The Effect of Drug Holidays on Metacognitive and Theory of Mind Skills of Children Diagnosed with Attention Deficit Hyperactivity Disorder: Preliminary Findings

Abstract

Purpose

This study aimed to evaluate the impact of summer drug holidays on metacognitive and theory of mind (ToM) skills in children with Attention-Deficit/Hyperactivity Disorder (ADHD) treated with methylphenidate (MPH).

Methods

A naturalistic, prospective study was conducted at a child psychiatry outpatient clinic. Fifty-six treatment-naïve children or those restarting treatment after >6 months (aged 9–17, 76.8% male) with ADHD were enrolled. Participants were assessed at baseline, after two months of MPH treatment, and after a two-month summer drug holiday. Assessments included the Clinical Global Impressions-Severity (CGI-S) scale, the DSM-IV-Based Screening Scale for Disruptive Behavior Disorders (D-S-DBD), the Faux-Pas Test (FPT), the Test of Perception of Affect via Nonverbal Cues (TPANC), and the Metacognition Questionnaire for Children (MCQ-C). Attrition was high; only 12 children (21.4% of original sample) completed the full protocol. Analyses were conducted for both completers and the full sample using the Last Observation Carried Forward (LOCF) method.

Results

MPH treatment significantly improved CGI-S, parent-rated hyperactive/impulsive and oppositional symptoms, and performance on FPT and affect perception tasks (all p<0.05). In the completer sample, the gains in ToM and affect perception were not significantly reversed after the two-month drug holiday. However, this finding is preliminary due to substantial attrition and limited statistical power. MCQ-C did not show significant change at any time point. Sensitivity analyses revealed that children lost to follow-up had higher baseline symptom and metacognitive problem scores, indicating potential selection bias.

Conclusion

Methylphenidate treatment was associated with significant improvements in ADHD symptoms, ToM, and affect perception. The results highlight the need for larger, multi-center studies to reliably assess the impact of drug holidays on higher-order cognitive and social skills in ADHD.

Introduction

Attention-Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder presenting with persistent, developmentally inappropriate, pervasive and impairing symptoms of inattention, hyperactivity, and impulsivity.¹ The global prevalence of ADHD is reported to be 8.0% with a male-to-female ratio of 2.0.² Symptoms of ADHD are best addressed with a personalized combination of pharmacological and psychosocial interventions although pharmacological interventions, especially stimulants, remain the first line of treatment.³

“Meta-cognition” refers to the awareness and regulation of one’s own cognitive processes. According to the literature, it comprises two primary components, namely metacognitive monitoring (ie, the ability to assess subjective cognitive performance) and metacognitive control (ie, the ability to regulate cognitive strategies based on the assessment).⁴ These processes are deemed to be crucial for effective learning, problem solving and decision making.^4,5 According to the literature, metacognitive abilities begin to emerge in early childhood and continue to develop through adolescence and early adulthood.^4,5 ADHD can also affect metacognitive skills in addition to the core symptoms of inattention, hyperactivity and impulsivity. Available studies suggest that children with ADHD may struggle with self-monitoring, planning and strategy regulation which may be related with problems in executive functioning (EF).^6,7 Although stimulants are effective for behavioral manifestations of ADHD, their effects on metacognitive skills are less clear with some studies suggesting improvement while others do not.^8,9 The consensus is that these children may require targeted behavioral interventions in addition to stimulant treatment to remedy problems in their metacognitive skills.⁷

Theory of Mind (ToM) refers to the awareness that others mental processes and content may differ from one’s own along with attributing mental states involving beliefs, desires, intentions and emotions to one’s own self and others.^10,11 It forms the basis of social interactions, communication and empathy.¹¹ ToM develops progressively from infancy through adolescence building upon joint attention, imitation and social referencing.^10,11 ToM skills in children with ADHD are also affected compared to their neurotypical peers, although not to the degree of autism spectrum disorder (ASD).^12,13 The effects are moderate, vary across tasks used and may be related with EF skills and language demands.¹³ The effects of stimulant treatment on ToM skills in ADHD are also controversial with some studies reporting improvement while others do not.^14,15

