Visual and Verbal Short-Term Memory Deficits in Childhood Leukemia Survivors After Intrathecal Chemotherapy 1

Transcription

1 Journal ofpediatric Psychology, Vol. 22, No , pp Visual and Verbal Short-Term Memory Deficits in Childhood Leukemia Survivors After Intrathecal Chemotherapy 1 Dina E. Hill, K. T. Ciesielski, 2 and Lisa Sethre-Hofstad Clinical Neuroscience Laboratory, University of New Mexico Marilyn H. Duncan School of Medicine, University of New Mexico Marguerite Lorenzi Clinical Neuroscience Laboratory, University of New Mexico Received March 15, 1996; accepted May 30, 1997 Assessed survivors of childhood lymphoblastic leukemia (ALL) treated with intrathecal chemotherapy, using the Wide Range Assessment of Memory and Learning (WRAML), compared to controls without cancer, matched as closely as possible in age, SES, and gender. Mild, but consistent, deficits were found in both visual-spatial and verbal single-trial memory tasks. In multitrial learning, only visual-spatial tasks resulted in deficient scores, while verbal learning was within the normal range. IQ results indicated scores points lower in the ALL group. Memory results are related to deficits in strategic planning and attentional distractiveness. The WRAML may be a useful clinical tool to evaluate differential memory deficits in children with ALL. KEY WORDS: acute lymphoblastic leukemia; intrathecal chemotherapy; single-trial visual-spatial and verbal memory; multitrial learning; WRAML. Partial support for this project was from the New Mexico Children's Cancer Fund. The authors thank the subjects and their families for their cooperation, Yolanda Vinajeras for her assistance in booking the patients, Barbara Brooks for assistance in the neuropsychological testing, and J. Scott Tonigan for assistance with the statistical analyses. 2 A11 correspondence should be sent to Kristina T. Ciesielski, Psychology Department, Logan Hall, University of New Mexico, Albuquerque, New Mexico l /97/120CM)86l$12.50/ Plenum Publishing Corporation

2 862 Hill, Ciesielski, Sethre-Hofctad, Duncan, and Lorenzi Medical advances in the field of pediatric oncology have resulted in an increasing number of children surviving acute lymphoblastic leukemia (ALL). Central nervous system (CNS) prophylaxis is considered crucial to this success (Poplack, 1989). The first successful regimens consisted of cranial or cranial/spinal irradiation combined with repeated dosages of intrathecal methotrexate (IT MTX) chemotherapy (Bleyer, 1988). More recently, protocols consisting of chemotherapy without irradiation have been utilized (Brown et al., 1992). As the prospects for a cure of ALL are excellent and more children are surviving, there is a growing concern about specific cognitive deficits induced by CNS prophylactic treatment. The long-term follow-up of children who received CNS prophylaxis with irradiation and chemotherapy has identified lower levels of psychometric intelligence (Moss, Nannis, & Poplack, 1981). Most consistently reported neuropsychological deficits include difficulties in memory, attention, and motor skills (Brouwers, Riccardi, Fedio, & Poplack, 1985; Mulhem, Wasserman, Fairclough, & Ochs, 1988). The mechanism of memory deterioration, one of the most prevalent problems in ALL survivors, has not been elaborated. Problems with short-term memory (STM) rather than deficits in long-term retrieval have been described mostly in those children who have received cranial irradiation and intrathecal chemotherapy (Cousens, Ungerer, Crawford, & Stevens, 1991). There is no consensus whether intrathecal chemotherapy alone affects memory. Copeland et al. (1988) found no impairment on memory performance for nonirradiated survivors of childhood leukemia. However, a study by Whitt, Wells, Lauria, Wilhelm, and McMillan (1984) found that intrathecal chemotherapy both with and without radiation was associated with dysfunctions in STM, visual-motor coordination, and concentration. Brown et al. (1992) compared cognitive functioning in older versus younger survivors of ALL who received chemotherapy alone and found that 70% of the younger survivors of ALL showed deficient performance in memory and visual-motor attention. Among antineoplastic drugs, methotrexate (MTX) has been described as the major cause of delayed clinical symptomatology (Boogerd, 1995). The pathogenesis of the specific histological changes underlying these behavioral symptoms is still unclear. Suggested factors include a direct damaging effect of MTX on endothelial cells (Phillips et al., 1987), oligodendrocytes, microglia, or the neurons themselves resulting in secondary demyelination (Gilbert, Harding, & Grossman, 1989) and possible loss of cells. The latter is indicated by MRI and CT studies showing dilated ventricles and atrophy or hypoplasia of cortical brain regions (Boogerd, 1995; Ciesielski, Yanofsky et al., 1994; Paakko, Vainionpaa, Lanning, Laitiness, & Pyhtiness, 1992). To date, studies on children who have received CNS prophylaxis with chemotherapy alone admit that the resulting cognitive deficits exist but may be more subtle than those associated with protocols involving radiation and chemo-

