Abstract
BACKGROUND:
The objective of this study was to determine the impact of seizure-related factors on neurocognitive, health-related quality of life (HRQOL), and social outcomes in survivors of childhood cancer.
METHODS:
Survivors of childhood cancer treated at St. Jude Children’s Hospital (n = 2022; 48.3% female; median age, 31.5 years; median time since diagnosis, 23.6 years) completed neurocognitive testing and questionnaires. The presence, severity, resolution, and treatment history of seizures were abstracted from medical records. Adjusting for the age at diagnosis, sex, and prior cancer therapy, multivariable models examined the impact of seizures on neurocognitive and HRQOL outcomes. Mediation analyses were conducted for social outcomes.
RESULTS:
Seizures were identified in 232 survivors (11.5%; 29.9% of survivors with central nervous system [CNS] tumors and 9.0% of those without CNS tumors). In CNS tumor survivors, seizures were associated with poorer executive function and processing speed (P < .02); in non-CNS tumor survivors, seizures were associated with worse function in every domain (P < .05). Among non-CNS survivors, seizure severity was associated with worse processing speed (P = .023), and resolution was associated with better executive function (P = .028) and attention (P = .044). In CNS survivors, seizure resolution was associated with improved attention (P = .047) and memory (P < .02). Mediation analysis revealed that the impact of seizures on social outcomes was mediated by neurocognitive function.
CONCLUSIONS:
Seizures in cancer survivors adversely affect long-term functional and psychosocial outcomes independently of cancer therapy. The resolution of seizure occurrence is associated with better outcomes. Seizure severity is associated with poorer outcomes and should be a focus of clinical management and patient education.
Keywords: brain tumor, cancer survivors, leukemia, neurocognitive, seizure
INTRODUCTION
Survivors of childhood cancer are a growing population bearing the burden of morbidities related to their cancer and/or its treatment. Among the estimated 500,000 survivors in the United States, 66% to 88% have some chronic health condition.1 A comorbidity that may be particularly grievous for normal brain development and social attainment is seizures. Neurocognitive dysfunction is a well-known outcome in noncancer patients with epilepsy. A large national study conducted in children diagnosed with epilepsy found that seizures increased the risk for developmental and physical comorbidities2 and working memory and processing speed deficits.3
The occurrence of seizures in survivors varies according to central nervous system (CNS) involvement and the intensity of CNS-directed therapy. However, it has been reported that 8% to 13% of leukemia survivors and 27% to 41% of CNS tumor survivors develop seizures,4,5 with some developing seizure recurrence more than 2 years after their initial withdrawal from antiseizure drug therapy.6 Although the etiology of the seizure may be a driver in neurocognitive outcomes, studies have demonstrated that both interictal spikes and seizure frequency may be independent risk factors for the development of cognitive problems.7–11 As CNS-directed cancer therapy can be neurotoxic, it is important to recognize independent contributions of seizures to neurocognitive outcomes, quality of life, and social attainment (ie, educational attainment, employment, income, and marital status), which may differ from those seen in survivors exposed to CNS-directed therapy who do not develop seizures.
Previous studies of childhood cancer survivors have demonstrated associations between on-therapy seizures and problems in attention, working memory, and processing speed as well as lower IQ.12,13 Use of anti-epileptic medications has been shown to adversely affect self-reported neurocognitive function in long-term survivors of childhood cancer.14 Studies in noncancer patients with epilepsy have shown that anti-epileptic medications have many adverse cognitive and behavioral effects,15,16 and neurocognitive outcomes may depend on seizure severity and control17. An elevated risk for poor health-related quality of life (HRQOL) and social attainment has been seen in cancer survivors.18,19 In noncancer patients, seizure frequency and lifelong anti-epileptic drug (AED) use have been found to be associated with poor HRQOL and reduced employment rates, whereas those who achieve seizure freedom may have HRQOL like that of the general population.20,21
Unfortunately, limited data are available in the literature to show the effect of a seizure diagnosis in adult survivors of childhood cancer on neurocognitive outcomes, HRQOL, and social attainment. The purpose of this study was to examine associations between seizures and neurocognitive outcomes among survivors of childhood CNS and non-CNS tumors, including impacts of seizure severity, seizure resolution, and anti-epileptic medication use. The direct and indirect impacts of seizures on HRQOL and social attainment were also examined.
MATERIALS AND METHODS
This study was approved by the institutional review board at St. Jude Children’s Research Hospital. All participants provided informed written consent.
Participants
Survivors treated at St. Jude Children’s Research Hospital who were enrolled in the institutional St. Jude Lifetime Cohort (SJLIFE) were potentially eligible for the study. The SJLIFE study design and characteristics have previously been described.22 To be eligible for SJLIFE, survivors had to be treated at St. Jude Children’s Research Hospital and to be at least 5 years from their diagnosis. As we were interested in long-term outcomes in adults, the current analyses were limited to those 18 years old or older who were 10 or more years from their cancer diagnosis. The exclusion criteria were seizures before and unrelated to the cancer diagnosis, developmental or genetic disorders associated with cognitive impairment, and neurologic injuries unrelated to the cancer diagnosis or treatment. We identified 4346 eligible survivors; 1645 did not participate in SJLIFE, and after the application of study-specific eligibility criteria, 679 were excluded. This left 2022 participants who were evaluable for the study (Fig. 1).
Figure 1.
Consolidated Standards of Reporting Trials diagram. EEG indicates electroencephalogram; SJLIFE, St. Jude Lifetime Cohort.