“Drug holidays” denote deliberately interrupting ADHD pharmacotherapy for a definite period and a specific clinical purpose and available guidelines suggest that treatments over a year may be accompanied by drug holidays on a personalized basis.¹⁶ These interruptions may aim to evaluate the necessity of continuing treatments, to manage adverse effects, to control the possibility of tolerance and to increase children’s and their parents’ awareness of the advantages of treatment.^16,17 Despite their widespread use in routine practice, knowledge concerning their beneficial or harmful effects on ADHD symptoms, adverse effects and associated problems is limited and mainly concerns drug holidays on weekends.¹⁷ There are only three previous studies on the effects of drug holidays among Turkish children with ADHD. According to one of these studies, drug holidays may not have an impact on growth rates in children with ADHD.¹⁸ The remaining studies evaluated the neuropsychological and emotion regulation-related effects of drug holidays over the summer and found that resistance to interference, emotion recognition and emotion regulation may be adversely impaired by drug holidays.^19,20

Methylphenidate (MPH) used in management of ADHD symptoms increases dopamine and norepinephrine within the prefrontal cortex (PFC) and striatum which are implicated in both metacognitive monitoring and social cognitive processes such as ToM.²¹ Given that MPH discontinuation during a drug holiday withdraws this neurochemical support, we hypothesize that metacognitive accuracy and ToM performance would decline following a drug holiday, consistent with evidence that most MPH-facilitated gains in social-emotional cognition—including emotion regulation and social cognitive measures—return to baseline upon medication interruption.^19,20 Due to controversial results reported and the preliminary evidence that certain neurocognitive gains may partially persist during drug-free periods, we aimed to evaluate the effects of drug holidays on metacognitive and ToM skills.^22,23

To the best of our knowledge, no study to date has evaluated the effects of drug holidays on metacognitive and ToM skills among children with ADHD using a naturalistic, prospective design. The purpose of this study was to assess the impact of drug holidays on metacognitive skills and ADHD symptoms in children receiving methylphenidate treatment for ADHD during their summer break. Change in clinician and parent-rated symptoms were evaluated in exploratory analyses.

Materials and Methods

Study Center, Design, Sampling and Ethics

This clinic-based, naturalistic, prospective study was conducted at the outpatient clinics of the Department of Child and Adolescent Psychiatry, Bolu Abant Izzet Baysal University in between November 2023–December 2024. The sample of the study consisted of children aged 9–17 years old, diagnosed with ADHD according to the DSM-5 criteria, being treatment-naïve (or restarting treatment after a period of ≥ six months) and providing written assent along with parental informed consent. Exclusion criteria included those younger than 9 or older than 17 years of age, continuing treatment for ADHD, presence of comorbid ASD, psychotic/mood/specific learning disorders, having chronic neurological/medical disorders requiring treatment and intellectual disability. ASD, psychotic/mood/specific learning disorders and intellectual disability were excluded due to their unique effects on metacognition and ToM.^13,21,24 Comorbid oppositional defiant (ODD), conduct (CD) and anxiety disorders were allowed due to high levels of comorbidity.^1,2 The hospital directorate (Date: 07.11.2023, No: E-33443051-903.99–228750100) and the Abant Izzet Baysal University Medical Faculty IRB (Date: 05.12.2023, No: 2023/403) approved of the study protocol and all the study procedures were in accordance with the Declaration of Helsinki and local laws and regulations. The potential, eligible, included and excluded patients, their reasons and numbers were illustrated in Figure 1.²⁵

Figure 1.

Sampling flowchart for the study to evaluate the effects of drug holidays on metacognitive and theory of mind skills of children with ADHD.

Abbreviations: tx., treatment; ASD, autism spectrum disorder; disord., disorder; f/u, follow-up; MPH, methylphenidate.