3 Chemotherapy Impact on Short-Term Memory 863 therapy and, therefore, may require more specific neuropsychological techniques to investigate them (e.g., Moore, Copeland, Ried, & Levy, 1992). The present work reports data from assessment of memory functioning and psychometric intelligence. The study assessed single-trial, short-term memory and multitrial learning in the visual and verbal domains using the Wide Range Assessment of Memory and Learning (WRAML; Sheslow & Adams, 1990). It was predicted, based on previous findings, that survivors of ALL treated with chemotherapy would show predominant deficits on visual short-term memory tasks. MATERIALS AND METHOD Participants Ten survivors of childhood ALL were included in the study. ALL survivors were consecutive cases recruited from the University of New Mexico Pediatric Oncology Program treated with identical cancer protocols. The program is located in the state's largest city and is the major provider of medical care for children diagnosed with cancer in New Mexico and southern Colorado. Therefore, it is believed that the sample of children tested was representative of the area. Eligible subjects met the following criteria: diagnosed as cases with standard risk ALL; 1-5 years of age at diagnosis (one case 5.5 years); at least 3 years post cancer treatment; CNS prophylaxis with intrathecal chemotherapy alone; and identical protocol of chemotherapy. The exclusion criteria included premorbid developmental, neurological, or psychiatric disorders, focal CNS lesions, history of CNS leukemia or systemic relapse, as well as premorbid sensory or motor disabilities. All children were classified as low risk, non-t, early B ALL cases and were treated between 1982 and The cancer treatment protocol, as previously reported (Foucar et al., 1991), included CNS prophylaxis with triple intrathecal treatment with MTX, hydrocortisone, and cytosine arabinoside. The children received approximately 20 intrathecal treatments over 3 years. The induction was achieved with vincristine and prednisone; consolidation with cyclophosphamide, L-asparaginase, and MTX; and maintenance with periodic vincristine/prednisone pulses and oral 6-mercaptopurine and MTX for a total of 3 years in remission. No intravenous MTX was used. Parents of all eligible children were contacted. Of the 17 eligible children, parents of 4 children refused due to the distance from the research facility, 1 was tested only recently and is not part of this study; 1 family was not located. No child was excluded due to premorbid developmental, neurological, or psychiatric disorders. Therefore, the sample consisted of 11 survivors of ALL and 11 control children. However, during the neuropsychological evaluation, it was revealed that 1 of the control