Procedures
Medical record abstraction was performed to capture diagnoses and cancer treatment data, AED use, and seizure characteristics from the records of survivors treated at St. Jude between 1962 and 2013. To control for differences in treatment regimens over time, patient-specific chemotherapy and radiation doses were abstracted and used in this analysis. The frequency and persistence of seizures and medication use for seizures were reviewed by 2 board-certified neurologists. Seizures were diagnosed on the basis of a verified electroencephalogram or a medical record review by a board-certified neurologist. The following seizure and AED characteristics were collected: type of seizure, number of past seizure events, number of current seizure events, past number of AEDs, maximum number of AEDs at any one time, current number of AEDs, and seizure resolution whether on or off AEDs. “Past” presence of a seizure was defined as having a seizure after cancer diagnosis and up to one year prior to their SJLIFE assessment. A current seizure was defined as any seizure occurring within the 12 months before the time of SJLIFE testing. Resolution of seizures was defined as having no seizures within the 12 months before SJLIFE testing time and being on no AEDs. Both seizure characteristics and AED characteristics were obtained from the past and current time definitions. For the 64% of survivors who were seizure-free at the time of their neuropsychological assessment, we used the past seizure characteristics for our final analysis. Otherwise, we used the current seizure characteristics. Seizure severity was classified with a data-driven approach. We explored using the type of seizure, the maximum number of seizures, the cumulative number of AEDs, and the types of AEDs and found that these metrics were highly correlated in our data set. We, therefore, decided to use a seizure severity score based on the cumulative seizure frequency for our analysis. The seizure frequency was later dichotomized as ≥3 or <3 on the basis of the significant difference in outcomes between groups with 1 to 2 seizures and ≥3 seizures. This methodology may underestimate the effect of seizure frequency in this analysis. This final severity scoring (>3 or ≤3 cumulative seizures), ever having a seizure after the cancer diagnosis (yes or no), and seizure resolution (yes or no) 12 months before SJLIFE testing with or without AED use were also used as the final categorical variables in association with neurocognitive outcomes, HRQOL outcomes, and social attainment. Those survivors with missing seizure severity scores (n = 3; 1.9%) or AED severity scores (n = 2; 1.2%) were excluded from the analysis.
Neurocognitive Evaluations
Neurocognitive evaluations were completed by licensed/certified examiners under the general supervision of a board-certified clinical neuropsychologist. The assessed neurocognitive domains (and instruments) included attention (Trail Making Test Part A,23 Conner’s Continuous Performance Test II,24 and Digit Span Forward from the Wechsler Adult Intelligence Scale III [WAIS-III]25), memory (California Verbal Learning Test-II26), processing speed (the coding and symbol search from WAIS-III), and executive function (Trail Making Test Part B,23 Controlled Oral Word Association Test,23 and Digit Span Backward from WAIS-III).
HRQOL
HRQOL was measured with patient-reported outcomes via the Medical Outcomes Survey 36-Item Short Form.27 The 36-Item Short Form is widely used and provides subscale scores for 8 domains of HRQOL: general health, role limitations physical, physical function, bodily pain, vitality, mental health, social function, and role limitations emotional. Age- and sex-specific norms were used, and t scores ≤ 40 were classified as impaired HRQOL.
Social Attainment
Survivors completed surveys during follow-up to provide information about marital status, employment, education, and household income. The marital status was categorized as single/never been married or ever married. Education was categorized as attaining less than a college degree or attaining at least a college degree. Employment was categorized as less than full-time employment or full-time employment. Household income was categorized as less than $40,000/y or at least $40,000/y.
Statistical Analysis
Descriptive statistics were computed for demographic and treatment variables. Age-adjusted z scores based on population norms were calculated for neurocognitive tests, and age- and sex-adjusted t scores were calculated for HRQOL outcomes. A larger z score or t score represented better performance. The analyses were stratified by diagnosis and categorized as having CNS or non-CNS involvement. Multiple linear regression analyses were conducted to examine the effects of the presence of seizures on neurocognitive and HRQOL outcomes, and modified Poisson regression analyses were performed to estimate the effect of seizures on social attainment outcomes.28 Covariates included age, sex, and brain radiation (yes or no) for the CNS stratum and age, sex, brain radiation (no, 0 to <24 Gy, or ≥24 Gy), high-dose intravenous methotrexate, and intrathecal methotrexate for the non-CNS stratum. The SAS macro %MEDIATION was used to conduct mediation analyses of social outcomes (ie, marital status, education, employment, and household income) and mediators (neurocognitive domains, namely attention, memory, executive function, and processing speed, and physical functioning), with adjustments made for age and sex.29 The neurocognitive domain score was the average z score of the individual tests of the domain. A separate analysis was conducted for each outcome and mediator. Within the survivors who had seizures, the effects of seizure resolution (yes or no) and the seizure severity score (<3 or ≥ 3) on neurocognitive and HRQOL outcomes were examined via multiple regression, and their effects on social attainment outcomes were evaluated via modified Poisson regression. The covariates were the same as those listed previously for the analyses in the entire study population.28 All analyses were conducted with SAS version 9.4.
Closed testing procedures were applied at each level of analysis to adjust for multiple-comparison testing.30
RESULTS
Survivor Characteristics
Among the 2022 survivors (48.3% female; median age at follow-up, 31.5 years [range, 18.4-65.9 years]; median time since diagnosis, 24 years [range, 10.4-51.1 years]), 232 (11.5%) had a seizure history (Table 1). A seizure history was identified more frequently in survivors of CNS tumors (n = 72; 29.9%) in comparison with non-CNS tumor survivors (n = 160; 9.0%; Supporting Table 1). Among CNS survivors with a seizure history, the most common tumor location was infratentorial (n = 129; 55.6%), and medulloblastoma was the most common diagnosis (n = 66; 28.4%) (Supporting Information Table 2). Patients who survive supratentorial tumors tend to have lower grade tumors, may be exposed to less CNS-directed therapy, and/or may have fewer complications, and this may be associated with the higher prevalence of seizures seen in the infratentorial tumor sample. No correlation was found between stroke (P = .2325), second neoplasm (P = .33), methotrexate (P = .13), or vincristine (P = .4) and seizures.
TABLE 1.
