Psychosocial and functional outcomes in young adults with childhood-onset epilepsy: a 10-year follow-up

Abstract

AIM

To compare long-term psychosocial and functional outcomes of young adults with uncomplicated childhood-onset epilepsy (COE) to population norm controls utilizing a controlled prospective cohort study.

METHOD

Psychosocial and functional outcomes were assessed at 10-year follow-up. Fifty-three young adults (27 males, 26 females) with COE (n=21 remission; 18y 1mo–30y 9mo; mean age 23y 4mo [SD 3y 4mo]; mean age of epilepsy onset 12y [SD 3y 2mo]) were compared to 55 (23 males, 32 females) first-degree cousin controls (18y 5mo–29y 8mo; mean age 23y 6mo [SD 3y]). Seizure remission status and baseline comorbidities (attention-deficit/hyperactivity disorder [ADHD], depressive disorders, anxiety disorders, and academic problems) were examined as possible risk factors for significant differences in functional outcomes.

RESULTS

Poorer functional outcomes, indicated by patient rated cognition and overall disability, were evident among young adults with epilepsy compared to controls (all p-values <0.05). These difficulties were due to baseline comorbid ADHD and academic problems. Remission status was not related to measured cognition and overall disability.

INTERPRETATION

Psychosocial outcomes of young adults with COE were similar to controls. In contrast, functional outcomes were worse in epilepsy across cognition and overall disability. Baseline comorbid ADHD and academic problems were identified as risk factors at 10-year follow-up suggesting that these early recognized comorbidities at or near diagnosis have long-term impacts.

Epilepsy is the fourth most common neurological disorder, affecting more than 50 million people globally and about 3.4 million adults and 470 000 children in the US.^1,2 Through population- and community-based studies, researchers in Finland, Japan, the Netherlands, Canada, the UK, and US have examined the long-term prognoses of childhood-onset epilepsy (COE) across wide domains of social and psychosocial outcomes (e.g. educational attainment, employment, marital status, living arrangements, and quality of life).^5–11 Research regarding the influence and interference of seizure-related variables on long-term social and psychosocial outcomes is varied, and at times conflicting. Seminal studies, such as Sillanpää et al.,¹² concluded that individuals with uncomplicated COE have favorable and long-term medical outcomes comparable to the general population, but poorer trajectories for employment, education, and marital status persist, despite seizure remission. In contrast, a recent investigation reported young adults with uncomplicated COE in remission have psychosocial outcomes comparable to their peers.¹³

In addition to investigating the influence of seizure-related variables, it is important to ascertain if comorbidities of epilepsy contribute to the prognostic understanding of COE. Children with epilepsy are at an increased risk of comorbid disorders, which precede, co-occur with, or occur after a diagnosis of epilepsy; putting children at greater risk of repeating a grade in school, poorer social competence, and unmet medical and mental health needs.^3,4 Investigators in Norway utilized nationwide registry data to compare comorbidity patterns in children with epilepsy in order to estimate the proportions of comorbidities in these children compared to the general child population.³ Nearly 80% of children with epilepsy had at least one or more comorbid disorders, of which 55% had physical disorders (e.g. gastrointestinal disorders, chronic lower respiratory disorders); 41% had neurological disorders (e.g. cerebral palsy, headache conditions); and 43% had developmental/psychiatric disorders (e.g. autism, attention-deficit/hyperactivity disorder [ADHD]). Review articles of neurological, psychological, and physical comorbidities concluded that comorbidities are not only common and variable, but may be more disabling than seizures alone.¹⁴ In the UK and US, researchers found comorbidities (neurodevelopment, psychiatric, and learning) to not only be strongly associated with long-term health-related quality of life in COE, but also associated with later educational and employment problems.^7,15

At 10-year follow-up, we compared psychosocial and functional outcomes of young adults with uncomplicated COE (i.e. those without intellectual disability or other neurological impairment) to population norm controls. The study was a two-fold design: first comparing psychosocial and functional outcomes between individuals with uncomplicated COE and controls; second, examining whether identified differences were attributable to seizure-related factors and/or the presence of baseline comorbidities.