4 864 Hill, Ciesielski, Sethre-Hofstad, Duncan, and Lorenzi Table I. ALL and Control Subject Characteristics Pairs of subjects: ALL/C Characteristics Age Grade Handedness SES" 7/6 1/1 9/3 8/9 3/4 6/2 8/10 5/5 6/3 10/8 4/3 2/2 10/10 5/4 1/2 11/11 5/5 6/3 11/10 5/4 6/3 11/11 6/5 2/3 14/14 8/8 L/L 6/9 13/12 8/7 3/6 "Hollingshead SES codes: 1 or 2 = major professions; 3 or 4 = managerial professions, 5 or 6 = skilled or semiskilled laborers; 7, 8, or 9 = unskilled or unemployed positions. children had organic problems, specifically early brain injury. Therefore, we eliminated that control child and the matched survivor of ALL. Ten controls with no history of cancer were included, with the authors attempting to make the groups as comparable as possible on age, handedness, school grade, socioeconomic status (SES), and gender. SES was determined by using Hollingshead occupational codes (1975). Control children were solicited through the families of the ALL survivors, including relatives and children from immediate neighborhoods. A total of 30 control families were referred. AH families were contacted and all agreed to participate. As the number of willing controls exceeded the number of controls needed, we chose those children who most closely matched the children with ALL to participate in the study. Controls met the same exclusion criteria as the ALL survivors. Table I presents demographic characteristics for ALL and control subjects. Prior to testing, all components of the study were explained to the children and their parents. We received written consent from the parents and verbal assent from the children. Procedures and Measurements All children were tested using a comprehensive battery of neuropsychological measures. Among them, the Wide Range Assessment of Memory and Learning (WRAML; Sheslow & Adams, 1990) was used to assess memory capacity and the Wechsler Intelligence Scale for Children (WISC-III; Wechsler, 1991) was used to assess psychometric intelligence. The Wide Range Assessment of Memory and Learning. The WRAML is a standardized psychometric tool designed to measure memory functioning in the pediatric population. The test consists of visual memory, verbal memory, learning memory, and delayed recall scales. Scores are based on a mean of 100 and a standard deviation of 15. The Visual and Verbal Memory Scales are each made up of three subtests designed to assess subjects' short-term memory capacity on a

5 Chemotherapy Impact on Short-Term Memory 865 single-trial presentation of visual-spatial and verbal information. The Learning Scale contains three verbal and visual learning subtests that evaluate subjects' performance after multitrial exposure to the materials, with each subtest including four presentations of the information. Please refer to the manual for further information on the procedures and normative data (Sheslow & Adams, 1990). Wechsler Intelligence Scale for Children-Ill. The WISC-III was employed to establish individual IQ values and to compare them with the memory scores. It was also given to compare our ALL group to other ALL groups described in the literature. All test results were transformed into z scores (M = 0, SD = 1) by comparing each child's results to the age-appropriate normative standards for each test (WRAML & WISC-III). Z scores were used as the most uniform mode of comparison between results of various tests in our neuropsychological battery, as some of the other measures in the battery were not based on a mean of 100. RESULTS Between-group / tests were conducted to determine whether the ALL group and the control group differed on age, grade, or SES level (as measured by the Hollingshead Occupational Scale, 1975). There were no significant differences between groups on age, f(18) =.15, ns; grade, f(18) =.57, ns; or SES, f(18) = 1.03, ns. Between-group t tests were conducted on two-family groups: WRAML composite scales and the WISC-III, allowing an alpha of.05 for each family. Thus, the resulting experiment-wise alpha is equal to Between-group t tests were conducted on the three composite scales from the WRAML (Verbal Memory Index, Visual Memory Index, and Learning Memory Index). There were significant differences (after Bonferroni adjustment: Family-wise a =.05/3 = ) between the groups on the short-term Verbal Memory Index, /(18) = -2.99, p =.008, and the short-term Visual Memory Index, /(18) = -3.91, p = 0.001, with higher memory performance by control subjects. Additionally, the results of the comparison of the Learning Memory Index revealed better performance by the control group, f(18) = 3.32, p =.004. The z-score means, as well as standard score means and standard deviations, for the WRAML Index Scales for the ALL and control groups are presented in Table II. To determine whether the ALL survivors differed not only from the control group but also from the norming population, a z test was conducted comparing the ALL survivors' WRAML test scores to the normative data (Harris, 1995, pp ; ). ALL survivors performed significantly below the normative mean on the Verbal Memory Index (z tesl = -4.17, p <.01) and on the

