Abstract
It is unclear whether the intermediate- to long-term use of psychostimulant medication (of at least 3 months duration) is associated with a similar level of, or a diminished improvement in, the core symptom domains of attention deficit hyperactivity disorder (ADHD), compared to short-term use of this medication in the published literature [1–3]. In addition, the intermediate- to long-term side effects of psychostimulant medication include affective symptoms, such as mood lability and dysphoria, which can have a delayed onset. These side effects may mimic the symptoms of ADHD [1–3].
A separate published literature records that ‘ADHD and anxiety’ has been associated with a poor response to short-term psychostimulant medication treatment in a number of studies [4–5]. Neuropsychological and neurophysiological markers have been associated with this poorly responding ‘ADHD and anxiety’ group [6–7]. These include a decreased ability to perform a serial addition digit span task [5], and an increased sensitivity to augmented noradranergic function [6–7]. Hence, ‘ADHD and anxiety’ has been proposed as a clinical ‘subtype’ of ADHD, although a variety of approaches to its definition have been used [5, [6], [8]].
We propose that anxiety will be increased in children with ADHD who are noted by clinicians to be not responding well to psychostimulant medication in the intermediate- to long-term, given that affective symptoms have emerged as side effects in the short term and intermediate- to long-term, and that ‘ADHD and anxiety’ has been associated with a poor short-term response to psychostimulant medication. A preliminary description of 20 children with ADHD who were referred to a specialised ADHD clinic because of a poor intermediate- to long-term response to psychostimulant medication is presented. All of the children had been treated with psychostimulant medication for 6 months or more and all had ongoing core symptoms of ADHD. An age- and intelligence quotient (IQ)-matched group of 20 medication-naive children with ADHD were used to compare the nature and frequency of anxiety phenomena in the two groups. The clinical and research implications of these findings are described and the limitations of these findings emphasised.
Benefits and side effects of psychostimulant medication
Short-term improvements
The efficacy of psychostimulant medication in achieving significant, rapid improvement in the essential symptom domains of ADHD has been demonstrated [9]. These symptom domains are hyperactivity, impulsivity and inattention [10]. Figures of 70–80% of children with significant improvement in these domains have been reported [11], with placebo response rates ranging from 2 to 39% and averaging 18% [12]. Cognitive function constructs such as working memory [5], task persistence, the flexibility of problem-solving, and the speed and accuracy of task response [13] have also been noted to significantly improve with psychostimulant medication.
Short-term side effects
The most frequent short-term side effects of psychostimulant medication are associated with the physiological effects of the activation of the noradranergic neurotransmitter system, among other neurotransmitter systems. These physiological effects include initial and middle insomnia, loss of appetite, stomach-aches, headaches, dizziness and daytime drowsiness [14]. An affective group of symptoms are also short-term side effects of psychostimulant medication. These symptoms include mood lability, dysphoria, sudden severe sadness, sudden crying, social withdrawal, and aggressive outbursts. The latter can mimic the impulsiveness of ADHD. Motor symptoms, such as tics, are a further group of short-term side effects of psychostimulant medication.
Intermediate- to long-term improvements
Gillberg et al. [15] reported a ‘remaining positive effect (of psychostimulant medication) in reducing inattention, hyperactivity and other disruptive behaviour problems’ in a 15 month, randomised, controlled psychostimulant medication trial. The sample studied consisted of ‘children with ADHD symptoms’, defined at clinical interview of the parent(s), using the DSM-III-R criteria. Schachar et al. [2–3] reviewed the randomised, controlled psychostimulant medication trials of 3 months' duration or longer. In addition, they published 4-month prospective data from a randomised, controlled psychostimtlant medication trial. The sample studied consisted of children with ‘a confirmed diagnosis of ADHD’ made on parent, teacher and child reports and clinician judgement. They determined that the essential symptom domains of ADHD were significantly improved, although there was also some evidence that the short-term improvements of psychostimulant medication may be attenuated [3–16].
Intermediate- to long-term side effects
Concern about growth problems and the development of drug tolerance and psychological and physiological dependence have not been demonstrated as intermediate- to long-term side effects of psychostimulant medication treatment [17]. Although other researchers [18] have attempted to define the mechanisms underlying short-term and long-term side effects, Schachar et al. [3] note that there has not been systematic empirical study of this latter area. They published data on the 4-month follow-up of 91 children with ADHD treated with methlyphenidate or placebo in a controlled setting. Physiological side effects, primarily loss of appetite and stomach-aches, affective side effects, primarily social withdrawal, sadness and crying, and a diminished weight gain were significant differences between the treatment and the control groups. Interestingly, the physiological side effects were evident from the commencement of the titration with methylphenidate, while the affective side effects were delayed in their onset until after the titration with methylphenidate was complete.
