Association of Psychopharmacological Medication Preference with Autistic Traits and Emotion Regulation in ADHD

Abstract

Background

This study intends to evaluate the relationship between medication switching and autistic traits, emotion dysregulation, and methylphenidate side effects in children with attention deficit hyperactivity disorder (ADHD).

Methods

Children with ADHD, ages 9–18, treated with methylphenidate (MTP) (n = 23), and switched to atomoxetine (ATX) (n = 20) were included. All participants were interviewed with K-SADS-PL to confirm ADHD diagnosis and exclude comorbid psychiatric disorders. The participants then completed Difficulty in Emotion Regulation Scale (DERS) and Autism-Spectrum Quotient (AQ) and their parents completed Autism Spectrum Screening Questionnaire (ASSQ) and Barkley Stimulant Side Effect Rating Scale(BSSERS).

Results

The MTP group scored higher than the ATX group in ASSQ, AQ, and the lack of emotional clarity subscale of DERS, while the ATX group had higher scores in the emotional non-acceptance subscale of DERS. No differences were found between the MTP and ATX groups in methylphenidate side-effect severity. Multiple regression analyses revealed that non-acceptance of emotions predicted the switch to ATX while lack of emotional clarity predicted the maintenance of MTP therapy, rather than autistic traits.

Conclusions

This study highlights emotion regulation difficulties and how different emotional profiles may influence medication selection in children with ADHD.

Introduction

Attention-deficit hyperactivity disorder (ADHD) is characterized by developmentally inappropriate levels of inattention, hyperactivity, and impulsivity, which result in disturbances in various cognitive and behavioural domains. A common neurodevelopmental disorder, ADHD affects nearly 5% of children worldwide with global prevalence estimates ranging from 2–7%^1,2 and is frequently comorbid with other psychiatric disorders.³ As such, prompt evaluation, detection of comorbidities and affected domains, and meticulous follow-up are crucial in reducing the burden this condition creates for individuals and their families.⁴

An important affected domain in ADHD is emotion dysregulation. Emotion dysregulation is described as the application of maladaptive strategies to regulate one’s emotions to achieve a set goal,⁵ and is thought to be the result of the integration of multiple different processes such as attention, cognition, and social behavior.⁶ Emotion dysregulation is reported to affect 25–45% of children and adolescents diagnosed with ADHD and is regarded to be a significant clinical characteristic due to its detrimental effects.^7,8 Emotion dysregulation is also frequently associated with autism and autistic symptoms.^9,10

As a neurodevelopmental disorder, ADHD is frequently complicated by autism spectrum disorder (ASD).¹¹ For children with ADHD, autistic traits are overrepresented compared to typically developing children and correlate with ADHD symptom severity.^12,13 One other important aspect of autism comorbidity in ADHD is the response to psychostimulants. Lower efficacy and an increase in psychostimulant side effects or adverse events are reported in the literature for the subgroup of patients with comorbid ADHD and ASD.^14,15

Atomoxetine is generally initiated in patients who cannot tolerate psychostimulants, as psychostimulants are widely accepted as an effective treatment for ADHD.¹⁶ Cited reasons for medication switching to atomoxetine include lack of efficacy of first-line treatments such as methylphenidate and amphetamines, adverse events, and targeting comorbid psychiatric disorders.^17,18 Atomoxetine is especially regarded as a safe and effective treatment of ADHD symptoms in the presence of ASD.^19,20 However, few studies investigate the association between atomoxetine use in children with ADHD and autistic traits.

This study aims to investigate the association of switching from methylphenidate to atomoxetine with other psychiatric comorbidities, autistic traits, and emotion dysregulation while also quantifying methylphenidate side effect severity in children with ADHD. To investigate this association, our hypotheses are threefold: (1) atomoxetine use is associated with higher autistic traits, (2) heightened emotion regulation difficulties, and (3) amplified methylphenidate side effects in children diagnosed with ADHD in a clinical setting.