METHOD

Participants

Participants with uncomplicated COE were initially recruited from pediatric neurology clinics in three medical centers (University of Wisconsin-Madison, Marshfield Clinic, and Dean Clinic), located in the mid-west of the US. Control participants eligible for participation were first-degree cousins closest in age to the participant with epilepsy. It has previously been demonstrated that there is a lack of association between cases and first-degree cousin performances on measures of cognition, behavior, and neuroimaging, the results suggested a non-significant genetic influence on control group performance, inferring that first-degree cousins serve as unbiased controls for cognitive, behavioral, and neuroimaging research in pediatric epilepsy.¹⁶ Procedures and study eligibility criteria have been detailed in previous studies.¹⁷ For this investigation, a minimum of 10-years’ follow-up from baseline assessment, conducted at or near the diagnosis of epilepsy, was required. One-hundred and eight participants (53 with epilepsy and 55 age-matched control participants) met the 10-year threshold. This evaluation was completed via online data collection. Any questionable responses were clarified with a follow-up telephone interview.

Research approval was obtained from the University of Wisconsin School of Medicine and Public Health Sciences Institutional Review Board. The 10-year follow-up consent forms were mailed to participants and online consent was obtained before survey completion and verbal informed consent was obtained on the day of the study interview from participants.

Baseline comorbidities

The Kiddie Schedule for Affective Disorders and Schizophrenia – Present and Lifetime Version is a semi-structured interview which directly assesses current and lifetime history of psychopathology in all participants.¹⁸ At baseline, the Kiddie Schedule for Affective Disorders and Schizophrenia – Present and Lifetime Version was administered to each participant and their accompanying parent. To ensure interrater reliability, interviews were recorded and randomly selected for review. A consensus diagnosis was achieved when participant and parent interviews were administered by different interviewers. Based on the baseline Kiddie Schedule for Affective Disorders and Schizophrenia – Present and Lifetime Version interview, the following comorbidities were selected for this study as they were the most prevalent: any lifetime psychiatric diagnosis, ADHD, depressive disorders, and anxiety disorders.

To determine rates of academic problems at baseline, parents were questioned via the structured interview about their child’s school progress, in particular about the specific educational services provided to address academic problems. These services included the traditional individualized educational plan or 504 Plan, as well as early childhood interventions including speech therapy, physical therapy, occupational therapy, mandatory summer school, grade retention, tutoring services, and other specific educational services. The interviewer was blind to the cognitive and behavioral results.

Psychosocial outcomes at 10-year follow-up

Demographic questionnaires assessed psychosocial outcomes, including educational attainment (high school diploma/General Educational Development credential or less vs some college/trade school or more), living arrangements (living at home with parents vs living independently), employment status (paid/self-employed, student, or unemployed), personal income (above Wisconsin state average vs below), and marital status (single, divorced, or widowed vs married or cohabitating).

The World Health Organization (WHO) Quality of Life Assessment - Brief version assessed quality of life. The WHO Quality of Life Assessment - Brief version is a 26-item abbreviated version of the WHO Quality of Life Assessment - 100, used to assess quality of life.¹⁹ Quality of life is assessed across four domains: physical heath, psychological health, social relationships, and environment. Each item is rated using a 5-point Likert-type scale (1=poor to 5=good). Higher scores indicate greater quality of life. A transformed scoring method outlined by the WHO was used to convert domain scores to a 0 to 100 scale. Cronbach’s alpha for the current study was 0.92.

