Psychiatric and personality factors in pediatric functional seizures: A prospective case-control study

Abstract

Purpose:

This study assessed psychiatric and personality characteristics in relation to pediatric functional seizures (FS).

Methods:

In a 1:1 prospectively matched-control study design, children with documented FS (confirmed via video EEG; ages 13–18) were matched to controls (MCs) on income, sex, race, and age. Primary outcomes were Behavior Assessment System for Children, Second Edition (BASC-2) and Millon Adolescent Clinical Inventory (MACI). Secondary measures included questionnaires assessing trauma, somatization, body awareness and quality of life (QOL). T-tests investigated differences between groups on T-scores. Due to lack of significant outcomes, an experimental analysis was conducted assessing differences in number of clinically elevated BASC-2 and MACI scores between groups. Binary logistic regressions determined the influences of clinically elevated scores on likelihood participants have FS. T-tests assessed differences on secondary measures.

Results:

Participants included 84 children, 42 with FS and 42 MCs (Children with FS: Mean_age=15.4, Interquartile Range_age=3; 73.5% female; 59.5% white). Children with FS had greater parent-reported somatization (t(23)=5.67, p<0.001) on BASC-2, greater somatization on CSSI-24 (t(35)=6.83, p<0.001), and poorer QOL (t(41)=−6.22; p<0.001) than MCs. There were no differences in clinically elevated BASC-2 or MACI scores compared to MCs and clinically elevated scores did not influence likelihood participants have FS.

Conclusions:

Children with FS had greater somatization and poorer QOL but similar rates of psychiatric symptoms, trauma, and maladaptive personality traits compared to MCs. Psychiatric or personality factors did not predict likelihood of FS. Explanations of pediatric FS should consider novel contributors to FS rather than relying solely on a psychiatric etiology.

INTRODUCTION

Children with functional seizures (FS; also known as psychogenic non-epileptic seizures or PNES, a subtype of functional neurological disorder or FND) experience seizure-like symptoms without associated EEG abnormalities.¹ In pediatric neurological services, FND incidence is at least 6 per 100,000 children under age 16,² and it has significant effects on psychosocial (i.e., quality of life, social experiences, school issues, etc.) and financial outcomes for patients and their families.^{3, 4}

Traditionally, FS have been explained as a physical response to stressors, trauma or psychiatric disorders, and based on this explanation, some providers still recommend targeting mood or trauma as the key to improving FS.⁵ However, etiological explanations for FND now rely on biopsychosocial formulations.^6–8 Notably, there is growing evidence suggesting that children and adults may have different risk factors for FS, with mood and trauma factors being less related to pediatric FS than adult FS.^9–11 Instead, family discord and school-related issues may be more common.^{11, 12} Pediatric FS studies have demonstrated clinically elevated scores on assessments of mood and maladaptive personality traits.^{13, 14} However, since these studies did not include a control group, it is unclear if these outcomes are unique to pediatric FS.

There are few controlled studies of pediatric FS. One study demonstrated greater anxiety sensitivity; increased personal illness, surgery, and medical procedures; and greater rates of anxiety, depression, and posttraumatic stress disorder (PTSD) in children with FS as compared to their siblings.¹⁰ However, another study found similar rates of anxiety, depression and alexithymia and greater somatization in children with FS compared to children with epilepsy.¹⁵ In a nationally representative study based on the Danish National Patient Register,⁹ children with FS were found to have an increased rate of diagnosed psychiatric disorders compared to those with epilepsy and controls matched on sex and year of birth. However, 60% of children with FS had no psychiatric disorder prior to the onset of FS. Further, only approximately 11% had a diagnosed emotional disorder, including anxiety, obsessive compulsive disorder, depression and other mood disorders prior to onset of FS and only 9.9% received incident diagnoses of emotional disorders in the two years after diagnosis.⁹ Assessment of comorbid mental health concerns and referral for psychotherapy, is recommended, if needed, following a diagnosis of FS.^{16, 17} However, the overall low rates of comorbid emotional disorders observed in this study suggest that non-psychiatric factors likely play a role in pediatric FS. Importantly, as this study relied on a national patient register based on ICD codes, additional research is needed due to the variability and inconsistency of diagnostic codes used in hospital settings when evaluating both FS and mood concerns.^{18, 19}

Additional research is needed to clarify the relationship between pediatric FS and psychiatric factors to help inform etiological understanding and effective treatment development in this age group.²⁰ Our aim is to evaluate the relationship between pediatric FS and both psychiatric and personality factors using the first study utilizing healthy controls matched on sex, age, race, and household income. We hypothesized that individuals with FS would have greater family discord, elevated school and social stress, and poorer quality of life (QOL) compared to matched controls (MCs), while all other psychiatric and personality factors would be similar.

