Adverse Childhood Experiences Are Associated With Seizures in Children

Abstract

Background and Objective

To assess the relationship between adverse childhood experiences (ACE/ACEs) and epilepsy.

Methods

We performed a cross-sectional retrospective cohort analysis using population-based data from the 2018 and 2019 National Survey of Children's Health to examine caregiver-reported ACE exposures and their relationship to caregiver-reported physician diagnoses of epilepsy or seizure disorder in children. ACEs elicited in the survey included questions about experience of violence, household dysfunction, and food and housing insecurity. Adjusting for age, race, and income level, we used logistic regression to test the relationships between cumulative ACE score and current seizure disorder or epilepsy diagnosis and to examine which specific ACEs were individually associated with current seizure disorder or epilepsy diagnosis.

Results

The study population consisted of 59,963 participants; 52.2% were female, and 47.8% were male. Participant ages ranged from 0 to 17 years. A current diagnosis of epilepsy or seizure disorder was reported in 377 (0.63%) participants, and 22,749 (37.9%) participants had one or more ACE exposures. As the number of ACEs increased, odds of current epilepsy or seizure disorder diagnosis increased by 1.14 (95% confidence interval 1.07–1.22). Five ACE exposures demonstrated a high association with a current diagnosis of epilepsy or seizure disorder: food/housing insecurity, witnessing domestic violence, household mental illness, neighborhood violence, and parent/guardian incarceration.

Discussion

Multiple ACE exposures were individually associated with reporting a diagnosis of epilepsy or seizure disorder. An increase in cumulative ACE exposures increased odds of having current diagnosis of epilepsy or seizure disorder.

An estimated 1.2% of the US population has epilepsy, including roughly 470,000 children.¹ Although the most commonly identified causes of epilepsy are structural, infectious, traumatic, genetic, metabolic, and vascular in nature, various environmental exposures may also be associated with the development and course of epilepsy. Exposure to adverse childhood experiences (ACE/ACEs), which include household dysfunction, physical and emotional abuse, and neglect, is widespread in the United States; it is estimated that approximately 45% of children in the United States experience at least 1 ACE during childhood or adolescence.^2,3 ACE exposure has been assessed using the National Survey of Children's Health (NSCH), a validated population-based survey administered by the US Census Bureau since 2003 that collects data on various childhood health issues.⁴ The associations between exposure to adverse childhood experiences and a variety of neurologic disorders, such as headache and ADHD, have been previously described in children.^5-11 By contrast, few studies have investigated the association between ACEs and epilepsy diagnosis and severity, despite previous research demonstrating that the prevalence of ACEs is high in children with epilepsy.² Here, we aim to (1) assess the relationship between ACEs and epilepsy and (2) determine whether there are specific ACEs that are more strongly associated with the diagnosis of epilepsy or seizure disorder in a pediatric cohort using population-based survey data.

Methods

Data Description

This retrospective cohort study used deidentified data from the 2018 and 2019 National Survey of Children's Health (NSCH). The NSCH is sponsored by the US Maternal and Child Health Bureau and administered by the US Census Bureau. The survey is stratified by state and sample type (landline, mobile, web-based, and mail questionnaire) with planned oversampling of children from birth to 5 years and children with special needs of all ages. Survey data are reweighted to allow for generalizability to the US pediatric population at both the state and national levels. Sampling frame includes addresses from households across the United States which is randomly selected each year, and only 1 child from each household is randomly selected to be the subject of the main topical questionnaire. Respondents are adult parents and/or primary caregivers who could answer questions about a child's health, health care utilization, and home and neighborhood environments. Each two-year combined data set also includes an adjusted probability weight variable that accounts for combining 2 years of data and addresses the concern for duplicate sampling.

To remain representative of a strictly population-based sample, the survey design implements a number of strategies to minimize nonresponse bias. These include incentives and other treatments to increase response and weights that account for differential nonresponse. After survey weights were applied, extensive nonresponse bias analyses for the 2018 and 2019 survey years performed by the NSCH indicated no systematic bias in estimates.

Data Analysis

Within the NSCH, respondents were asked 1 joint question about epilepsy or seizure disorder; the diagnoses were not separately elicited. To capture participants who carried a diagnosis of epilepsy or seizure disorder at the time of survey administration, the following question was asked: “What is the parent-rated severity of child's current epilepsy or seizure disorder?” Original survey answer choices were does not currently have condition, current condition mild, or current condition moderate or severe. Answer choices were recoded into a dichotomous categorical variable: ‘1’ yes current epilepsy or seizure disorder diagnosis (comprised current condition mild and current condition moderate or severe) vs ‘0’ does not currently have condition.