Study Procedure

The primary author evaluated the children with ADHD with the Kiddie Schedule for Affective Disorders and Schizophrenia for School Age Children (K-SADS-PL), Test of Perception of Affect via Nonverbal Cues (TPANC), Faux-Pas-Test (FPT) at baseline, after two months of treatment with methylphenidate (MPH) and after two months of drug holiday. The children completed the Metacognition Questionnaire for Children (MCQ-C), while the parents completed the DSM-IV-Based Screening Scale for Disruptive Behavior Disorders (D-S-DBD) at the same time points. In the summer holiday, the primary author contacted the parents with telephone to inquire whether they practiced drug holiday and, if they did, its duration. Children who interrupted their treatment for two months during the summer holiday were invited to the department along with their parents and completed TPANC, FPT and MCQ-C while their parents completed D-S-DBD. The clinicians primarily treating the children who are blind to study hypotheses completed the CGI-S at all time points. The tests were completed at a fixed order. Methylphenidate (MPH) formulations were converted to immediate-release MPH dose equivalents.^26,27

Study Measures

The Sociodemographic Data Form

This form was developed by the researchers and collected information on children’s age, sex, grade, past medical/psychiatric history, family structure, parents’ ages, education levels and vocations. Parental physical and mental disorders requiring treatment were also recorded.

The Kiddie Schedule for Affective Disorders and Schizophrenia for School Age Children (K-SADS-PL)

The K-SADS-PL was designed to evaluate current and lifetime psychopathology among 6–18 years old children according to the Diagnostic and Statistical Manual of Mental Disorders (DSM)-III and IV criteria and was later revised to reflect changes in DSM-5.^28,29 The Turkish version was previously found to be valid and reliable.³⁰ Parents and children in this study participated in semi-structured interviews using the K-SADS-PL, and present and lifetime psychopathologies were determined by consensus from both sources. The authors were trained and certified in K-SADS-PL application from the Turkish Association of Child and Adolescent Mental Health.

The Clinical Global Impressions Scale (CGI)

This clinician completed scale evaluates the symptom severity (CGI-S), the extent of benefit from treatments and the severity of side effects that may arise after the treatments.³¹

The Test of Perception of Affect via Nonverbal Cues (TPANC)

The TPANC is a 56-item tool to evaluate the perception of six basic emotions (ie, happiness, sadness, anger, surprise, fear and disgust) presented via auditory, visual (static pictures and dynamic videos) and textual stimuli. It includes 24 pictures of facial affect, 8 videos demonstrating affective motions/postures, 12 speech recordings consisting of prosodically conveyed affects and 12 stories involving social/affective situations. The test lasts approximately 35 minutes to complete, and the participants responded verbally with the primary researcher recording them as true/false. The scale was developed to evaluate affect perception in Turkish children and was previously used to evaluate the development of affect perception in children with SLD.^32,33 Permission to use the test was granted by the original researchers. Guttman’s λ-3 was found to be 0.92 for the current sample.

The Faux-Pas Test (FPT)

The Faux-pas test was developed by Baron-Cohen et al³⁴ to evaluate recognition of social blunders by children. The FPT is thought to tap into advanced levels of mentalizing.³⁴ In the blunder tasks, one of the characters in a story expresses a socially inappropriate/inconvenient statement in the context of the conversation. The story was followed by a set of queries for the participant to identify this as socially inappropriate. The Turkish version of the FPT was found to be valid and reliable and was used in clinical studies.³⁵ Guttman’s λ-3 was found to be 0.95 for the current sample.

The Metacognition Questionnaire for Children (MCQ-C)

MCQ-C was developed by Bacow et al³⁶ and was later revised for intelligibility for children younger than 13 years old. MCQ-C consists of 24, four point Likert-type items grouped under four factors (ie, “positive” and “negative meta-worry”, “cognitive monitoring” and “beliefs of superstition, punishment and responsibility”, BICS) Total scores vary between 24–96 with elevated scores reflecting greater negative metacognitive activity. The Turkish version was found to be valid and reliable for 8–17 years old children previously.³⁷ Cronbach alpha for the MCQ-C in the current sample was 0.91.

DSM-IV-Based Screening Scale for Disruptive Behavior Disorders (D-S-DBD)

The D-S-DBD was developed by Turgay³⁸ to screen symptoms of disruptive behavior disorders (ie, ADHD, ODD and CD) according to the DSM-IV diagnostic criteria³⁸ The D-S-DBD yields symptom scores for inattention (IA), hyperactivity/impulsivity (HIP), ODD and CD, while the items scored “much” or “very much” are summed to provide symptom counts. The Turkish version was found to be valid and reliable.³⁹ D-S-DBD is widely used in child and adolescent psychiatry outpatient departments in Turkey and it was previously found to be valid also for DSM-5 criteria, leading to its use in the current study.⁴⁰ Cronbach's alpha for the current sample was 0.94.

