Journal of Anxiety Disorders 30 (2015) 48–58

Contents lists available at ScienceDirect

Journal of Anxiety Disorders

Review

Neuropsychological differences between obsessive-compulsive washers and checkers: A systematic review and meta-analysis Rebecca Leopold a,∗ , Matthias Backenstrass a,b a b

Department of General Psychiatry, Center for Psychosocial Medicine, University Hospital Heidelberg, Vossstrasse 4, D-69115 Heidelberg, Germany Institute of Clinical Psychology, Center for Mental Health, Priessnitzweg 24, D-70374 Stuttgart, Germany

a r t i c l e

i n f o

Article history: Received 10 September 2014 Received in revised form 22 December 2014 Accepted 25 December 2014 Available online 9 January 2015 Keywords: Obsessive-compulsive disorder Symptom dimensions Washers Checkers Neuropsychology Meta-analysis

a b s t r a c t Inconsistent results in neuropsychological research of obsessive-compulsive disorder (OCD) may be the result of the heterogeneous nature of OCD symptoms. The most frequently investigated symptoms are contamination/cleaning and doubt/checking. The aim of this review was to determine whether OCD washers and checkers differ in their neuropsychological performance. We conducted a meta-analysis of 13 studies (including 535 patients) comprising tests in 10 different neuropsychological domains. Washers showed significant better task performance than checkers in 8 of 10 cognitive domains. Large effect sizes were found in planning/problem solving and response inhibition. Effect size in set shifting was medium, whereas effect sizes in attention, processing speed, encoding, verbal memory and nonverbal memory were small. Limitations consisted in a relatively small number of primary studies. In line with current neurobiological findings, the results provide further evidence for the validity of different symptom dimensions in OCD. Clinical and theoretical implications are discussed. © 2015 Elsevier Ltd. All rights reserved.

Contents 1. 2.

3.

4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Search strategy and study selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Data extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Study selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Study characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Participants’ characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Neuropsychological functioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Discussion of the results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

1. Introduction Abbreviation: PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-analyses Statement. ∗ Corresponding author. Present address: The City of Karlsruhe’s Youth and Drug Counseling Service, Kaiserstrasse 64, D-76133 Karlsruhe, Germany. Tel.: +49 721 133 5394; fax: +49 721 133 5489. E-mail addresses: [email protected] (R. Leopold), [email protected] (M. Backenstrass). http://dx.doi.org/10.1016/j.janxdis.2014.12.016 0887-6185/© 2015 Elsevier Ltd. All rights reserved.

Obsessive-compulsive disorder is increasingly considered as a heterogeneous mental disorder (Mataix-Cols, do Rosario-Campos, & Leckman, 2005) comprising, as cited in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM5; American Psychiatric Association, 2013), specific symptom

R. Leopold, M. Backenstrass / Journal of Anxiety Disorders 30 (2015) 48–58

dimensions that differ from each other in numerous ways. The most widely investigated OCD symptom dimensions are the contamination/washing dimension, and the doubt/checking dimension (Fontenelle, Mendlowicz, Marques, & Versiani, 2004; Rasmussen & Eisen, 1989). The washing dimension contains intrusive thoughts of contamination, emotions of fear and disgust, and ritualized washing and cleaning behavior. The checking dimension includes thoughts of harm avoidance, pathological doubt, and repeated checking of, for example, locked doors and turned-off stoves or irons. In both symptom dimensions, the obsessive thoughts cause marked distress, and the time-consuming compulsive behaviors lead to significant impairment in daily life. An OCD patient can suffer either from washing or checking symptoms or from both. Thus, for instance, an OCD patient may score higher on the washing dimension and lower on the checking dimension (or vice versa or even on both dimensions to a similar degree). As obsessive-compulsive disorder symptoms vary on a continuum of severity, categorical attempts to classify OCD patients according to distinct subgroups of non-overlapping symptoms have been criticized (Leckman, Mataix-Cols, & do Rosario-Campos, 2005). Nevertheless, in several studies OCD researchers compared washers with checkers (Horesh, Dolberg, Kirschenbaum-Aviner, & Kotler, 1997; Khanna & Mukherjee, 1992 Rasmussen & Eisen, 1989). In our review, we will use ‘symptom dimensions’ as a superordinate term referring to ‘checkers’ as OCD patients either with exclusive checking symptoms or with predominant checking symptoms amongst multiple OCD symptoms, whereas we regard ‘washers’ as OCD patients either with exclusive washing symptoms or with predominant washing symptoms amongst multiple OCD symptoms. Studies in various fields of research such as genetic, neuroimaging and treatment studies provided evidence of the washing and the checking symptom dimension in OCD symptomatology. Factor analyses of symptom scales such as the symptom checklist of the Yale-Brown Obsessive Compulsive Scale (Y-BOCS; Goodman, Price, Rasmussen, & Mazure, 1989) and self-report measures such as the Obsessive-Compulsive Inventory-Revised (OCI-R; Foa et al., 2002) consistently generated both the washing and checking dimensions (Bloch, Landeros-Weisenberger, Rosario, Pittenger, & Leckman, 2008; Foa et al., 2002; Pinto et al., 2007). The symptom dimensions are stable over the time (Mataix-Cols, Rauch et al., 2002) and across different cultures as the two dimensions have been found in Western (Leckman, Grice, Boardman, & Zhang, 1997; Mataix-Cols et al., 2005) and Asian OCD patient samples (Kim, Lee, & Kim, 2005; Matsunaga et al., 2008). Explorative studies found specific genetic influences with regard to the contamination/washing dimension in contrast to the doubt dimension (Katerberg et al., 2010). Likewise, molecular genetic associations of the contamination/washing dimension with the serotonin transporter polymorphism have been found (Kim et al., 2005). In a large twin study, van Grootheest, Boomsma, Hettema, and Kendler (2008) reported that the contamination dimension was influenced by specific genetic and environmental factors. With respect to the checking dimension, particular genetic findings are still missing. Neuroimaging studies showed that the washing and the checking dimension differ in structural brain features as reported by van den Heuvel et al. (2009). The authors found that the washing dimension correlated negatively with gray matter volume in the bilateral dorsal nucleus caudatus, while the checking dimension correlated negatively with gray matter and white matter volumes of both temporal lobes. Results from a functional magnetic resonance imaging (fMRI) study using a symptom provocation paradigm revealed that the contamination/washing dimension was related to increased cerebral activation in prefrontal brain regions and the right nucleus caudatus, whereas the checking dimension was related to increased

49

activation in the putamen and thalamus (Mataix-Cols et al., 2004). In a recent fMRI study also applying a symptom provocation paradigm, Murayama et al. (2013) reported that checking rituals were associated with subcortical brain regions such as the nucleus caudatus and anterior cingulate cortex, whereas washing rituals were associated with large cortical brain regions including the cerebellum. The OCD symptom dimensions also vary with regard to comorbidity and treatment response. Patients with predominant washing symptoms seem to be at increased risk of eating disorders whereas patients with predominant checking behavior and aggressive obsessions are at higher risk for comorbid major depression and generalized anxiety disorder (Hasler et al., 2005). The OCD symptom dimensions are regarded as predictors of treatment response to serotonin reuptake inhibitors (Mataix-Cols, Rauch, Manzo, Jenike, & Baer, 1999) and cognitive behavioral therapy (Mataix-Cols, Marks, Greist, Kobak, & Baer, 2002), since patients with predominant washing symptoms need more therapy sessions when they primarily feel disgust compared to checkers as stated in the review by Williams, Mugno, Franklin, and Faber (2013). Although these prior studies provide evidence that OCD patients with either predominant washing or checking symptoms differ from each other in genetic, structural and functional neurobiological features as well as in comorbidity and treatment response, it remains unclear whether they also differ in their neuropsychological functioning. Neuropsychological research of OCD in general draws a heterogeneous picture. Two recent meta-analyses reported that OCD patients exhibited significant but moderate deficits in the domains of attention (Abramovitch, Abramowitz, & Mittelman, 2013), executive functions, verbal and nonverbal memory, visuospatial abilities, processing speed, and working memory compared to healthy controls (Abramovitch et al., 2013; Shin, Lee, Kim, & Kwon, 2014). In a critical review, Kuelz, Hohagen, and Voderholzer (2004) reported visuospatial memory deficits and using more complex tasks, also verbal memory deficits in OCD, both presumably due to underlying encoding deficits. Findings regarding attention and executive functions were somewhat contradictory across the included studies: Kuelz et al. (2004) reported that sustained attention seemed to be unaffected. Moreover, some studies found impaired set shifting abilities in the WCST, whereas other studies did not (Kuelz et al., 2004). However, a substantial part of research did not find significant differences in neuropsychological performance between OCD patients and healthy participants in response inhibition (Moritz et al., 2008), set shifting (Abbruzzese, Ferri, & Scarone, 1995; Simpson et al., 2006), verbal memory (Moritz, Kloss, von Eckstaedt, & Jelinek, 2009) and nonverbal memory (Moritz, Kloss et al., 2009; Simpson et al., 2006). In summary, the presented outcomes regarding possible neuropsychological deficits in OCD patients are inconsistent. A different conceptualization of OCD symptomatology may bring light into the divergent findings in neuropsychological research. Based on the outcome of their meta-analysis, Abramovitch et al. (2013) suggested considering the symptomatic heterogeneity of OCD in order to find out whether differences in subsamples moderate cognitive performance and whether symptom heterogeneity in the OCD samples accounts for the underestimation of cognitive deficits in OCD patients in general. While none of the aforementioned reviews considered any of the OCD symptom dimensions, two systematic reviews took at least the checking dimension into account. First, Woods, Vevea, Chambless, and Bayen (2002) found significant medium effect sizes for visual free recall, verbal cued recall and recall of actions, and significant small effect sizes for verbal free recall, visual recognition and working memory, indicating moderate memory impairment of checkers compared to non-checkers. Second, as a result of their

