You are on page 1of 15

+ model

ARTICLE IN PRESS
European Neuropsychopharmacology (2006) xx, xxx — xxx

www.elsevier.com/locate/euroneuro

REVIEW

Antipsychotic augmentation of serotonergic antidepressants in treatment-resistant obsessive—compulsive disorder: A meta-analysis of the randomized controlled trials
Petros Skapinakis a,b,*, Tzeni Papatheodorou a, Venetsanos Mavreas a
a b

Department of Psychiatry, University of Ioannina, School of Medicine, Ioannina 45110, Greece Academic Unit of Psychiatry, University of Bristol, UK

Received 30 March 2006; received in revised form 29 June 2006; accepted 4 July 2006

KEYWORDS
Obsessive—compulsive disorder/drug therapy; Antipsychotic agents/ therapeutic use; Antidepressive agents/ therapeutic use; Serotonin uptake inhibitors/therapeutic use; Controlled clinical trials; Meta-analysis

Abstract This study aimed to determine the effectiveness of antipsychotic augmentation of serotonergic antidepressants in the management of treatment-resistant obsessive compulsive disorder by carrying out a meta-analysis of all randomized controlled trials. Studies selected through a literature search conducted in March 2006. Ten trials comparing antipsychotic drugs versus placebo met inclusion criteria (haloperidol [n = 1], risperidone [n = 3], olanzapine [n = 2], quetiapine [n = 4]). A total of 157 patients were randomized to study drug and 148 were randomized to placebo. Response occurred more often among patients randomized to antipsychotic drugs. The weighted combined response rate ratio by random effects metaanalysis was 3.31 (95% CI 1.40—7.84). Significant between studies heterogeneity was partly explained by the definition of refractoriness, the type and dose of the drug used and the inclusion or exclusion of patients with tic disorders. The study supports the use of antipsychotic drugs as an augmentation strategy but more and larger trials are needed.

D 2006 Elsevier B.V. and ECNP. All rights reserved.

1. Introduction
Obsessive—compulsive disorder (OCD) is a relatively common condition in the general population. Epidemiological data from the UK have shown a prevalence of 1.2% in the general population (Jenkins et al., 1997) and similar figures have been reported from the US and elsewhere (Horwath

* Corresponding author. Department of Psychiatry, University of Ioannina, School of Medicine, Ioannina 45110, Greece. Tel.: +30 26510 97748; fax: +30 26510 97049. E-mail address: pskapin@cc.uoi.gr (P. Skapinakis).

0924-977X/$ - see front matter D 2006 Elsevier B.V. and ECNP. All rights reserved. doi:10.1016/j.euroneuro.2006.07.002