The cumulative ACE score provided in the data set ranged from 0 to 9 and was the summation of each ‘Yes’ answer to any of the following 9 exposures elicited: having a divorced parent or guardian; having a parent or guardian die; having a parent or guardian in jail; seeing an adult slap, hit, kick, or punch another; being the victim of neighborhood violence; living with a person with an alcohol or drug abuse; living with a person with mental illness; unfair treatment because of race; and difficulty covering basics such as food or housing. All ACE exposures were elicited as binary responses with answer choices as ‘Yes’ or ‘No’ to exposure except for the exposure question, “How often is it hard to cover basics such as food or housing,” which had 4 possible responses: ‘no,’ ‘rarely,’ ‘somewhat often,’ and ‘very often.’ Any response other than ‘No’ for this question was counted toward the cumulative ACE score. ACE scores were separated into the following groupings: score of 0, 1, 2, 3, or 4 or greater. Respondents were excluded from the analysis if questions regarding seizure/epilepsy diagnosis or ACE exposures had missing values.

Regarding behavioral and cognitive diagnoses, caregivers were asked separately whether the child had current ADD/ADHD diagnosis; 1 joint question on whether the child was taking medication for ADD/ADHD, autism/ASD, or difficulties with emotions, concentration, or behavior; and 1 joint question on whether the child had a mental, emotional, developmental, or behavioral (MEDB) problem; the latter diagnoses were not separately elicited. All these questions were elicited as binary responses with answer choices as ‘Yes’ or ‘No.’

Univariate analyses were conducted between seizure/epilepsy diagnosis status and the following demographic factors such as age, sex, race, ethnicity, income level, and insurance status. Participants had the following options for race and/or ethnicity: Hispanic, White, Black, Asian, American Indian/Alaska Native, Native Hawaiian/Other Pacific Islander, or Multiple Race. We used the following clustering groups for race/ethnicity given small sample sizes for certain groups: Hispanic, non-Hispanic White, non-Hispanic Black, and multiracial and/or non-Hispanic other race. Significant associations were found for income level, race, and age, and these demographic factors were controlled for in our multivariate logistic regression model. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated through logistic regression to identify which specific ACEs were individually associated with current seizure disorder or epilepsy diagnosis. Analyses were conducted using STATA version 17 (StataCorp, College Station, TX).

Data Availability

All data within this article are available from the Child and Adolescent Health Measurement Initiative [2018–2019] National Survey of Children's Health (NSCH) [Stata Constructed Data Set] and were retrieved [10/10/21] from childhealthdata.org. Data Resource Center for Child and Adolescent Health is supported by Cooperative Agreement U59MC27866 from the US Department of Health and Human Services, Health Resources and Services Administration (HRSA), and Maternal and Child Health Bureau (MCHB). Further anonymized data can be made available to qualified investigators on request to the corresponding author.

Standard Protocol Approvals, Registrations, and Patient Consents

This study was exempt from review by the Children's Hospital of Philadelphia Institutional Review Board.

Results

The study population consisted of 59,963 participants, of which 52.2% were female and 47.8% were male. Participant age ranged from 0 to 17 years, and 11.9% of participants self-identified as Hispanic, 6.4% Black (non-Hispanic), 69.2% White (non-Hispanic), and the remaining 12.5% multiracial and/or non-Hispanic (Table 1). The prevalence of individual ACEs in the sample is presented in Table 2 with parental divorce as the most common ACE in the sample (22.3%). A current diagnosis of epilepsy or seizure disorder was reported in 377 (0.63%) participants, and 22,749 (37.9%) respondents reported one or more ACE exposures (Table 3). As the number of ACE exposures increased, the odds of current epilepsy or seizure disorder diagnosis increased by 1.14 (95% CI 1.07–1.22) (Table 4). After adjusting for the demographic factors of age, race, and income level, the ACE exposures that demonstrated individual associations with current diagnosis of epilepsy or seizure disorder were (1) difficulty covering basics such as food or housing: ‘rarely’ (OR 1.67, 95% CI 1.30–2.13), ‘somewhat often’ (OR 2.48, 95% CI 1.82–3.38), and ‘very often’ (OR 3.69, 95% CI 2.35–5.80); (2) witnessing domestic violence (OR 1.54, 95% CI 1.08–2.21); (3) having a household member with mental illness (OR 1.53, 95% CI 1.14–2.05); (4) being the victim of or witnessing neighborhood violence (OR 1.60, 95% CI 1.06–2.40); and (5) having a parent or guardian who served time in jail (OR 1.52, 95% CI 1.09–2.11) (Table 4).