50

R. Leopold, M. Backenstrass / Journal of Anxiety Disorders 30 (2015) 48–58

review including 67 studies with 1519 OCD patients and 236 subclinical checkers, Cuttler and Graf ([Cuttler and Graf, 2009]2009, p. 406) stated that ‘deficits in retrospective memory are not specific to checkers” but “that deficits in prospective memory are specific to checkers’. In both reviews, checkers were compared to ‘non-checkers’; the latter comprised healthy subjects, OCD patients without checking compulsions, and patients with other psychiatric disorders than OCD. Because of the heterogeneity of the nonchecking samples, appropriate conclusions of the findings cannot be drawn without difficulties. Aside from reviews of OCD in general and those of checkers versus non-checkers, there have been several single studies published comparing OCD washers and checkers with regard to their neuropsychological performance. Bouvard, Dirson, and Cottraux (1997), Cha et al. (2008) and Nakao et al. (2009) reported that checkers were significantly more impaired in nonverbal memory than washers. In contrast, Ceschi, van der Linden, Dunker, Perroud, and Brédard (2003) did not find significant nonverbal memory differences between washers and checkers. Studies examining executive functions in OCD patients demonstrated that checkers performed significantly worse than washers in planning tasks (Dittrich & Johansen, 2013; Nedeljkovic et al., 2009), while Omori et al. (2007) and Penadés et al. (2007) observed that checkers experienced deficits in response inhibition. Altogether, these studies yielded inconsistent results that have not been summarized so far. To the best of our knowledge, a systematic review and meta-analysis comparing particularly OCD washers and checkers in their neuropsychological functions has not been published yet. Our objective was to find out whether OCD patients with predominant washing versus checking symptoms differed from each other with regard to neuropsychological impairment. Particularly, we wanted to know whether washers and checkers differ from each other in five cognitive domains according to the neuropsychological classification of Abramovitch et al. (2013); namely, attention, executive functioning, memory, processing speed, and working memory. We therefore conducted a systematic review and metaanalysis.

2. Method 2.1. Search strategy and study selection The review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Statement (PRISMA; Moher, Liberati, Tetzlaff, & Altman, 2009). Prior to the systematic search of literature, we established a review protocol (available on request from the first author). As the meta-analysis of OCD washers and checkers was part of a comprehensive umbrella review investigating the neuropsychological performance of patients of any DSM-5 OCD symptom dimension as well as compulsive hoarding, merely a part of the literature search was used for this meta-analysis. Potential articles were identified through a three-step search procedure for articles published between 1948 and January 2014, including electronic database search, manual search, and contact to study authors. First, the database search included PsychInfo, MEDLINE, PubMed, Google Scholar, ProQuest Dissertations & Theses, and Scirus, using the search terms (‘wash*’ OR ‘check*’ OR ‘hoard*’ OR ‘obsessive compulsive’ OR ‘OCD’) AND (‘neuropsych*’ OR ‘neurocogn*’ OR ‘information processing’ OR ‘memory’ OR ‘attention’ OR ‘executive’ OR ‘set shifting’ OR ‘cognitive *function’ OR ‘cognitive deficit’ OR ‘frontal *function’ OR ‘frontal deficit’). Second, we examined reference lists of review articles for other relevant studies. Third, leading scientists in the field of neuropsychology of OCD were contacted and asked for unpublished studies and additional data. Of 17

research groups, 14 responded, and three of them (Exner, Martin, & Rief, 2009; Jang et al., 2010; Tumkaya et al., 2013) provided unpublished data that was included in the meta-analysis. The studies had to meet the following inclusion criteria: (1) Adult patients (18 years or older) diagnosed with OCD according to the Diagnostic and Statistical Manual of Mental Disorders, third edition (DSM-III), fourth edition (DSM-IV), or fifth edition (DSM-5), or the International Classification of Diseases, ninth revision (ICD-9), or tenth revision (ICD-10); (2) comparison of washing and checking symptom dimensions; (3) at least one dependent variable consisted of neuropsychological test performance; (4) studies reported statistics (means, standard deviations and sample size or t, F, r and exact p) that were convertible to effect sizes, (5) published in English, French, or German. Studies were excluded when they met one of the following exclusion criteria: (1) studies with sample size smaller than 10, (2) studies that did not specify washing and checking symptom dimensions, (3) studies with subclinical samples (e.g. students), (4) studies examining patients with obsessive compulsive symptoms due to a General Medical Factor (DSM-5) (e.g. due to brain lesions, infections or demential processes), (5) studies reporting only graphical information or nonparametric tests with study authors not providing exact data until June 2014. After removing duplicates, titles and abstracts of articles were reviewed to exclude studies that failed the inclusion criteria. The remaining studies were retrieved in full text and all studies that did not fulfill inclusion criteria were finally excluded from the meta-analysis. 2.2. Data extraction Using a modified version of the data collection form by Burt, Zembar, and Niederehe (1995), variables were recorded in a structured manner as follows: (1) study characteristics such as year of publication, country in which the study was conducted, publication status, and methodological study quality; (2) participants’ characteristics such as OCD washers’ and checkers’ sample sizes, percentage of females in each group, mean scores on measures of intelligence, years of education, mean duration of OCD in years, mean age of OCD onset, percentage of OCD washers and checkers receiving psychotropic medication, mean scores on measures of OCD severity (e.g. Y-BOCS), depression (e.g. BDI or BDI-II), and state anxiety (e.g. STAI-S); (3) means and standard deviations or r, t, F, and p statistics of at least one neuropsychological measure. To minimize the risk of bias in data extraction, two independent raters (the first author and a trained postgraduate psychologist) recorded 20 randomly selected studies, using the standardized data collection form and a coding manual. Differences were resolved by consensus. Hereafter, one rater (first author) recorded all variables of the included studies. Further, we developed a 10-item rating scale to rate the methodological quality of the included studies. For this purpose, we modified the ‘Psychotherapy Outcome Study Methodology Rating Scale’ (Öst, 2008) and also used items derived from The Cochrane Collaboration (Deutsches Cochrane Zentrum, 2013). Our study quality rating scale included the following items: (1) ethical approval, (2) clarity of sample description, (3) assessment of OCD severity, (4) representativeness of the samples concerning psychiatric comorbidity, (5) reliability of the diagnosis in question, (6) reliability and validity of neuropsychological tests, (7) number of test supervisors, (8) training of test supervisors, (9) control of pharmacological treatment, (10) statistical analyses and data presentation. Each item was rated with 0 = poor, 1 = fair, or 2 = good, and each value had a verbal description. The interrater-reliability for the ten items was good to excellent (intraclass correlation coefficients, according to Shrout and Fleiss (1979), ranged from ICC (2,1) = .75, p = .001 to ICC (2,1) = .94, p < .001) and internal consistency of the scale was acceptable (Cronbach’s alpha = .697). The summary score of the study quality rating scale

R. Leopold, M. Backenstrass / Journal of Anxiety Disorders 30 (2015) 48–58 Table 1 Neuropsychological tests subsumed in five cognitive domains and ten subdomains. Domain