In summary, short-term use of psychostimulant medication is associated with improvement in the essential symptom domains of ADHD. In contrast, intermediate- to long-term use of psychostimulant medication may be associated with either a similar level of improvement or a diminished improvement in the core symptoms of ADHD. The short-term findings are based on studies from 3 to 6 weeks in duration while the intermediate- to long-term findings are based on studies of at least 3 months in duration.
The side effects of the short-term use of psycho-stimulant medication include the affective symptoms of mood lability, dysphoria, sadness, crying and social withdrawal, physiological symptoms and motor symptoms. The intermediate- to long-term side effects of psychostimulant medication include the affective symptoms of sadness, crying and social withdrawal and the physiological symptoms of loss of appetite, stomach-aches and decreased weight gain. The affective symptoms have a delayed onset in this group, while the physiological symptoms have an immediate onset.
Short-term psychostimulant treatment response and ‘ADHD and anxiety’
During the 1970s, studies were completed which compared good responders and poor responders to short-term psychostimulant medication. For example, Zahn et al. [20] investigated children with minimal brain dysfunction (MBD) and Swanson et al. [21] studied children with ‘hyperactivity’ [19]. Both reported that poor responders had associated ‘high anxiety’ relative to the good responders. Subsequent comparative studies with more rigorous definition of subjects and of responders/non-respon-ders confirmed these earlier findings. Taylor et al. [22–24] reported that an ‘absence’ of ‘overt emotional disorder’ is associated with a ‘good response’ to psychostimulant medication. Plizka [25] reported that approximately two-thirds of children with ‘ADHD and anxiety’ have an ‘ineffective’ response to psychostimulant medication. Approximately one-third responded well to psychostimulant medication. This led to an ‘overall non-significant improvement’ due to treatment in this group of children. In 1992, Plizka [8] again reported that children with ‘ADHD and anxiety’ have a ‘less robust response’ to psychostimulant medication. Buitelaar et al. [4] reported that ‘low rates of anxiety’ are associated with a ‘strong’ stimulant medication ‘response’.
Studies completed recently have suggested neu-ropsychological markers of the ‘ADHD and anxiety’ group. Tannock et al. [5] recorded that elevated levels of anxiety are associated with a lesser response to short-term psychostimulant medication. They reported that ADHD subjects with anxiety did not show an improved performance on a serial addition digit span task when medicated with ‘stimulants’, unlike ADHD subjects without anxiety. Other studies have tried to identify neurophysiological markers of the ‘ADHD and anxiety’ group. Plizka et al. [6] suggested that this group of children with ‘ADHD and anxiety’ have a ‘higher tonic activity’ of the nor-adranergic system. This, in part, may account for this different response to short-term psychostimulant medication. Urman et al. [7] reported that children with ‘ADHD and anxiety’ have a significantly greater rise in diastolic blood pressure when medicated short term with ‘stimulants’. This was interpreted as evidence for an increased sensitivity to augmented noradranergic function in this group.
In summary, ‘ADHD and anxiety’ has been associated with a poor response to short-term psycho-stimulant medication treatment in a number of studies with varying methodological rigour. Neuro-psychological and neurophysiological markers have been associated with the ‘ADHD and anxiety’ group when treated with psychostimulant medication. These findings have been consistently reported despite the following methodological problems in the published literature. The definition of ‘treatment response’ based on behavioural outcome alone, that is, a measured reduction in the level of hyperactivity, rather than on a behavioural and a cognitive response. The heterogeneity of the samples studied, which in part relates to the dramatic changes in the nosological systems used such as the significant change from the Diagnostic and Statistical Manual of Mental Disorders, third edition, to the third edition, revised.
Parent and/or child definition of anxiety
A variety of approaches to defining the anxiety in ‘ADHD and anxiety’ have been used. These include, among others, a structured clinical interview of a parent alone [5] or a structured clinical interview of the child with confirmatory evidence from a child self-report measure [6, [8]]. Each approach has addressed the variation in ADHD and in anxiety symptoms in different situations [10] and the use of categorical or continuous variables to define the ‘presence’ or ‘absence’ of ADHD and of anxiety.