METHODS

Design Overview

Children with FS were paired prospectively with demographically similar MCs in a 1:1 matched study design. All participants completed a study visit assessing mood, behavior, emotional functioning, maladaptive personality traits, somatization, psychosocial functioning, and health-related QOL, with the Millon Adolescent Clinical Inventory (MACI) and the Behavior Assessment System for Children, Second Edition (BASC-2) as the primary outcome measures.

Participants

Participants with FS were prospectively enrolled from the University of Alabama at Birmingham’s (UAB’s) Functional Neurological Disorder Treatment Clinic beginning in November 2016 until March 2020. Patients with documented FS (confirmed via video EEG²¹) were eligible to participate after diagnosis if they were between the ages of 13–18 years, and they came to the laboratory to complete all study assessments prior to starting FS treatment at UAB. Comorbid epilepsy was acceptable if they had not had an epileptic seizure for at least 6 months. Exclusion criteria for the study included a history of active substance use, psychosis, or intellectual disability. However, no screened participants were excluded for these reasons. The study was approved by UAB’s Institutional Review Board.

Participants and a parent completed informed consent/assent. Once a participant with FS enrolled in the study, a MC was identified using a research participant database of healthy adolescents maintained by the laboratory, recruitment flyers displayed in public areas throughout the greater Birmingham area (i.e., schools, YMCAs, etc.), and/or ads posted on social media. MCs had no parent-reported mental health or medical diagnoses and were matched to children with FS based on an exact match of sex and race as well as age within ±1 year and household income within ±1 of the following intervals: 1) Below $20,000, 2) $20,000–39,999, 3) $40,000–59,999, 4) $60,000–79,999, 5) $80,000–99,999, and 6) Above $100,000. These intervals were chosen to allow for sensitivity in the assessment of household income and allow for the evaluation of income below the poverty threshold. All participants were reimbursed $40 for time spent at laboratory visits.

Measures

Demographics.

Parents reported age, sex, race, income, and medications through a demographics survey.

Psychiatric and Personality Assessment

Millon Adolescent Clinical Inventory (MACI)

Millon Adolescent Clinical Inventory (MACI) is a 160-item adolescent-specific self-report personality inventory used to identify maladaptive personality patterns and early signs of mental health disorders. Literature demonstrates an internal consistency of individual subscales ranging from alpha=0.73–0.91 with test-retest reliability 0.57–0.92.²² Eleven of the subscales were chosen a priori to be used as primary outcomes based on previous literature in adults and children with FS.^{13, 23} Scores >75 indicate clinically significant concern for maladaptive personality patterns, expressed concerns, and clinical syndromes.

Behavior Assessment System for Children, Second Edition (BASC-2)

Behavior Assessment System for Children, Second Edition (BASC-2) is a norm-referenced questionnaire with both parent- and self-report formats that assess adolescent psychiatric and emotional functioning. BASC-2 has strong construct validity and correlates well with other tools assessing adolescent behavior and self-perception.²⁴ Six self-report and three parent-report subscales were chosen a priori to be used as primary outcomes based on previous literature suggesting relationships between pediatric FS and mood, peer stress, school problems and relationship with parents.^{10, 13} T-scores for all subscales have a mean of 50 and standard deviation of 10 and were determined via comparisons with general combined population norms. Scores of ≥70 indicate clinically significant concerns.

Trauma and Neglect

Childhood Trauma Questionnaire (CTQ)

Childhood Trauma Questionnaire (CTQ) is a 28-item self-report measure of physical, sexual, and emotional abuse and physical and emotional neglect. Participants respond to scenarios described in each item using a Likert scale ranging from 1 (Never True) to 5 (Very Often True). Test-retest reliability and internal consistency and reliability across a range of samples are good.²⁵

Somatization

Children’s Somatic Symptoms Inventory-24 (CSSI-24)

Children’s Somatic Symptoms Inventory-24 (CSSI-24) is a 24-item measure of somatization severity in children. Participants indicate how much each symptom listed in the questionnaire has bothered them in the last 2 weeks using a Likert scale ranging from 0 (not at all) to 4 (a whole lot). Internal reliability and validity were found to be adequate, and test-retest reliability was high in a clinical sample (alpha=0.90).²⁶