Subdomain

Attention

Sustained attention

Executive functions

Memory

Processing speed Working memory

Tests

K-T cancellation test, SAT, WMS-R mental control Planning/problem TOH, solving CANTAB SOC (initial latency) Response inhibition FWIT (reaction time), color word Stroop (interference, reaction time and errors), Go/No-Go (commission errors and reaction time) Set-shifting/cognitive CANTAB IED (trial level), flexibility WCST (number of trials and total errors), TMT-B (reaction time and subtracted score), letter fluency A/Fu (Japanese version) RCFT (copy and organization Encoding strategies score), SWM (strategy score) Verbal memory WAIS-R LM (II), BEM 144 (verbal total), CVLT, HVLT Nonverbal memory BEM 144 (visuel), RCFT (delayed recall), WMS-R visual reproduction (II) Processing speed WAIS-R digit symbol, TMT–A WMS-R block span, Spatial working memory CANTAB SWM (errors), spatial N-Back task spatial working memory task Verbal working WAIS-R digit span, memory verbal working memory task

BEM, Batterie d’Efficience Mnésique 144; CANTAB, Cambridge Automated Neuropsychological Test Battery with subtests: IED, Intra-Extra-Dimensional Shift; SOC, Stockings of Cambridge; SWM, Spatial Working Memory; CVLT, California Verbal Learning Test; FWIT, Farbe-Wort-Interferenztest; HVLT, Hopkins Verbal Learning Test; K-T Cancellation Test; LM I and II, Logical Memory Form I and Form II; RCFT, Rey-Osterrieth Complex Figure Test; SAT, Situation Awareness Test; TMT-A and B, Trail-Making Test, Part A and B; TOH, Tower of Hanoi; WAIS-R, Wechsler Adult Intelligence Scale-Revised; WCST, Wisconsin Card Sorting Task; WMS-R, Wechsler Memory Scale-Revised.

later was entered as a potential moderator in meta-regression analysis. Align with the grouping of neuropsychological tests and domains of Abramovitch et al. (2013), we divided the neuropsychological test measures into five domains and ten subdomains of cognitive functioning (see Table 1): attention (sustained attention), executive functions (planning/problem solving, response inhibition, set shifting/cognitive flexibility, encoding strategies), memory (verbal memory, nonverbal memory), processing speed (processing speed), and working memory (verbal working memory, visuospatial working memory).

51

relationship between symptom dimension and cognitive deficit, the larger is the group difference. Thus, we calculated Hedges’ g not only from means and standard deviations but also from Pearson’s product moment correlations reported in four studies. Descriptions of the simultaneous use of correlation coefficients and group differences in a meta-analysis are given in Borenstein, Hedges, Higgins, and Rothstein (2009a) and Rosenthal and DiMatteo (2001). We combined correlational data by Fisher’s Z transformation (Silver & Dunlap, 1987) and followed the guidelines for the calculation of effect sizes by Dunst, Hamby, and Trivette (2004) to convert Pearson’s r into Cohen’s d and finally d into Hedges’ g. If studies reported more than one outcome for one neuropsychological test, we used the higher-ranking outcome, e.g. the total test score. If studies reported more than one test for one cognitive domain, we calculated a synthetic effect size for that domain and corrected the variance of the synthetic effect size by taking the correlation among the non-independent outcomes into account according to Borenstein, Hedges, Higgins, and Rothstein (2009b). In line with Cohen (1992), we classified effect sizes as small (0.2), medium (0.5) or large (0.8). Given the great variety of neuropsychological constructs and tests, we used random-effects models to calculate combined effect sizes across the studies. To assess heterogeneity, we calculated the Q and I2 statistics. In case of sufficient homogeneity, we also calculated combined effect sizes according to the fixed-effect model. Publication bias was examined by Egger’s regression intercept (Egger, Smith, Schneider, & Minder, 1997) and the funnel plot. If necessary, the trim-and-fill method (Duval & Tweedie, 2000) was used to adjust for publication bias. Furthermore, we calculated Orwin’s fail-safe N (Orwin, 1983) that indicates how many unpublished studies are necessary to reduce the calculated combined effect to a trivial size which we assumed to be g = 0.10. To control for publication bias, we contacted study authors and asked for unpublished data. We used meta-regression analyses to assess the influence of the following variables on effect sizes: standardized mean differences in age, intelligence/years of education, methodological study quality (summary score of the study quality rating scale), severity of OCD (Y-BOCS score), percentage of patients on psychotropic medication, comorbid depression (standardized mean differences of BDI, BDI-II, HDRS, MADRS), comorbid state anxiety (standardized mean differences of STAI-S, HARS, BAI), OCD duration, and participants’ age of OCD onset. We also performed a subgroup analysis to control for possible differences in the data entry format (means and standard deviations versus Pearson’s correlations). The Comprehensive Meta-Analysis Version 2.0 software (CMA; Biostat, Inc., USA) was used for all meta-analytic calculations. All significance tests were two-tailed with alpha = .05, except of the meta-regression analyses. Since CMA does not allow for simultaneous testing of multiple predictor variables in meta-regression, we performed several meta-regression analyses consecutively and therefore set alpha = .01 to adjust for multiple testing.

2.3. Statistical analysis We calculated the effect size Hedges’ g which is defined as the difference in means divided by the pooled standard deviation and weighted for sample size (Hedges & Olkin, 1985) for the neuropsychological outcome variables and the cognitive domains. By not only considering the categorical but also the coexistent dimensional nature of washing and checking symptoms, we made the following assumption: If in a sample of participants a systematic relationship (correlation) between OCD symptom dimensions and cognitive deficits exists, then two subsamples scoring either high vs. low on a particular symptom dimension (e.g. checking) will also differ in neuropsychological performance; the stronger the

3. Results 3.1. Study selection The search resulted in 5790 potentially eligible studies after the exclusion of duplicates. A total of 498 articles appeared to be potentially relevant and were retrieved as full text. Of these, 485 articles were excluded as they did not meet inclusion criteria. We then identified 13 studies that compared the neuropsychological performance of OCD patients with either washing or checking symptoms and that further met all inclusion criteria (Bouvard,

52

R. Leopold, M. Backenstrass / Journal of Anxiety Disorders 30 (2015) 48–58

Records identified through database searching: k = 6 037

Records identified through other sources: manual search in reference lists: k = 20 contact with study authors: k = 12

Total records: k = 6 069 after duplicates removed: k = 5 790

Records excluded (title + abstract): k = 5 292 Records screened (title and abstract): k = 5 790

reasons for exclusion: animal studies, k = 155 + 34 pediatric studies, k = 363 + 69 no OCD, k = 2 766 + 249 subclinical OC symptoms, k = 37 + 76 acquired OC symptoms, k = 100 + 9 no symptom dimensions, k = 0 + 74 no neuropsychological tests, k = 353 + 638 methodological reasons, k = 64 + 305

Full text articles assessed for eligibility: k = 498 Full text articles excluded: k = 485

Studies included in meta-analysis: k = 13

reasons for exclusion: pediatric studies, k = 1 no OCD, k = 9 subclinical OC symptoms, k = 18 no symptom dimensions, k = 280 no neuropsychological tests, k = 54 other symptom dimensions than washing and checking, k = 37 methodological reasons, k = 86

Fig. 1. Search strategy used for selection of studies included in the meta-analysis.

Dirson, & Cottraux, 1997; Ceschi et al., 2003; Cha et al., 2008; Dittrich & Johansen, 2013 Exner et al., 2009 Hashimoto et al., 2011 Jang et al., 2010 Kraft, 2006 Nakao et al., 2009; Nedeljkovic et al., 2009 Omori et al., 2007 Penadés et al., 2007; Tumkaya et al., 2013). A flow diagram illustrates the study selection process (see Fig. 1). 3.2. Study characteristics Of the 13 studies, three were conducted in South Korea, three in Germany, two in Japan, and the remaining studies in each of the following countries: Australia, France, Spain, Switzerland,

and Turkey. One study comprised an unpublished doctoral thesis (Kraft, 2006). Four studies reported correlation coefficients, whereas nine studies reported means and standard deviations. All studies made OCD diagnoses according to DSM–IV or DSM–III. Twelve studies measured severity of OCD with the Y-BOCS. All studies assessed comorbid symptoms with at least one established psychometric instrument. The methodological quality of the included studies, measured with the aforementioned 10-item scale (range of values from 0 to 20), ranged from 11 to 17 with a mean study quality of 14.00 (SD = 2.00). Table 2 illustrates the study characteristics.