6 866 Hill, Ciesielski, Sethre-HoCstad, Duncan, and Lorenzi Table II. Means, Standard Deviations, and z-score Means for Memory Index Scales and WISC-III IQ Scales ALL survivors Controls Measures z M SD z M SD WRAML scores Verbal MI Visual MI Learning MI WISC-III scores FSIQ VIQ PIQ " "p <.01, two-tailed tests. Visual Memory Index (z test = 2.41, p <.05). The z test of the Learning Memory Index failed to reach the critical value of 1.96 (z test = 1.48, ns]. These results suggest that the significantly worse scores in the ALL survivors were not determined by the higher performance of our control group, but resulted from the true cognitive deficits in the ALL group. To follow up on these differences, fully post hoc, exploratory tests were conducted to determine the nature of the short-term memory and learning deficits seen in the ALL group. Group differences in learning within the visual and verbal modalities from the multitrial tasks were examined. The Learning Memory Index is composed of three subtests including a visual multitrial task and a verbal multitrial task; each subtest is repeated four times. Between-group differences on the visual and verbal subtests of the Learning Memory Index for the ALL and control groups are presented in Figure 1. Between-group / tests on the multitrial visual and verbal learning tasks were conducted. Significant group differences were found on the visual learning task, r(18) = -3.73, p <.002, with the ALL group performing worse. No statistically significant group differences were found between groups on the verbal task, f(18) = -2.24, ns. Therefore, although both groups performed normally on the verbal multitrial task, the ALL children continued to perform significantly below age norms and below the control children on the visual multi-trial task. Between-group t tests were also conducted on the VIQ, PIQ, FSIQ scales from the WISC-III, with group (ALL vs. controls) as the independent variable. Z-score means, as well as standard score means and standard deviations, for the WISC III for the ALL and control groups are presented in Table II. The results revealed significantly lower levels of psychometric intelligence in the ALL group (after Bonferroni adjustment: Family-wise a =.05/3 =.0167). On all three psychometric intelligence measures, controls performed significantly better, VIQ: /(18) = -3.70, p =.002; PIQ: r(i8) = -3.24, p =

7 Chemotherapy Impact on Short-Term Memory 867 Visual Learning I ALL I Controls Verbal Learning Fig. 1. The results of multitrial learning for the visual and verbal subtests in ALL and control children..005; and FSIQ: f(18) = -3.70, p = Based on the above findings, the question was raised whether the observed memory abnormalities in the ALL group were related to the changes in psychometric intelligence scores. Pearson correlations were computed on the memory scores and the FSIQ scores for each group. All the correlations between intelligence measures and memory measures were positive. The correlations within the ALL group were generally higher than in the control group. Correlations between visual single-trial memory and FSIQ (strongest) as well as between single-trial verbal memory and FSIQ were significant for the ALL group. No other significant correlations were obtained. A post hoc power analysis was conducted to identify the confidence we could have in retaining the null hypothesis in cases of nonsignificant findings. Assuming Type I error rate of.05, with an averaged effect size which considered all six tests (d = 1.22; range: ) and sample cells of 10, Type II error rate equaled.16. This estimate of power was based on an assumption of a nondirectional test. DISCUSSION We have shown that survivors of childhood ALL who received triple intrathecal therapy for CNS prophylaxis have mild but consistent deficits in singletrial visual and verbal short-term memory. The pattern of STM deficits of comparable severity on both visual and verbal STM tasks resembles results from prior

8 868 Hill, Ciesielski, Sethre-Hofctad, Duncan, and Lorenzi studies of ALL survivors who received chemotherapy and irradiation. For example, Cousens et al. (1991) found verbal STM deficits in ALL survivors when compared to children with solid tumors and sibling controls. Our study elucidates the nature of the memory deficit by describing the differential effects of chemotherapy on visual and verbal multitrial learning. There were similar degrees of abnormalities in visual-spatial learning after repeated exposure as in single-trial memory, while the results of verbal learning were within normal range and similar to controls. Visual-spatial memorization appears to be, therefore, more vulnerable than verbal memorization to the deleterious effects of chemotherapy. Some earlier studies of ALL survivors found the predominant cognitive deficit in visually mediated tasks. Pfefferbaum-Levine et al. (1984) described memory deficits for spatial material, but with no group differences on languagebased measures of memory. Similarly, Copeland et al. (1985) identified both visual-spatial skills and nonverbal memory as major areas of impairment in ALL survivors treated with CNS prophylaxis. To the best of our knowledge, it has not been reported that the same children (ALL survivors) show deficits in verbal single-trial memory but no problems on verbal multiple exposure learning tasks, while visually mediated learning and memory are both abnormal. Consistent with previous reports (e.g., Packer, Meadows, Rorke, Goldwein, & D'Angio, 1987), the two groups were significantly different on IQ measures with the ALL group points lower. It is unlikely that the decrease in IQ is caused by the general memory problems, for example difficulty in memorizing instructions, since no difficulty in following task procedures was observed in the children. Nor does the literature support any clear causal low IQlow memory relationship, and indeed low IQ may go in parallel with high memory performance (e.g., Wang & Bcllugi, 1994). It is possible, however, that deficits in a more basic function, and common to both measures such as attentional distractibility or application of strategic planning may be responsible for the depressed IQ and memory scores. The significant correlations between low IQ and poor single-trial memory scores but no significant relation between multitrial learning and IQ favors involvement of attentional and planning abilities. The studies reporting susceptibility to attentional distractiveness and planning deficits (e.g., Britton, Morris, Kernachan, & Craft, 1992; Peckham, Meadows, Bartel, & Marrero, 1988; Whitt et al., 1984) and our electrophysiological studies on selective attention in ALL survivors (Ciesielski, Hill et al., 1994) are consistent with the attentional and strategic planning explanations. The generalization of the present conclusions to the population of ALL survivors at large must be done with caution as it is a retrospective, small sample study. Because pretreatment measures of neuropsychological functioning were not available, we rigorously reviewed parental, school, and medical reports to determine how children were progressing prior to the onset of ALL; the review