Functional outcomes 10-year follow-up

The WHO Disability Assessment Schedule 2.0 (WHODAS 2.0) is a 36-item, generic, self-report instrument measuring perceived individual functioning and disability across six life domains in the last 30 days: cognition (understanding and communication), mobility (getting around), self-care, getting along (getting along with people), life activities (work and/or school), and participation (participation in society). ²⁰ Each item is rated using a 5-point Likert-type scale (1=none to 5=extreme or cannot do). Higher scores indicate higher levels of disability and functional limitation. Item response theory-based scoring was applied. This approach codes each item response individually and then applies an algorithm to determine summary scores by a differential weighting of items and corresponding levels of severity.¹⁹ The algorithm provided by the WHO involves three steps: (1) summarizing recoded items within each domain; (2) summing all six domain scores and the additional overall disability score (i.e. average general disability score); and (3) converting summary scores into a metric ranging from 0 (no disability) to 100 (full disability).²⁰ Cronbach’s alpha for the present study was calculated to be 0.93.

Seizure course at 10-year follow-up

Individuals with epilepsy were considered to be in remission if they had been seizure-free for 5 or more years and were no longer taking antiepileptic drugs. This information was obtained through self-report of participants and independent proxy-report (i.e. completed by a participant’s significant other or parent).

Statistical analyses

Statistical analyses were performed using SPSS software version 26 (IBM Corp., Armonk, NY, US). Bivariate analyses, including χ² and Mann–Whitney U tests, first examined the relationship between group (young adults with epilepsy vs controls) across all psychosocial and functional outcomes. Additional χ² tests examined the relationship between group and the presence of baseline comorbidities. A series of Kruskal–Wallis analyses of variance (ANOVA) on ranks were employed to investigate whether significant differences existed between three subgroups (controls without baseline comorbidities, young adults with epilepsy without baseline comorbidities, and young adults with epilepsy with baseline comorbidities) with regard to WHODAS domain scores. Bonferroni and Games-Howell post hoc tests determined at which level significant differences existed.

RESULTS

Sample characteristics

The total sample included 58 females and 50 males. With respect to ethnicity, 98 (90.7%) were white, four (3.7%) black, four (3.7%) multiracial, one (0.9%) Hispanic, and one (0.9%) Native American. Of the 108 participants at 10-year follow-up, the sample consisted of 55 controls (18y 5mo–29y 8mo; mean age 23y 6mo [SD 3y]) and 53 young adults with COE (18y 1mo–30y 9mo; mean age 23y 3.6mo [SD 3y 3.6mo]). Mean age at diagnosis was 12 years 0mo (SD 3y 2mo). Twenty-one participants (39.6%) were considered to be in remission, 14 (26.4%) with localization-related syndromes (e.g. Rolandic epilepsy, temporal lobe epilepsy) and seven (13.2%) with generalized syndromes (e.g. juvenile myoclonic epilepsy, absence epilepsy). Thirty-two participants (60.4%) were not in remission, 15 (28.3%) with localization-related syndromes and 17 (32.1%) with generalized syndromes.

Several factors were examined as potential predictors of non-response at 10-year follow-up: mother’s age at childbirth, biological mother’s educational attainment, group (epilepsy vs control), sex, parental baseline marital status, baseline full-scale IQ score, localization-related versus generalized epilepsy syndrome, specific epilepsy syndrome (e.g. Rolandic epilepsy, temporal lobe epilepsy, absence epilepsy), baseline or lifetime Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) diagnosis, and baseline current DSM-IV diagnosis. All factors were determined to be insignificant except a diagnosis of epilepsy (p=0.012) any lifetime (p=0.027) and/or current DSM-IV diagnosis at baseline (p=0.015), and higher full-scale IQ score (p=0.003).

Psychosocial outcomes: epilepsy versus control

As shown in Table 1 and Table 2, young adults with epilepsy and controls did not differ significantly across educational attainment, living arrangement, employment status, marital status, personal income, and quality of life (WHO Quality of Life Assessment - Brief version).