R. Leopold, M. Backenstrass / Journal of Anxiety Disorders 30 (2015) 48–58

53

Table 2 Characteristics of the studies included in the meta-analysis. Study, country

Sample characteristics

Assessment of diagnoses and comorbidity

Cognitive domain and assessment of intelligence

Bouvard et al. (1997) France Ceschi et al. (2003) Switzerland Cha et al. (2008) South Korea Dittrich et al. (2013) Germany *◦ Exner et al. (2009) Germany

11 washers, 11 checkers

DSM-III Y-BOCS BDI DSM-IV Y-BOCS BDI-II, STAI DSM-IV Y-BOCS, BDI, HDRS, HARS DSM-IV Y-BOCS MADRS, STAI DSM-IV Y-BOCS BDI, BSI

Attention, memory RAPM, WAIS verbal Memory

◦

Hashimoto et al. (2011) Japan ◦ Jang et al. (2010) South Korea *Kraft (2006) Germany Nakao et al. (2009) Japan

16 washers, 16 checkers 23 washers, 24 checkers 6 washers, 7 checkers 19 OCD patients with multiple symptoms

63 OCD patients with multiple symptoms 144 OCD patients with multiple symptoms 36 OCD patients with multiple symptoms 14 washers, 10 checkers

Nedeljkovic et al. (2009) Australia

12 washers, 11 checkers

Omori et al. (2007) Japan Penadés et al. (2007) Spain *◦ Tumkaya et al. (2013) Turkey

26 washers, 27 checkers 10 washers, 7 checkers 42 OCD patients with multiple symptoms

DSM-IV Y-BOCS BDI-II, STAI DSM-IV Y-BOCS BDI, BAI DSM-IV Y-BOCS, BDI DSM-IV Y-BOCS HDRS, STAI DSM-IV Y-BOCS BDI, BAI, HDRS, HARS DSM-IV Y-BOCS BDI-II, STAI DSM-IV MOCI BDI DSM-IV Y-BOCS, MOCI HDRS, HARS

Executive functions, memory WAIS-R verbal Executive functions NART Executive functions, memory, processing speed, working memory Executive functions, memory, processing speed Executive functions, memory WAIS-R Working memory MWT-B Attention, executive functions, memory, working memory WAIS-R Executive functions, working memory NART Executive functions, memory, processing speed Executive functions WAIS-III Attention

Studies marked with an asterisk * are unpublished studies or authors provided additional unpublished data. Studies marked with a circle ◦ are correlational studies. BAI, Beck Anxiety Inventory; BDI, Beck Depression Inventory; BDI-II, Beck Depression Inventory-Revised, second edition; BSI, Brief Symptom Inventory; DSM-III, Diagnostic and Statistical Manual of Mental Disorders, third edition; DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, fourth edition; HARS, Hamilton Anxiety Rating Scale; ´˚ HDRS, Hamilton Depression Rating Scale; MADRS, Montgomery–Asberg Depression Rating Scale; MOCI, Maudsley Obsessive Compulsive Inventory; MWT-B, MehrfachwahlWortschatz-Test, part B; NART, National Adult Reading Test; MOCI, Maudsley Obsessive Compulsive Inventory; OCI-R, Obsessive-Compulsive Inventory-Revised; RAPM, Raven Advanced Progressive Matrices; STAI, State-Trait Anxiety Inventory; WAIS, Wechsler Adult Intelligence Scale; WAIS-R,Wechsler Adult Intelligence Scale-Revised; WAIS-III, Wechsler Adult Intelligence Scale, third edition; Y-BOCS, Yale-Brown Obsessive Compulsive Scale.

3.3. Participants’ characteristics In total, 535 participants were included in the 13 studies. Of the 535 participants, 118 showed predominant washing symptoms, 113 patients predominantly displayed checking symptoms, 304 participants suffered from multiple OCD symptoms and scored to varying degrees on each of the washing and the checking dimension. As far as data were available, patients seemed well matched with regard to age, gender distribution, percentage on medication, intelligence, education, duration of OCD, age of onset, OCD severity measured by Y-BOCS total score, depression measured by BDI or BDI-II scores, and state anxiety measured by STAI-S scores. Demographic and clinical characteristics of the participants are displayed in Table 3. 3.4. Neuropsychological functioning Effect sizes of washers’ and checkers’ neuropsychological performance in the respective tasks are shown in Fig. 2. Positive

effect sizes indicate that washers performed better than checkers, whereas negative effect sizes point out that washers performed poorer than checkers. Washers achieved better results than checkers in 38 of 41 neuropsychological tasks. The largest significant effects were found in the Tower of Hanoi (TOH; g = 1.79, p = .004) and the Stockings of Cambridge task (SOC; g = 1.76, p = .000), followed by performance in the CANTAB SWM (g = 1.32, p = .003) and the WMS-R visual reproduction (g = 0.94, p = .027). In the WMS-R Logical Memory, a large effect size (g = 0.82 p = .050) contrasted three medium effect sizes that did not reach significance. The medium effect sizes for the WCST, color word Stroop task, the BEM 144 verbal and visual total scores and the TMT-A also failed statistical significance. Likewise, small and non-significant effects were found for the digit symbol test, spatial N-Back task, and the digit span. Contrary to these results, washers performed significantly poorer than checkers in a spatial working memory task (g = −1.07, p = .027) and a verbal working memory task (g = −1.08, p = .027). In 8 of 10 cognitive subdomains, combined effect sizes significantly greater than zero provided evidence of washers’ better

54

R. Leopold, M. Backenstrass / Journal of Anxiety Disorders 30 (2015) 48–58 Group by Domain/Subdomain

Study name

Outcome

Hedges's g and 95% CI

01 Sustained attention

Bouvard 1997

KT cancellation test

0,39

0,323

01 Sustained attention

Nakao 2009

Mental control

0,66

0,108

01 Sustained attention

Tumkaya 2014

SAT

0,45

0,166

0,49

0,022 0,004

Hedges's g

01 Sustained attention

p-Value

02 Planning/problem solving

Dittrich 2013

TOH

1,79

02 Planning/problem solving

Nedelj kovic 2009

CANTAB SOC

1,76

0,000

1,77

0,000

02 Planning/problem solving 03 Response inhibition

Exner* 2009

FWIT RT

0,36

0,462

03 Response inhibition

Hashimoto 2011

Color word Stroop

0,37

0,599

03 Response inhibition

Omori 2007

Combined Stroop, Go/No-Go

1,16

0,000

03 Response inhibition

Penadés 2007

Combined Stroop, Go/No-Go

0,92

0,063

0,89

0,000

03 Response inhibition 04 Set-shifting/flexibility

Nakao 2009

WCST TE

0,56

0,167

04 Set-shifting/flexibility

Nedelj kovic 2009

CANTAB IED trial level

0,23

0,566

04 Set-shifting/flexibility

Exner* 2009

Combined WCST, TMT-B

1,07

0,050

04 Set-shifting/flexibility

Hashimoto 2011

Combined TMT-B, fluency

0,75

0,097

04 Set-shifting/flexibility

Omori 2007

Combined WCST, TMT-B

0,66

0,018

04 Set-shifting/flexibility

Cha 2008

WCST no of trials

0,30

0,306

0,53

0,000

04 Set-shifting/flexibility 05 Encoding

Nedelj kovic 2009

CANTAB SWM strategy score

0,69

0,097

05 Encoding

Cha 2008

RCFT copy

0,68

0,021

05 Encoding

Jang 2010

RCFT organization

0,28

0,167

0,44

0,004

05 Encoding 06 Verbal memory

Bouvard 1997

BEM 144 verbal total

0,50

0,206

06 Verbal memory

Nakao 2009

WMS-R LM II

0,82

0,050

06 Verbal memory

Hashimoto 2011

WMS-R LM II

0,47

0,277

06 Verbal memory

Omori 2007

WMS-R LM

0,45

0,104

06 Verbal memory

Cha 2008

HVLT delayed recall

-0,18

0,536

06 Verbal memory

Exner 2009

WMS-R LM II

0,70

0,171

06 Verbal memory

Ceschi 2003

CVLT total recall

0,09

0,784

0,32

0,017

06 Verbal memory 07 Nonverbal memory

Bouvard 1997

BEM 144 visuel total

0,68

0,089

07 Nonverbal memory

Nakao 2009

WMS-R visual reproduction

0,94

0,027

07 Nonverbal memory

Omori 2007

WMS-R visual reproduction

0,36

0,189

07 Nonverbal memory

Cha 2008

RCFT delayed recall

0,70

0,018

07 Nonverbal memory

Jang 2010

RCFT delayed recall

0,17

0,324

07 Nonverbal memory

Exner 2009

WMS-R visual reproduction

0,12

0,798

0,39

0,001

07 Nonverbal memory 08 Processing speed

Exner* 2009

TMT-A

0,54

0,278

08 Processing speed

Hashimoto 2011

WAIS-R digit symbol

0,42

0,548

08 Processing speed

Omori 2007

WAIS-R digit symbol

0,43

0,117

0,45

0,047

08 Processing speed 09 Spatial working memory

Nakao 2009

Spatial n-back

0,40

0,319

09 Spatial working memory

Nedelj kovic 2009

SWM errors

1,32

0,003

09 Spatial working memory

Exner 2009

WMS-R block span

0,60

0,229

09 Spatial working memory

Kraft* 2006

Spatial working memory

-1,07

0,027

0,32

0,506

09 Spatial working memory 99 Verbal working memory

Exner 2009

WMS-R digit span

0,50

0,327

99 Verbal working memory

Kraft* 2006

Verbal working memory

-1,08

0,027

-0,30

0,706

99 Verbal working memory

-2,00

-1,00

0,00

1,00

2,00

WASH < CHECK WASH > CHECK

Fig. 2. Forest plot of Hedges’ g for differences between washers and checkers, according to the random effects model. Effect sizes greater than zero indicate washers performing better than checkers. Studies marked with an asterisk * contain unpublished data.