Table 1.

Descriptive Statistics for Participants Included in Analysis (n = 59,963)

Table 2.

Prevalence of Adverse Childhood Experiences (ACEs) in Analytic Sample (n = 59,963)

Table 3.

Descriptive Statistics for Variables Included in Analysis (n = 59,963)

Table 4.

Association of Current Epilepsy or Seizure Disorder Diagnosis With Demographic Variables, Cumulative ACE Score, and Individual ACEsa (n = 59,963)

In our adjusted model of cumulative ACE score, associations between current epilepsy or seizure diagnosis and age, race, and income level were found (Table 4). Higher age was associated with slightly increased odds of reporting epilepsy or seizure diagnosis (OR 1.09, 95% CI 1.01–1.18, p < 0.05). Black race, but not White or multiracial/other race, was strongly associated with increased odds of reporting epilepsy or seizure diagnosis compared with Hispanic race (OR 2.26, 95% CI 1.46–3.51, p < 0.001). Higher income was associated with a decreased odds of reporting seizure or epilepsy diagnosis, with incomes 200–399% of federal poverty level and 400% of federal poverty level having odds ratios of 0.56 (CI 0.41–0.77, p < 0.001) and 0.58 (CI 0.42–0.79, p < 0.01), respectively. Neither sex nor insurance status demonstrated a statistically significant association with epilepsy or seizure disorder diagnosis.

Within the epilepsy/seizure disorder, population is a high incidence of associated comorbidities, including behavioral issues, ADHD, and cognitive limitations. The odds of current epilepsy or seizure disorder diagnosis was very strongly associated with current ADD/ADHD diagnosis (OR 3.07; CI 2.39–3.94, p < 0.001), taking medication for ADD/ADHD, autism/ASD or other emotional/behavioral difficulties (OR 3.32; CI 2.60–4.24, p < 0.001), or mental, emotional, developmental, or behavioral problems (OR 9.22; CI 7.22–11.77, p < 0.001), adjusting for aforementioned demographic factors (Table 5). For this reason, further analyses were performed on these select comorbidities to investigate whether they are more proximal factors associated with adverse childhood experiences than seizures themselves. However, as the number of ACE exposures increased, the odds of current epilepsy or seizure disorder diagnosis still increased by 1.14, even when controlling for current ADD/ADHD diagnosis (CI 1.07–1.22, p < 0.001), taking medication for ADD/ADHD, autism/ASD or other emotional/behavioral difficulties (CI 1.07–1.21, p < 0.001), or mental, emotional, developmental, or behavioral problems (CI 1.06–1.21, p < 0.001) (Table 6).

Table 5.

Association of Current Epilepsy or Seizure Disorder Diagnosis With Current ADD/ADHD; Taking Medication for ADD/ADHD, Autism/ASD or Other Emotional/Behavioral Difficulties; and Mental, Emotional, Developmental, or Behavioral Problems

Table 6.

Association of Current Epilepsy or Seizure Disorder Diagnosis With Cumulative ACE Score When Controlling for These^a (n = 59,963)

Discussion

This study adds population-based evidence for the association between cumulative ACE exposure and a diagnosis of epilepsy or seizure disorder. Moreover, of the demographic characteristics examined, notably Black race was associated increased odds of epilepsy or seizure diagnosis (OR 2.26, 95% CI 1.46–3.51, p < 0.001). Although it is unclear what is mediating the increased odds of epilepsy among Black children in this pediatric cohort, racism at either or both the individual and structural level may be playing an important role in disease development and burden. This study additionally demonstrates that 5 of the 9 individual ACE exposures are independently associated with an epilepsy or seizure disorder diagnosis. These individual ACE exposures include (1) food and housing insecurity; experiencing or witnessing (2) domestic or (3) community violence; and having (4) an incarcerated household member or (5) a household member with mental illness (including suicidality or severe depression).

Our findings are in line with other studies of ACE exposure and epilepsy/seizure diagnosis in childhood, although we used novel methods by studying this association through a large national survey database.⁴ One recent study performed at Kaiser Permanente Northern California demonstrated the association of ACE exposure and epilepsy in childhood and adolescence.² In a primary care clinic setting, children were screened for ACE exposure and asked about a wide array of health issues. Epilepsy/seizure disorder was found to be associated with exposure to any ACE (OR 3.92, 95% CI 1.77–8.67, p < 0.01) and even more strongly associated with exposure to 3 or more ACEs (OR 4.40, 95% CI 1.21–16.05, p < 0.01). Individual-level ACE exposure associations were not described in this study. Another study, conducted in Germany, interviewed 120 adult patients with epilepsy and found that, compared with matched controls, patients with epilepsy had higher emotional and sexual maltreatment scores.¹² Our study adds to these findings by using an unexamined large population-based nationally representative sample and adding insight into individual ACE exposure associations.