Descriptive Characteristics
| Variable | Level | Entire Cohort (n = 2022) | Survivors with Seizures (n = 232) | ||
|---|---|---|---|---|---|
| CNS survivors (n = 241) n(%) | Non CNS survivors (n=1781) n(%) | CNS survivors (n=72) n(%) | Non CNS survivors (n=160) n(%) | ||
| Sex | Female | 103(42.7) | 873(49.0) | 30(41.7) | 77(48.1) |
| Male | 138(57.3) | 908(51.0) | 42(58.3) | 83(51.9) | |
| Race | White | 201(83.4) | 1515(85.1) | 62(86.1) | 150(93.8) |
| Non-white | 40(16.6) | 266(14.9) | 10(13.9) | 10(6.3) | |
| Diagnosis | Leukemia | 997(56.0) | 126(7.1%) | ||
| Brain Tumor | 241(100.0) | 72(29.9%) | |||
| Lymphoma | 339(19.0) | 14(0.8%) | |||
| Solid Tumor | 445(25.0) | 20(1.1%) | |||
| Mean Age at Diagnosis (range) | Years | 8.9 (0.3-22.7) | 8.2 (0.0-23.6) | 8.6(0.3-19.8) | 5.0(0.2-21.8) |
| Mean Age at Follow up (years) | Years | 26.7 (18.9-53.1) | 33.6(18.4-65.9) | 27.6((19.1-53.1) | 34.9(18.9-55.1) |
| Mean Time Since Diagnosis (range) | Years | 17.8 (10.7-41.8) | 25.4(10.4-51.1) | 18.9(11.0-41.8) | 27.6(11.4-43.5) |
| Number of IT MTX | 0 | 236(97.9) | 720(40.4) | 69(95.8) | 32(20.0) |
| 1 | 5(2.1) | 945(53.1) | 3(4.2) | 112(70.0) | |
| 2 | 108(6.1) | 14(8.8) | |||
| >=3 | 8(0.4) | 2(1.3) | |||
| Cranial Radiation | Yes | 162(67.2) | 741(41.6) | 40(55.6) | 108(67.5) |
| No | 79(32.8) | 1039(58.4) | 32(44.4) | 52(32.5) | |
| >=24Gy | 160(66.4) | 433(24) | 39(54.2) | 74(46.3) | |
| <24 Gy | 2(0.8) | 308(17) | 1(1.4) | 34(21.3) | |
| Marital Status | Single, never married | 164(70.4) | 592(34.3) | 47(67.1) | 58(36.5) |
| Married, living as married, Separated, divorced, widowed | 69(29.6) | 1136(65.7) | 23(32.9) | 101(63.5) | |
| Education | < college graduate | 185(76.8) | 1140(64.0) | 52(72.2) | 108(67.5) |
| >= college graduate | 56(23.2) | 641(36.0) | 20(27.8) | 52(32.5) | |
| Employment | Part time or unemployed | 146(63.5) | 665(38.8) | 48(68.6) | 81(51.6) |
| Full time | 84(36.5) | 1049(61.2) | 22(31.4) | 76(48.4) | |
| Household income | < $40,000 | 108(60.3) | 672(45.1) | 32(58.2) | 70(49.6) |
| $40,000 or more | 71(39.7) | 818(54.9) | 23(41.8) | 71(50.4) | |
IT, intrathecal; MTX, methotrexate; CNS, central nervous system; N, number; AED, anti-epileptic drug; SJL, St. Jude Life follow-up visit. Seizure severity and AED severity were the graded based on cumulative number greater than 12 months prior to St. Jude life testing.
Seizure type, seizure severity, AED severity, Max N of AED at 1 time obtained in past seizure history means seizures recorded from time of seizure diagnosis until prior to 1 year before SJLIFE testing
Most of the survivors with seizures (n = 164; 70.1%: 55.6% of CNS survivors and 67.5% of non-CNS survivors) had a resolution of their seizures 1 year before longterm follow-up. Non-CNS tumor survivors were more likely to not experience resolution if they had a secondary neoplasm (n = 12; P = .03). No association between seizure resolution and demographic or treatment effects was identified in the CNS group. More than half of the survivors with seizures (n = 135; 58.2%) had a seizure severity score ≥ 3 (75.0% of those with CNS tumors and 51.6% of those with non-CNS tumors). Among non-CNS tumor survivors, seizures were most common in patients with leukemia (n = 126; 78.8%), and this is consistent with previously published reports.31 Non-CNS survivors with seizures had more intrathecal methotrexate injections (80.0% vs 59.6%) and cranial radiation (67.5% vs 41.6%) than those without seizures. Moreover, seizure resolution was not associated with sex, age, or treatment. Finally, non-CNS survivors had correlations between seizures and stroke (P = .015) and anxiety (P = .005), but not depression.
Neurocognitive Outcomes
A history of seizures was associated with a clinically significantly lower z score (age-adjusted standard score change of 1/2 standard deviation) for cognitive flexibility in survivors of CNS tumors (effect size [ES], −0.74; P = .02) after we controlled for cranial radiation and in non-CNS tumor survivors (ES, −0.68; P < .001) after we controlled for cranial radiation and methotrexate (Fig. 2). The presence of seizures in survivors of CNS tumors was associated with clinically significantly lower scores in focused attention (ES, −1.09, P < .001), variability (ES, −0.52; P = .003), short-term memory (ES, −0.63; P = .001), long-term memory (ES, −0.58; P = .003), and visual-motor processing speed (ES, −0.50; P = .001). In non-CNS survivors with seizures, all domains (executive function, attention, memory, and processing speed) were associated with statistically significantly lower scores (P < .001). When we assessed the impact of seizure resolution and seizure severity (>3 or ≥3 cumulative seizures) on neurocognitive and HRQOL measures, only those outcomes that were significantly associated with a past seizure history were examined. CNS survivors with seizure resolution, compared with those with current seizures, were associated with clinically significantly higher scores in focused attention (ES, 1.20; P = .047), new learning (ES, 0.72; P = .017), short-term recall (ES, 1.02; P = .005), and long-term recall (ES, 1.01; P = .010) (Table 2). Seizure severity had no impact on neurocognitive outcomes in CNS survivors. Non-CNS survivors with a resolution of seizures had higher scores in working memory (ES, 0.41; P = .028) and focused attention (ES, 0.66; P = .044). In non-CNS survivors, worse seizure severity (>3 cumulative seizures) was associated with poorer visual-motor (ES, −0.43; P = .023) and visual processing speed (ES, −0.47; P = .019). The severity score of AEDs (0-1 vs ≥2) was significantly associated with worse efficiency (mean, −0.7678 vs −1.2012; P = .0151) and worse memory (mean, −0.5343 vs −0.9405; P = .0151) in non-CNS groups and worse processing speed (mean, −0.4383 vs −0.7815; P = .0100) in CNS groups. No other significant associations were seen.
Figure 2.
Presence of seizure effect on neurocognitive measures. The models have been adjusted for age and sex for CNS and non-CNS tumors in addition to what is listed in the table. Non-CNS tumors have also been adjusted for intrathecal methotrexate (yes or no). CNS tumors have been adjusted for radiation therapy. Coefficient estimates reflect the change in the age-adjusted z score (mean, 0; standard deviation, 1.0) associated with each variable (a larger estimate or z score represents a better performance). Bars represent 95% confidence intervals. The boundary of clinically significant estimates is shown on the x-axis with arrows. CNS indicates central nervous system.
