RESEARCH ARTICLE Open Access Antipsychotic medications and cognitive functioning in bipolar disorder: moderating effects of COMT Val 108/158 Met genotype Baer Arts, Claudia JP Simons 12, Marjan Drukker and Jim van Os 1,3 Abstract Background: There is a negative association between the use of antipsychotics and cognitive functioning in bipolar patients, which may be mediated by altered dopamine signaling in selected brain areas, and moderation thereof by genetic sequence variation such as COMT Val 108/158 Met. The interaction between antipsychotic drug use and the COMT Val108/158 Met genotype on two-year cognitive functioning in bipolar patients was examined. Methods: Interaction between the COMT Val 108/158 Met and antipsychotics on a composite cognitive measure was examined in 51 bipolar patients who were assessed 12 times at two-monthly intervals over a period of two years (379 observations). Results: There was a significant negative effect of the interaction between antipsychotic medications and Val allele load on the composite cognitive measure in bipolar patients (p < 0.001). Conclusions: The negative effects of antipsychotics on cognitive functioning in bipolar disorder may be moderated by the COMT Val 108/158 Met genotype, with a negative effect of Val allele load. If replicated, the results may be indicative of pharmacogenetic interactions in bipolar disorder. Keywords: Bipolar disorder, Cognition, Antipsychotics, COMT Background Meta-analyses of neuropsychological functioning in euthymic bipolar patients suggest that generalized, rather than specific, cognitive impairments may exist, char- acterized by substantial heterogeneity that is not fully explained by demographic, illness and medication variables [1-4]. Nevertheless, reviews point to a possible role for antipsychotics, indicating that antipsychotics may have detrimental impact on cognitive functioning in bipo- lar patients [5]. For example, Jamrozinski and colleagues (2008) reported no differences between euthymic bipolar patients not exposed to antipsychotic treatment and healthy controls on any neuropsychological measure, whereas a significant underperformance was apparent in *Correspondence: b.arts@maastrichtuniversity.nl 'Department of Psychiatry and Psychology, School for Mental Health and Neuroscience, European Graduate School of Neuroscience (EURON), South Limburg Mental Health Research and Teaching Network (SEARCH), Maastricht University Medical Centre, P.O. Box 616 (DRT 12), Maastricht, MD 6200, The Netherlands Full list of author information is available at the end of the article BioMed Central the bipolar group treated with antipsychotics [6]. A more recent study showed dose-independent deficits in several cognitive tasks in euthymic bipolar patients treated with quetiapine, olanzapine or risperidone, with worse per- formance in patients on second generation antipsychotics compared to untreated euthymic patients [7]. A recent 2-year naturalistic study on cognitive functioning in bipo- lar patients showed significant variation of cognitive functioning over time, largely independent of clinical factors, with the exception of antipsychotic drug use impacting negatively on tasks indexing speed of informa- tion processing [8]. The suggested negative effects of antipsychotics on cognition in bipolar patients contrast with the apparently positive, cognition-enhancing effects of these drugs in the treatment of schizophrenia and schizoaffective disorder [9-12]. However, cognition- enhancing effects in schizophrenia may at least in part be attributable to practice effects [13,14]. Cognitive effects of antipsychotics may be mediated by alterations in dopa- mine signaling in selected brain areas. It may be 2013 Arts et al. licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. hypothesized that in altered and/or hyperdopaminergic states, which may underlie schizophrenia symptoms [15], antipsychotic drugs improve cognition, whereas in bipolar patients without hyperdopaminergia, antipsychotic treat- ment may induce suboptimal cognitive functioning [6]. Antipsychotic effects impacting on dopamine signaling may be modulated by genetic sequence variation. The COMT (catechol-O-methyltransferase) gene Val108/158 Met polymorphism modulates dopaminergic function in frontostriatal circuitry, and may impact information processing efficiency, due to its critical role in the enzym- atic degradation of dopamine. The Val/Val genotype is associated with greater activity of the enzyme and hence with lower concentrations of dopamine in the prefrontal cortex [16,17]. Furthermore, in patients with schizophre- nia, an interaction of the COMT Vall08/158Met genotype and antipsychotic treatment on cognitive functioning has been reported, Met allele load predicting better cognitive performance [18-21], especially in tasks requiring effortful cognitive control [22]. The aim of the present study was to examine whether COMT Val108/158 Met genotype modulates the effect of antipsychotics on a composite cognitive measure, indexing effortful cognitive control, in a sample of bipolar patients, hypothesizing a detrimental effect of Val allele. load. Methods Subjects Individuals were participants in the BIPOLCOG (BIPO- Lar and COGnition) study [23], a study on cognitive functioning in bipolar disorder (BD) in which three groups were investigated: (i) patients with bipolar dis- order, (ii) healthy first-degree relatives of patients with bipolar disorder, and (iii) healthy control participants. All subjects were white, between the ages of 18 and 60 years, fluent in Dutch, had an IQ> 70 and were without a history of neurological disorders such as epilepsy and concussion with loss of consciousness. For the purpose of the current report, only the bipolar patient group was studied, with the healthy control group as reference. A representative cohort of successively attending patients with a diagnosis of bipolar spectrum disorder according to DSM-IV [24] were recruited through the in-patient and out-patient mental health facilities in the geographically defined catchment area of South Lim- burg. In addition, patients were recruited through the local association of bipolar patients and their families, in order to also include patients not currently in contact with services. The computer program OPCRIT was used to confirm DSM-IV diagnosis on the