Statistical Analysis

Statistical analyses were conducted using SPSS version 23.0 (IBM Corporation, Armonk, NY, USA) for Windows ^TM and JASP (JASP Team (2023). JASP (Version 0.16.4). [Computer software] (Retrieved from https://jasp-stats.org) as well as JAMOVI (The jamovi Project (2023). Jamovi (version 2.3) [Computer software] (Retrieved from https://www.jamovi.org). Thirty-two patients (57.1% of the original sample) attended the second and 12 (21.4% of the original sample) attended the last visit. Normality at those visits was evaluated with the Shapiro–Wilk test (with pairwise deletion of missing variables). Apart from MCQ-C at all visits, Total D-S-DBD score at first and FPT/TPANC scores at last visit, none of the variables were distributed normally among completers. Mardia’s test revealed multivariate non-normality.⁴¹

Sensitivity analyses revealed that patients lost at second and third visits did not differ significantly in terms of sex (χ2 = 2.3, p = 0.128, r = 0.20), while those lost at second visit had significantly higher CGI-S and (Z = −2.1, p = 0.034, effect size [ES] = 0.3), MCQ-C negative meta-worry (Z = −3.3, p = 0.001, E.S. = 0.4), cognitive monitoring (Z = −2.4, p = 0.016, E.S. = 0.3) and BICS (Z = −2.4, p = 0.018, E.S. = 0.3) scores while they had lower FPT (Z = −2.1, p = 0.037, E.S. = 0.3) scores compared to those attending the second visit. All effect sizes were small to moderate. Parent rated IA, HIP, ODD and CD symptoms did not differ significantly between children lost to follow-up and children attending the second visit (all Mann–Whitney U-test). Children attending third visit and those lost in between second and third visits did not differ significantly in any of the variables (Mann–Whitney U-test, p > 0.05). Therefore, last observation carried forward (LOCF) method was used, and the analyses are reported for both completers and with LOCF (Figures 2 and 3).

Figure 2.

Parent rated symptoms of inattention (IA), hyperactivity/impulsivity (HIP), oppositional defiant disorder (ODD) and conduct disorder (CD) among children attending the second visit and those lost to follow-up.

Abbreviation: CI, Confidence interval,

Figure 3.

Parent rated symptoms of inattention (IA), hyperactivity/impulsivity (HIP), oppositional defiant disorder (ODD) and conduct disorder (CD) among children attending the third visit and those lost to follow-up.

Abbreviation: CI, Confidence interval.

Nominal variables are summarized as counts and frequencies, while quantitative variables are summarized as means and standard deviations (SD) or medians and inter-quartile ranges (IQR) depending on normality and outliers. Univariate (Z ≥ ± 3.3) and multivariate (Mahalanobis Distance, p < 0.001) outliers formed 1.8 (n = 1) and 5.4% (n = 3) of the sample, respectively.⁴¹ Among the LOCF sample, Jr-MAI-A total scores at first and third visits (p = 0.430 and 0.100, respectively) and baseline D-S-DBD (p = 0.117) were distributed normally. Therefore, we used Friedman non-parametric variance analysis for repeated measures to evaluate measures prior to treatment, after treatment and after drug holiday and employed Durbin-Conover test for post-hoc pairwise comparisons. Holm-Bonferroni sequential correction was used to control family-wise error (Gaetano J. Holm-Bonferroni Sequential Correction: An EXCEL Calculator. Doi: 10.13140/RG.2.1.4466.9927. https://www.researchgate.net/publication/322569220_HolmBonferroni_sequential_correction_An_Excel_calculator_13). P was set at 0.05 (two-tailed), and effect sizes are also reported along with statistical significance.