This study presented here demonstrates an association between ACEs and epilepsy/seizure disorder diagnosis and that 5 of the 9 specific ACEs were independently associated with increased risk of current epilepsy or seizure disorder diagnosis. This suggests either that certain forms of household dysfunction and violence may confer higher risk for developing epilepsy or seizure disorders or that epilepsy/seizure disorders may influence or predispose to the experience of certain ACEs. Interestingly, there seemed to be a dose response direct relationship between seizure disorder/epilepsy diagnosis and how frequently a family experienced difficulty covering basics such as food or housing, with odds ratios increasing incrementally from ‘rarely’ experiencing difficulty to ‘somewhat frequent’ difficulty to ‘very frequent’ difficulty (1.67 vs 2.48 vs 3.69, respectively). This indicates that food and housing insecurity may be an especially important ACE risk factor regarding epilepsy or seizure disorder diagnosis in children. Identification of specific higher risk individual ACE epilepsy associations may enable more targeted and prioritized approaches for intervention and future research such as targeted screenings/referrals or initiatives that address food and housing security for particularly vulnerable patients.

A critical limitation of this study is imprecision of the "seizure disorder and epilepsy diagnosis” measure. This includes absence of distinct information regarding the child's specific diagnosis (e.g., epilepsy, isolated seizure, provoked seizure, and psychogenic nonepileptic events) as well as potential reporter unreliability. Caregivers—who are inherently less reliable proxies for reporting medical diagnoses than medical professionals—were surveyed for these data, which increase potential for diagnostic reporting error; it is possible that patients without a true diagnosis of epilepsy (recurrent unprovoked seizures) and/or those with a history of isolated seizure were included in the epilepsy/seizure disorder diagnosis category based on caregiver interpretation of their child's diagnosis. Error may also present based on caregiver interpretation of the question, which queried epilepsy/seizure disorder diagnosis in a single combined survey item without option to discern between above diagnoses—a potential source of confusion. To minimize these potential confounders, we intentionally used the question that asked about current epilepsy or seizure disorder diagnosis to exclude past diagnoses. Namely, inclusion of past diagnoses may have included a disproportionate number of participants with a history of febrile seizures, a common but often self-limited disorder that young patients outgrow. Still, given these limitations in measure precision, epilepsy and seizure disorder were analyzed together as a single unit, although definitions, etiologies, and implications for management and prognosis vary greatly across the spectrum of seizure disorder/epilepsy diagnoses.

Given that the survey asked caregivers to report diagnosis of seizure disorder, the diagnosis of psychogenic nonepileptic events (PNEE) is particularly important to consider. Categorized as a type of functional neurologic disorder, PNEE resemble epileptic seizures but lack concurrent epileptiform activity. It is important to distinguish between epileptic seizures and PNEE as the underlying biologic mechanisms differ greatly, as may their respective potential associations with ACE exposure. Given that people with exposure to trauma, particularly sexual trauma, have increased risk of experiencing PNEE,^13-16 a relationship between ACEs and PNEE may be unsurprising. Therefore, to study the relationship between ACEs and epileptic seizures, as this study aims to do, PNEE must be scrutinized as a possible confounding variable.

Although PNEE cannot be separated from other epilepsy/seizure disorder diagnoses in the analyses, prevalence data may help assess the influence of PNEE on the results found here. Few studies have parsed the prevalence of PNEE from that of non-PNEE epilepsy or seizure disorder, even fewer in the pediatric population, making accurate estimate of exact prevalence difficult. The epidemiologic study is in part limited by multiple diagnostic challenges. One challenge is that initial misdiagnosis of epilepsy in people with isolated PNEE is common, which can cause diagnostic delay and confusion when the correct diagnosis is made. Another is that patients can have both active epilepsy and PNEE, and most data do not stratify to exclude comorbidity. Nevertheless, the estimated worldwide prevalence of PNEE is 3.1 per 100,000 people,^17-19 and estimated worldwide prevalence of active epilepsy is 621.5 per 100,000 people.²⁰ Comorbid epilepsy has been reported in 5–20% of patients with PNEE,^19,21 and it is estimated that PNEE are diagnosed in 20–30% of people seen at epilepsy centers for intractable seizures (presumably including those with and without comorbid epilepsy).¹⁸ Together, these statistical data suggest that, even if not accounting for comorbid duplicates, the estimated population prevalence of active epilepsy is 200-fold higher than that of PNEE. Therefore, while a proportion of the children captured in the NSCH sample as “having an epilepsy or seizure disorder” may have PNEE alone (e.g., without comorbid epilepsy), this proportion is likely a very small minority and thus minimally confounding. Nevertheless, additional confounding factors that may influence the presented findings may remain unaccounted for within the NSCH survey scope. It may be appropriate to add additional survey items in future iterations of the NSCH survey to assess their association.