neuropsychological performance than checkers. Large effect sizes were found for planning/problem solving (g = 1.77, p < .001) and response inhibition (g = 0.89, p < .001). A medium effect size emerged for set shifting/flexibility (g = 0.53, p < .001), while small effect sizes turned out for sustained attention (g = 0.49, p = .022), encoding strategies (g = 0.44, p = .004), verbal memory (g = 0.34, p = .027), nonverbal memory (g = 0.39, p = .001) and processing

speed (g = 0.45, p = .047). No significant differences between washers and checkers were found with regard to spatial working memory (g = 0.32, p = .506) and verbal working memory (g = −0.30, p = .706). Tests of heterogeneity indicated that the diverging test scores of the studies differed significantly in two cognitive subdomains, i.e. in spatial working memory and verbal working memory (Table 4). In

R. Leopold, M. Backenstrass / Journal of Anxiety Disorders 30 (2015) 48–58

55

Table 4 Mean effect sizes according to both the fixed effect model and the random effects model and measures of heterogeneity for the 10 cognitive subdomains. Cognitive subdomain

Number of studies

Hedges’ g (fixed) (random)

95% CI

Q

df (Q)

p (Q)

I2

Sustained attention

3

2

.880

0.00

2

0.00

1

.968

0.00

2.63

3

.453

0.00

Set shifting/flexibility

6

2.65

5

.754

0.00

Encoding strategies

3

1.66

2

.437

0.00

Verbal memory Nonverbal memory

7

5.93

6

.431

0.00

5.37

5

.372

6.97

Processing speed

3

0.04

2

.981

0.00

Spatial working memory Verbal working memory

4

0.07–0.91 0.07–0.91 1.03–2.52 1.03–2.52 0.47–1.31 0.47–1.31 0.24–0.82 0.24–0.82 0.14–0.75 0.14–0.75 0.06–0.58 0.26–0.58 0.15–0.60 0.15–0.63 0.01–0.90 0.01–0.90 −0.09–0.80 −0.63–1.27 −1.01–0.36 −1.84–1.25

0.26

Planning/problem solving Response inhibition

0.49 0.49 1.77 1.77 0.89 0.89 0.53 0.53 0.44 0.44 0.32 0.32 0.38 0.39 0.45 0.45 0.36 0.32 −0.32 −0.30

13.51

3

.004

77.80

4.44

1

.035

77.46

4

6

2

eight other subdomains, the variance explained by between-study heterogeneity, measured by the I2 -statistic, was between zero and 6.97% and the Q-tests failed to reach significance, thus, we retained the null hypothesis of between-study homogeneity. As effect sizes within subdomains were of sufficient homogeneity, contrary to our a priori assumption of heterogeneity, we integrated effect sizes using both the random effects model and the fixed effect model. Differences between the two integration models were minimal as shown by the combined effect sizes for each cognitive subdomain. Due to the small number of included studies, we assessed the risk of publication bias for the total effect across all 13 studies (g = 0.35 in the fixed effect model and g = 0.42 in the random effects model). Visual inspection of the funnel plot showed a moderate asymmetry. After adjusting by means of the trim-and-fill method, the overall effect decreased but still remained larger than zero (g = 0.16 for the fixed effect model and g = 0.17 for the random effects model). Orwin’s fail-safe N was 36. That is, based on 13 published studies, almost three times as many unpublished studies with effect sizes smaller than 0.10 would have been necessary to decrease the observed effect to a trivial size. Egger’s regression intercept revealed no significant publication bias (t = 1.06, df = 12, p = .156). Meta-regression analyses assessing the influence of methodological study quality, group differences in age, gender distribution, Table 3 Participants’ demographical and clinical characteristics of studies included in the meta-analysis. Participants‘ characteristic

Washers M (SD)

Checkers M (SD)

Number of studies

Age % female % on medication IQ Education (years) Duration of illness (years) Age of onset (years) Y-BOCS total BDI BDI-II STAI-S

33.91 (3.84) 58.6 (15.3) 70.1 (28.9) 108.41 (5.82) 13.68 (1.58) 9.60 (2.77)

34.66 (4.87) 56.6 (17.3) 70.8 (29.8) 107.17 (5.46) 13.83 (1.41) 9.90 (2.62)

13 11 12 6 8 8

19.98 (1.50) 24.85 (5.72) 14.01 (3.24) 22.30 (6.77) 51.16 (5.98)

20.52 (2.54) 24.10 (5.14) 13.71 (4.54) 19.99 (3.74) 50.97 (5.78)

4 12 7 3 5

Y-BOCS, Yale-Brown Obsessive Compulsive Scale; BDI, Beck Depression Inventory, BDI-II, Beck Depression Inventory-Revised, second edition; IQ, intelligence quotient; STAI-S, State-Trait Anxiety Inventory, subtest state anxiety.

intelligence/years of education, OCD severity, percentage of patients on psychotropic medication, illness duration, and state anxiety were not of statistical significance. As only two studies reported the participants’ age of OCD onset, we were unable to assess a possible influence of age differences at OCD onset on differences in neuropsychological performance. Differences in severity of comorbid depression did not significantly contribute to neuropsychological effect sizes in set shifting, encoding, verbal memory and nonverbal memory. Due to lack of a sufficient number of primary studies, it was not possible to assess the influence of comorbid depression on intergroup differences with regard to sustained attention, planning/problem solving, response inhibition, processing speed, and verbal working memory. However, differences in severity of comorbid depression significantly accounted for differences in spatial working memory impairment (beta = −1.42, intercept = 0.69, p < .001, number of studies = 3). A subgroup analysis of differences in the data entry format (means and standard deviations versus Pearson’s correlations) demonstrated smaller effect sizes in the correlational studies, yet this difference failed statistical significance for both the fixed effect model (Q = 0.29, df = 1, p = .589) and the random effects model (Q = 0.20, df = 1, p = .655). A random effects model sensitivity analysis revealed that the largest effect sizes remained unchanged or even increased once the correlational studies were excluded: planning/problem solving (g = 1.77, p < .001), response inhibition (g = 1.10, p < .001), set shifting/flexibility (g = .46, p = .006), encoding (g = .68, p = .005), nonverbal memory (g = .61, p < .001). Effect sizes in sustained attention (g = .52, p = .068) and spatial working memory (g = .23, p = .728) increased but failed statistical significance and the found effect in verbal memory (g = .28, p = .095) turned out to be smaller and non-significant. The missing statistical significance can be due to the smaller sample sizes. When correlational studies were excluded, only one study remained for the analysis of processing speed and verbal working memory so a meta-analysis could not be calculated for these two subdomains. 4. Discussion 4.1. Discussion of the results In this meta-analysis, we investigated whether OCD patients with predominant washing versus checking symptoms differed from each other with regard to their neuropsychological