Table 1:

Psychosocial outcomes of young adults with epilepsy compared to controls

	Controls (n=55)	Cases (n=53)	χ2	p
Education			1.17	0.279
HS/GED and less	8 (14.5)	12 (22.6)
Some college/trade+	47 (85.5)	41 (77.4)
Living arrangement			0.01	0.909
Lives at home	14 (25.5)	14 (26.4)
Not at home	41 (74.5)	39 (73.6)
Employment status			0.22	0.817
Paid/self-employed	31 (56.4)	32 (60.4)
Student	20 (36.4)	17 (32.1)
Unemployed	4 (7.3)	4 (7.5)
Personal incomea			0.07	0.795
Above average	51 (92.7)	41 (77.4)
Below average	3 (5.5)	3 (5.7)
Marital status			0.01	0.905
Single/divorced/widowed	40 (72.7)	38 (71.7)
Married/cohabitating	15 (50.0)	15 (28.3)
Baseline comorbidities
Any baseline diagnosis	26 (47.3)	41 (77.4)	10.37	0.001
Baseline academic problems	11 (20)	25 (47.2)	8.99	0.003
Baseline ADHD	4 (7.3)	14 (26.4)	7.12	0.008
Baseline depression	2 (3.6)	11 (20.8)	7.47	0.006
Baseline anxiety	14 (25.5)	15 (28.3)	0.11	0.739

Data are n (%) unless otherwise stated. Significant p-values (<0.05) are in bold type.

^aTen participants preferred not to answer, thus controls were n=54 and cases n=44. HS/GED, high school/general educational development; ADHD, attention-deficit/hyperactivity disorder.

Table 2:

WHODAS and WHOQOL outcomes of young adults with epilepsy compared to controls

	Controls (n=55)			Cases (n=53)
	Mean	SD	Median	Mean	SD	Median	z-score	p
WHODAS
Cognition	12.37	13.28	8.33	19.73	13.82	16.67	3.15	0.002
Mobility	4.36	8.05	0.00	7.08	13.57	0.00	0.15	0.884
Self-care	3.07	8.41	0.00	3.07	5.15	0.00	0.78	0.439
Getting along	10.82	13.19	5.00	13.40	13.33	10.00	1.29	0.196
Life activities	14.09	10.81	12.50	17.94	13.94	15.63	0.92	0.357
Participation	11.19	13.85	6.25	16.86	16.47	12.50	2.12	0.034
Overall disability score	9.30	8.90	8.09	12.88	9.71	9.34	2.07	0.039
WHOQOL
Physical health	16.71	1.93	17.00	16.21	2.63	16.00	−0.72	0.471
Psychological	14.89	2.82	15.00	14.43	2.96	15.00	−0.96	0.336
Social relationships	15.29	3.02	16.00	15.25	2.79	15.00	−0.15	0.884
Environment	16.02	2.13	16.00	15.45	2.64	16.00	−1.05	0.295

Significant p-values (<0.05) are in bold. WHODAS, World Health Organization Disability Assessment Schedule; WHOQOL, World Health Organization Quality of Life Assessment.

Functional outcomes: epilepsy versus control

Using the WHODAS as a measure of functional limitations, Mann–Whitney U tests determined that individuals with epilepsy had greater measured disability, as indicated by the following domains: cognition (p=0.002), participation (p=0.034), and overall disability (p=0.039) compared to controls. As summarized in Table 3, Mann–Whitney U tests further indicated that epilepsy remission status was not related to the WHODAS domain scores of cognition, participation, and overall disability.

Table 3:

World Health Organization Disability Assessment Schedule domain scores of young adults with epilepsy by remission status

	Remission (n=21)			Non-remission (n=32)
	Mean	SD	Median	Mean	SD	Median	z-score	p
Domain
Cognition	18.65	11.76	16.67	20.44	15.17	18.75	−0.28	0.776
Participation	15.33	14.62	9.38	17.87	17.74	15.63	−0.36	0.722
Overall disability score	12.62	9.60	8.79	13.05	9.93	9.53	−0.21	0.834

χ² tests revealed young adults with epilepsy had increased baseline rates of having any of the targeted comorbidity diagnoses (77.4% vs 47.3%; p=0.001), with increased baseline rates specifically for academic problems (47.2% vs 20%; p=0.003), ADHD (26.4% vs 7.3%; p=0.008), and depression (20.8% vs 3.6%, p=0.006) compared to controls. Baselines rates of anxiety were not different between groups (Table 1).

