, and visual motor integration, but the exact etiology of these deficits has yet to be established. Exposure to IT medications is a possible contributing factor. Both IT methotrexate and IT cytarabine are associated with potential acute neurotoxicity, including stroke-like syndromes, myelopathy, encephalopathy, and seizures,16,17 and also may have long-term effects. A compelling question is whether the addition of cytarabine and hydrocortisone to IT methotrexate is associated with worse long-term toxicity.
We evaluated neurocognitive functioning in patients previously randomly assigned to IT methotrexate versus triple IT therapy on legacy CCG 1952, from which Matloub et al8 analyzed event-free survival and overall survival outcomes. We hypothesized that triple IT therapy is associated with greater neurocognitive impairment than IT methotrexate, especially in processing speed, attention, memory, and visual motor integration, and that younger age and female sex affect the association between IT preparation and neurocognitive functioning.
We conducted a cross-sectional study of patients at limited institutions who were previously enrolled and randomly assigned on CCG protocol 1952, which was open between May 1996 and February 2000. Eligible patients had B-lineage or T-lineage ALL that satisfied National Cancer Institute criteria for standard risk (ie, age at diagnosis 1.0-9.99 years and presenting WBC < 50 × 109/L).18 The therapeutic protocol consisted of a 2 × 2 factorial design, in which participants were assigned randomly to either IT methotrexate or triple IT therapy as CNS-directed therapy and to either oral mercaptopurine or thioguanine as the maintenance thiopurine. Maintenance duration was 20 months for girls and 32 months for boys; thus, girls received 19 IT injections, and boys received 23 injections during the entire treatment course. For all patients, prednisone was given in induction and maintenance; dexamethasone was given in the delayed-intensification phases.8
Patients were eligible for participation in the neurocognitive follow-up study if they were diagnosed and enrolled on CCG 1952 at one of the 22 designated limited-institution sites (Appendix, online only). Additional eligibility requirements included the following: continuous remission, no history of CNS leukemia (thus, no cranial radiation), ≥ 1 year since cessation of therapy, age at evaluation of 6 to 16.99 years, no history of pre-existing developmental disorders (eg, trisomy 21, developmental delay), and no history of very low birth weight (< 1,500 g). The age restriction was chosen to correspond to the validated age range of the standardized neuropsychological instruments used in the evaluation. In addition, individuals were excluded if they had been nonrandomly assigned to more intensive therapy because of unfavorable cytogenetic findings or a slow response after induction.
Five hundred twenty-seven patients were enrolled on the therapeutic study at one of the participating sites and were confirmed to meet the inclusion and exclusion criteria for this study. Of these, 171 patients consented and completed the evaluation. Of the remaining 356 patients, 144 were lost to follow-up and could not be traced, 184 refused, and 28 were never offered participation. The 171 participants were similar to the 356 eligible nonparticipants in terms of age at diagnosis, elapsed time since diagnosis, sex, and therapeutic random assignments (Table 1).
The institutional review board of each participating center approved the protocol and study documents. Informed consent, and assent if indicated, was obtained from all participants.
Participants underwent a comprehensive, half-day, neurocognitive assessment supervised by a licensed psychologist. This evaluation was paid by research funds and was at no cost to the patient. The test battery was based on a previous CCG study, which successfully utilized previous editions of almost all the same tests in a longitudinal study of the neurobehavioral effects of intermediate risk ALL and its therapy.19 The neurocognitive functioning evaluation included, among others, the following tests: Wechsler Intelligence Scale for Children, fourth edition (WISC-IV); Wechsler Individual Achievement Test, second edition, abbreviated (WIAT-II-A); Beery Developmental Test of Visual Motor Integration; Conners' Continuous Performance Test II (CPT II); and Children's Memory Scale (CMS). Table 2 details the subsets administered and the scores analyzed.
Parents of participants completed a demographic and medical history survey. Parents were asked about their marital status, education, and income. This questionnaire confirmed that the child was developing normally before the ALL diagnosis as an additional check of eligibility for this study. Also, parents were asked about neurologic events, special education services, and psychotropic drug use during and after ALL therapy.
Characteristics such as age, sex, and therapy history were summarized and were compared between participants and nonparticipants by using t test, χ2 test, and Fisher's exact test to evaluate the potential for response bias. These characteristics also were compared between those randomly assigned to IT methotrexate versus triple IT therapy to determine comparability of exposed groups.
Multiple linear regression was used to evaluate the differences in neurocognitive outcomes between the IT groups, and analysis was adjusted for sex, age at diagnosis, and elapsed time since diagnosis. Adjusted least squares means and standard errors (SEs) for the neurocognitive outcomes, as well as 95% CIs for treatment group differences in the means, are presented. In addition, the proportion of patients with standard scores at or worse than one standard deviation less than the norm was compared by using χ2 test or Fisher's exact test.
