Background

There has been recent interest in the role of the enzyme Catechol-O-Methyltransferase (COMT) and prefrontal cognition. COMT together with monoamine oxidase catalyzes degradation of catecholamines [1] with COMT being responsible for the majority of dopamine catabolism in the frontal cortex. The COMT gene is located on chromosome 22q11 [2]. Most molecular genetic interest in COMT has focused on a Val 158 Met polymorphism which affects the functional characteristics of the soluble isoform of COMT, with the Met variant being associated with lower activity and thermostability than the Val variant [3]. It is unclear whether the polymorphism affects the function of the membrane form of COMT, the major form in brain, in a similar manner.

A study by Egan et al (2001) [4] demonstrated that patients with schizophrenia, unaffected siblings and controls with the Met allele (low activity) performed better on the Wisconsin Card Sorting Test (WCST), [5] a neurocognitive test of prefrontal cognition [4]. This effect was more prominent when an individual had two copies of the Met allele. These findings have subsequently been replicated in a normal volunteer sample [6] in which fewer perseverative errors on the WCST were made by subjects homozygous for the Met allele. In another study of adult volunteers tested using the Attention Network Test (ANT), there was no association of the COMT variant with overall reaction time or alerting but there was a trend towards higher executive attention scores in those with the Met allele [7]. A more recent study of patients with schizophrenia [8] also reported evidence of association of this COMT variant with neurocognitive function, but with a measure of processing speed and attention, rather than executive function. Extending the sample used by Egan et al.,[4] Goldberg et al. (2003)[9] set out to identify which specific neurocognitive components within the WCST were related to COMT Val 158 Met differences within Schizophrenic patients, their siblings and controls. Using the n-back task, Goldberg et al. found that genotype differences arose from measures of loading onto working memory rather than general attentional deficits (measured by the Continuous Performance task).

Attention Deficit Hyperactivity Disorder (ADHD) is a highly heritable disorder [9] that affects between 2 and 6% of school children.[11] ADHD is characterised by overactivity, impulsivity and inattention and is accompanied by neurocognitive deficits, including those involving executive function [12, 13]. Neuropsychological [14] and functional brain imaging studies [15, 16] have implicated the involvement of the prefrontal cortex and fronto-striatal pathways in the aetiology of ADHD. Given the importance of the dopaminergic system in ADHD and the associations between the neurocognitive abnormalities and the Val 158 Met polymorphism, COMT is clearly a compelling candidate gene for ADHD. However, most groups, including our own, have failed to find evidence for association between polymorphisms in this gene and ADHD [17–21]. However, one study [22] has reported an association between the high activity COMT allele (Val) and a subtype of ADHD; DSM-IV hyperactive-impulsive type.

In this study, we have tested the hypothesis that the Val 158 Met COMT polymorphism is associated with specific aspects of the ADHD phenotype, namely neurocognitive function as assessed by tasks that measure attention, response inhibition and, to some extent, working memory in children with ADHD.

Methods

Subjects

Families of children with suspected ADHD were recruited from Child and Adolescent Psychiatry clinics from Greater Manchester and Cheshire and invited to participate in a genetic study. The study was approved by the North West Multicentre Research Ethics Committee and informed written consent and assent was obtained from the families. Families of 124 children meeting ICD-10 criteria for Hyperkinetic Disorder or DSM-III-R/IV criteria for ADHD, who undertook the neurocognitive test battery and who had been genotyped for the COMT variant were included in the study. The sample were of British Caucasian origin (including the grandparents of the index child), drug-naïve, aged between 6 and 16 years (mean age 9.2 years, standard deviation 1.8 years). 114 were male and 10 were female. Exclusion criteria included children with an IQ test score below 70, as assessed by the Wechsler Intelligence Scale for Children-Third Edition UK (WISC-IIIUK), [23] Tourette's syndrome, fragile X syndrome, epilepsy, pervasive developmental disorder, or other major neurological disorders.

Assessment

A research diagnostic interview; the Child and Adolescent Psychiatric Assessment (CAPA) [24] was completed with the parents of the children by trained interviewers all of whom had a degree in Psychology. To determine the diagnostic criterion of pervasiveness of symptoms across situations, the Child ADHD Teacher Telephone Interview (CHATTI) [25] was used to also assess ADHD symptoms and impairment at school. Interviews were audiotaped and tapes were checked for consistency between interviewers. Good inter-rater reliability was found [26]. Diagnoses were then assigned according to ICD-10, DSM-IV and DSM-III-R criteria. Further details of the assessment and diagnostic process and sample characteristics have been published previously [26].