Results

Fifty-six children (n = 43, 76.8% male) with a mean age of 11.0 (SD = 2.6) years were enrolled in the study. Thirty-two (n = 28, 87.5% male) and twelve (n = 11, 91.7% male) children with mean ages of 11.3 (SD = 2.6) and 10.3 (SD = 2.1) years attended the second and third visits, respectively. According to the K-SADS-PL most (n = 42, 75.0%) children were diagnosed with ADHD-Combined type while the remainder were diagnosed with inattentive type. Half of the children (n = 28, 50.0%) received treatment for ADHD in the past. Most common comorbidities were ODD (n = 5, 8.9%), separation anxiety (n = 3, 5.4%), specific phobia (n = 2, 3.6%) and CD (n = 1, 1.8%), respectively. Sociodemographic features of children and their parents are presented in Table 1.

Table 1.

Sociodemographic Features of Children with ADHD Participating in the Study to Evaluate the Effects of Drug Holidays on Metacognitive and Theory of Mind Skills

		N	%
Maternal education	Primary school or lower	15	26.8
	Secondary school	8	14.3
	High school	26	46.4
	College or higher	7	12.5
Maternal vocation	Housewife	29	51.8
	Worker	15	26.8
	Civil servant	5	8.9
	Artisan	4	7.1
	Retired	3	5.4
Maternal chronic physical disorder†		8	14.3
Maternal psychopathology†		1	1.8
Paternal education	Primary school or lower	17	30.4
	Secondary school	8	14.3
	High school	18	32.1
	College or higher	13	23.2
Paternal vocation	Jobless	1	1.8
	Worker	34	60.7
	Civil servant	7	12.5
	Artisan	11	19.6
	Retired	3	5.4
Paternal chronic physical disorder†		6	10.7
Paternal psychopathology†		2	3.6
Family structure	Intact/nuclear	41	73.2
	Extended/traditional	5	8.9
	Separate/Divorced/widower	10	17.9
Family SES††	Low	12	21.4
	Middle	38	67.9
	High	6	10.7

Note: ^†self-reported, requiring treatment, ^††according to minimum worker wage/month for 2024 [https://www.verginet.net/dtt/1/asgari-ucret.aspx, accessed on 01.02.2024]; lower: 1–2 X minimum wage, middle: 3–5 X minimum wage, upper: ≥ 6 X minimum wage.

Abbreviations: ADHD, Attention deficit hyperactivity disorder; SES, socioeconomic status.

Most children were living in nuclear families with middle-to-low socioeconomic status. Their fathers were mostly blue-collar workers, while mothers were mostly housewives. One of the mothers and two of the fathers reported receiving treatment for major depressive disorder. Median dose of MPH at treatment initiation and at the second visit were, 10.0 (IQR = 6.3) and 20.0 (IQR = 2.7) mg/day respectively.

Clinician and parent rated symptom levels of children at baseline, after MPH treatment and after drug holidays are presented in Table 2.

Table 2.

Clinician and Parent Rated Symptom Levels of Children with ADHD at Baseline, After Two Months of MPH Treatment and After Drug Holidays

Median (IQR)	Baseline		At 2nd Month of tx		After Drug Holiday		P*		E.S.
	C. (n = 56)	LOCF (n = 56)	C. (n = 32)	LOCF (n = 56)	C. (n = 12)	LOCF (n = 56)	C.	LOCF	C.	LOCF
CGI-S	4.0 (1.0)	-	3.0 (1.0)	4.0 (1.0)	3.0 (0.0)	4.0 (1.0)	<0.001	<0.001	0.7	0.4
D-S-DBD-IA	14.0 (11.0)	-	12.5 (8.8)	13.0 (8.8)	11.0 (7.3)	12.5 (8.8)	0.502	0.218	-	-
D-S-DBD-HIP	11.0 (11.8)	-	9.0 (12.5)	9.0 (11.8)	8.0 (6.3)	8.5 (9.8)	0.127	0.014	-	0.1
D-S-DBD-ODD	6.5 (13.8)	-	6.0 (11.0)	6.5 (12.5)	4.5 (8.8)	6.5 (12.8)	0.115	0.009	-	0.1
D-S-DBD-CD	0.5 (4.8)	-	0.0 (2.0)	0.0 (4.0)	0.0 (1.8)	0.0 (3.8)	0.163	0.025	-	0.1

Note: *Friedman test. Italics indicate statistically significant p-values (p < 0.05).