Methods

This study was conducted at Ege University Faculty of Medicine, Department of Child and Adolescent Psychiatry, Adolescents and ADHD Outpatient Clinics. Ege University Faculty of Medicine Ethical Committee approved the study regarding ethical principles (Decision number: 20-10.1T/36). Children and their parents were verbally informed about the study’s design. Verbal assent was sought from the children involved in the study, and appropriate written informed consent forms were obtained from the children and their parents.

The inclusion criteria for the present study were being in the 9–18 age group and a clinical diagnosis of ADHD, and the exclusion criteria were a clinical psychiatric diagnosis other than ADHD, or psychotropic mediation use other than a stimulant or non-stimulant (such as an antipsychotic or antidepressant) for symptom control or augmentation. Seventy-two children diagnosed with ADHD and no other psychiatric comorbidities between the ages of 9 and 18 who were followed up in the Adolescents and ADHD Outpatient Clinic who were only on either a stimulant or non-stimulant medication were informed about the study, and the 55 children and their parents who agreed to participate were included in this study. Four participants were unable to complete the interviews or questionnaires and were excluded. The remaining 51 children were assessed with the Schedule for Affective Disorders and Schizophrenia for School-Age Children—Present and Lifetime Version (K-SADS-PL) and DSM-5 criteria to confirm Attention Deficit Hyperactivity Disorder diagnosis, and children with any other accompanying comorbidities were excluded (n = 8). A total of 43 children diagnosed with ADHD alone, who were using atomoxetine and had a history of side effects or ineffectiveness against methylphenidate formed the Atomoxetine Group (ATX Group, n = 20), And children diagnosed with ADHD using methylphenidate formed the Methylphenidate Group (MTP Group, n = 23), (Figure 1).

Figure 1.

Study Flow Diagram

Inclusion criteria for the ATX Group were: aged between 9–18 and had a diagnosis of ADHD, no history or diagnosis of other psychiatric disorders recorded in the semi-structured K-SADS-PL interview, having no history of neurological disease or neurosurgical intervention, having an average intellectual functioning, having a history of change from methylphenidate to atomoxetine. The inclusion criteria for the MTP Group were: aged between 9–18 and diagnosed with ADHD, having no history of neurological disease or neurosurgical intervention, no history or diagnosis of other psychiatric disorders recorded in the in the semi-structured K-SADS-PL interview, and having a clinically normal intellectual capacity. Exclusion criteria for the case groups were as follows: having a psychiatric disorder other than ADHD including other neurodevelopmental disorders such as autism spectrum disorder, or other psychiatric comorbidities (anxiety, psychotic and affective disorders including major depressive and bipolar disorders) and having a history of neurological disease or neurosurgical intervention, or intellectual disability.

Participants were assessed with the Autism Spectrum Screening Questionnaire and Autism Spectrum Quotient-Adolescent Version were filled out by parents to assess autistic traits. The side effect profile after methylphenidate use was evaluated with the Barkley Stimulant Side Effect Rating Scale filled by the parents. Difficulties in Emotion Regulation Scale was filled out by the participants to evaluate emotion regulation.

Schedule for Affective Disorders and Schizophrenia for School-Aged Children – Present and Lifetime Version (K-SADS-PL): It is a semi-structured interview form developed to determine past and present psychopathologies of children and adolescents according to DSM-V diagnostic criteria.²¹ This comprehensive semi-structured interview involves separate interviews with both the child and their caregiver. The interview tool is structured into two parts, aligned with the collection of demographic and developmental information. These sections consist of a screening interview and a diagnostic supplement designed for psychiatric conditions, including autism spectrum disorder. The data gathered from both the child and the caregiver is coded independently, with predetermined thresholds for a positive initial assessment. If a positive diagnostic indication is identified, the relevant diagnostic supplement is administered to confirm the diagnosis. The adaptation of K-SADS-PL to Turkish norms was conducted and the interview was found to be valid and reliable for the Turkish population.²²