Owing to its retrospective and observational design, this study is limited in its inability to investigate directionality or causality between ACE exposure and seizure/epilepsy diagnosis. The lack of detailed time course information surrounding ACE exposures, including age of ACE onset, frequency of ACE episodes, and relation to seizure disorder onset, further limits insight on directionality or causality. One potential pathophysiologic mechanism could be that increased ACEs cause overactivation of the hypothalamic-pituitary-adrenal (HPA) axis, which is activated during stress, resulting in an increased release of cortisol. Such HPA axis hyperactivity and subsequently elevated levels of glucocorticoids have been found to change the structure and function of several brain regions, including the hippocampus.²² The changes in the hippocampus can affect neurogenesis and thus increase susceptibility to neuronal damage, which has been postulated to contribute to temporal lobe epilepsy.²² On the contrary, multiple studies have shown that children with epilepsy are likely to have poor social outcomes in adulthood—less education, lower income, and underemployment compared with age-matched controls without history of childhood epilepsy.^23,24 Data on more proximal social outcomes in children with epilepsy are limited; one could posit that epilepsy diagnosis in childhood predisposes that child for subsequent ACE exposures, akin to the predisposition for poor adult social outcomes on shorter timescale. Further investigation is warranted to elucidate the pathophysiologic relationship between ACE exposure and seizure disorders and prospectively assess causality.

While we await further investigation of directionality and causality, we need not wait to ascertain the implications of the clear association between current epilepsy/seizure diagnosis and ACEs. An estimated 56.1% of our pediatric epilepsy population has at least 1 ACE exposure based on the data presented here and population estimates published by the NSCH (weighted to represent the child population of the United States).²⁵ The estimated odds of epilepsy/seizure disorder diagnosis in a child with no ACEs is 1 in 205. When a child has even 1 ACE, those odds become 1 in 115, equating to an estimated 55,130 additional children diagnosed with epilepsy. While it is unclear how many of these 55,130 children may be spared epilepsy if ACEs (food/housing insecurity, neighborhood safety, racism, and the like) were addressed or how addressing ACEs could mitigate disease outcomes, these estimates provide insight into the vast impact ACEs have on a likely majority of our patients. Neurologists often aim for seizure control as their epilepsy treatment goal; however, the work presented here suggests that social intervention, namely ACE prevention and screening, must be better prioritized if we want to prevent and treat seizure disorders adequately. With strong associations demonstrated in a large-scale data set, this study provides novel insight into the influence and need for such ACE screening and intervention referrals in epilepsy clinics.

Our study showed that multiple ACE exposures, particularly those related to household dysfunction and violence, were associated with epilepsy/seizure disorder diagnosis in a cohort of pediatric patients in the United States. An increase in cumulative ACE exposure increased odds of having a present diagnosis of epilepsy or seizure disorder. Future prospective cohort studies that examine the time course of childhood adversity and development of epilepsy will better determine causality and potential pathophysiologic mechanisms, informing and improving psychosocial and medical management of seizure disorders.

Appendix. Authors

Study Funding

C.L. Szperka is funded by the National Institutes of Health's National Institute of Neurologic Disorders and Stroke (K23NS102521).

Disclosure

C.L. Szperka and her institution have received compensation for her consulting work for Eli Lilly; Lundbeck; Teva Pharmaceutical Industries Ltd.; and Upsher-Smith Laboratories, LLC. She has received grant support from the National Institutes of Health's National Institute of Neurologic Disorders and Stroke (K23NS102521). S.L. Massey received the NIH, T32-NS061779, TA, 6/30/2015 - 6/29/2017. Dr. Massey has given expert testimony for a legal proceeding involving a case unrelated to the current manuscript. This occurred in 2018–2019. M. Anto received the 2021 Clinical Fellowship Award from the American Headache Society; the award was supported by an unrestricted educational grant from Amgen, Inc. S.C. Shipley has no disclosures. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.