56

R. Leopold, M. Backenstrass / Journal of Anxiety Disorders 30 (2015) 48–58

performance. Our results show that washers performed better than checkers in the majority of neuropsychological tasks and in the majority of cognitive subdomains. Our first finding that washers performed better than checkers is supported by 38 of 41 positive effect sizes in a variety of tests such as the Tower of Hanoi and the RCFT. Our second finding that washers performed better than checkers is demonstrated by positive effect sizes in 8 of 10 cognitive subdomains. More precisely, washers showed significantly higher performance in sustained attention, planning/problem solving, response inhibition, set shifting/cognitive flexibility, encoding, verbal memory, nonverbal memory and processing speed. It is remarkable that checkers did not outperform washers in any cognitive subdomain. Compared to washers, checkers showed the largest performance deficits in planning/problem solving, response inhibition and set shifting and, thus, overall in executive functioning. OCD patients in general were reported to show the largest deficits in set shifting and processing speed (Abramovitch et al., 2013) and nonverbal memory (Abramovitch et al., 2013 Shin et al., 2014) compared to healthy controls, although the authors appraised the effects as too small to be clinically meaningful. The two reviews of Woods et al. (2002) and Cuttler and Graf (2009) have shown that checkers suffered merely from small to moderate memory impairment when compared to non-checkers. Five studies in our meta-analysis have matched washers and checkers each with healthy controls and revealed conflicting results of washers’ neuropsychological performance. In three studies washers did not differ significantly from healthy controls in planning/problem solving (Nedeljkovic et al., 2009), set shifting (Cha et al., 2008; Nedeljkovic et al., 2009), verbal and nonverbal memory (Bouvard et al., 1997; Cha et al., 2008) and spatial working memory (Nedeljkovic et al., 2009). In contrast, two studies suggested that washers showed lower performance than healthy controls in verbal memory (Ceschi et al., 2003), verbal working memory and nonverbal working memory (Kraft, 2006). Due to lack of studies, we cannot draw a conclusion about possible response inhibition differences in washers and healthy controls. Taken the aforementioned findings together, we conclude that verbal and nonverbal memory performance differences between checkers and washers, indicated by small effect sizes, do not appear to be clinically meaningful. However, checkers experience significantly more impairment in planning/problem solving, response inhibition and set shifting than washers and healthy controls, the latter apparently being unimpaired in planning/problem solving and set shifting. When treating OCD washers and checkers as a homogenous group, it is likely that executive deficits of OCD checkers are underestimated, while the executive deficits of the washers are overestimated (Abramovitch et al., 2013). We assume that checkers’ greater impairment of executive functioning corresponds to particular neurobiological features. Previous neuroimaging findings revealed that healthy controls have largely intact neurobiological structures, whereas OCD patients taken in general show frontostriatal abnormalities (Menzies et al., 2008; Milad & Rauch, 2012), and, moreover, only OCD checkers exhibit additional abnormal amygdala activation (Via et al., 2014). In an fMRI study of OCD patients, Via et al. (2014) used an emotional facematching task and found that the checking dimension, in contrast to other OCD symptom dimensions, was related to increased amygdala activation. This finding indicates an increased involvement of neuronal fear circuitry. Da Victoria, Nascimento, & Fontenelle (2012) and Moritz, von Mühlenen, Randjbar, Fricke, and Jelinek (2009) reported that checkers showed an attentional bias to subjective threat-related information (e.g. an open door) and that this attentional bias is related to the reduced ability to disengage and shift attention from anxiogenic to task-relevant stimuli. Disengaging attention requires inhibitory abilities, whereas shifting of

attention requires cognitive flexibility, both are regarded as executive functions (Eysenck & Keane, 2005). This notion is align with the assumption that fear is associated with an attentional bias directed to anxiogenic stimuli. The main finding of our meta-analysis that OCD checkers are more impaired than washers in their executive functioning performance contributes to the given debate in the research area of interest. Impairment in applying organizational strategies to the encoding of verbal and nonverbal information has widely been suggested as an explanation of poorer memory performance in OCD patients (Deckersbach, Otto, Savage, Baer, & Jenike, 2000; Kashyap, Kumar, Kandavel, & Reddy, 2013; Kuelz et al., 2004 Penadés, Catalán, Andrés, Salamero, & Gastó, 2005; Sawamura, Nakashima, Inoue, & Kurita, 2005). In line with these findings are the found significant group differences between OCD washers and checkers in the use of organizational strategies of encoding of our meta-analysis. Furthermore, deficits in neuropsychological tests assessing response inhibition were proposed to be linked to deficits in suppressing obsessive thoughts and compulsive behaviors (Chamberlain, Blackwell, Fineberg, Robbins, & Sahakian, 2005). In line with these outcomes, we found that checkers experience a large impairment in response inhibition compared to washers. Moreover, the neuropsychological function of planning/problem solving seems of particular importance, because planning/problem solving was the cognitive subdomain in our meta-analysis in which checkers exhibited the largest impairment. To date, the differences between washers and checkers in planning/problem solving are not well investigated, reflected by the only two studies in this field. Therefore, more studies dedicated to the comparison of OCD washers and checkers in this neuropsychological subdomain are clearly needed. 4.2. Limitations The number of included studies as well as the sample size of each study was small; hence, the statistical power was low. Nevertheless, the presented meta-analysis demonstrated significant results. Some studies had to be excluded due to methodological reasons; e.g. some authors reported non-significant correlations between symptom dimensions and cognitive performance but did not provide any exact data to calculate effect sizes (a list of excluded studies is available on request from the first author). However, leading scientists were asked for unpublished studies (i.e. grey literature). Statistical analyses of publication bias were not significant. Thus, the neuropsychological differences between OCD washers and checkers seem to be robust. A larger number of studies might have enabled a more detailed moderator analyses. Due to lack of sufficient studies involving comorbidity, it was not possible to investigate the possible influence of comorbid depression on each cognitive domain. About two-thirds of patients were on psychotropic medication in the studies reporting medication status. However, percentage of medication was equal in washers and checkers and the meta-regression analysis did not find a significant effect of medication on participants’ performance. This finding corresponds well with previous studies demonstrating that medication status had little or no significant effect on neuropsychological test performance (Mataix-Cols, Alonso, Pifarré, Menchón, & Vallejo, 2002; Roh et al., 2005; Simpson et al., 2006). Besides, since all included studies were cross-sectional it remains unclear whether the observed cognitive deficits in checkers compared to washers represent state or trait deficits. The heterogeneous results across studies were found with regard to spatial working memory and verbal working memory. This may be due to the small number of primary studies with contradictory results and two outliers. In order to include all relevant studies, we treated washing and checking as distinct symptom

R. Leopold, M. Backenstrass / Journal of Anxiety Disorders 30 (2015) 48–58

subgroups in our meta-analysis and, therefore, dichotomized correlational data. However, an approach treating washing and checking as continuous symptom dimensions might be more accurate in describing the experiences of many OCD patients. The effect sizes of the comparative and correlational studies in our meta-analysis did not differ significantly from each other, even though dichotomization of data is generally linked to a loss of information and statistical power (Cohen, 1983). Excluding the correlational results led to even larger cognitive differences between washers and checkers. 4.3. Conclusions In summary, the present meta-analysis indicated that OCD checkers showed significantly more impairment than washers in sustained attention, planning/problem solving, response inhibition, set shifting/cognitive flexibility, encoding, verbal memory, nonverbal memory and processing speed, whereas differences between washers and checkers in verbal and nonverbal working memory failed to reach significance. The differences in executive functioning, especially in planning/problem solving and response inhibition were large enough to be interpreted as clinically meaningful, whereas the attention and memory differences between washers and checkers appeared too small to be clinically relevant. With particular regard to executive functions, our results provide further evidence for the validity of different symptom dimensions in OCD. In future studies, we suggest to compare OCD washers and checkers particularly with regard to their abilities in planning/problem solving as we found merely two studies that had investigated this subdomain. We also propose to investigate differences between OCD washers and checkers in their working memory performance, because of the inconsistent findings of our meta-analysis in this regard. Bearing in mind the continuous structure of the multiple overlapping OCD symptom dimensions, prospective studies should rather use a correlational design instead of a categorical approach to OCD symptom subgroups. Moreover, future research undertakings should consider severity of depression with respect to neuropsychological test performance. To find out whether checkers’ poorer cognitive performance represents state or trait deficits, longitudinal research designs are required to test for possible changes in neuropsychological impairment and treatment effects of neuropsychological training. Altogether, we recommend taking the larger executive impairment of OCD checkers compared to washers into account which is related to symptom-specific behavioral and neurobiological differences. This approach will be useful in future research of OCD, especially when combining neuropsychological testing with neuroimaging studies. Acknowledgments The authors are indebted to Diana Byrd Burt for providing her data coding form. We also thank all study authors who responded to our enquiry. Especially, we wish to thank Cornelia Exner, Joon Hwang Jang and Selim Tumkaya for providing unpublished information. We also thank Verena Oehmann for her support in data coding in order to determine interrater-reliability and are grateful to Alexandra Tietz for her language editing. References Abbruzzese, M., Ferri, S., & Scarone, S. (1995). Wisconsin Card Sorting Test performance in obsessive-compulsive disorder: no evidence for involvement of dorsolateral prefrontal cortex. Psychiatry Research, 58(1), 37–43. Abramovitch, A., Abramowitz, J. S., & Mittelman, A. (2013). The neuropsychology of adult obsessive-compulsive disorder: a meta-analysis. Clinical Psychology Review, 33(8), 1163–1171.