Difficulty in Emotion Regulation Scale (DERS): Evaluates emotion regulation on a 5-point Likert scale. DERS is valid and reliable for adolescents.²³ The Dysregulation in Emotion Regulation Scale encompasses several subscales that evaluate various aspects of emotion regulation difficulties, including nonacceptance of emotional responses, difficulty engaging in goal-directed behavior, impulse control difficulties, lack of emotional awareness, limited access to emotion regulation strategies, and lack of emotional clarity. Turkish psychometric evaluations of the scale for adolescents demonstrated this instrument to be reliable and valid.²⁴

Autism Spectrum Screening Questionnaire (ASSQ): It is a screening test completed by the primary caregivers, used to detect autism spectrum disorder symptoms with an emphasis on behavioral and communicative disturbances to screen the disorder at an early stage.²⁵ The Turkish version was found to be valid and reliable.²⁶

Autism-Spectrum Quotient (AQ): The Autism Quotient is a self-administered questionnaire designed to assess the presence of autism-related traits in adults and adolescents. The test consists of five different domains and totals up to 50 questions. The five areas are poor social skills, poor communication skills, poor imagination, extraordinary attention to detail, and inability to shift attention/strong focus of attention.²⁷ The reliability and psychometric properties of the Turkish form of the questionnaire were conducted.²⁸

Barkley Stimulant Side Effect Rating Scale (BSSERS): Side effects related to stimulant treatment are questioned in terms of frequency and severity. On the scale, 17 symptoms are questioned, and the questions are answered by the parents on a 10-point Likert scale (0 = never, 9 = almost always). Scores above seven indicate that the side effect is ‘severe’.²⁹

Analyses were made using the IBM – SPSS (The Statistical Package for Social Sciences) 25.0 package program. Whether the quantitative variables were suitable for normal distribution was examined with the Kolmogorov-Smirnov test. Variables that did not show normal distribution were expressed as the median (25–75p). Regarding non-normally distributed variables, the groups were compared with the Mann-Whitney U Test. The relationship between the qualitative variables was examined with the chi-square test. Exploratory correlation analyses were conducted with the Spearman correlation test. Binary logistic regression analysis was used to identify potential predictors of switching to atomoxetine. Model fit was assessed with Hosmer and Lemeshow test. The p < 0.05 value was taken as a reference for the findings to be considered statistically significant.

Results

The comparison of the case groups in terms of sociodemographic data characteristics is given in Table 1. No statistically significant difference was found between the two groups except maternal and paternal age; the groups were similar in terms of age and gender. (Table 1).

Table 1. Comparison of ATX Group and MTP Group in Terms of Sociodemographic Data Characteristics.

		ATX Group (N = 20)		MTP Group (N = 23)		Test
Gender
Female		2 (10)		4 (17.4)		χ2 = 0.487 df = 1	p = 0.669
Male		18 (90)		19 (82.6)
Age		14 (12–16)		15 (14, 25–17)		Z = –1.345	p = 0.179
Mother’s age		40 (38–44)		45 (39–48, 75)		Z = –2.199	p = 0.028*
Father’s age		45 (42–48)		49.5 (43, 25–55)		Z = –2.517	p = 0.012*
Maternal psychiatric disorder		5 (25)		1 (4.3)		χ2 = 3.800 df = 1	p = 0.081
Paternal psychiatric disorder		0 (0)		2 (8.7)		χ2 = 1.824 df = 1	p = 0.491
Siblings psychiatric disorder		5 (25)		5 (21.7)		χ2 = 0.064 df = 1	p = 0.999

^aBonferroni corrected p value is presented. χ2 = Chi-square test, Z = Mann-Whitney U test. *: p < 0.05.