Method
Subjects
Forty children aged from 6 to 12 years were identified in a specialised ADHD clinic in the Eastern region of metropolitan Melbourne, Victoria. All of the children met the criteria for a DSM-IV diagnosis of ADHD [10], and the standard ‘cut-offs’, based on Australian normative data, for the subscale scores of the core symptom domains of ADHD using the parent or the teacher forms of the Child Behaviour Checklist [26].
Twenty children were medication-naive and not responding to usual clinical management approaches, and 20 children had been medicated with psycho-stimulant medication for 6 months or more and had had an initial response to their symptoms but remained impaired and disabled by their symptoms.
Subject characteristics of the primary school-aged children with attention deficit hyperactivity disorder (ADHD)
Measures
The Child Behaviour Checklist (CBCL) [26] consists of 118 behaviour problem items which are rated by the parent or teacher on a three-point scale as to how well each describes the child. Parent and teacher forms are used. A total behaviour problems-scaled score, including subscale scores of all the symptom domains of ADHD, is generated. All scores form continuous variables. The CBCL is well researched and frequently used in treatment evaluations.
The Anxiety Disorders Interview Schedule for children (A-DISC) [27] was used. It is a structured diagnostic interview schedule based on DSM-IV criteria with child and parent versions. It is frequently used internationally by researchers to diagnose childhood anxiety disorders and a range of other disorders, including ADHD, based on the DSM-IV criteria. It yields a categorical variable of the presence or absence of a given disorder. Research findings support the clinical utility, reliability and validity of the A-DISC.
The Revised Child Manifest Anxiety Scale (R-CMAS) [28] consists of 37 items designed to assess anxiety. The R-CMAS also contains a scale to assess social desirability or lying. All scores form continuous variables. Good test-retest reliability and concurrent validity have been demonstrated.
The third edition of the Wechsler Intelligence Scale for Children (WISC-3) [29] was used. This provides verbal, performance and full scale scores of measured intelligence. The third edition of the Wide Range Achievement Test (WRAT-3) [30] provides a measure of spelling and arithmetic abilities. The revised version of the Neale Analysis of Reading Ability [31] provides a measure of reading accuracy and comprehension. All of the above tests are well established with Australian normative data.
Procedure
The child was assessed by a psychologist who performed the WISC-3, the WRAT-3, and the Neale Analysis of Reading Ability Revised. A Fellow in Child and Adolescent Psychiatry concurrently administered the A-DISC parent form. After completing the psychological test battery, the child was then administered the A-DISC child form and the child completed the R-CMAS.
Analysis
To compare the two groups, Chi-squared tests were performed on the dependent variables which were categorical and t-tests were performed on the combination of nominal dependent and continuous independent variables.
Methods
The prevalence of anxiety disorders in the combined sample of 40 children according to the child form was 40% (16:40 subjects). The parent form yielded a prevalence figure of 63% (25:40 subjects).
The A-DISC was used to demonstrate the number of children who received anxiety disorder diagnoses on either the parent and/or child forms of the instrument.
In the 20 children who were medication-naive, the parent report led to an increased number of subjects with anxiety disorder diagnoses (11:5, respectively). It also led to a greater average number of anxiety diagnoses in a given subject (27:20 subjects vs 10:20 subjects, respectively).
In the 20 children who were medicated for 6 months or more, the parent report again led to an increased number of subjects with anxiety disorder diagnoses (14:11, respectively). It also led to a greater average number of anxiety diagnoses in a given subject (31:20 subjects vs 26:20 subjects, respectively).
A gross comparison of the medication-naive group of children with the medicated group shows a clear increase in the summed total of the number of subjects with anxiety disorder (16:25, respectively) and the average number of anxiety diagnoses in a given subject (37:20 subjects vs 57:20 subjects, respectively) in the medicated group. This is primarily due to the increased rate of anxiety diagnoses in the A-DISC child report of the number of subjects with anxiety disorder (5:11) and the increased rate of the average number of anxiety diagnoses in a given subject (10:26) in the medicated group. In contrast, there was a moderate increase in the number of subjects with anxiety disorder (11:14) and the average number of anxiety diagnoses in a given subject (27:31) in the medicated group, according to the A-DISC parent report.