57

American Psychiatric Association, M. H. (2013). Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5). Arlington, VA: American Psychiatric Association. Bloch, Landeros-Weisenberger, A., Rosario, M. C., Pittenger, C., & Leckman, J. F. (2008). Meta-analysis of the symptom structure of obsessive-compulsive disorder. The American Journal of Psychiatry, 165(12), 1532–1542. Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. R. (2009a). Chapter 7. Converting among effect sizes. In: M. Borenstein, L. Hedges, J. Higgins, & H. R. Rothstein (Eds.), Introduction to meta-analysis (pp. 45–49). Chichester: Wiley. Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. R. (2009b). Chapter 24. Multiple outcomes or time-points within a study. In: M. Borenstein, L. Hedges, J. Higgins, & H. R. Rothstein (Eds.), Introduction to meta-analysis (pp. 225–238). Chichester: Wiley. *Bouvard, M., Dirson, S., & Cottraux, J. (1997). Etude de la mémoire de sujets obsessionnels compulsifs laveurs et vérificateurs et de sujets contrôles. European Review of Applied Psychology/Revue Européenne de Psychologie Appliquée, 47(3), 189–196. Burt, D. B., Zembar, M. J., & Niederehe, G. (1995). Depression and memory impairment: a meta-analysis of the association, its pattern, and specificity. Psychological Bulletin, 117(2), 285–305. *Ceschi, G., van der Linden, M., Dunker, D., Perroud, A., & Brédart, S. (2003). Further exploration memory bias in compulsive washers. Behaviour Research and Therapy, 41(6), 737–748. *Cha, K. R., Koo, M.-S., Kim, C.-H., Kim, J. W., Oh, W.-J., Suh, H. S., et al. (2008). Nonverbal memory dysfunction in obsessive-compulsive disorder patients with checking compulsions. Depression & Anxiety, 25(11), E115–E120. Chamberlain, S. R., Blackwell, A. D., Fineberg, N. A., Robbins, T. W., & Sahakian, B. J. (2005). The neuropsychology of obsessive compulsive disorder: the importance of failures in cognitive and behavioural inhibition as candidate endophenotypic markers. Neuroscience and Biobehavioral Reviews, 29(3), 399–419. Cohen, J. (1983). The cost of dichotomization. Applied Psychological Measurement, 7(3), 249–253. Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155–159. Cuttler, C., & Graf, P. (2009). Checking-in on the memory deficit and metamemory deficit theories of compulsive checking. Clinical Psychology Review, 29(5), 393–409. Da Victoria, M. S., Nascimento, A. L., & Fontenelle, L. F. (2012). Symptom-specific attentional bias to threatening stimuli in obsessive-compulsive disorder. Comprehensive Psychiatry, 53(6), 783–788. Deckersbach, T., Otto, M. W., Savage, C. R., Baer, L., & Jenike, M. A. (2000). The relationship between semantic organization and memory in obsessive-compulsive disorder. Psychotherapy and Psychosomatics, 69(2), 101–107. Deutsches Cochrane Zentrum (2013). www.cochrane.de. http://www.cochrane. de/de/so-finden-sie-uns-freiburg Retrieved 14.5.2013. *Dittrich, W. H., & Johansen, T. (2013). Cognitive deficits of executive functions and decision-making in obsessive-compulsive disorder. Scandinavian Journal of Psychology, 54(5), 393–400. Dunst, C. J., Hamby, D. W., & Trivette, C. M. (2004). Guidelines for calculating effect sizes for practice-based research syntheses. Centerscope, 3(1), 1–10. Duval, S., & Tweedie, R. (2000). Trim and fill: a simple funnel plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics, 56(2), 455–463. Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in metaanalysis detected by a simple, graphical test. British Medical Journal, 315(7109), 629–634. *Exner, C., Martin, V., & Rief, W. (2009). Self-focused ruminations and memory deficits in obsessive-compulsive disorder. Cognitive Therapy and Research, 33(2), 163–174. Eysenck, M. W., & Keane, M. T. (2005). Cognitive psychology: a student’s handbook (5th edition). Hove: Psychology Press. Foa, E. B., Huppert, J. D., Leiberg, S., Langner, R., Kichic, R., Hajcak, G., et al. (2002). The obsessive-compulsive inventory: development and validation of a short version. Psychological Assessment, 14(4), 485–495. Fontenelle, L. F., Mendlowicz, M. V., Marques, C., & Versiani, M. (2004). Trans-cultural aspects of obsessive-compulsive disorder: a description of a Brazilian sample and a systematic review of international clinical studies. Journal of Psychiatric Research, 38(4), 403–411. Goodman, W. K., Price, L. H., Rasmussen, S. A., & Mazure, C. (1989). The Yale-Brown obsessive compulsive scale: I. Development, use, and reliability. Archives of General Psychiatry, 46(11), 1006–1011. *Hashimoto, N., Nakaaki, S., Omori, I. M., Fujioi, J., Noguchi, Y., & Murata, Y. (2011). Distinct neuropsychological profiles of three major symptom dimensions in obsessive compulsive disorder. Psychiatry Research, 187(1–2), 166–173. Hasler, G., LaSalle-Ricci, V. H., Ronquillo, J. G., Crawley, S. A., Cochran, L. W., Kazuba, D., et al. (2005). Obsessive-compulsive disorder symptom dimensions show specific relationships to psychiatric comorbidity. Psychiatry Research, 135(2), 121–132. Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando: Academic Press. Horesh, N., Dolberg, O. T., Kirschenbaum-Aviner, N., & Kotler, M. (1997). Personality differences between obsessive-compulsive disorder subtypes: washers versus checkers. Psychiatry Research, 71(3), 197–200. *Jang, J. H., Kim, S. H., Ha, T. H., Shin, N. Y., Do-Hyung, A., Jung-Seok, C., et al. (2010). Nonverbal memory and organizational dysfunctions are related with distinct symptom dimensions in obsessive-compulsive disorder. Psychiatry Research, 180(2–3), 93–98.