TABLE 2.
Among survivors with seizure, seizure severity related to neurocognitive and health-related quality of life outcomes.
| Outcomes Measure | CNS survivors | Non-CNS survivors | ||||||
|---|---|---|---|---|---|---|---|---|
| Seizures resolved | Seizure Severity | Seizures resolved | Seizure Severity | |||||
| Est | P | Est | P | Est | P | Est | P | |
| Neurocognitive Domain | ||||||||
| Executive function | ||||||||
| Cognitive Flexibility | 0.84 | 0.17 | 0.47 | 0.49 | 0.68 | 0.076 | −0.40 | 0.26 |
| Cognitive fluency | - | - | - | - | 0.19 | 0.39 | −0.37 | 0.065 |
| Working memory | - | - | - | - | 0.41 | 0.028 | −0.12 | 0.49 |
| Attention | ||||||||
| Span | - | - | - | - | - | - | - | - |
| Sustained | - | - | - | - | 0.27 | 0.15 | 0.17 | 0.32 |
| Focused | 1.20 | 0.047 | 0.33 | 0.63 | 0.66 | 0.044 | −0.53 | 0.079 |
| Variability | 0.53 | 0.14 | −0.33 | 0.42 | 0.31 | 0.25 | −0.34 | 0.17 |
| Detectability | - | - | - | - | −0.06 | 0.74 | −0.03 | 0.86 |
| Memory | ||||||||
| New learning | 0.72 | 0.017 | 0.33 | 0.32 | 0.40 | 0.11 | −0.41 | 0.075 |
| Short-term recall | 1.02 | 0.005 | 0.09 | 0.83 | 0.34 | 0.21 | −0.38 | 0.12 |
| Long-term recall | 1.01 | 0.010 | 0.18 | 0.68 | 0.53 | 0.057 | −0.51 | 0.051 |
| Processing speed | ||||||||
| Visual motor | 0.37 | 0.19 | 0.26 | 0.42 | 0.27 | 0.18 | −0.43 | 0.023 |
| Visual | 0.65 | 0.053 | 0.11 | 0.77 | 0.31 | 0.15 | −0.47 | 0.019 |
| Health-Related Quality of Life | ||||||||
| Physical Health | ||||||||
| Physical functioning | - | - | - | - | 2.66 | 0.21 | −1.08 | 0.58 |
| Physical role | - | - | - | - | 0.70 | 0.74 | −0.62 | 0.76 |
| Bodily pain | - | - | - | - | 1.08 | 0.60 | 1.85 | 0.34 |
| General health | - | - | - | - | 5.76 | 0.010 | 3.33 | 0.11 |
| Overall physical | - | - | - | - | 2.42 | 0.22 | −0.19 | 0.92 |
| Mental Health | ||||||||
| Vitality | 3.92 | 0.14 | −2.49 | 0.42 | 2.61 | 0.21 | 2.12 | 0.27 |
| Social functioning | - | - | - | - | 2.15 | 0.32 | 0.98 | 0.62 |
| Emotional role | - | - | - | - | 0.47 | 0.85 | 1.78 | 0.44 |
| Mental health | - | - | - | - | 2.44 | 0.29 | 2.01 | 0.35 |
| Overall mental | - | - | - | - | 1.88 | 0.43 | 2.72 | 0.22 |
Seizures resolved refers to yes vs. no. Seizure severity was scored as ≥3 (most severe) vs. <3 (less severe) in the past. For CNS, covariates included age, sex, brain radiation yes vs no; for non-CNS, covariates included age, sex, brain radiation dose group (none, <24 Gy, ≥24 Gy), high dose methotrexate, and intrathecal methotrexate. Est = standardized estimate, reflecting the change in age-adjusted z-score (neurocognitive) or t-score (health-related quality of life). A larger estimate represents a better performance. Dash line (−) denotes test not significantly associated with past or current history of seizure, therefore not tested in this model. Bold indicates significant difference.
HRQOL Outcomes
Non-CNS survivors with seizures had poorer physical functioning (ES, −2.72; P = .001), physical role limitations (ES, −2.56; P = .003), and general health (ES, −2.59; P = .006) in comparison with non-CNS survivors without seizures (Fig. 3). Interestingly, CNS survivors with seizures reported higher vitality (ES, 3.49; P = .016) in comparison with CNS survivors without seizures. The resolution of seizures was associated with better general health in CNS tumor survivors (ES, 5.76; P = .010; Table 2). Seizure severity did not affect HRQOL in CNS or non-CNS survivors.
Figure 3.
Presence of seizure effect on health-related quality of life. The models have been adjusted for age and sex for both CNS and non-CNS tumors in addition to what is listed above. Non-CNS tumors have also been adjusted for intrathecal methotrexate (yes or no). CNS tumors have been adjusted for radiation therapy. Estimates reflect the change in the age-adjusted z score (mean, 0; standard deviation, 1.0) associated with each variable (a larger estimate or t score represents a better performance). Bars represent 95% confidence intervals. The boundary of clinically significant estimates is shown on the x-axis with arrows. CNS indicates central nervous system.
Social Attainment
In CNS survivors, the presence of a seizure history conveyed a 9% greater risk of being unemployed or employed part-time (relative risk [RR], 1.09; 95% confidence interval [CI], 1.01-1.18; Table 3). Likewise, non-CNS survivors with a seizure history had an 11% increased risk for part-time employment or unemployment (RR, 1.11; 95% CI, 1.05-1.17). In survivors of non-CNS tumors, the resolution of seizures was associated with a lower risk for less than full-time employment (RR, 0.83; 95% CI, 0.74-0.93). There was a trend for decreased risk in CNS tumor survivors, but it did not reach statistical significance (RR, 0.89; 95% CI, 0.79-1). CNS survivors with seizure resolution had a lower risk of being single or never married (RR, 0.85; 95% CI, 0.76-0.97) in comparison with those without resolution. No associations were found between a seizure history or seizure resolution and education or household income. Seizure severity had no association with social attainment in either CNS or non-CNS survivors.
Table 3.