Abbreviations: IQR, inter-quartile range; C, completer; LOCF, last observation carried forward; E.S., effect size (Kendall’s W); CGI-S, Clinical Global Impression-Severity; D-S-DBD, DSM-IV-Based Screening Scale for Disruptive Behavior Disorders; IA, inattention; HIP, hyperactivity/impulsivity; ODD, oppositional defiant disorder; CD, conduct disorder.

Among completers, the median CGI-S scores differed significantly across visits. Pairwise comparisons revealed that scores at baseline and after treatment for two months, baseline and after drug holidays and after treatment for two months and after drug holidays differed significantly from each other (p≤ 0.01 for all), reducing at each visit. In the LOCF sample CGI-S, HIP, ODD and CD scores differed significantly across visits. Pairwise comparisons revealed that the median CGI-S scores were reduced significantly between baseline and after two months of treatment and between baseline and after drug holidays (p < 0.001 for both) while scores at the second month of treatment and after drug holidays did not differ significantly (p = 0.210). Parent rated HIP, ODD and CD symptoms were reduced similarly between baseline and after two months of treatment (p = 0.014, 0.006 and 0.014) and between baseline and after drug holidays (p = 0.008, 0.008 and 0.022) but scores at the second month of treatment did not differ from those after drug holidays (p > 0.05). Apart from CGI-S (moderate to large effect size) all effect sizes were small.

Metacognitive and ToM skills of children at baseline, after MPH treatment and after drug holidays are presented in Table 3.

Table 3.

Metacognitive and Theory of Mind Skills of Children with ADHD at Baseline, After Two Months of MPH Treatment and After Drug Holidays

Median (IQR)	Baseline		At 2nd Month of tx		After Drug Holiday		P*		E.S.
	C. (n= 56)	LOCF (n = 56)	C. (n = 32)	LOCF (n = 56)	C. (n = 12)	LOCF (n = 56)	C.	LOCF	C.	LOCF
FPT	1.0 (4.8)	–	2.0 (7.8)	2.0 (7.8)	6.0 (4.5)	3.5 (8.0)	0.010	<0.001	0.4	0.2
MCQ-C-Pmw	12.0 (7.0)	-	13.0 (6.5)	12.0 (5.8)	10.0 (8.8)	11.5 (6.0)	0.404	0.542	-	-
MCQ-C-Nmw	13.0 (7.3)	-	15.0 (7.5)	13.0 (5.0)	17.0 (5.8)	13.0 (6.8)	0.211	0.368	-	-
MCQ-C-Cm	15.0 (6.0)	-	16.5 (5.8)	16.0 (6.8)	16.5 (7.8)	15.5 (7.0)	0.909	0.927	-	-
MCQ-C-BICS	12.0 (7.8)	-	15.5 (5.8)	12.0 (7.0)	15.5 (6.8)	14.0 (7.0)	0.754	0.733	-	-
TPANC-F	18.0 (5.0)	-	19.0 (5.0)	19.0 (5.0)	19.0 (3.0)	19.0 (5.0)	0.741	0.001	-	0.1
TPANC-Mo.	6.0 (2.8)	-	6.0 (2.0)	6.0 (3.0)	6.5 (2.0)	6.0 (3.0)	0.018	0.009	0.3	0.1
TPANC-Au.	6.0 (4.0)	-	8.0 (3.3)	7.0 (5.0)	7.0 (2.5)	7.0 (4.8)	0.042	<0.001	0.3	0.2
TPANC-S	8.0 (3.0)	-	8.0 (2.0)	8.0 (3.0)	9.5 (3.8)	8.0 (3.8)	0.115	0.260	-	-

Note: *Friedman test. Italics indicate statistically significant p-values (p < 0.05).

Abbreviations: IQR, inter-quartile range; C, completer; LOCF, last observation carried forward; E.S., effect size (Kendall’s W); FPT, The Faux-Pas Test; MCQ-C, The Metacognition Questionnaire for Children; Pmw, positive meta-worry; Nmw, negative meta-worry; Cm, cognitive monitoring; BICS, beliefs of superstition; punishment and responsibility; TPANC, The Test of Perception of Affect via Nonverbal Cues; F, Faces; Mo, movement; Au, auditory; S, story.