Body Awareness

Multidimensional Assessment of Interoceptive Awareness (MAIA)

Multidimensional Assessment of Interoceptive Awareness (MAIA) is a 32-item self-report measure that assesses aspects of body awareness including body listening and trusting. Participants respond to a list of statements, indicating how each statement applies to their daily life using a 0 (Never) to 5 (Always) Likert scale. Internal consistency and construct validity are strong.²⁷

QOL

Pediatric Quality of Life Inventory (PedsQL) Generic Core Scales²⁸

Pediatric Quality of Life Inventory (PedsQL) Generic Core Scales²⁸ is a 23-item measure that assesses health-related QOL using 6 self-report subscales including Psychosocial Health, Physical Functioning, Emotional Functioning, Social Functioning, School Functioning and Total Health-Related QOL. Children indicate the frequency of problems within each domain over the last month using a Likert scale ranging from 0 (never) to 4 (almost always). This measure has been shown to have strong internal consistency and construct validity.²⁸

All questionnaire data were entered two times by independent researchers and compared for differences. Overall, statistical differences were <3% when comparing 1^st and 2^nd entry. All differences were evaluated and corrected.

Power Analysis

A power analysis for a paired samples t-test was conducted in G*Power (v3.1). Our primary outcomes are MACI and BASC-2. Effect sizes for studies using MACI to determine differences between other disorders and using MCMI (the adult version of the MACI) to determine differences between adults with FS and epilepsy ranged from 0.6 to 0.85.^{23, 29} Effect sizes ranged from 0.6 to 2.5 in studies comparing children with other disorders using the BASC-2.^{30, 31} Thus, a conservative effect size of 0.6 was chosen for the analysis. When testing at an alpha level of 0.0025 and assuming a within subject autocorrelation of 0.5, an overall total sample size of N=41, indicating at least 21 per group, was required to achieve a power of 80% for the study to be sufficiently powered to detect differences between children with FS and MCs on MACI and BASC-2.

Data Analysis

Normal distribution of variables was assessed prior to all final analyses. Paired samples t-tests evaluated differences between groups on BASC-2 and MACI T-scores. As there were 20 t-tests for the primary outcomes, a Bonferroni-Holm correction was conducted in which the lowest p-value was assessed at the 0.05/20=0.0025 level, the second lowest p-value was assessed at the 0.05/(20–1)=0.0026 level, the third lowest p-value was assessed at the 0.05/(20–2)=0.0028 level, and so on until reaching an insignificant outcome. Given the lack of significance in our primary outcomes and the lack of clinically meaningful differences for these measures, as an exploratory analysis, scores on BASC-2 self- and parent-reports and MACI were coded as clinically elevated or not based on normed cutoffs, and the number of participants with clinically elevated scores for each subscale were compared between groups with McNemar chi-square tests. Three binary logistic regressions were used to determine the influences of the total number of clinically elevated BASC-2 and MACI scores on the likelihood that participants have FS. Paired samples t-tests were also used to compare differences in secondary outcomes including trauma and neglect, somatization, body listening, and QOL. All data were analyzed using complete case analysis for each variable of interest.

RESULTS

Participants included 84 children: 42 with FS and 42 MCs matched on age, race, gender, and income (Children with FS: Mean_age=15.4, Interquartile Range_age=3; 73.5% female; 59.5% white; Table 1). Questionnaires included in the initial study were the PedsQL and CSSI-24. After the first 16 participants with FS and first 16 MCs were enrolled, the CTQ, BASC-2, and MACI were added to the study, resulting in differences in the sample sizes that were also impacted by missing data resulting from incomplete questionnaires, improper responding (i.e., selecting more than one multiple choice option), and other related factors (See Supplementary Figure 1 for the participant enrollment diagram). The sample sizes for the primary outcomes were 46 for BASC-2 and 42 for MACI (Mean_age=15.2, 70% female, 52% white). Psychotropic medication use was similar between children with FS and MCs. Of the 42 participants with FS, 9.5% had parent-reported comorbid epilepsy. Notably, most children with epilepsy outgrow the condition,³² and of the four individuals with previous epilepsy diagnoses, three were no longer on anticonvulsant medications. The participant who was still on medication had not had an epileptic seizure in over one year. Further, the individuals with a comorbid diagnosis of epilepsy did not differ from those without on demographics or outcome measures. Therefore, these participants were not analyzed separately from the other participants with FS.