NEUPSY-09912; No of Pages 15

Data extraction. Inclusion and exclusion criteria We selected studies that met all the following criteria: a) study design: randomized controlled trial.. 2. Kaplan and Hoolander. this source by also searching EMBASE (1980—2005).. sex. social phobia and agoraphobia. 1990. Experimental procedures 2.stakes. 2002). we recorded the number of respondents at the . McDougle et al.1.. Regarding the former. statistical analysis). These reviews. Search strategy We searched PubMed for English and non-English medical literature published from 1966 to January 2006. Since the duration of the studies might differ. are commonly diagnosed in patients with OCD (Bartz and Hollander. A range of augmentation strategies have been proposed. type and doses of antidepressant drugs). 2002... in combination with SSRIs. in the present paper we aimed to carry out a meta-analysis of all randomized controlled trials that examined the effectiveness of antipsychotic drugs as augmenting agents in patients with OCD non-responsive to monotherapy with antidepressants. the disorder is under-recognized and under-treated (Stein. 2005). number of patients responding in active and control arms. Despite its prevalence. outcomes and results (number of patients entering and completing the study. 2003). P.. diagnosis. An adequate response is generally achieved by 40% to 60% of patients taking SSRIs and this leaves a lot of patients with partial or poor response (Pallanti et al. 2004). Their findings support the use of these drugs. In addition. 1989) or general assessment of clinical response with the Clinical Global Impressions-Improvement Scale (CGI-I) (Guy. 2. mean change and standard deviation of change from baseline scores. Fineberg and Gale. control intervention: placebo. 1997). definition of treatment resistance.. 1999. 2006). We used the following search string (string 1) in PubMed: (bObsessive—Compulsive DisorderQ[MeSH]) AND ((bClinical TrialsQ[MeSH]) OR (bSingle-Blind MethodQ[MeSH]) OR (bDouble-Blind MethodQ[MeSH]) OR (bPlacebosQ[MeSH])) AND (bAntipsychotic AgentsQ[Pharmacological Action]) Because the use of the MeSH terms does not return records that have been supplied by the publishers or are in the process of indexing. 2005. b) participants: adult patients with obsessive—compulsive disorc) d) e) der non-responsive to previous antidepressant treatment. We manually checked the reference lists of prior reviews. 2002). 2006. Early reports for the usefulness of typical antipsychotic drugs in treatment-resistant OCD (Delgado et al. 1997). did not attempt to combine the results of the available randomized controlled trials to give an estimate of the overall effect of antipsychotic drugs in treatment-resistant OCD. or a combination of both.. We supplemented Our primary outcome measure was the number of patients who were classified as responders based on pre-defined relative score improvements on the Y-BOCS scale. 2005). 1988). three of them specifically reviewing the use of antipsychotic drugs in treatment-resistant OCD (Fineberg et al. systematic reviews and trials. Skapinakis et al. mainly selective serotonin reuptake inhibitors (SSRIs) (Fineberg and Gale. Pharmacological treatment of OCD is based on the use of serotonergic antidepressants. the use of antipsychotic drugs has been suggested for at least two reasons: a) the degree of insight into obsessions differs in OCD. 2005. b) participants: patients with comorbid schizophrenia or other c) psychotic disorder. baseline and end of study mean Y-BOCS scores and standard deviations. methods (design. number and reasons for dropouts and withdrawals. Kim et al.2. outcome measurement: assessment of symptoms with the YaleBrown Obsessive Compulsive Scale (Y-BOCS) (Goodman et al. during the past two years new randomized controlled trials on this issue were published that were only included in a very recent systematic review (Fineberg et al.fi/). 1976). Patients may be reluctant to report their symptoms to doctors and depression. in OCD patients with poor or partial response to SSRIs. the Cochrane Controlled Trials Register (2006. Steketee. outcome measurement and assessment of methodological quality Data extracted included information on: a) b) c) d) patients (age. with or without panic disorder. both pharmacological and non-pharmacological. 2005. definition of treatment response. we also used the following sensitive string (string 2) to identify new studies not yet officially indexed: (Obsess* OR Compuls*) AND (Antipsychot* OR Haloperidol OR Risperidone OR Olanzapine OR Quetiapine OR Pimozide OR Chlorpromazine OR Sulpiride OR Amisulpride OR Clozapine OR Aripiprazole OR Ziprasidone) AND ((publisher [sb]) OR (in process [sb])). Some patients do not regard their symptoms as senseless or unreasonable and their obsessions resemble more overvalued or delusional ideas than anxiety worries (Matsunaga et al.ARTICLE IN PRESS 2 and Weissman. 1988. 2000. At least five systematic reviews of the evidence behind the pharmacotherapy of OCD have been published in the past three years. interventions (type and doses of antipsychotic drugs used to augment the antidepressant). Karno et al. the global improvement subscale of the CGI. duration of study. 2.. experimental intervention: antipsychotic augmentation of antidepressant monotherapy. Skoog and Skoog. Given that most of the studies were small and probably underpowered this is an important limitation that has not been addressed. however. For these reasons. Keuneman et al. 2006). duration of study: less than 4 weeks. for these patients (Fineberg and Gale.. comorbid conditions.3. The course is usually chronic and the disorder may lead to considerable disability without treatment (Karno et al. Steketee.. 2003. We excluded studies from our analysis if they met at least one of the following criteria: a) study design: crossover study.. Sareen et al. b) an involvement of the dopamine system has been postulated in the pathophysiology of OCD (Denys et al. issue 1) and the PsiTri database (http://psitri. 1990) were followed by more rigorous randomized controlled trials of both typical and newer antipsychotic drugs..