9 Chemotherapy Impact on Short-Term Memory 869 pointed to healthy, normal physical and psychological development. Since there was also no significant between-group difference in SES markers, we assumed with a reasonable degree of confidence that the consistent deficits in short-term memory in our ALL group resulted from the deleterious impact of the chemotherapy and possible impact of systemic disease itself. Our conclusions are reassured by past research which suggested that there are no differences between children with ALL and normal controls at the time of diagnosis (Copeland et al., 1988). However, in relation to the latter, there are methodological concerns for using the results of performance of newly diagnosed children already influenced by first chemotherapy dose and psychological stress. A carefully designed prospective and longitudinal study would be of great value. In summary, the present study elucidates the pattern of memory deficits in survivors of childhood leukemia treated with intrathecal chemotherapy, who were mostly under 5 years of age at diagnosis and at least 3 years posttreatment. The study demonstrates consistent single-trial visual-spatial and verbal shortterm memory deficits, while the multitrial learning was deficient only in the visual-spatial domain, but not in verbal. This specific pattern of memory-learning deficits has been identified through the use of the WRAML. We suggest that the WRAML may be a useful clinical tool to evaluate differential memory deficits in children. REFERENCES Bleyer, W. A. (1988). Central nervous system leukemia. Pediatric Clinics of North America, 35, Boogerd, W. (1995). Neurological complications of chemotherapy. In F. A. de Wolff (Ed.), Handbook of clinical neurology: Vol 21. Intoxications of the nervous system, Part II (pp ). Amsterdam: Elsevier Science. Britton, J., Morris, R. G., Kemachan, J., & Craft, A. W. (1992). Memory after treatment for acute lymphoblastic leukaemia. Archives of Diseases of Childhood, 67, Brown, R. T., Madan-Swain, A., Pais, R., Lambert, R. G., Baldwin, K., Casey, R., Frank, N., Sexson, S., & Ragab, A. (1992). Cognitive status of children treated with central nervous system prophylactic chemotherapy for acute lymphocytic leukemia. Archives of Clinical Neuropsychology, 7, Brouwers, P., Riccardi, R., Fedio, R., & Poplack, D. (1985). Long-term neuropsychological sequelae of childhood leukemia: Correlations with CT brain scan abnormalities. Journal of Pediatrics, 106, Ciesielski, K. T., Hill, D. E., Jung, R. E., Knight, J. E., Prince, R. J., Duncan, M. H,, & Hart, B. L. (1994). Cerebellar-frontal lobe abnormalities do not fully explain autism symptomology: MRI, ERP, and neuropsychological findings. Society for Neuroscience Abstracts, 168, 10. Ciesielski, K. T, Yanofsky, R., Ludwig, R. N., Hill, D. E., Hart, B. L., Astur, R. S., & Snyder, T. (1994). Hypoplasia of the cerebellar vermis and cognitive deficits in survivors of childhood leukemia. Archives of Neurology, 51, Copeland, D. R., Dowell, R. E. Jr., Fletcher, J. M., Sullivan, M. P., Jaffe, N., Cangir, A., Frankel, L. S., & Judd, B. W. (1988). Neuropsychological test performance of pediatric cancer patients at diagnosis and one year later. Journal of Pediatric Psychology, 13, Copeland, D. R., Fletcher, J. M., Pfefferbaum-Levine, B., Jaffe, N., Ried, H., & Maor, M. (1985).