Table 1.

Demographics for children with FS and MCs (controls matched on age, sex, race and household income)

Characteristics		% FS	% MC
Sex	Female	73.5	73.5
Race	White	59.5	59.5
	Black	23.8	26.2
	Other	16.7	14.3
Income	Below $20,000	21.4	11.9
	$20,000–$39,999	21.4	28.6
	$40,000–$59,999	2.4	11.9
	$60,000–$79,999	28.6	11.9
	$80,000–$99,999	9.5	14.3
	Above $100,000	16.7	21.4
		Mean ± SD	Mean ± SD
Age (yrs)		15.4 ± 2.9	16.0 ± 2.0

In order to assess significance of the outcomes using the Bonferroni-Holm correction, the p-values were arranged in order from lowest to highest. There were 7 significant outcomes, and the 8^th, with a p-value of 0.008, was not significant at the 0.05/(20–7)=0.0038 level. Therefore, testing was stopped. There were no differences between groups on any of the MACI or BASC-2 self-report subscales (Table 2). BASC-2 parent-report revealed greater somatization in children with FS compared to MCs (t(21)=5.67, p<0.001), consistent with the greater somatization reported by children with FS on the CSSI-24 (t(35)=6.83, p<0.001). Further, while BASC-2 parent-report of anxiety was not significantly different following Bonferroni-Holm correction, the relationship trended toward significance (t(22)=2.13, p=0.008) with parents of children with FS trending toward reporting greater anxiety symptoms in their children as compared with MCs. No other significant differences or trends regarding parent-reported BASC-2 scores were noted.

Table 2.

Average of self- and parent-report questionnaires and test statistics for paired samples t-tests comparing adolescents with FS and matched controls. Higher scores signify greater endorsement of characteristic. After Bonferroni correction for multiple comparisons, results at alpha level of 0.0025 were determined to be statistically significant.

Characteristic	n	FS M (SD)	n	Matched Controls M (SD)	t	p
MACI
Family Discord	21	53.8 (23.8)	21	57.2 (26.4)	−0.39	0.70
Childhood Abuse	21	29.2 (20.2)	21	29.6 (24.3)	−0.07	0.95
Peer Insecurity	21	59.1 (23.7)	21	63.5 (21.3)	−0.70	0.49
Anxious Feelings	21	84.8 (17.0)	21	76.0 (13.7)	1.68	0.11
Depressive Affect	21	66.4 (29.0)	21	77.9 (19.1)	−1.56	0.13
Suicidal Tendency	21	31.2 (25.3)	21	36.8 (25.5)	−0.86	0.40
Borderline Tendency	21	27.9 (18.5)	21	37.6 (21.4)	−1.54	0.14
Inhibited	21	59.7 (19.1)	21	63.6 (14.0)	−0.84	0.41
Submissive	21	71.2 (14.4)	21	65.7 (17.2)	0.97	0.34
Introversive	21	57.5 (20.9)	21	64.9 (22.2)	−1.23	0.23
Dramatizing	21	58.0 (20.6)	21	51.2 (17.7)	1.29	0.21
BASC-2 Self-Report
Anxiety	23	59.6 (14.4)	23	51.4 (14.4)	1.73	0.10
Depression	23	53.9 (11.6)	23	54.7 (13.3)	−0.22	0.83
Social Stress	21	50.4 (14.0)	21	52.4 (13.4)	−0.44	0.67
Attitude to School	23	44.1 (12.1)	23	51.4 (10.7)	−2.55	0.02
Somatization	23	64.1 (13.6)	23	53.0 (13.3)	2.89	0.01
Relations with Parents	23	54.4 (11.7)	23	49.1 (11.7)	1.41	0.17
BASC-2 Parent-Report
Anxiety	23	64.22 (14.37)	23	51.96 (11.66)	2.13	0.008
Depression	22	62.86 (14.10)	22	54.86 (13.87)	2.00	0.06
Somatization	22	62.55 (11.66)	22	47.14 (8.82)	5.67	<0.001
CTQ
Physical Abuse	21	9.1 (4.9)	21	9.4 (1.8)	−0.29	0.77
Sexual Abuse	21	7.8 (3.3)	21	9.0 (0.0)	−1.74	0.10
Emotional Abuse	21	7.5 (5.0)	21	7.7 (3.6)	−0.16	0.87
Physical Neglect	21	8.0 (3.7)	21	9.8 (1.7)	−2.42	0.03
Emotional Neglect	21	7.4 (3.7)	21	10.0 (1.9)	−2.98	0.01
Total	21	39.8 (18.7)	21	45.9 (5.9)	−1.60	0.13
CSSI-24 Self-Report
Teen Total	36	38.5 (17.4)	36	12.7 (11.3)	6.83	<0.001
MAIA
Noticing	39	3.1 (1.3)	39	2.7 (1.3)	1.16	0.25
Not-Worrying	39	2.6 (1.0)	39	2.6 (0.9)	−0.31	0.76
Body Listening	38	1.8 (1.5)	38	1.9 (1.4)	−0.25	0.80
Trusting	38	2.7 (1.7)	38	3.5 (1.3)	−2.52	0.02
PedsQL Self-Report
Psychosocial Health	42	55.3 (15.7)	42	76.0 (18.1)	−5.53	<0.001
Physical Functioning	42	58.8 (27.4)	42	87.0 (12.5)	−5.58	<0.001
Emotional Functioning	42	51.7 (26.9)	42	71.1 (21.5)	−3.80	<0.001
Social Functioning	42	70.1 (21.6)	42	82.5 (22.4)	−2.44	0.02
School Functioning	40	43.7 (19.7)	40	75.5 (19.7)	−7.04	<0.001
Total Score	42	56.6 (17.2)	42	79.8 (14.6)	−6.22	<0.001