The MTP group had higher ASSQ (p = .008) and AQ total scores (p = .032) compared to the ATX group. Further subscale testing revealed that the MTP group also had significantly higher scores in AQ attention switching (p = .018), DERS lack of emotional awareness (p = .020), DERS lack of emotional clarity (p = .002) subscales, while the ATX group had scored higher in DERS non-acceptance of emotions (p = .008). After a Bonferroni correction was applied for multiple testing, only the DERS lack of emotional clarity and non-acceptance of emotions subscales remained significant favoring the MTP and ATX groups respectively. The methylphenidate side effect score assessed by BSSERS, did not differ among the MTP or ATX groups. The group differences in ASSQ, AQ, DERS and BSSERS total and subscale scores are summarized in Table 2.

Table 2. Differences between ASSQ, AQ, DERS, BSSERS Scales and Subscales between ATX and MTP Groups.

ATX Group (N = 20)

MTP Group (N = 23)

Test statistic (Z)

P value

Mean (SD)

Mdn (Min-Max)

Mean (SD)

Mdn (Min-Max)

ASSQ Total Score

5.55 (6.09)

3.5 (0–25)

9.91 (5.69)

9 (1–19)

–2.636

.008*

AQ Total Score

14.05 (7.77)

12 (3–31)

18.35 (6.43)

19 (2–29)

–2.146

.032*

Social

2.25 (1.65)

2 (0–5)

3.39 (3.93)

2 (0–18)

–.594

.553

Attention switching

3.55 (1.88)

3 (2–8)

4.65 (1.80)

5 (1–8)

–2.360

.090a

Attention to details

3.90 (2.85)

3.5 (0–10)

5.26 (2.09)

5 (2–9)

–1.914

.056

Communication

2.3 (1.78)

2.0 (0–6)

3.13 (1.86)

3 (0–7)

–1.399

.162

Imagination

2.05 (1.67)

2 (0–6)

3.30 (1.64)

3 (0–6)

–2.447

.014

DERS Total Score

83.8 (21.85)

83 (54–117)

81.17 (18.88)

80 (46–130)

–.365

.715

Non-acceptance

14.15 (11.27)

11.5 (6–60)

9.52 (5.0)

8 (6–25)

–2.633

.048*

Goals

14.75 (5.34)

14 (7–23)

12.96 (4.42)

13 (5–21)

–.952

.341

Impusivity

15.45 (6.58)

14.5 (6–26)

13.0 (4.37)

13 (6–23)

–1.199

.230

Awareness

14.2 (2.82)

14.5 (8–18)

18.13 (5.30)

18 (10–26)

–2.322

.120a

Strategy

15.5 (4.99)

14 (8–24)

15.65 (6.77)

13 (8–34)

–.440

.660

Clarity

9.70 (2.41)

9 (6–16)

11.91 (2.25)

12 (7–17)

–3.062

.012*

BSSERS

26 (11.53)

24 (8–55)

30.39 (19.07)

24 (0–72)

–.670

.503

Bonferoni correction was applied to subscale scores, corrected p values are presented. ASSQ: Autism Spectrum Screening Quostionnaire; AQ: Autism Spectrum Quotient—Adolescent Version; DERS: Difficulties in Emotion Regulation Scale; BSSERS: Barkley Stimulant Side Effect Rating Scale; Z: Mann-Whitney U Test; *: p < 0.05.

The variables of interest were thus identified as ASSQ and AQ total scores and DERS non-acceptance and clarity subscales. Exploratory correlation analyses were conducted to assess for multicollinearity. While ASSQ and AQ showed a moderate correlation in the ATX group, no other variable of interest were found to be correlated in either the ATX or the MTP groups. The correlation coefficients for the scales and subscale scores in ATX and MTP groups are summarized in Tables 3 and 4, respectively.

Table 3. Exploratory Correlation Analysis of ASSQ, AQ, DERS and BSSERS Scales and Subscales in the ATX Group (n = 20).