Kruskal–Wallis ANOVA rank analyses

Kruskal–Wallis ANOVA rank analyses revealed a statistically significant difference between WHODAS cognitive domain scores and the baseline presence of: any targeted comorbidity diagnosis (H[2]=10.02, p=0.007); academic problems (H[2]=14.95, p=0.001); ADHD (H[2]=15.13, p=0.001); and depression (H[2]=10.62, p=0.005). Analyses also revealed statistically significant differences between WHODAS participant domain scores and the baseline presence of academic problems (H[2]=6.49, p=0.039) and ADHD (H[2]=6.98, p=0.030). Further, there were statistically significant differences between WHODAS overall disability scores and the baseline presence of academic problems (H[2]=7.54, p=0.023) and ADHD (H[2]=7.15, p=0.028).

Dunn-Bonferroni post hoc analyses

Post hoc analysis results revealed significant group differences within the WHODAS cognitive domain to be between the following groups with epilepsy and controls without comorbidity: individuals with epilepsy and any targeted baseline comorbidity (p=0.007), individuals with epilepsy with academic problems (p=0.001), individuals with epilepsy with ADHD (p=0.001), and individuals with epilepsy without depression (p=0.004). The following groups with epilepsy were found to be significantly different from controls without comorbidity across the overall disability domain: those without academic problems (p=0.031) and those with ADHD (p=0.040).

DISCUSSION

This investigation yielded two key findings. First, at 10-year follow-up of a cohort with newly diagnosed epilepsy, young adults with epilepsy had similar psychosocial outcomes compared to controls in educational attainment, living arrangement, employment status, personal income, marital status, and quality of life. At first glance, these findings appear incongruent with several seminal studies that reported markedly disparate outcomes between intellectually average adults with uncomplicated COE and controls, including reduced rates of educational attainment, employment, marriage, and cohabitation.^21–23 However, the results are consistent with findings from a population-based study in Japan, and a community-based cohort in the US, in which adults with COE had similar health, educational, and employment outcomes to control comparisons.^10,24 While representing a cohort older than those examined in this study, a recent 50-year follow-up determined education, work, and leisure activities to be minimally affected, if at all, in adults diagnosed with uncomplicated COE.²⁵

Second, independent of seizure remission status, differences in the measurement of functional limitations at 10-year follow-up between young adults with epilepsy and controls were detected in cognitive and overall scores, as indicated by the WHODAS domain scores. Greater cognitive disability in individuals with epilepsy was attributable to the independent presence of both baseline ADHD and baseline academic problems. Academic problems also appeared to be the driving force between greater overall disability measurement among individuals with epilepsy compared to controls. The impact of specific baseline comorbidities on perceived disability is comparable to earlier investigations that have suggested factors other than seizure-related variables, including chronic comorbidities such as psychiatric disorders and learning problems, to be associated with poorer psychosocial functioning and quality of life outcomes in young adults with COE.^5–7,11,15 Noting that adolescents with epilepsy are more likely to have social and academic difficulties than their general population peers, and that prevalence rates of ADHD are two to three times higher in persons with epilepsy,^26,27 the importance of early identification and interventions at the outset of epilepsy may contribute to improved psychosocial and academic outcomes in adulthood and across the lifespan.^4,28,29