Among survivors with seizure, effect of seizure resolution and seizure severity on social attainment outcomes.
| Outcomes Measure | CNS survivors | Non-CNS survivors | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Seizure Presence | Seizures resolved | Seizure Severity | Seizure Presence | Seizures resolved | Seizure Severity | |||||||
| RR | 95% CI | RR | 95% CI | RR | 95% CI | RR | 95% CI | RR | 95% CI | RR | 95% CI | |
| Marital Status Single or Never married | 1.02 | 0.95,1.09 | 0.85 | 0.76, 0.97 | 0.93 | 0.81,1.05 | 1.03 | 0.97,1.08 | 0.95 | 0.85,1.07 | 1.03 | 0.92, 1.15 |
| Married/Living as married | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | ||||||
| Education (< college graduate) | 0.98 | 0.92,1.05 | 0.95 | 0.84,1.08 | 0.91 | 0.80,1.02 | 1.02 | 0.97,1.07 | 0.99 | 0.89,1.09 | 1.04 | 0.94,1.15 |
| College graduate | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | ||||||
| Employment (part time or unemployed) | 1.09 | 1.01,1.18 | 0.88 | 0.78,1.00 | 0.94 | 0.81,1.09 | 1.11 | 1.05,1.17 | 0.83 | 0.74,0.93 | 1.02 | 0.91,1.15 |
| Full time employment | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | ||||||
| Household income (<$40,000) | 0.98 | 0.88,1.08 | 1.01 | 0.85,1.19 | 0.95 | 0.79,1.13 | 1.02 | 0.96,1.08 | 0.93 | 0.82,1.06 | 0.98 | 0.87,1.10 |
| Household income (≥$40,000) | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | ||||||
Seizures resolved refers to yes vs. no. Seizure severity was scored as ≥3 (most severe) vs. <3 (less severe) in the past. For CNS, covariates included age, sex, brain radiation yes vs no; for non-CNS, covariates included age, sex, brain radiation dose group (none, <24 Gy, ≥24 Gy), high dose methotrexate, and intrathecal methotrexate. Bold indicates significant relative risk.
Mediation Analysis
In CNS survivors, the impact of seizure history on marital status was partially mediated by problems with memory (odds ratio [OR], 1.32; 95% CI, 1.00-1.74), attention (OR, 1.33; 95% CI, 1.03-1.71), and processing speed (OR, 1.41; 95% CI, 1.02-1.94), and the impact of seizure history on employment was partially mediated by problems with memory (OR, 1.49; 95% CI, 1.02-2.20), attention (OR, 1.71; 95% CI, 1.12-2.61), and processing speed (OR, 1.56; 95% CI, 1.04-2.35; Fig. 4 and Table 4). In non-CNS survivors, the risk for less than full-time employment was mediated by poor executive function (OR, 1.36; 95% CI, 1.22-1.51), memory (OR, 1.30; 95% CI, 1.19-1.43), attention (OR, 1.49; 95% CI, 1.32-1.68), processing speed (OR, 1.34; 95% CI, 1.21-1.48), and overall physical health (OR, 1.19; 95% CI, 1.06-1.33). Marital status was mediated by executive function (OR, 1.18; 95% CI, 1.09-1.28), memory (OR, 1.17; 95% CI, 1.08-1.27), attention (OR, 1.21; 95% CI, 1.10-1.33), and processing speed (OR, 1.17; 95% CI, 1.08-1.27).
Figure 4.
Mediation analysis of seizures related to social attainment. Mediation analysis measured the direct impact of seizures on social attainment and the indirect impact of seizures on social attainment via neurocognitive outcomes. HRQOL indicates health-related quality of life.
TABLE 4.
In the entire cohort, presence of seizure related to social attainment, adjusted for age and sex.
| Outcomes Measure | Mediator | CNS survivors | Non-CNS survivors | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Direct | Indirect | Direct | Indirect | ||||||
| OR | 95% CI | OR | 95% CI | OR | 95% CI | OR | 95% CI | ||
| Marital status | Executive function | 0.92 | 0.45,1.86 | 1.15 | 0.91,1.47 | 1.25 | 0.85,1.83 | 1.18 | 1.09,1.28 |
| Memory | 0.79 | 0.39,1.62 | 1.32 | 1.00,1.74 | 1.26 | 0.86,1.86 | 1.17 | 1.08,1.27 | |
| Attention | 0.83 | 0.40,1.69 | 1.33 | 1.03,1.71 | 1.20 | 0.82,1.77 | 1.21 | 1.10,1.33 | |
| Processing speed | 0.77 | 0.36,1.64 | 1.41 | 1.02,1.94 | 1.25 | 0.85,1.83 | 1.17 | 1.08,1.27 | |
| PCS | 1.04 | 0.51,2.12 | 1.04 | 0.88,1.23 | 1.43 | 0.98,2.09 | 1.01 | 0.98,1.04 | |
| Employment | Executive function | 1.66 | 0.84,3.31 | 1.19 | 0.89,1.59 | 1.50 | 1.05,2.13 | 1.36 | 1.22,1.51 |
| Memory | 1.36 | 0.68,2.74 | 1.49 | 1.02,2.20 | 1.55 | 1.09,2.21 | 1.30 | 1.19,1.43 | |
| Attention | 1.15 | 0.57,2.33 | 1.71 | 1.12,2.61 | 1.39 | 0.97,1.98 | 1.49 | 1.32,1.68 | |
| Processing speed | 1.20 | 0.58,2.48 | 1.56 | 1.04,2.35 | 1.52 | 1.07,2.17 | 1.34 | 1.21,1.48 | |
| PCS | 1.81 | 0.89,3.69 | 1.08 | 0.76,1.53 | 1.72 | 1.20,2.47 | 1.19 | 1.06,1.33 | |
OR, odds ratio; CI, confidence interval. Executive function, memory, attention, and processing speed: average z-scores of the individual tests in the domain. PCS, overall physical health of HRQOL. Bold indicates significant increased odds.