Variable		M		SD		1		2		3		4		5		6		7		8		9		10		11		12		13		14
1. ASSQ		5.55		6.09		–
2. AQ Total		14.05		7.77		.51*		–
3. AQ Social		2.25		1.65		.51*		.84***		–
4. AQ Attention Switch		3.55		1.88		.22		.87***		.67***		–
5. AQ Attention Detail		3.90		2.85		.32		.51**		.15		.40		–
6. AQ Communication		2.3		1.78		.53*		.76***		.68***		.67**		.14		–
7. AQ Imagination		2.05		1.67		.37		.85***		.80***		.77***		.25		.68**		–
8. DERS Total		83.8		21.85		.09		–.14		.20		–.21		–.42		.17		–.13		–
9. DERS Non-acceptance		14.15		11.27		–.10		–.25		–.05		–.31		–.19		–.06		–.31		.81***		–
10. DERS Goals		14.75		5.34		.13		–.02		.37		–.16		–.50*		.21		–.01		.85***		.63**		–
11. DERS Impulsivity		15.45		6.58		.12		–.11		.33		–.23		–.56*		.11		–.04		.82***		.51*		.92***		–
12. DERS Awareness		14.2		2.82		.22		.50*		.45*		.42		.16		.59**		.54*		.28		.11		.13		.25		–
13. DERS Strategy		15.5		4.99		.13		.04		.42		–.03		–.37		.17		.05		.75***		.56*		.86***		.86***		.19		–
14. DERS Clarity		9.70		2.41		.24		.27		.35		.25		–.24		.50*		.20		.53*		.26		.48*		.48*		.54*		.58**		–
15. BSSERS		26		11.53		–.06		–.31		–.04		–.27		–.13		–.35		–.30		.33		.45		.25		.32		–.15		.32		–.22

*p < .05; **p < .01; ***p < .001 ASSQ : Autism Spectrum Screening Quostionnaire; AQ : Autism Spectrum Quotient—Adolescent Version; DERS : Difficulties in Emotion Regulation Scale; BSSERS : Barkley Stimulant Side Effect Rating Scale.

Table 4. Exploratory Correlation Analysis of ASSQ, AQ, DERS and BSSERS Scales and Subscales in the MTP Group (n = 23).

Variable		M		SD		1		2		3		4		5		6		7		8		9		10		11		12		13		14
1. ASSQ		9.91		5.69		–
2. AQ Total		18.35		6.43		.34		–
3. AQ Social		3.39		3.93		.19		.51*		–
4. AQ Attention Switch		4.65		1.80		.19		.61**		.24		–
5. AQ Attention Detail		5.26		2.09		–.09		.26		–.11		–.15		–
6. AQ Communication		3.13		1.86		.18		.73***		.51*		.56**		–.01		–
7. AQ Imagination		3.30		1.64		.10		.29		.37		–.10		–.03		.13		–
8. DERS Total		81.17		18.88		.35		–.09		–.30		.13		.22		–.34		–.47*		–
9. DERS Non-acceptance		9.52		5.0		.10		.29		.18		.21		.43*		.15		–.27		.59**		–
10. DERS Goals		12.96		4.42		.46*		–.15		–.25		.15		–.08		–.32		–.39		.78***		.25		–
11. DERS Impulsivity		13.0		4.37		.51*		.16		–.03		.20		.14		.15		–.29		.63**		.51*		.39		–
12. DERS Awareness		18.13		5.30		–.44*		–.72***		–.48*		–.29		–.06		–.72***		–.27		.28		–.14		.07		–.09		–
13. DERS Strategy		15.65		6.77		.46*		.11		–.16		.29		.19		–.09		–.38		.88***		.63**		.65**		.69***		.07		–
14. DERS Clarity		11.91		2.25		.06		.20		–.27		.06		.46*		–.06		–.15		.47*		.41		.41		.22		–.20		.36		–
15. BSSERS		30.39		19.07		.27		.53*		.46*		.30		.19		.48*		.12		.08		.41		.07		.47*		–.53*		.26		.14

*p < .05; **p < .01; ***p < .001. ASSQ: Autism Spectrum Screening Quostionnaire; AQ: Autism Spectrum Quotient—Adolescent Version; DERS: Difficulties in Emotion Regulation Scale; BSSERS: Barkley Stimulant Side Effect Rating Scale.