Each child (n = 124) also completed the WISC-IIIUK to assess IQ and to obtain a measure of working memory derived from the Arithmetic and Digit Span subtests. The forwards and backwards digit span subtest of the WISC III was used as a measure of both attention and working memory. At a second, school based, visit, a battery of neurocognitive tasks were administered, including the Matching Familiar Figures Test (MFFT), [27] the Continuous Performance Test (CPT) [28] and the Stop and Go no Go tasks from the MARS battery (Maudsley Attention and Response Suppression task battery).[14] Data were obtained for 114 children (10 children did not complete the task battery due to administrative difficulties in setting up a second visit at school). The MFFT measures impulse control from reaction time (RT) for correct and incorrect responses, and the number of pictures identified correctly [27]. The CPT (identical pairs) measures inattention, based on the number of targets missed (omission errors), and sustained attention and impulse control, from the frequency of incorrectly pressing a button (commission errors) [28]. The display duration was 1 second with a inter-stimulus delay of 1.5 seconds. The Stop task and Go no Go task are measures of response inhibition. The number of successful inhibitions and mean reaction times (MRT) were obtained for these tests [14].

Genotyping

DNA was obtained from venous blood or mouthwash samples using standard techniques. Polymerase chain reactions (PCRs) were performed using the primers and conditions described by Norton et al (2002) [29]. The Val 158 Met polymorphism was genotyped by single nucleotide primer extension using a template-directed dye-terminator incorporation assay with fluorescence polarization detection [30] based upon AcycloPrime™ reagents (Perkin Elmer Life Science Products, Boston, MA, USA) according to the manufacturers recommendations. Analyses were performed using a LJL Biosystems Analyst™ platform.

Statistics

Children were divided into 3 groups according to genotype (Val/Val, Val/Met and Met/Met) and task performance was compared using one way analyses of variance (ANOVA). A post hoc Bonferroni test was conducted for multiple comparison. All reported p-values have been corrected for multiple testing using Bonferroni statistics.

The groups were compared on all the task measures together using ordinal regression analysis. Power calculations revealed that our sample provides over 80% power to detect an effect size of 0.20 [31].

Results

Data from the WISC-III-R were available for 124 children and the CPT, MFFT and MARS battery were completed on 114 children. The groups did not differ in terms of gender or age. The mean scores on measures for each of the genotypic groups are shown in table 1. The results show no significant difference in attention, working memory or performance on any of the neurocognitive tasks between the three genotypic subgroups. These findings remained when data were analysed for males only (data available from authors). Similarly, where age was significantly associated with neuropsychological measures, this factor was used as a covariate in univariate ANOVAs. Again this did not alter the results (data available from authors). No neurocognitive measure accounted for a significant proportion of the variance in logistic regression analysis.

Table 1 Comparison of neuropsychological test scores for each genotype and results of ANOVA
Full size table

Discussion

We set out to examine the association between the functional Val 158 Met polymorphism in COMT and neurocognitive task performance in a sample of children with ADHD. There was no evidence of differences in performance according to the genotypic group. No differences were found between those homozygous for the Met allele and those with other genotypes (details available from corresponding author on request). These findings contrast with those of previous studies that have suggested an association between the COMT variant and prefrontal cognitive functioning [4, 6]. Two studies found an association of COMT genotype and performance on the Wisconsin Card Sorting task [4, 6] in patients with schizophrenia and in normal adult controls. However in another study of patients with chronic schizophrenia [8], the COMT variant was associated with a measure of processing speed and attention and in a different study of adult volunteers, there was a trend toward association with executive attention scores [7]. Given the postulated link between the COMT variant and prefrontal cognition and the findings of Bilder et al, 2002, [8] we had primarily been interested in differences in the measures of response inhibition, attention and working memory.

Our results suggest that in children with ADHD, not only is there no evidence of association of the functional COMT variant with ADHD [17–21] but that there is also no association with neurocognitive task performance. Our findings suggest that although ADHD is characterised by neurocognitive deficits, in children with ADHD there are no effects of COMT on neurocognitive functioning. However there are alternative explanations for the results.

First, the sample size for which we have neurocognitive data available is small and thus our negative findings could be attributed to a lack of statistical power to detect small effects; in the study by Egan et al, 2001, [4] the COMT gentoype explained only 4% of the variance in the frequency of perseverative errors and a larger sample than ours would be needed to detect small effects. Second, previous studies have been based on patients with schizophrenia (and their unaffected siblings) or normal adults. It is plausible that the mechanisms that lead to impaired cognitive functioning are different in ADHD and that important developmental factors lead to contrasting findings in children and adults. Moreover, our sample consists of children who are drug naïve. This is a strength, but again findings could vary depending on exposure to medication. Finally, the Wisconsin Card Sorting Test was not used in this study and some (but not all) of the previously replicated findings were based on performance on this test specifically. It has not yet been resolved how COMT may be involved with neurocognitive functioning insofar as a more recent study found no evidence of association with the WCST but rather with processing speed and attention [8]. Nevertheless, it could be argued that our measures did not specifically assess prefrontal cognition in the same way that the Wisconsin Card Sorting Test does. In future studies, it would be advantageous to use the WCST.

Conclusion

We find no evidence of an association between the COMT Val 158 Met genotype and performance on a variety of neurocognitive tasks in children with ADHD. Family-based studies of the COMT gene in this sample may shed further light on any association and are currently underway. Our findings indicate that if the COMT gene is involved in ADHD, or a specific phenotypic manifestation of the disorder, as has been suggested, such associations do not lie in differential performance on the neurocognitive tasks undertaken in this study.