Subgroup analyses also were conducted to determine whether the differences in neurocognitive outcomes between treatment groups were modified by sex and/or age at diagnosis (< 3 years v ≥ 3 years). Post-hoc analyses were conducted to assess the impact of the formulation of the thiopurine (ie, mercaptopurine v thioguanine) on neurocognitive functioning.
The sample size of 82 and 89 patients in the two IT groups provided 80% power at the two-sided .05 significance level to detect a difference of 0.4 standard deviations between the two group means. For example, there was adequate power to detect a difference of six IQ points between groups.
Data were analyzed with the SAS software package, version 9.1 (SAS Institute, Cary, NC), with two-sided tests at the .05 significance level. The data were analyzed by the Biostatistics and Study Design Core of the Yale Center for Clinical Investigation in collaboration with the COG Biostatistics Committee.
There were 171 participants with neurobehavioral data available for analysis. The participants of this study were similar to the 1,856 patients enrolled on CCG 1952 in terms of sex, age at diagnosis, corticosteroid random assignment, and mercaptopurine random assignment. However, the current sample was more likely to be white (81% v 67% in CCG 1952; P < .001).
Table 3 lists the characteristics of participants, stratified by IT random assignment. Patients in the different IT treatment groups were similar in terms of sex, elapsed time since diagnosis, ethnicity, education of primary caregiver, and family income. Patients who received IT methotrexate were slightly older at ALL diagnosis (4.6 v 4.0 years; P = .04) and were more likely to have a married primary caregiver (91% v 78%; P = .03).
The thiopurine groups were equally distributed between those randomly assigned to IT methotrexate and triple IT therapy. Post-hoc analyses comparing neurocognitive performance between the mercaptopurine and thioguanine showed no differences for any of the neurocognitive domains (data not shown).
Table 4 lists the least squares means of neurocognitive test results among patients randomly assigned to IT methotrexate or to triple IT therapy, as adjusted for sex, age at diagnosis, and elapsed time since diagnosis. Patients who received IT methotrexate scored 3.6 points lower, or about one fourth of a standard deviation, on Processing Speed Index (P = .04). Otherwise, the groups performed similarly in tests of full-scale IQ, academic achievement, attention/concentration, memory, and visual motor integration. Caregiver marital status and caregiver education were not significant predictors of neurocognitive functioning for any of the domains tested.
The distribution of scores in the two treatment groups then was examined by comparing the proportion of patients with standardized scores at or worse than one standard deviation less than the norm (Table 5). For Processing Speed Index, children randomly assigned to IT methotrexate were more likely to score in the below-average range than those who received triple IT therapy (19.5% v 6.9%; P = .02). No differences were observed for the other domains of neurocognitive functioning.
Potential interactions were tested. The association of treatment with the different measures of cognitive functioning was not modified by sex or age at diagnosis.
The frequency of neurologic events reported during therapy was higher among patients randomly assigned to IT methotrexate (23.1%) than among those who received triple IT therapy (8.1%; P = .01), largely because of more reports of loss of limb function and altered mental status. In the post-therapy period, there were no group differences in history of special education services, neurologic events, or use of psychotropic medications (Table 6).
In general, there were no significant differences in long-term neurocognitive and academic performance in children with ALL previously randomly assigned to IT methotrexate versus triple IT therapy in this multisite, cross-sectional study. The exception was that patients who received IT methotrexate scored slightly worse—about one fourth of a standard deviation worse—on a test of processing speed. The IT methotrexate group also had a higher proportion of patients with Processing Speed Index scores at or worse than one standard deviation less than the mean (19% v 7%). There was no difference between IT methotrexate and triple IT therapy groups in neurologic complications, psychotropic drug use, or special education services post-therapy by parent report. A substantial number of patients (approximately 25%), however, received special education services after completing therapy for ALL, which indicated the elevated need for these services in this population. This is the first study, to our knowledge, to compare long-term neurocognitive functioning in ALL patients according to the formulation of IT therapy that they received.