DISCUSSION
Our results indicate that after adjustments for age, sex, and prior cancer treatments known to impair functional outcomes in cancer survivors, a diagnosis of seizures adversely affected neurocognitive outcomes, HRQOL, and social attainment in long-term survivors of CNS and non-CNS cancer. In CNS cancer survivors with a seizure history, seizure resolution was associated with better attention and memory, marital status, and employment. In non-CNS cancer survivors with a seizure history, seizure resolution was associated with better executive function and attention outcomes. In contrast to CNS cancer survivors, in non-CNS cancer survivors, seizure resolution was associated with better general health and employment status. Similarly, seizure severity was associated with worse neurocognitive function in non-CNS survivors but not in CNS survivors. This may indicate that in CNS survivors, the underlying etiology has a greater impact on neurocognitive outcomes than seizure control. Finally, seizure severity had no direct impact on HRQOL or social attainment in either CNS or non-CNS survivors. However, the impact of seizures on neurocognitive outcomes does appear to mediate the effect of seizures on HRQOL and social attainment.
Neurocognitive dysfunction is a well-known outcome in noncancer patients with epilepsy. In noncancer pediatric populations, this manifests as working memory and processing speed deficits.3 Attention-deficit/hyperactivity disorders are also often seen in children with epilepsy.32,33 Our data suggest that pediatric cancer survivors, who are already at risk for neurocognitive impairment because of their cancer treatment history, experience additional risk if they develop a seizure diagnosis, particularly in the domains of executive function, attention, memory, and processing speed. In patients without cancer, seizure severity affects neurocognitive outcomes with respect to the frequency of seizures, the duration of seizures, and higher use of AEDs.17,34 However, in our study, the severity of past seizures was negatively associated with worse neurocognitive function in non-CNS survivors but not in CNS survivors even though 18% of CNS survivors had intractable seizures in comparison with 6% of non-CNS survivors. This discrepancy between CNS and non-CNS tumor survivors may indicate that the underlying etiology for the seizure disorder is the main driver for neurocognitive impairments in CNS tumor survivors.34 However, in non-CNS survivors, there is a clear association between seizure frequency and neurocognitive impairment independent of the underlying etiology, and this implies that survivors of non-CNS tumors may be a more vulnerable population with a lower threshold for impact from seizures of any severity.
Seizure resolution was the single most common predictor of better neurocognitive outcomes among those survivors with a seizure history. Specifically, seizure resolution was associated with improved attention and memory in CNS survivors and with improvements in executive function, attention, and memory in non-CNS survivors in comparison with survivors without seizure resolution. Although most of our patients had resolution of their seizures at the long-term follow-up, 36% had unresolved seizures. Our findings are consistent with reports of the importance of seizure resolution in patients without cancer.35 Seizure resolution in noncancer patients with epilepsy surgery secondary to refractory seizures results in improved IQ and verbal memory in comparison with presurgery IQ and in comparison with those who do not have surgery.9,36 Achieving seizure resolution appears to be crucial to improved neurocognitive outcomes in childhood cancer survivors.
A younger age at diagnosis, female sex, and cranial radiation therapy exposure are known risk factors for diminished HRQOL in cancer survivors.19,37,38 Adjusting for these known risks, we found that non-CNS survivors with a past history of seizures reported worse HRQOL in physical functioning, physical role, and general health. Seizure resolution was associated with improved general health in non-CNS survivors, but seizure severity was not associated with HRQOL in CNS survivors. These findings may suggest that the experience of seizures, regardless of severity, is equally impactful on perceived quality of life, perhaps because of imposed functional limitations, but once the survivor has seizure resolution and the limitations are lifted, quality of life is perceived to be significantly better.
Adult survivors of childhood cancer have impaired neurocognitive function, which leads to lower social attainment, such as being less likely to graduate from college and gain full-time employment.39–41 Physical limitations in adult childhood cancer survivors may also lead to less than full-time employment and reduced education and annual income.19 Interestingly, neither CNS nor non-CNS cancer survivors with seizures had any increased risk of decreased education attainment in contrast to what is seen in noncancer patients with epilepsy.42 This may be related to the emphasis placed on early education interventions (individualized education plans and 504 plans) by the Children’s Oncology Group (https://www.childrensoncologygroup.org/). However, mediation analysis demonstrated that in CNS survivors with seizures, impaired attention, memory, and processing speed affected a survivor’s odds of having full-time employment and marital status. These results are similar to those seen in noncancer patients with epilepsy, who have reduced employment and marriage rates.43–45 We found that non-CNS survivors with seizures had impaired neurocognitive and physical function, which was associated with reduced social attainment (marital status and employment). Achieving seizure resolution is the most important prognostic factor for employment outcomes in non-CNS survivors and for improving marital status and employment in CNS survivors. The mediation analysis suggests that improvements in neurocognitive and overall physical performance outcomes may result in improved employment. A similar effect of improving neurocognitive function may help with marital status and employment in CNS survivors. Seizures directly affect employment outcomes in non-CNS survivors, and this suggests that aggressive management of seizure control, focused early interventions, and/or vocational rehabilitation may help this subgroup. Therefore, we propose that early neurocognitive interventions may be warranted in both CNS and non-CNS survivors with seizures, and early physical fitness interventions in non-CNS survivors with seizures may help to improve social outcomes.46,47 Our results are similar to those of a report demonstrating that seizure freedom in noncancer patients with epilepsy led to improved social attainment, but they differ in that seizure severity did not affect social outcomes.21
Seizure management in patients with cancer has evolved as newer medications with better efficacy and fewer medication interactions have become available. Historically, the choice of AEDs was limited because of interactions with chemotherapy agents. Thus, participants may have had higher seizure severity scores before being switched to a newer agent that offered better control. This may account for the lack of associations between HRQOL, social attainment, and seizure severity in this study. This also means that we did not have enough variation in AEDs to fully investigate the roles that different AEDs played in the assessed outcomes. Moreover, because this study includes a population from a single institution, it may not be representative of other institutional populations; however, no other site follows survivors for this duration with objective evaluation outcomes. Additionally, the historical nature of this cohort limited our ability to apply clinically based seizure severity rating scales, and thus we used a data-driven approach for the analysis. However, data-driven approaches have been shown to be useful in identifying previously unidentified associations that are missed with clinically based scales. Finally, our study did not have a control group of noncancer patients with seizures. Nonetheless, aside from seizure severity, our results match published associations in noncancer populations.