A comparison of the presence or absence of anxiety from a structured clinical interview with the child in the medication-naive and the medicated children with attention deficit hyperactivity disorder
A comparison of the standardised scores of the total anxiety subscale of the Revised Child Manifest Anxiety Scale (RCMAS) in the medication-naive and the medicated children with attention deficit hyperactivity disorder
Interestingly, corresponding non-parametric Chisquared analysis of a categorical variable formed from the A-DISC parent form failed to show the difference between the two groups (X2=0.96; df=1.00; p=0.33). No difference was also found using a continuous variable of anxiety/depression from the parent form of the CBCL (medication-naive group; 68.24 ++ 12.50; medicated group; 65.4 ++ 9.47; statistic-t; value-0.66; p=0.51).
Discussion
The prevalence of ‘ADHD and anxiety’ was 63% using the A-DISC parent form and 40% using the A-DISC child form. This is approximately twice the average prevalence figure of 25% if both clinical and epidemiological studies are reviewed [32]. It highlights the ‘referral bias’ evident in this sample compared to an epidemiological sample: this group of children is relatively more impaired and disabled than other children in the regional Child and Adolescent Mental Health Service (CAMHS), aside from those children with ‘ADHD and anxiety’ who fail to be detected and referred to the service. The data emphasise the requirement for an epidemiological sample to be studied which will lead to an accurate representation of ‘ADHD and anxiety’.
The parent and the child reports showed that the number of children with anxiety disorders and the average number of anxiety disorders per subject were increased in the children treated with psychostimulant medication compared to the medication-naive children. This general increase was primarily due to the child report increasing in both of these factors, while the parent report remained relatively constant. Interestingly, both categorical and dimensional measures of anxiety were significantly increased in the medicated group of children.
This result could be interpreted in a number of ways. First, the recognition of anxiety phenomena in these two groups of children with ADHD is important. Anxiety and its management in children with ADHD are poorly understood. Plizka [25–33] has advocated the use of tricyclic antidepressant medication in children with ADHD and comorbid anxiety, but he states that more careful empirical investigation is required. At present, there are no proven specific therapeutic interventions for children with ‘ADHD and anxiety’. Rather, a multimodal treatment approach is suggested [34–35]. Further definitional work and treatment intervention studies are required.
Second, the nature of the anxiety in children with ADHD remains unclear. The significant difference between the medicated and the medication-naive groups may reflect anxiety as a side effect of psy-chostimulant medication in the intermediate- to long-term. This would be consistent with the findings of Schachar et al. [3] of delayed onset affective side effects which were noted above. The data presented in this paper cannot answer this question. A placebo-controlled study design is required to test this hypothesis.
Third, the significant increase in the anxiety phenomena of the medicated children with ADHD may be a potential marker of diminished response to psy-chostimulant medication in this intermediate- to long-term period. This would be consistent with the replicated finding that premorbid ‘ADHD and anxiety’ predicts a lesser response to psychostimulant medication. However, a prospective, longitudinal within-subject design is required to test this hypothesis.
Fourth, the difference or dissonance between parent and child reports of anxiety, in this particular group of children, is noteworthy. It has been noted before in the literature albeit with no clear explanation of its nature [8]. It may reflect a degree of ‘denial’ being evident in the parent, the child or their relationship: conflictual family relationships and the associated difficulty of forming ‘relatively objective’ interpretations are a ubiquitous context for children with ADHD.
Conclusion
Anxiety is more prevalent in children with ADHD treated with psychostimulant medication in the intermediate- to long-term who are noted by clinicians to have ongoing core symptoms of ADHD. This statistically significant finding is evident with both categorical and dimensional measures of anxiety from the child's perspective.
From a clinical perspective, it suggests that ‘ADHD and anxiety’ is unrecognised as most clinicians did not note that anxiety was present. Its recognition may lead to better management of this group of children with ADHD who have not responded well to treatment. For example, the application of cognitive-behavioural therapy for anxiety and alternative medication strategies have not been rigorously studied and used in this group of children [33]. Specific psychoeducation of the parent(s), teachers, and other carers of the child may also improve management.
Footnotes
Acknowledgements
This work was supported by the 1997 Eli Lilly Psychiatry Fellowship which was awarded to Dr Vance. The authors acknowledge the help of Dr Peter Birleson, Director, Maroondah Child and Adolescent Psychiatry Service, Melbourne, Australia.