basis of current and lifetime recorded symptomatology listed in the Op- erational Criteria Checklist for Psychotic Illness, scored by the clinical researcher on the basis of all interview and historical case note data (OCCPI) [25]. Control subjects were recruited from the general population using a random mailing sampling procedure and were clinically and diagnostically interviewed with The Comprehensive Assessment of Symptoms and His- tory (CASH) [26] and OPCRIT criteria to exclude those with a past or current diagnosis of BD or psychotic dis- order. Healthy controls were additionally interviewed with the Family Interview for Genetic Studies (FIGS) [27] in order to confirm the absence of a family history of psychotic or bipolar disorder. The initial sample consisted of 81 patients and 61 healthy control subjects. Three patients were excluded because data on diagnosis were missing. Neuropsycho- logical testing data were missing for two patients. The last filter concerned incomplete or missing genetic data, leaving a final risk set for analysis of 51 patients and 50 healthy controls. Procedure As cognitive alterations in bipolar disorder largely develop after onset of illness [28], longitudinal assessment is ne- cessary to adequately capture the phenotype. Thus, patients were examined at 2-monthly intervals over a period of 2 years, yielding a maximum of 12 assessments. At all time points, neuropsychological testing and psychi- atric interviewing took place and questionnaires were completed (regarding social functioning, medication, use of drugs etc.). Genetic material was collected at the first visit. During the baseline interview, basic demographic in- formation was collected as was information on illness characteristics. The study was performed in conformity with the Dec- laration of Helsinki, and approved by the Ethics Com- mittee of the Maastricht University and Academic Hospital. All subjects gave written informed consent prior to participation. Neuropsychological tests and psychiatric interviews were conducted by trained psychologists, each interview occasion taking approximately 2 hours to complete. Healthy controls, in whom cognition is more stable than in patients, were tested twice at two monthly intervals. Data on healthy controls were used as reference to calcu- late standardized z-scores. Neurocognitive assessment Neurocognitive tests were administered by computer, using E-prime for Windows on a 15-inch monitor Toshiba Tecra laptop. The test battery included tasks measuring various neurocognitive domains, guided by previous evidence of impaired performance in these domains in bipolar patients [1,2]. Three subtests were selected a priori from the original test battery, described elsewhere [8], representing tasks with high cognitive load that previous work suggests are most sensitive to moderation by sequence variation in COMT [29-31]. Overall intellectual functioning was estimated at base- line using three Groningen Intelligence Test (GIT) subtests (Mental Rotation, Word Analogies and Mental Arithmetic) [32], yielding results that are comparable to those of the Wechsler Adult Intelligence Scale III [33]. Verbal learning and memory was assessed with the standardized Dutch version of the visually-presented Ver- bal Learning Test [34,35]. In three consecutive trials, 15 monosyllabic non-related words had to be memorized and reproduced. Delayed recall was measured after a 20-minute delay. Parallel versions of this test were used, in order to avoid test-retest-effects. The Flanker CPT (Cogtest plc, London) [36] is a measure of selective visual control of attention. Subjects are instructed to respond by pressing the right or left mouse button depending on whether the middle element in a display of five lines has an arrowhead pointing to the right or left. There are three trial types: (i) neutral trials in which the flankers are just horizontal lines with- out arrowheads, (ii) congruent trials in which all flankers have an arrowhead pointing in the same direction as the target, and (iii) incongruent trials, in which flankers are pointing in the opposite direction from the target. The incongruent condition involves more cognitive effort, because the flankers are associated with a response that needs to be suppressed (measure of response inhibition). One-half of the trials was presented with the stimuli above the fixation cross and the other half were presented below fixation, in order to prevent the subjects from keeping their gaze fixed in one position. The test consisted of 144 trials of neutral, congruent and incongruent flankers, which were presented randomly. Outcome measure was the mean reaction time for correct responses (RT) in the incongruent condition. Finally, Digit Span Backward of the Wechsler Intelligence Scale III [37] was used as measure of working memory. All 3 cognitive measures were standardised, higher scores reflecting better performance. In order to calcu- late a measure of global cognitive functioning, raw test scores were converted into standardized z-scores against the means and standard deviations of the healthy control group. The final composite measure of neurocognition was based on the means of the three domain scores, representing effortful information processing (verbal memory, selective attention/response inhibition and working memory). Genotyping Buccal swab samples were obtained followed by SNP ana- lysis. For the current analysis, the COMTVal 108158Met rs4680 was selected a priori because this is the only SNP that is consistently associated with antipsychotics in pharmacogenetic models [18,19], making it the only cred- ible candidate. Although more SNPs had been determined (n = 184), these formed part of a standard set for the study of gene-environment interactions (GxE), based on published findings up to April 2009 (for the rational underlying this selection process and an overview of selected SNPs see [38]). The 184 SNPs had been chosen a priori (i.e. were not selected from a larger set of genome- wide markers) and selectively determined by Sequenom (Hamburg, Germany) using the Sequenom Mass ARRAY iPLEX platform at the facilities of the manufacturer. In ac- cordance with a priori quality control criteria of the GROUP study, SNPs with more than 10% genotyping errors were excluded, as were SNPs in marked Hardy- Weinberg disequilibrium (p