Table 3 reports the number and percentage of children with FS and MCs whose subscale scores on BASC-2 and MACI reached clinical significance. However, none of the comparisons between MCs and children with FS reached significance, even at the 0.05 alpha level. Binary logistic regression models found no significant influences of clinically elevated scores on BASC-2 self- (p=0.23) and parent-report (p=0.12) or MACI (p=0.97) on the likelihood of having FS.

Table 3.

Number of children with FS and MCs reporting clinically elevated concerns on the BASC-2 and MACI and p-values of McNemar chi-square tests. There were no significant differences between groups.

	FS #(%)	MC #(%)	p
Frequency of maladaptive personality patterns (MACI ≥ 75)	n=21	n=21
Borderline Tendency	1 (4.8%)	2 (9.5%)	1.00
Inhibited	6 (28.6%)	4 (19.0%)	0.69
Submissive	9 (42.9%)	6 (28.6%)	0.51
Introversive	4 (19.0%)	6 (28.6%)	0.73
Dramatizing	5 (23.8%)	3 (14.3%)	0.69
Frequency of expressed concerns (MACI ≥ 75)	n=21	n=21
Family Discord	6 (28.6%)	8 (38.1%)	0.75
Childhood Abuse	1 (4.8%)	2 (9.5%)	1.00
Peer Insecurity	8 (38.1%)	10 (47.6%)	0.69
Frequency of clinical syndromes (MACI ≥ 75)	n=21	n=21
Anxious Feelings	16 (76.2%)	10 (47.6%)	0.15
Depressive Affect	11 (52.4%)	12 (57.1%)	1.00
Suicidal Tendency	3 (14.3%)	2 (9.5%)	1.00
# of children reporting at least one clinically elevated MACI score	20 (95.2%)	20 (95.2%)
Mean (SD) # of clinically elevated concerns reported on MACI	3.3 (1.68)	3.1 (1.78)
	FS #(%)	MC #(%)
Frequency of clinically elevated self-reported behavior concerns (BASC-2 SR T-score ≥ 70)	n=23	n=23
Anxiety	6 (26.l%)	2 (8.7%)	0.29
Depression	1 (4.3%)	3 (13.0%)	0.63
Social Stress (n=22)	2 (9.1%)	3 (13.6%)	1.00
Attitude to School	0 (0%)	1 (4.3%)	1.00
Somatization	7 (30.4%)	2 (8.7%)	0.13
Relations with Parents	0 (0%)	0 (0%)	1.00
# children reporting at least one clinically elevated BASC-2 score	11 (47.8%)	7 (30.4%)
Mean (SD) # of clinically elevated concerns reported on BASC-2	0.70 (0.88)	0.48 (0.79)
	FS #(%)	MC #(%)
Frequency of clinically elevated parent behavior concerns (BASC-2 PRS T-score ≥ 70)	n=23	n=23
Anxiety	7 (30.4%)	1 (4.3%)	0.07
Depression	7 (30.4%)	4 (17.4%)	0.51
Somatization	5 (21.7%)	0 (0%)	0.22
# of parents reporting at least one clinically elevated BASC-2 score	10 (43.5%)	5 (21.7%)
Mean (SD) # of clinically elevated concerns reported on BASC-2	0.83 (1.11)	0.22 (0.42)