Further findings from the present study not only corroborate the significance of these specific neurobehavioral comorbidities in the broader spectrum of literature, but also demonstrate their pervasive impact at 10-year follow-up, as indicated by a previous follow-up investigation with the present cohort conducted at 2-years after diagnosis in the context of ADHD and academic problems.³⁰ The study by Hermann et al. yielded four key findings: (1) the presence of neurobehavioral comorbidities, specifically ADHD and academic problems, was significantly associated with worse baseline cognitive function, particularly executive functions; (2) children with epilepsy with ADHD and academic problems exhibited lower cognitive growth over the 2-year interval than controls and children with epilepsy without comorbidities; (3) children with epilepsy without ADHD and academic problems were comparable and equivalent to controls in all areas of cognitive and academic achievement at baseline and 2-year follow-up; and (4) the presence of ADHD and academic problems in children with epilepsy was independent of epilepsy-related factors (e.g. age at onset, epilepsy syndrome), other medical conditions (e.g. asthma), and use of medications for other associated medical conditions. These findings highlighted the critical contribution of ADHD and academic problems at the time of epilepsy diagnosis in children as an indicator of abnormal cognitive development antecedent to and after the onset of epilepsy This is noteworthy in the context of the present study, which found individuals with epilepsy to have greater levels of cognitive and overall disability and functional impairment 10 years after epilepsy diagnosis.

Limitations

This study has several limitations. First the sample is relatively young. As most participants were in their early twenties, several significant life events (e.g. post-secondary graduation, marriage, and living independently) may have yet to occur and, therefore, are likely to have contributed to the lack of differences between participants. This study can only be considered a snapshot of the otherwise long-term trajectory and development of participants as they move into adulthood; laying the foundation for future research endeavors that monitor changes in psychosocial outcomes as participants develop and reach what may be considered adult milestones. This supposition is supported by researchers in Japan¹⁰ who found overall reduced rates of marriage among their epilepsy group. However, the authors acknowledged that the participants were disproportionately aged between 20 and 24 years and therefore the low rates were possibly due to a higher proportion of students in the sample. Second, participant bias may exist in the characteristics of the participants who chose to participate at follow-up. Baseline factors that were identified as significant contributors to non-response at 10-year follow-up for all participants included a diagnosis of epilepsy (p=0.012) and any lifetime (p=0.027) and/or current DSM-IV diagnosis at baseline (p=0.015). Among participants with epilepsy, a higher IQ was associated with participation at 10-year follow-up (p=0.003). Third, it is recognized that the sample size for this study may be considered small and therefore limited in power and generalizability of results.

CONCLUSION

This study offers several unique contributions. First, it followed new onset cases of children with epilepsy into young adulthood 10 years after diagnosis. Second, the identification of psychiatric comorbidity was particularly notable for two reasons: (1) psychiatric diagnoses consistent with the DSM-IV were identified through structured interviews rather than self-report, thus reducing threats to internal validity, and (2) these disorders were identified at or near the time of epilepsy diagnosis. Finally, this study offers additional insight into how the presence of comorbid conditions at baseline may influence the trajectory of psychosocial outcomes in uncomplicated COE.

The long-term impact of baseline comorbidities on functional limitations is not only a key consideration in the medical treatment of COE but also in the development and implementation of rehabilitation interventions during the transition from childhood to young adulthood. As the differences in outcomes between controls and young adults with epilepsy was tentatively discerned to be additive, the effect of the interaction between epilepsy and comorbidities, addressing both a primary index diagnosis in addition to secondary or tertiary conditions, is important, both for comprehensive care and understanding observed life differences when epilepsy is well-controlled. Findings from this study suggest professionals should not only focus on the perceived functional limitations of COE, but also on antedated cognitive and psychiatric comorbid conditions early in the course of epilepsy.

What this paper adds.

• Young adults with childhood-onset epilepsy (COE) and controls have similar psychosocial outcomes 10 years after diagnosis.

• Young adults with COE report greater limitations in cognition and overall disability than controls.

• Baseline presence of attention-deficit/hyperactivity disorder and academic problems significantly affect cognitive and overall disability scores.

ABBREVIATIONS

COE

Childhood-onset epilepsy

WHO

World Health Organization

WHODAS

World Health Organization Disability Assessment Schedule