Though the neurocognitive outcomes have not been compared previously, IT medications have long been known to be potentially neurotoxic. Acutely, IT methotrexate and—less so, IT cytarabine—are associated with stroke-like syndromes, myelopathy, encephalopathy, and seizures.16,17 The mechanism for methotrexate-associated toxicity is not known. Some hypothesize that methotrexate, by inhibiting dihydrofolate reductase, transiently elevates CNS levels of homocysteine, which in turn results in small-vessel vasculopathy. Other investigators believe methotrexate effects are mediated by disruption of brain amines20 or β-oxidation of fatty acids in the cerebrospinal fluid.21 It is unclear if higher cumulative IT methotrexate doses are associated with worse outcomes. In a small study of 21 children, the number of IT methotrexate doses was positively associated with neurocognitive impairment.22 However, in a larger study of 121 patients with ALL, cumulative IT dose did not predict cognitive functioning.23
Among all children who participated in the randomized, controlled study (N = 1,018), grades 3 to 4 CNS toxicity on therapy, as reported by treating physicians, occurred in 5.8% on IT methotrexate and in 6.7% of those on triple IT therapy.8 In our cross-sectional sample, parents of patients reported a higher frequency of acute CNS toxic events of any severity during therapy in the IT methotrexate group. There was no difference between IT groups in this study during the post-therapy period. This study was not designed to explain why there was similar long-term neurocognitive functioning in the two groups. Future studies could potentially explore if hydrocortisone interacts with methotrexate or if the neurotoxicity of cytarabine is limited to the acute treatment period.
The major advantage of this study was that the participants of this report had been enrolled on a randomized, controlled, treatment study. Thus, any potential adjuvant therapy or patient-related confounders that could affect neurocognitive functioning likely would be equally distributed between the treatment groups. Patients who received triple IT therapy were more likely to have an unmarried or less-educated caregiver. However, caregiver marital status and caregiver education were not significant predictors of neurocognitive functioning for any of the domains tested. All participants were similar in terms of National Cancer Index standard risk status and lack of cranial radiation. Chemotherapy exposures were identical except for the thiopurine random assignment. To verify that the thiopurine preparation did not confound the association between corticosteroid and neurocognitive functioning, we did post-hoc analyses to compare neurocognitive performance between the mercaptopurine and thioguanine groups. We found no differences for any of the neurocognitive domains.
These results must be interpreted in the setting of several limitations. We enrolled only 171 of the eligible, traceable 355 patients at the participating institutions who were offered the study. Therefore, there is a potential for participation bias in terms of factors related to neurocognitive functioning. However, the participants were similar to the nonparticipants in terms of age, sex, elapsed time since diagnosis, and—most importantly—IT random assignment. The total number of participants was adequate to address clinically important differences, as determined by sample size calculations that were based on the t test statistic for two independent samples (Data Analysis section). The patients in the IT methotrexate group were more likely to have a married primary caregiver and trended toward having more educated primary caregivers. However, we do not believe this led to selection bias, because the worse processing speed was observed in the IT methotrexate group. We acknowledge that the lack of parent and teacher questionnaires in our analysis is a potential limitation. Such reports may identify subjective deficits that are not detected by patient performance instruments but that still can affect quality of life.
We used a cross-sectional study design, so causality between treatment exposure and neurocognitive outcome can be inferred but can not be known with certainty. However, longitudinal, prospective studies are more expensive and are difficult to complete with reasonable sample sizes. Furthermore, they often are not feasible to conduct in patients who are ill and/or receiving intensive therapy. Neuropsychological assessments administered at diagnosis in some longitudinal studies have yielded lower than expected cognitive performance scores.24,25 Longitudinal studies also can introduce practice effect if done frequently because of general improvement in performance with repeated testing26; methodologic difficulties can result from changing to test instruments appropriate for older ages and/or newer edition forms.
In conclusion, this study did not show any clinically meaningful differences in cognitive functioning between patients previously randomly assigned to IT methotrexate or triple IT therapy. These results do not support modifying IT therapy for ALL because of differential effects on neurocognitive outcome. A small difference in processing speed was seen between children randomly assigned to triple IT therapy versus IT methotrexate, but is not of a magnitude to support differential follow-up of either group. Therefore, clinicians, teachers, and parents should monitor children treated with any IT therapy for the need for accommodation in this domain. For example, extra time for examinations may be needed. The Children's Oncology Group Long-Term Follow-Up Guidelines (http://www.survivorshipguidelines.org/) represent an important resource for evidence- and consensus-based care recommendations for childhood cancer survivors in terms of neurocogntive functioning and overall.27 Our data demonstrated that there is substantial individual variation in neurobehavioral outcome. Detailed investigations of host/drug interactions must be conducted to additionally define the source of this variation.