A binary logistic regression analysis was conducted with ASSQ and AQ total scores and DERS non-acceptance and clarity subscales as predictors and ATX vs. MTP groups as the dependent variables. ASSQ and AQ were both included in the analyses as their correlation in the ATX group was moderate and they measured different domains of autistic traits: ASSQ focuses on observable autistic behavior and is a screening questionnaire completed by the caregivers, while AQ measures subjective autistic traits to identify where the individual falls on the autistic spectrum and is a self-report questionnaire. The binary logistic regression analyses revealed that one point increase in non-acceptance of emotions significantly predicted switching to atomoxetine (OR: 1.346 [1.075 − 1.687], p = .010) while a point increase in lack of emotional clarity predicted continuing methylphenidate treatment (OR: .539 [.332 − .857], p = .012). Autistic traits measured by both ASSQ and AQ were not found to be significant in medication switching (all p > .05). (Table 5).

Table 5. Multiple Logistic Regression Analysis for Prediction Of Switching to ATX (n = 43, χ²: 21.177, df: 4, p < .001, Hosmer and Lemeshow test p = .141).

		B		Standard Error		Wald		DF		P		Odds Ratio		95% CI
														LL		UL
ASSQ Total		–.125		.083		2.290		1		.130		.882		.750		1.038
AQ Total Scores		.018		.066		.075		1		.784		1.018		.895		1.159
DERS Non-acceptance		.294		.115		6.685		1		010		1.346		1.075		1.687
DERS Emotional Clarity		–.616		.247		6.255		1		.012		.539		.332		.875

BSSERS: Barkley Stimulant Side Effect Rating Scale; AQ: Autism-Spectrum Quotient; DERS: Difficulty in Emotion Regulation Scale.

Discussion

The present study compares children with ADHD who are treated with methylphenidate (MTP) and children who had switched from methylphenidate to atomoxetine (ATX) in terms of autistic traits, emotional dysregulation, and methylphenidate side effects. No differences in age or gender were observed between the MTP and ATX groups. The groups were also similar in terms of familial psychiatric diagnoses. There are several studies in the literature that report maternal psychiatric diagnoses to be associated with increased symptom severity³⁰ and comorbidities in children with ADHD,³¹ although we did not observe this relationship in our study.

The MTP group had scored higher in autistic traits as measured by ASSQ and AQ, while reporting similar scores in methylphenidate side effects, contrary to our hypotheses. This interesting result may indicate that autism symptoms or traits may not play a major role in reported methylphenidate side effects and subsequent medication switching to atomoxetine. No differences between children with ADHD and ADHD with ASD symptoms in terms of side effects or medication efficacy were found in a Swedish cohort study, although they had not distinguished between different medications.³² Of course, the possibility of recall and proxy biases should also be considered as BSSERS was completed not during but after methylphenidate trials by the participants’ caregivers. There are conflicting reports as the presence of ASD is associated with lower medication efficacy and higher adverse events in autistic individuals,³³ but the same association was not found for individuals with ADHD and high-functioning autism or autistic traits.^32,34 More studies are needed to better establish the association between psychostimulant side effects and autistic traits.

The two groups were found to be similar in terms of emotion dysregulation total scores. However, subscale analyses revealed that the ATX group scored significantly higher than the MTP group in the emotional non-acceptance subscale of DERS, and the MTP group scored higher in lack of emotional clarity than the ATX group. Considering AQ and ASSQ’s effects on switching were nullified in the regression analysis, this could imply disturbances in different subdomains in emotion regulation in two distinct subgroups of ADHD, one of which does not benefit from methylphenidate as much as the other, independent of autistic traits. Indeed, one study identified two distinct clusters in an adult sample with ADHD. Both clusters demonstrated emotion dysregulation; however, one group had more severe disturbances in emotion dysregulation, which also coincided with increased comorbidities.³⁵ Also, methylphenidate is reported to improve emotion dysregulation,³⁶ which may have also confounded our results.