Children with FS had poorer QOL than MCs including lower psychosocial health (t(41)=−5.53, p<0.001), physical functioning (t(41)=−5.58, p<0.001), emotional functioning (t(41)=−3.80, p<0.001), school functioning (t(39)=−7.04, p<0.001) and total health (t(41)=−6.22, p<0.001; Table 2). There were no differences in physical, emotional, or sexual abuse or physical or emotional neglect on the CTQ and no differences on the MAIA between children with FS and MCs.

DISCUSSION

This is the first study using healthy controls prospectively matched on age, race, gender, and income to assess psychiatric and personality factors related to pediatric FS. Previous research has suggested incidence of psychiatric diagnoses is higher in children with FS than healthy controls,⁹ and diagnoses of anxiety, depression and posttraumatic stress disorder were found do be greater in children with FS compared to their siblings.¹⁰ While it is important to interpret the non-significant findings with caution, this study adds to growing evidence that although psychiatric comorbidities are present in children with FS at higher rates than the general population,⁹ psychiatric symptoms, personality factors, and trauma may not be the main factor associated with pediatric FS.^{10, 15} Specifically, non-significant findings have been reported in past studies comparing 1) rates of physical and sexual abuse in children with FS compared to siblings¹⁰ and 2) rates of psychiatric symptoms between children with FS and children with epilepsy.¹⁶ The present study adds to this body of research by identifying only parent-reported somatization as significantly different between children with FS and MCs on the BASC-2, MACI and CTQ. Beyond parent-reported somatization, no significant differences in clinically elevated psychiatric symptoms and personality outcomes were reported. Further, in the present study, the existence of any clinically elevated psychiatric symptoms or maladaptive personality factors did not influence the likelihood of participants having FS, and there was consistency observed in the self- and parent-report of outcomes for these questionnaires.

Previous research has reported clinically elevated levels of maladaptive personality traits in pediatric FS as measured using the MACI (scores >75).¹³ However, when comparing these levels to MCs, none of the maladaptive personality patterns, expressed concerns, or clinical syndromes assessed by MACI were significantly higher for children with FS. This is a contrast between pediatric and adult FS, as adults with FS have been found to have greater anxiety, dysthymia, and borderline personality traits on the adult version of the MACI, the Millon Clinical Multiaxial Inventory, as compared to adults with epilepsy.²⁹

There were also no BASC-2 subscale differences in self- or parent-reported anxiety or depression compared to MCs. However, the difference in BASC-2 parent-reported anxiety indicated a trend of elevated anxiety symptoms in children with FS compared to MCs, and the Cohen’s d effect size was large, suggesting parents of children with FS may report greater anxiety in their children than MCs. This could be due to parents’ interpretation of FS symptoms as anxiety. Since this sample consisted of children who had been previously diagnosed with FS, parents could be biased to interpret FS symptoms as anxiety if a healthcare provider linked FS to anxiety when the diagnosis was explained to the family. Alternatively, given that children with FS report significantly lower emotional functioning on the PedsQL - a 5-question assessment of general feelings of sadness, anxiety, anger, trouble sleeping and worry - parents may be responding to their child’s emotional distress related to having functional symptoms they are unable to control. Finally, it could also indicate a slightly higher level of anxiety in children with FS. However, only 7 children with FS had parent-reported levels of anxiety above the cutoff for clinical significance on the BASC-2, and this was not significantly different between children with FS and MCs. Further, self-reported anxiety across BASC-2 and MACI outcomes was not significantly different for children with FS as compared with MCs.