A.B. Chandler Medical Center, University of Kentucky, Lexington, KY: Martha Greenwood (institutional oncologist [IO]), Michelle Mattingly (institutional neuropsychologist [IN]); C.S. Mott Children's Hospital, Ann Arbor, MI: Rajen Mody (IO), Eileen Mollen (IN); Children's Health Care, Minneapolis, Minneapolis, MN: Bruce Bostrom (IO), Karen Wills (IO), Jonathan Miller (IN); Children's Health Care, St. Paul, St Paul, MN: Yoav Messinger (IO), Bonnie Carlson-Green (IN); Children's Hospital & Regional Medical Center, Seattle,WA: Doug Hawkins (IO), David Brieger (IN); Children's Hospital Los Angeles, Los Angeles, CA: Paul Gaynon (IO), Sharon O'Neil (IN); Children's Hospital of Columbus, Columbus, OH: Amanda Termuhle (IO), Tami Young-Saleme (IN); Children's Hospital of Oakland, Oakland, CA: Joseph Torkildson (IO), Seth Ubogy (IN); Children's Hospital of Philadelphia, Philadelphia, PA: Jill Ginsberg (IO), Alyssa Rodriguez (IN); Children's Hospital of Pittsburgh, Pittsburgh, PA: Jean Tersak (IO), Hilary Feldman (IN), Christy Emmons (IN); Christiana Care Health/A.I. DuPont Institute, Wilmington, DE: Gregory C. Griffin (IO), Rochelle Glidden (IN); Doernbecher Children's Hospital–OHSU, Portland, OR: Kelly Anderson (IO), Robert Butler (IN); DeVos Children's Hospital, Grand Rapids, MI: David Dickens (IO), Steve Pastyrnak (IN); Indiana University–Riley Children's Hospital, Indianapolis, IN: Terry Vik (IO), Stephen J. Pongonis (IN); Loma Linda University Cancer Institute, Loma Linda, CA: Antranik Bedros (IO), Thomas Kaleita (IN); Long Beach Miller Children's Hospital, Torrance, CA: Jerry Z. Finklestein (IO), Teddi Softley (IN); M. D. Andersen Cancer Center, Houston, TX: Joann Ater (IO), Bartlet Moore (IN); Phoenix Children's Hospital, Phoenix, AZ: Jessica Boklan (IO), Michael Lavoie (IN); Primary Children's Medical Center, Salt Lake City, UT: Phil Barnette (IO), Paule Colte (IN); The Children's Hospital Denver, Denver, CO: Kelly Maloney (IO), Greta Wilkening (IN); University of Minnesota, Minneapolis, MN: Jospeh Neglia (IO), Alicia Kunin-Batson (IN); Vanderbilt University, Nashville, TN: Jennifer Domm (IO), Frances Niarhos (IN); Yale University, New Haven, CT: Tara McPartland, Sui Tsang; Coordinating Center: Megan Nicoletti (study coordinator).
Supported in part by Grant No. RSGPB-03-167-01-PBP from the American Cancer Society (J.N.); Clinical and Translational Science Award Grant No. KL2 RR024138 from the National Center for Research Resources, a component of the National Institutes of Health (NIH) and NIH Roadmap for Medical Research (N.K.-L.); the Children's Oncology Group Chair Grant No. U10 CA98543; the Statistics and Data Center Grant No. U10 CA98413; and the CCOP Grant No. U10 CA95861.
Presented in part at the 10th International Conference on Long-Term Complications of Treatment of Children and Adolescents and Cancer, June June 6-7, 2008, Niagara-on-the-Lake, Canada; and at the 44th Annual Meeting of the American Society of Clinical Oncology, May 30-June 3, 2008, Chicago, IL.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Clinical trial information can be found for the following: NCT00085176.
The author(s) indicated no potential conflicts of interest.
Conception and design: Nina S. Kadan-Lottick, Pim Brouwers, David Breiger, Thomas Kaleita, Lu Chen, Linda Stork, Joseph P. Neglia
Financial support: Joseph P. Neglia
Administrative support: Megan Nicoletti, Joseph P. Neglia
Provision of study materials or patients: Thomas Kaleita, Bruce Bostrom, Linda Stork, Joseph P. Neglia
Collection and assembly of data: Nina S. Kadan-Lottick, David Breiger, Thomas Kaleita, Megan Nicoletti
Data analysis and interpretation: Nina S. Kadan-Lottick, Pim Brouwers, David Breiger, Thomas Kaleita, James Dziura, Veronika Northrup, Joseph P. Neglia
Manuscript writing: Nina S. Kadan-Lottick, Pim Brouwers, David Breiger, Thomas Kaleita, James Dziura, Veronika Northrup, Bruce Bostrom, Linda Stork, Joseph P. Neglia
Final approval of manuscript: Nina S. Kadan-Lottick, Pim Brouwers, David Breiger, Thomas Kaleita, Bruce Bostrom, Joseph P. Neglia