In the binary logistic analysis, emotional non-acceptance was found to predict switching to atomoxetine, while lack of emotional clarity predicted staying on methylphenidate. This variance was attributed to a subset of ADHD patients encountering challenges in acknowledging emotions, possibly leading to heightened distress and disruptive behaviors. Consequently, the clinician opted to switch to atomoxetine rather than due to the influence of methylphenidate side effects, as the patient and their caregivers expressed higher dissatisfaction with the medication. In contrast, the decision to persist with methylphenidate treatment was predicted by a deficiency in emotional clarity, likely through a similar mechanism. The absence of emotional clarity might cause personal confusion, though it is seldom linked with behavioral disruptions. Consequently, the group on methylphenidate treatment could have perceived a greater effectiveness in managing ADHD symptoms, given their reduced instances of behavioral disturbances due to their inherent emotional regulation characteristics.

This finding strengthens the possibility that methylphenidate side effect profiles in ADHD are affected by different processes. Emotional non-acceptance represents the inability to acknowledge emotions, which results in distress and behavioral disturbances. As such, this group of children with ADHD may have had perceived less behavioral benefit from methylphenidate due to comprising a non-emotion accepting intrinsic phenotype. Another explanation may be methylphenidate’s positive effect on emotion dysregulation and emotional lability.³⁶ Adolescents with ADHD reported lower scores in the impulsivity subscale and total scores of DERS after 6-month treatment with methylphenidate.³⁷ Methylphenidate was also reported to reduce emotion dysregulation in both participants with ADHD and ADHD comorbid ASD without adverse effects.³⁸ As such, the group who switched to atomoxetine may have lost methylphenidate’s positive effect on emotion dysregulation hence scoring higher on emotional non-acceptance.

There are several limitations to our study. Although the inclusion of a semi-structured interview by seasoned child and adolescent psychiatrists is a significant strength, the modest sample size resulting from the inclusion and exclusion criteria is a major limitation, as such, the results lack generalizability. Another limitation is the cross-sectional naturalistic design which introduces bias both from the observers and the participants. Studies with prospective designs, including larger cohorts, should be devised to effectively study the relationship between medication side effects/switching, autistic traits and emotion regulation.

Conclusions

Switching to atomoxetine in children with ADHD was found to be associated non-acceptance of emotions while a lack of emotional clarity predicted staying on methylphenidate in our study. However, atomoxetine use was not associated with autistic traits, or methylphenidate side effect severity, contrary to our hypotheses. This was thought to be due to methylphenidate being percieved ineffective in a subgroup of ADHD patients struggles with accepting emotions, which may result in increased distress and behavioral disturbances which resulted in switching to atomoxetine rather than methylphenidate side effects driving the medication switch in this group. Lack of emotional clarity, on the other hand, predicted staying on methylphenidate treatment, probably through a similar mechanism. A lack of emotional clarity may be confusing for the individual, however, it is seldom associated with behavioral disturbances. As such, the MTP group may have had percieved greater efficacy of methylphenidate in controlling ADHD symtpoms, since they experienced fewer behavioral disturbances due to their intrinsic emotion regulation profiles. In conclusion, ATX and MTP groups had differences in emotion regulation subdomains, raising the possibility that different mechanisms in emotion regulation could interplay with perceived methylphenidate efficacy, which could be further investigated in future studies.

Contrary to our hypothesis, methylphenidate side effect severity did not differ between the groups and autistic traits did not significantly predict medication switching. Both these findings highlight the importance of addressing emotion regulation difficulties in children with ADHD as they may influence medication selection and treatment efficacy.