These outcomes are contrary to findings based on the Danish National Patient Register, which found increased rates of emotional disorders, including anxiety and depression, compared to MCs.⁹ Notably, although not significantly greater compared to MCs, the present study reports greater endorsement of clinically elevated anxiety and depression among children with FS on self- and parent-reports of the BASC-2 (>26%) compared to the rates of emotional disorders found in the nationally representative Danish sample prior to their FS diagnosis (~11%).⁹ This inconsistency could be due to a difference in the measures of psychiatric illness used, as data from the Danish National Patient Register relied on retrospective history of FS and psychiatric diagnoses acquired via a register-based cohort study⁹ while this study prospectively enrolled children with FS and MCs and used validated emotional, behavioral and personality measures. The retrospective use of diagnostic history identified via patient registers as an indicator of psychiatric illness is limited by treatment seeking behavior, access to healthcare professionals, and the low reliability of accurately coding of mood-related diagnoses using ICD codes in certain healthcare settings.¹⁹ However, in the 2 years after FS diagnosis, 9.9% of the Danish sample received a new diagnosis of an emotional disorder, which may indicate an overall incidence more similar to our outcomes. As individuals with somatic symptoms often have increased interactions with the health care system^33–35 and diagnosis of pediatric FS takes, on average, about 3.5 years from the first FS episode,³⁶ it is possible that individuals with FS in the nationally representative Danish sample had increased opportunity for psychiatric assessment and diagnosis prior to and following their FS diagnosis as compared with MCs. Additionally, the Danish study matched controls only on age and sex,⁹ while the current study matched on age, sex, race, and household income. Given that race and household income are significantly associated with mental health in children and related to access to health care services,³⁷ this could also contribute to these differences, although this is unclear given the differences in methods. However, given that both this study and the Danish study found rates of anxiety and depression to be much lower than 50%, this emphasizes that many children with FS do not have anxiety or depression.

Importantly, the similarity between groups on the MACI and BASC-2 observed in this study is because both groups had similar rates of the concerns and not because the concerns were not endorsed by children with FS and their parents. While data from the MACI and BASC-2 indicated statistically similar outcomes for children with FS and MCs, clinically significant scores were prevalent in the sample, with 95% of children with FS reporting at least one clinically elevated score on the MACI and 48% reporting at least one clinically significant score on the BASC-2. This is inconsistent with other studies which found low endorsement of clinically elevated symptoms among children with FS using self- and parent-reported measures.^{15, 38} This also contradicts suggestions that children with FS have poor insight into their psychological functioning and that the use of rating scales to assess psychiatric symptoms provide limited sensitivity for children with FS.^{15, 16, 38} However, further research is needed to assess the validity of self- and parent report using rating scales as compared with clinical interview for comorbid psychiatric diagnoses for children with FS. Additionally, it is important to highlight that even though psychiatric comorbidities may not be directly related to FS, children with FS have higher rates of psychiatric diagnoses and stressors than the general population.⁹ Therefore, it is important that children with FS are comprehensively assessed for psychiatric comorbidities and appropriate treatments are offered as needed.

Parent-report of the BASC-2 and self-report of the CSSI-24 indicated significantly greater somatization in FS than MCs. This is similar to research comparing children with FS to those with epilepsy¹⁵ and is consistent with research demonstrating increased somatization in adults with FND.³⁹ Somatization in pediatric FS could be the result of increased attention to physical symptoms, impaired inhibition to physical sensations, somatosensory amplification (catastrophic symptom expectations) or impairments in sensory motor gating,^{6, 40} and further research is needed to further understand the causal relationship between somatization and pediatric FS. This finding may serve as a novel target for prevention and treatment.

Several studies have found school-related difficulties, bullying and family discord to be common in pediatric FS.^{3, 10, 12} Conversely, our results show no statistical difference between children with FS and MCs in the tendency to feel alienated, hostile, or dissatisfied toward school based on the BASC-2 attitude to school subscale, and, in fact, there was a medium effect size for MCs trending towards reporting higher rates of these factors as compared with children with FS .There was also no difference between groups in feeling lonely, isolated, or “picked on” in social situations (social stress subscale), and less than 10% of children with FS reported clinically elevated scores for social stress and attitude to school. Further, children with FS and MCs had similar perceptions of their value within the family, their relationship with their parents, and parental trust and concern (relations with parents subscale). This suggests increased rates of these factors are not uniquely related to FS.

Total QOL and the physical, emotional, psychosocial, and school QOL subscales were impaired compared to MCs. This emphasizes the substantial effect of FS across all aspects of children’s lives. As the average length of time to diagnose FS in children ranges from weeks to up to 3.5 years³⁶ and there are few providers comfortable providing treatment for FS,⁴¹ this may represent a significant secondary treatment target. Consistent with a study which compared children with FS to an unmatched control group,⁴² there is no difference in social functioning between children with FS and MCs. Social functioning may serve as a strength and coping mechanism in pediatric FS and may be beneficial to utilize in FS treatment. Of importance is that the effects of FS on QOL may be different in children than they are in adults – a series of adult studies that compared patients with FS and epilepsy indicated that the QOL is driven by the presence or absence of mood problems and not by the diagnosis⁴ which is opposite to the findings of the present study.