10 870 Hill, Ciesielski, Sethre-Hofstad, Duncan, and Lorenzi Neuropsychological sequelae of childhood cancer in long-term survivors. Pediatrics, 75, Cousens, P., Ungerer, J. A., Crawford, J. A., & Stevens, M. M. (1991). Cognitive effects of childhood leukemia therapy: A case for four specific deficits. Journal ofpediatric Psychology, 16, Foucar, K., Duncan, M. H., Stidley, C. A., Wiggins, C. L., Hunt, W. C, & Key, C. R. (1991). Survival of children and adolescents with acute lymphoid leukemia: A study of American Indians and Hispanic and non-hispanic whites treated in New Mexico (1969 to 1986). Cancer, 67, Gilbert, M. R., Harding, B. L., & Grossman, S. A. (1989). Methotrexate neurotoxicity: In vitro studies using cerebellar explants from rats. Cancer Research, 49, Harris, M. B. (1995). Basic statistics for behavioral science research. Boston: Allyn & Bacon. Hollingshead, A. B. (1975). Four factor index of social status. New Haven, CT: Yale University Department of Sociology. Moore, B. D., Copeland, D. R., Ried, H., & Levy, B. (1992). Neurophysiological basis of cognitive deficits in long-term survivors of childhood cancer. Archives of Neurology, 49, Moss, H. A., Nannis, E. D., & Poplack, D. G. (1981). The effects of prophylactic treatment of the central nervous system on the intellectual functioning of children with acute lymphocytic leukemia. American Journal of Medicine, 71, Mulhem, R., Wasserman, A., Fairclough, D., & Ochs, J. (1988). Memory function in disease-free survivors of childhood acute lymphoblastic leukemia given CNS prophylaxis with or without 1,800 cgy cranial irradiation. Journal of Clinical Oncology, 6, Paakko, E. L., Vainionpaa, M., Lanning, J., Laitiness, J., & Pyhtiness, J. (1992). White matter changes in children treated for acute lymphoblastic leukemia. Cancer, 70, Packer, R. J., Meadows, A. T., Rorke, L. B., Goldwein, J. L. & D'Angio, G. (1987). Long-term sequelae of cancer treatment on the central nervous system in childhood. Medical and Pediatric Oncology, 15, Peckham, V. C, Meadows, A. T, Bartel, N., & Marrero, D. (1988). Educational late effects in long-term survivors of childhood acute lymphocytic leukemia. Pediatrics, 81, Pfefferbaum-Levine, B., Copeland, D. R., Fletcher, J. M., Ried, A. L., Jaffe, N., & McKinnon, W. R. (1984). Neuropsychological assessment of long-term survivors of childhood leukemia. American Journal of Pediatric HematologylOncology, 6, Phillips, P. C, Dhawan, V., Strother, S. C, Stieltis, J. J., Evans, A. C, Allen, J. C, & Rottenberg, D. A. (1987). Reduced cerebral glucose metabolism and increased brain capillary permeability following high-dose methotrexate chemotherapy: A positron emission tomographic study. Annals of Neurology, 21, Poplack, D. G. (1989). Acute lymphoblastic leukemia. In P. Pizzo & D. Poplack (Eds.), Principles and practices of pediatric oncology. Philadelphia, PA: Lippincott. Sheslow, D., & Adams, W. (1990). Wide range assessment of memory and learning: Administration manual. Wilmington, DE: Jastak Assessment Systems. Wang, P. P., & Bellugi, U. (1994). Evidence from two genetic syndromes for a dissociation between verbal and visual-spatial short-term memory. Journal of Clinical and Experimental Neuropsychology, 16, Wechsler, D. (1991). Wechsler Intelligence Scale for Children (3rd ed.). New York: Harcourt Brace Jovanovich. Whitt, J. K., Wells, R. J., Lauria, M. M., Wilhelm, C. L., & McMillan, C. W. (1984). Cranial radiation in childhood acute lymphoblastic leukemia. American Journal of Diseases of Children, 138,