Our results demonstrate that seizures negatively affect neurocognitive outcomes, and seizure resolution is the most important variable associated with improved outcomes in HRQOL and social attainment in CNS and non-CNS survivors. This implies the importance of close follow-up for seizure recurrence and optimal seizure control in cancer survivors with the goal of complete seizure control being actively pursued. Regular assessments of cognitive function and HRQOL along with appropriate interventions may reduce this indirect impact of seizures on social attainment. Finally, adherence to antiepileptic therapy wanes as parental fear and stress about future health complications secondary to treatment increase.48 Education regarding the beneficial aspects of therapy could be a necessary factor for interventions designed to improve adherence outcomes in young children with newly diagnosed epilepsy.49
Supplementary Material
FUNDING SUPPORT
This work was supported by grants from the National Cancer Institute (T32 CA225590 [principal investigator Kevin R. Krull], U01 CA195547 [principal investigators Melissa M. Hudson and Leslie L. Robison], and P30 CA021765 [principal investigator C. Roberts]) and from the American Lebanese Syrian Associated Charities.
We thank Jennifer Zebrowski, RN, the St. Jude Lifetime Cohort, and the families for participating in this study.
Footnotes
Additional supporting information may be found in the online version of this article.
CONFLICT OF INTEREST DISCLOSURES
Deokumar Srivastava reports consulting fees from General Dynamics Information Technology. The other authors made no disclosures.
REFERENCES
- 1.Phillips SM, Padgett LS, Leisenring WM, et al. Survivors of childhood cancer in the United States: prevalence and burden of morbidity. Cancer Epidemiol Biomarkers Prev. 2015;24:653–663. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Russ SA, Larson K, Halfon N. A national profile of childhood epilepsy and seizure disorder. Pediatrics. 2012;129:256–264. [DOI] [PubMed] [Google Scholar]
- 3.Nickels KC, Zaccariello MJ, Hamiwka LD, Wirrell EC. Cognitive and neurodevelopmental comorbidities in paediatric epilepsy. Nat Rev Neurol. 2016;12:465–476. [DOI] [PubMed] [Google Scholar]
- 4.Maytal J, Grossman R, Yusuf FH, et al. Prognosis and treatment of seizures in children with acute lymphoblastic leukemia. Epilepsia. 1995;36:831–836. [DOI] [PubMed] [Google Scholar]
- 5.Wells EM, Ullrich NJ, Seidel K, et al. Longitudinal assessment of late-onset neurologic conditions in survivors of childhood central nervous system tumors: a Childhood Cancer Survivor Study report. Neuro Oncol. 2018;20:132–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Khan RB, Morris EB, Pui CH, et al. Long-term outcome and risk factors for uncontrolled seizures after a first seizure in children with hematological malignancies. J Child Neurol. 2014;29:774–781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Glass HC, Glidden D, Jeremy RJ, Barkovich AJ, Ferriero DM, Miller SP. Clinical neonatal seizures are independently associated with outcome in infants at risk for hypoxic-ischemic brain injury. J Pediatr. 2009;155:318–323. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Korman B, Krsek P, Duchowny M, Maton B, Pacheco-Jacome E, Rey G. Early seizure onset and dysplastic lesion extent independently disrupt cognitive networks. Neurology. 2013;81:745–751. [DOI] [PubMed] [Google Scholar]
- 9.Helmstaedter C, Elger CE, Vogt VL. Cognitive outcomes more than 5 years after temporal lobe epilepsy surgery: remarkable functional recovery when seizures are controlled. Seizure. 2018;62:116–123. [DOI] [PubMed] [Google Scholar]
- 10.Kleen JK, Scott RC, Holmes GL, et al. Hippocampal interictal epileptiform activity disrupts cognition in humans. Neurology. 2013;81:18–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Krauss GL, Summerfield M, Brandt J, Breiter S, Ruchkin D. Mesial temporal spikes interfere with working memory. Neurology. 1997;49:975–980. [DOI] [PubMed] [Google Scholar]
- 12.Nassar SL, Conklin HM, Zhou Y, et al. Neurocognitive outcomes among children who experienced seizures during treatment for acute lymphoblastic leukemia. Pediatr Blood Cancer. 2017;64:e26436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Armstrong GT, Conklin HM, Huang S, et al. Survival and long-term health and cognitive outcomes after low-grade glioma. Neuro Oncol. 2011;13:223–234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Brinkman TM, Zhang N, Ullrich NJ, et al. Psychoactive medication use and neurocognitive function in adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. Pediatr Blood Cancer. 2013;60:486–493. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Loring DW, Williamson DJ, Meador KJ, Wiegand F, Hulihan J. Topiramate dose effects on cognition: a randomized double-blind study. Neurology. 2011;76:131–137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Meador KJ, Loring DW, Hulihan JF, Kamin M, Karim R. Differential cognitive and behavioral effects of topiramate and valproate. Neurology. 2003;60:1483–1488. [DOI] [PubMed] [Google Scholar]
- 17.Vingerhoets G Cognitive effects of seizures. Seizure. 2006;15:221–226. [DOI] [PubMed] [Google Scholar]
- 18.Huang IC, Quinn GP, Krull K, et al. Head-to-head comparisons of quality of life instruments for young adult survivors of childhood cancer. Support Care Cancer. 2012;20:2061–2071. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ness KK, Gurney JG, Zeltzer LK, et al. The impact of limitations in physical, executive, and emotional function on health-related quality of life among adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. Arch Phys Med Rehabil. 2008;89:128–136. [DOI] [PubMed] [Google Scholar]
- 20.Birbeck GL, Hays RD, Cui X, Vickrey BG. Seizure reduction and quality of life improvements in people with epilepsy. Epilepsia. 2002;43:535–538. [DOI] [PubMed] [Google Scholar]
- 21.Schneider-von Podewils F, Gasse C, Geithner J, et al. Clinical predictors of the long-term social outcome and quality of life in juvenile myoclonic epilepsy: 20-65 years of follow-up. Epilepsia. 2014;55:322–330. Accessed August 20, 2021. https://pubmed.ncbi.nlm.nih.gov/24417603/ [DOI] [PubMed] [Google Scholar]