Finally, children with FS and MCs reported similar rates of physical, sexual and emotional abuse in our study. While a trend toward higher rates of physical (medium effect size) and emotional (large effect size) neglect was observed in MCs as compared with children with FS, there were no statistically significant differences between groups with regard to these variables. This is consistent with outcomes of previous research comparing children with FS and their siblings¹⁰ and suggests that these factors may not be uniquely related to pediatric FS.

This study provides additional evidence that pediatric FS should not be oversimplified as solely related to psychiatric symptoms, maladaptive personality traits, and trauma.^{9, 15} Notably, somatization was the only psychiatric, personality or trauma-related outcome that significantly differed between children with FS and MCs in this study. This suggests that not all children with FS may need treatment for comorbid psychiatric or psychosocial issues. Instead, some children may benefit from treatment targeting novel predisposing factors for FS other than mood (i.e., sense of control, symptom expectations, impairments in habituation, selective attention, cognitive inhibition).^43–47 There could be many benefits of such treatment, including expanded access to treatment,⁴⁸ faster recovery, and limited effects of FS on school, work, and social activities for families. However, additional research is needed to evaluate the efficacy of targeting these factors in treating pediatric FS.⁴³

Our results suggest that somatization and factors other than psychiatric disease, personality traits, and trauma contribute to pediatric FS. Therefore, when discussing the diagnosis of FS with children and their parents, we recommend that a biopsychosocial model of FS be discussed^{6, 7} and discourage overreliance on a purely psychiatric explanation. Explaining FS as the direct result of stress or mood to families whose children do not have significant psychiatric symptoms may make them less likely to accept the diagnosis, possibly leading to delayed treatment and worse prognosis.⁴⁹ Alternatively, it may lead them to pursue treatment for mood exclusively, which could result in an iatrogenic mental health disorder due to a search for a psychiatric etiology for their symptoms rather than resulting in effective FS treatment. Children often report that strong emotions, such as anxiety, can be a trigger for FS, which can be confusing to families when distinguishing FS from mood. However, this occurrence may be explained as the result of classical conditioning in which the physical symptoms of anxiety caused by having a FS are paired with FS symptoms, and then the physical symptoms of anxiety due to another situation (e.g. school stress) trigger the FS symptoms.⁶ In this case, it may be beneficial for treatments for pediatric FS to work to extinguish emotional triggers as a conditioned response.

The results from this study also underscore a need for several new research directions for pediatric FS. Although previous research has often focused on psychiatric comorbidities of FS, outcomes from this and other studies support the notion that, while some children with FS may have comorbid psychiatric symptoms, this may not be this may not be the main factor related to FS in children. Therefore, future studies should further investigate other potential causal mechanisms for pediatric FS and explore novel predisposing factors, such as abnormal attentional focus, catastrophic symptom expectations, impaired habituation, network differences or proteomic, epigenetic, or genetic differences.^{43, 50, 51,44} This will help inform an experimental therapeutics approach to the development of effective treatments for FS in which the mechanism by which FND symptoms improve with treatment can be confirmed. Additionally, targeting somatization could result in the prevention or effective treatment of FS. Finally, research should continue to address improving QOL for children with FS.

Limitations of our study include a cross-sectional design, absence of a clinical interview to further assess psychiatric symptoms, no control for symptom similarity (e.g. epilepsy), use of measures not previously validated in children with FS and some variable sample size among different measures. Overall, the study population was also primarily comprised of individuals identifying as white and female. However, this is typical for FND patient samples.⁵² Finally, because psychiatric symptoms are often elevated in relation to ongoing medical problems⁵³ and we are comparing a group with acute medical distress to healthy individuals matched on age, sex race and household income, the elevation in somatic symptoms should be interpreted with caution, as it is unclear from this study if this elevation is specific to FS or due to having an acute medical condition. However, the finding is strengthened by past research which demonstrates elevated somatization in children with FS as compared with children with epilepsy.¹⁵

Conclusions

This study demonstrates that psychiatric symptoms and maladaptive personality traits are not specific to pediatric FS. These findings highlight a need for a change in the etiological explanation provided to children with FS and their families and provides an opportunity for future research evaluating other causal mechanisms in the development of FS.