Among completers, the median FPT scores increased significantly between baseline and at two months of treatment (p = 0.036) and between baseline and after drug holiday (p = 0.001), while scores at the second month of treatment did not differ significantly from scores after drug holiday (p > 0.05). TPANC movement (p = 0.049) and auditory (p = 0.065) scores tended to increase between baseline and at the second month of treatment while they increased significantly between baseline and after drug holiday (p = 0.003 and 0.012, respectively). For both, scores in the second month of treatment did not differ significantly from scores after drug holiday. For all these results effect sizes were moderate. Among the LOCF sample, the median scores at the FPT test increased significantly between baseline and at the second month of treatment (p < 0.001) and between baseline and after drug holiday (p < 0.001) while scores at the second month of treatment were similar to end scores. For TPANC-Faces median number of correct responses increased significantly between baseline and at the second month of treatment (p = 0.001) and baseline and after drug holiday (p = 0.001) while scores at the second month were similar to end scores. For both TPANC-movement and auditory tests, scores at baseline increased significantly between baseline and at the second month of treatment (p = 0.022 and <0.001, respectively) and between baseline and after drug holiday (p = 0.003 and <0.001, respectively) while scores at the second month of treatment were similar to end scores. Effect sizes for these results were small.

According to Holm-Bonferroni sequential corrections all analyses for completers and for the LOCF sample remained significant (p’ < 0.05).

Discussion

This single-center, naturalistic, clinic-based, prospective study evaluated the impact of drug holidays on metacognitive skills and ADHD symptoms in children receiving methylphenidate treatment for ADHD during their summer break and the results should be deemed preliminary due to elevated levels and the selective nature of attrition. Change in clinician and parent-rated symptoms were evaluated in exploratory analyses.

The primary result of our study is that MPH treatment over two months was associated with significant improvements in clinician-rated symptom severity, parent-rated hyperactive/impulsive and oppositional symptoms along ToM (ie, FPT) and affect perception (ie, TPANC) tasks Our results are in accordance with previous studies reporting improvement in ToM and emotion recognition tasks in children with ADHD being treated with MPH.^12,14 Additionally, in the small subset of children (n = 12) who completed the drug holiday protocol, there was no significant reversal of the gains in ToM and affect perception skills after two months without medication. However, this finding should be deemed preliminary, hypothesis generating and be interpreted with caution due to the elevated rate of attrition (ie, 78.6% of the original sample), potential for selection bias and limited statistical power. Supporting the hypothesis on selection bias, sensitivity analyses revealed that clinicians evaluated children lost at second visit as having more severe symptoms and the children had significantly elevated metacognitive problems in self-report scales at baseline. Therefore, children who dropped out of our study early had elevated baseline symptom severity and poorer metacognitive profiles and the completer sample may represent a subgroup with a more favorable clinical trajectory, limiting the generalizability of our results. According to the post-hoc power analysis with twelve completers, the power to detect a medium-sized effect (d = 0.5) at a p level of 0.05 was approximately 15.0–20.0%.⁴² Therefore, our completer sample lacked power to detect all but very large effects rendering the non-significant findings for metacognitive measures uninterpretable and limiting the confidence in the positive findings. Although we employed LOCF to overcome this limitation, this method provides a best-case estimate, which likely minimized the potential worsening of symptoms after the drug holiday. Indeed, past studies suggest that LOCF may consistently underestimate within group changes, while completer analyses may overestimate within group changes.^43,44 Alternative approaches such as mixed model for repeated measures (MMRM) and multiple imputation may have affected our results, although they require specific features for missing data (ie, randomness). Instead, we used both completer and LOCF analyses with sensitivity analysis to place our results in context. Regardless, our results on effects of drug holidays should be replicated with further studies from multiple centers and employing larger samples.