- 22.Hudson MM, Ness KK, Nolan VG, et al. Prospective medical assessment of adults surviving childhood cancer: study design, cohort characteristics, and feasibility of the St. Jude Lifetime Cohort Study. Pediatr Blood Cancer. 2011;56:825–836. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Strauss E, Sherman EM, Spreen O. Attention. In: A Compendium of Neuropsychological Tests: Administration, Norms, and Commentary. 3. Oxford University Press; 2006:655. [Google Scholar]
- 24.Conners CK. Conners’ Continuous Performance Test (CPT). In: Encyclopedia of Clinical Neuropsychology. New York, NY: Springer; 2011. doi: 10.1007/978-0-387-79948-3_1535 [DOI] [Google Scholar]
- 25.Wechsler D Wechsler Adult Intelligence Scale-Fourth Edition (WAIS-IV). 3rd ed. APA Psych Tests; 2008. doi: 10.1037/t15169-000 [DOI] [Google Scholar]
- 26.Woods SP Delis DC, Scott JC, Kramer JH, Holdnack JA. The California Verbal Learning Test--second edition: test-retest reliability, practice effects, and reliable change indices for the standard and alternate forms. Arch Clin Neuropsychology. 2006;21:413–420. Accessed August 20, 2021. https://pubmed.ncbi.nlm.nih.gov/16843636/ [DOI] [PubMed] [Google Scholar]
- 27.Ware JE, Kosinski M, Gandek B. SF-36 Health Survey: Manual and Interpretation Guide. Quality Metric Incorporated; 2003. Accessed August 20, 2021. https://www.semanticscholar.org/paper/SF-36-health-survey%3A-Manual-and-interpretation-Ware/c4262cefae0217aee75dbc23400fc74d3ad416f6 [Google Scholar]
- 28.Zou G A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159:702–706. [DOI] [PubMed] [Google Scholar]
- 29.Valeri L, Vanderweele TJ. Mediation analysis allowing for exposure-mediator interactions and causal interpretation: theoretical assumptions and implementation with SAS and SPSS macros. Psychol Methods. 2013;18:137–150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Marcus R, Peritz E, Gabriel KR. Closed testing procedures with special reference to ordered analysis of variance. Biometrika. 1976;63:655–660. doi: 10.1093/biomet/63.3.655 [DOI] [Google Scholar]
- 31.Singh G, Rees JH, Sander JW. Seizures and epilepsy in oncological practice: causes, course, mechanisms and treatment. J Neurol Neurosurg Psychiatry. 2007;78:342–349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Dunn DW, Austin JK. Behavioral issues in pediatric epilepsy. Neurology. 1999;53:S96–S100. [PubMed] [Google Scholar]
- 33.Dunn DW, Austin JK, Harezlak J, Ambrosius WT. ADHD and epilepsy in childhood. Dev Med Child Neurol. 2003;45:50–54. [PubMed] [Google Scholar]
- 34.Bourgeois BF, Prensky AL, Palkes HS, Talent BK, Busch SG. Intelligence in epilepsy: a prospective study in children. Ann Neurol. 1983;14:438–444. [DOI] [PubMed] [Google Scholar]
- 35.Neyens LG, Aldenkamp AP, Meinardi HM. Prospective follow-up of intellectual development in children with a recent onset of epilepsy. Epilepsy Res. 1999;34:85–90. [DOI] [PubMed] [Google Scholar]
- 36.Bjornaes H, Stabell KE, Henriksen O, Roste G, Diep LM. Surgical versus medical treatment for severe epilepsy: consequences for intellectual functioning in children and adults. A follow-up study. Seizure. 2002;11:473–482. doi: 10.1016/s1059-1311(02)00134-6 [DOI] [PubMed] [Google Scholar]
- 37.Fernandez-Pineda I, Hudson MM, Pappo AS, et al. Long-term functional outcomes and quality of life in adult survivors of childhood extremity sarcomas: a report from the St. Jude Lifetime Cohort Study. J Cancer Surviv. 2017;11:1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Nolan VG, Krull KR, Gurney JG, Leisenring W, Robison LL, Ness KK. Predictors of future health-related quality of life in survivors of adolescent cancer. Pediatr Blood Cancer. 2014;61:1891–1894. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Prasad PK, Hardy KK, Zhang N, et al. Psychosocial and neurocognitive outcomes in adult survivors of adolescent and early young adult cancer: a report from the Childhood Cancer Survivor Study. J Clin Oncol. 2015;33:2545–2552. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Brinkman TM, Krasin MJ, Liu W, et al. Long-term neurocognitive functioning and social attainment in adult survivors of pediatric CNS tumors: results from the St Jude Lifetime Cohort Study. J Clin Oncol. 2016;34:1358–1367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Jacola LM, Edelstein K, Liu W, et al. Cognitive, behaviour, and academic functioning in adolescent and young adult survivors of childhood acute lymphoblastic leukaemia: a report from the Childhood Cancer Survivor Study. Lancet Psychiatry. 2016;3:965–972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Besag FM. Cognitive and behavioral outcomes of epileptic syndromes: implications for education and clinical practice. Epilepsia. 2006;47(suppl 2):119–125. [DOI] [PubMed] [Google Scholar]
- 43.Dansky LV, Andermann E, Andermann F. Marriage and fertility in epileptic patients. Epilepsia. 1980;21:261–271. [DOI] [PubMed] [Google Scholar]
- 44.Sillanpää M, Jalava M, Kaleva O, Shinnar S. Long-term prognosis of seizures with onset in childhood. N Engl J Med. 1998;338:1715–1722. [DOI] [PubMed] [Google Scholar]
- 45.Sillanpaa M, Schmidt D. Long-term outcome of medically treated epilepsy. Seizure. 2017;44:211–216. doi: 10.1016/j.seizure.2016.09.002 [DOI] [PubMed] [Google Scholar]
- 46.Phillips NS, Howell CR, Lanctot JQ, et al. Physical fitness and neurocognitive outcomes in adult survivors of childhood acute lymphoblastic leukemia: a report from the St. Jude Lifetime Cohort. Cancer. 2020;126:640–648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Krull KR, Hardy KK, Kahalley LS, Schuitema I, Kesler SR. Neurocognitive outcomes and interventions in long-term survivors of childhood cancer. J Clin Oncol. 2018;36:2181–2189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Loiselle K, Rausch JR, Modi AC. Behavioral predictors of medication adherence trajectories among youth with newly diagnosed epilepsy. Epilepsy Behav. 2015;50:103–107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Modi AC, Guilfoyle SM, Rausch J. Preliminary feasibility, acceptability, and efficacy of an innovative adherence intervention for children with newly diagnosed epilepsy. J Pediatr Psychol. 2013;38:605–616. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.