58

R. Leopold, M. Backenstrass / Journal of Anxiety Disorders 30 (2015) 48–58

Kashyap, H., Kumar, J. K., Kandavel, T., & Reddy, Y. C. J. (2013). Neuropsychological functioning in obsessive-compulsive disorder: are executive functions the key deficit? Comprehensive Psychiatry, 54(5), 533–540. Katerberg, H., Delucchi, K. L., Stewart, S. E., Lochner, C., Denys, D. A., Stack, D. E., et al. (2010). Symptom dimensions in OCD: item-level factor analysis and heritability estimates. Behavior Genetics, 40(4), 505–517. Khanna, S., & Mukherjee, D. (1992). Checkers and washers: valid subtypes of obsessive compulsive disorder. Psychopathology, 25(5), 283–288. Kim, S. J., Lee, H. S., & Kim, C.-H. (2005). Obsessive-compulsive disorder, factoranalyzed symptom dimensions and serotonin transporter polymorphism. Neuropsychobiology, 52(4), 176–182. *Kraft, S. (2006). Arbeitsgedächtnis bei Zwangsstörungen. Homburg/Saar: Universität des Saarlandes. Unpublished doctoral thesis. Kuelz, A. K., Hohagen, F., & Voderholzer, U. (2004). Neuropsychological performance in obsessive-compulsive disorder: a critical review. Biological Psychology, 65(3), 185–236. Leckman, J. F., Grice, D. E., Boardman, J., & Zhang, H. (1997). Symptoms of obsessivecompulsive disorder. The American Journal of Psychiatry, 154(7), 911–917. Leckman, J. F., Mataix-Cols, D., & do Rosario-Campos, M. C. (2005). Reply to Taylor: combined dimensional and categorical perspectives as an integrative approach to OCD. In: J. S. Abramowitz, & A. C. Houts (Eds.), Concepts and controversies in obsessive-compulsive disorder (pp. 43–47). New York, US: Springer Science+Business Media. Mataix-Cols, D., Alonso, P., Pifarré, J., Menchón, J. M., & Vallejo, J. (2002). Neuropsychological performance in medicated vs: unmedicated patients with obsessive compulsive disorder. Psychiatry Research, 109(3), 255–264. Mataix-Cols, D., do Rosario-Campos, M. C., & Leckman, J. F. (2005). A multidimensional model of obsessive-compulsive disorder. The American Journal of Psychiatry, 162(2), 228–238. Mataix-Cols, D., Marks, I. M., Greist, J. H., Kobak, K. A., & Baer, L. (2002). Obsessive-compulsive symptom dimensions as predictors of compliance with and response to behaviour therapy: results from a controlled trial. Psychotherapy and Psychosomatics, 71(5), 255–262. Mataix-Cols, D., Rauch, S. L., Baer, L., Eisen, J. L., Shera, D. M., & Goodman, W. K. (2002). Symptom stability in adult obsessive-compulsive disorder: data from a naturalistic two-year follow-up study. The American Journal of Psychiatry, 159(2), 263–268. Mataix-Cols, D., Rauch, S. L., Manzo, P. A., Jenike, M. A., & Baer, L. (1999). Use of factor-analyzed symptom dimensions to predict outcome with serotonin reuptake inhibitors and placebo in the treatment of obsessive-compulsive disorder. The American Journal of Psychiatry, 156(9), 1409–1416. Mataix-Cols, D., Wooderson, S., Lawrence, N., Brammer, M. J., Speckens, A., & Phillips, M. L. (2004). Distinct neural correlates of washing, checking, and hoarding symptom dimensions in obsessive-compulsive disorder. Archives of General Psychiatry, 61(6), 564–576. Matsunaga, H., Maebayashi, K., Hayashida, K., Okino, K., Matsui, T., Iketani, T., et al. (2008). Symptom structure in Japanese patients with obsessive-compulsive disorder. The American Journal of Psychiatry, 165(2), 251–253. Menzies, L., Chamberlain, S. R., Laird, A. R., Thelen, S. M., Sahakian, B. J., & Bullmore, E. T. (2008). Integrating evidence from neuroimaging and neuropsychological studies of obsessive-compulsive disorder: the orbitofronto-striatal model revisited. Neuroscience and Biobehavioral Reviews, 32(3), 525–549. Milad, M. R., & Rauch, S. L. (2012). Obsessive-compulsive disorder: beyond segregated cortico-striatal pathways. Trends in Cognitive Sciences, 16(1), 43–51. Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. British Medical Journal, 339, 332–336. Moritz, S., Fischer, B.-K., Hottenrott, B., Kellner, M., Fricke, S., Randjbar, S., et al. (2008). Words may not be enough! No increased emotional Stroop effect in obsessive-compulsive disorder. Behaviour Research and Therapy, 46(9), 1101–1104. Moritz, S., Kloss, M., von Eckstaedt, F. V., & Jelinek, L. (2009). Comparable performance of patients with obsessive-compulsive disorder (OCD) and healthy controls for verbal and nonverbal memory accuracy and confidence: time to forget the forgetfulness hypothesis of OCD? Psychiatry Research, 166(2–3), 247–253. Moritz, S., von Mühlenen, A., Randjbar, S., Fricke, S., & Jelinek, L. (2009). Evidence for an attentional bias for washing- and checking-relevant stimuli in obsessive-compulsive disorder. Journal of the International Neuropsychological Society, 15(3), 365–371.

Murayama, K., Nakao, T., Sanematsu, H., Okada, K., Yoshiura, T., Tomita, M., et al. (2013). Differential neural network of checking versus washing symptoms in obsessive-compulsive disorder. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 40, 160–166. *Nakao, T., Nakagawa, A., Nakatani, E., Nabeyama, M., Sanematsu, H., Yoshiura, T., et al. (2009). Working memory dysfunction in obsessive compulsive disorder: A neuropsychological and functional MRI study and functional MRI study. Journal of Psychiatric Research, 43(8), 784–791. *Nedeljkovic, M., Kyrios, M., Moulding, R., Doron, G., Wainwright, K., Pantelis, C., et al. (2009). Differences in neuropsychological performance between subtypes of obsessive-compulsive disorder. Australian and New Zealand Journal of Psychiatry, 43(3), 216–226. *Omori, I. M., Murata, Y., Yamanishi, T., Nakaaki, S., Akechi, T., Mikuni, M., et al. (2007). The differential impact of executive attention dysfunction on episodic memory in obsessive-compulsive disorder patients with checking symptoms vs. those with washing symptoms. Journal of Psychiatric Research, 41(9), 776–784. Orwin, R. G. (1983). A fail-safe N for effect size in meta-analysis. Journal of Educational Statistics, 8(2), 157–159. Öst, L.-G. (2008). Efficacy of the third wave of behavioral therapies: a systematic review and meta-analysis. Behaviour Research and Therapy, 46(3), 296–321. Penadés, R., Catalán, R., Andrés, S., Salamero, M., & Gastó, C. (2005). Executive function and nonverbal memory in obsessive-compulsive disorder. Psychiatry Research, 133(1), 81–90. *Penadés, R., Catalán, R., Rubia, K., Andrés, S., Salamero, M., & Gastó, C. (2007). Impaired response inhibition in obsessive compulsive disorder. European Psychiatry, 22(6), 404–410. Pinto, A., Eisen, J. L., Mancebo, M., Greenberg, B. D., Stout, R. L., & Rasmussen, S. (2007). Taboo thoughts and doubt/checking: a refinement of the factor structure for obsessive-compulsive disorder symptoms. Psychiatry Research, 151(3), 255–258. Rasmussen, S. A., & Eisen, J. L. (1989). Clinical features and phenomenology of obsessive compulsive disorder. Psychiatric Annals, 19(2), 67–73. Roh, K. S., Shin, M. S., Kim, M.-S., Ha, T.-H., Shin, Y.-W., & Lee, K. J. (2005). Persistent cognitive dysfunction in patients with obsessive-compulsive disorder: a naturalistic study. Psychiatry and Clinical Neurosciences, 59(5), 539–545. Rosenthal, R., & DiMatteo, M. R. (2001). Meta analysis: recent developments in quantitative methods for literature reviews. Annual Review of Psychology, 52, 59–82. Sawamura, K., Nakashima, Y., Inoue, M., & Kurita, H. (2005). Short-term verbal memory deficits in patients with obsessive-compulsive disorder. Psychiatry and Clinical Neurosciences, 59(5), 527–532. Shin, N. Y., Lee, T. Y., Kim, E., & Kwon, J. S. (2014). Cognitive functioning in obsessivecompulsive disorder: a meta-analysis. Psychological Medicine, 44(6), 1121–1130. Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: uses in assessing rater reliability. Psychological Bulletin, 86(2), 420–428. Silver, N. C., & Dunlap, W. P. (1987). Averaging correlation coefficients: should Fisher’s z transformation be used? Journal of Applied Psychology, 72(1), 146–148. Simpson, H. B., Rosen, W., Huppert, J. D., Lin, S.-H., Foa, E. B., & Liebowitz, M. R. (2006). Are there reliable neuropsychological deficits in obsessive-compulsive disorder? Journal of Psychiatric Research, 40(3), 247–257. *Tumkaya, S., Karadag, F., Mueller, S. T., Ugurlu, T. T., Oguzhanoglu, N. K., Ozdel, O., et al. (2013). Situation awareness in obsessive-compulsive disorder. Psychiatry Research, 209(3), 579–588. van den Heuvel, O. A., Remijnse, P. L., Mataix-Cols, D., Vrenken, H., Groenewegen, H. J., & Uylings, H. B. M. (2009). The major symptom dimensions of obsessivecompulsive disorder are mediated by partially distinct neural systems. Brain: A Journal of Neurology, 132(4), 853–868. van Grootheest, D. S., Boomsma, D. I., Hettema, J. M., & Kendler, K. S. (2008). Heritability of obsessive-compulsive symptom dimensions. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 147B(4), 473–478. Via, E., Cardoner, N., Pujol, J., Alonso, P., López-Sola, M., Real, E., et al. (2014). Amygdala activation and symptom dimensions in obsessive-compulsive disorder. The British Journal of Psychiatry, 204(1), 61–68. Williams, M. T., Mugno, B., Franklin, M., & Faber, S. (2013). Symptom dimensions in obsessive-compulsive disorder: phenomenology and treatment outcomes with exposure and ritual prevention. Psychopathology, 46(6), 365–376. Woods, C. M., Vevea, J. L., Chambless, D. L., & Bayen, U. J. (2002). Are compulsive checkers impaired in memory? A meta-analytic review. Clinical Psychology: Science and Practice, 9(4), 353–366. References marked with an asterisk indicate studies included in the metaanalysis.