Our results should be evaluated within the context of their limitations. Firstly, the results are valid only for the 9–17 years old children diagnosed with ADHD-combined or inattentive types at the study center within the time-frame who are treatment-naïve (or starting treatment after a period of ≥six months) and may not be valid for children with ADHD-hyperactive-impulsive type, older and/or younger children, those being followed at different centers, those under current stimulant treatment or children initiating treatment with non-stimulant medications. Second, we aimed to increase within group homogeneity at the cost of patient heterogeneity, and the results may not be valid for children with comorbid ASD, psychotic/mood/specific learning disorders, having chronic neurological/medical disorders requiring treatment and intellectual disability. Also, rates of comorbid ODD, CD and anxiety disorders were very low in our sample precluding subgroup analyses. Further studies should evaluate the effects of disruptive behavior disorders comorbid with ADHD on ToM and affect perception tasks and whether they moderate the beneficial effects of MPH treatment and/or interact with drug holiday. Third, the clinicians primarily following the children determined the MPH doses according to risk-benefit calculations along with parents’ and children’s wishes and this led to daily doses at the lower limits of suggested ranges (ie, 60.0–72.0 mg/day).^45,46 Therefore, the selected dose range may have affected our results. Fourth, we depended on self-reports for metacognitive, affect perception and ToM skills and parent reports for symptoms and this may be a source of recall and reporting bias as well as shared method variance. Further studies evaluating the effects of MPH treatment on metacognitive, affect perception and ToM skills may employ multiple informants (ie, parents, teachers, children, peers) and more varied tasks with greater ecological validity (ie, dynamic video clips of social interactions, interactive digital avatars, academic/organizational tasks, virtual/augmented reality environments).⁴⁷ Fifth, our results are valid for the two month’s duration after treatment initiation and after drug holiday and studies with longer follow-up durations may be needed to evaluate long term effects.²³ Also, many families practice shorter drug holidays (ie, over weekends or for a few weeks) and the effects of drug holidays over shorter term may differ from our results. Sixth, although the clinicians using CGI-S were blinded to hypotheses, the children’s and their parents’ responses may have been affected by expectancy bias. Furthermore, clinician evaluations may reflect increased familiarity with children and parents, regression to the mean, maturation or the lack of academic demands within the summer. Lastly, our results may have been affected by attrition bias and Type I error risk. Future studies may enroll children whose parents do not employ drug holidays, children whose parents employ short-term drug holidays and those who employ long term (ie, ≥two months) drug holidays and compare them in terms of ToM, affect perception and metacognitive skills. Regardless of its limitations, our study is the first one to evaluate the effects of MPH treatment and drug holiday on ToM, metacognitive and affect perception skills in a naturalistic, longitudinal design, employing instruments with high levels of reliability. As a result, we found that MPH treatment may improve those skills while reducing symptom severity and that drug holiday may not worsen the evaluated skills at least in the short term. Drug holidays should be determined by the primary clinician treating children with ADHD on an individual basis and their effects (including on quality of life and family functioning) should be evaluated with further studies.

Data Sharing Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to (restrictions, e.g. their containing information) that could compromise the privacy of research.

Ethical Statement

The hospital directorate (Date: 07.11.2023, No: E-33443051-903.99-228750100) and the IRB (Date: 05.12.2023, No: 2023/403) approved of the study protocol and all the study procedures were in accordance with the Declaration of Helsinki and local laws and regulations.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. Design of study: Merve KÖKSAL, Yusuf ÖZTÜRK, Gonca ÖZYURT, Ali Evren TUFAN, Mesut SARİ, Yasemin İMREK. Literature searches and analyses: Merve KÖKSAL, Mesut SARİ, Yasemin İMREK, Yusuf ÖZTÜRK, Ali Evren TUFAN, Gonca ÖZYURT. Statistical analyses: Merve KÖKSAL, Yasemin İMREK, Yusuf ÖZTÜRK, Gonca ÖZYURT, Ali Evren TUFAN, Mesut SARİ. Interpretation of data: Merve KÖKSAL, Mesut SARİ, Yusuf ÖZTÜRK, Gonca ÖZYURT, Ali Evren TUFAN, Yasemin İMREK. Manuscript writing: Merve KÖKSAL, Mesut SARİ, Yasemin İMREK, Yusuf ÖZTÜRK, Gonca ÖZYURT, Ali Evren TUFAN.

Disclosure

The authors report no conflicts of interest in this work.