Quality assessment was made by two investigators (PS and VM).. Kapur et al. mainly the direct dopamine-D2 blockade (Ramasubbu et al. Risperidone 2 mg/day (Kapur et al.... 1996) to assess study quality but given that most studies were small.ARTICLE IN PRESS Antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD time-point defined by the original investigators for the measurement of their primary outcome.b). We combined results on the rate ratio using fixed or random effect models. 3 e) may be more important for the antipsychotic effect. 2002). Sensitivity analysis aimed to investigate the influence of study quality as measured by the modified Jadad scale (2 or less was recorded as low quality). 8 weeks or more). 1998). Using the Review Manager software we calculated the response rate ratios (ratios of the number of patients who responded divided by the number of patients initially randomized to the respective group) and their 95% confidence intervals.. 1999. we also measured the total number of dropouts in each arm to assess the safety of these drugs in patients with treatmentresistant OCD. 2003) by one investigator (PS). according to whether the authors had included or excluded patients with comorbid tic disorders. 2000a.. a random effects model was used. 1994). Olanzapine 10 mg/day (Kapur et al.b. 3. 1).. 1999) and Quetiapine 400 mg/day (Kapur et al. For this reason we stratified studies into two categories. we additionally deducted one point if losses to follow-up were greater than 20%. In the presence of significant heterogeneity. The number of respondents in each study was recorded according to the intention to treat principle. . Rate ratios greater than 1 indicate a better response for the antipsychotic medication group. Sareen et al.. Studies and patients characteristics Ten trials were included in the meta-analysis and their characteristics are summarized in Table 1.1. No other trials were identified from other sources including the reference lists of previous systematic reviews. Subgroup analyses Causes of heterogeneity were examined by the following predefined subgroup analyses: 3.to 8-week durations) a) the type of antipsychotic drug. because we think that results are more readily interpretable from a clinical perspective. Previous PET studies have shown that the dose needed to achieve this D2 occupancy is for Haloperidol 2—4 mg/day (Kapur et al.. the dose of antipsychotic medication used: antipsychotic drugs are given in OCD for their effect on the dopaminergic system. It should be noted that the figure for Quetiapine is estimated at two hours after administration and not at twelve hours. Although. Based on these measurements we classified studies into two categories of low or standard/high dose if the mean dose reported was below or above the cutoff to achieve adequate D2 occupancy respectively... the focus of this review was the effectiveness of antipsychotic drugs. b) the definition of refractoriness: studies were grouped into two c) categories depending on whether they had used the 25% or the 35% reduction in the Y-BOCS criterion for refractoriness. 2004). 3 risperidone studies (6. 1994). To investigate the degree of between-trial heterogeneity. We used improvement in OCD symptoms measured as a dichotomous outcome (response vs. measured as a continuous outcome.2 software (The Cochrane Collaboration. 2. Nordstrom et al. Search flow Our search strategy in Pubmed yielded 67 abstracts (49 abstracts from string 1 and 18 abstracts from string 2). The data were extracted onto hard copy data sheets. Remington et al.4. from which 10 (9 and 1 respectively) were retained after consideration of the inclusion and exclusion criteria (Fig. nonresponse) and not reduction in the severity of symptoms. 2000. 1993).4. 2. Statistical analysis Data from the data sheets were entered into the Review Manager 4. the inclusion or exclusion of patients with comorbid tic disorders: previous literature has suggested that antipsychotic drugs may be more effective in treatment resistant OCD in the presence of comorbid tic disorders (McDougle et al. We used the Jadad scale (Jadad et al. 1996). 1998.. the duration of the trial: studies were grouped into two categories according to the duration of the randomized trial (less than 8 weeks. A total of 11 citations (10 out of 49 and 1 out of 18) were retrieved as likely placebo-controlled trials. The study excluded was that of Li et al. the chi squared test was performed and I squared was calculated (Higgins and Thompson.. Data were extracted by one investigator (TP) and checked by another investigator (PS). These trials included 1 haloperidol study (4-week duration) (McDougle et al.1.2. It has been suggested that these drugs can exert their antipsychotic effect only if they have blocked more than 60% to 65% of the dopamine D2 receptors (Kapur et al. (2005) which was a crossover trial. 2000a. but the authors have suggested that for this drug the transient effect d) Figure 1 Trials identification and selection. Studies were then classified into two categories as either of low (a score of 2 or less on the modified Jadad scale) or acceptable quality (a score of 3 or more). Results 3.

(2004)/USA Shapira et al.5 Olanzapine 11.9 (4. 8 imipramine. (2005)/Italy Bystritsky et al.8 (10.0) P: 24.7 (120.6 (3.5.3. Y-BOCS: Yale-Brown Obsessive Compulsive Scale. (2004)/USA Atmaca et al.8 (9.3 (2.5) P: 28.9 (11. SD: standard deviation.6) P: 27.7) P: 29.5) 2 6 2 Olanzapine 6.8 (12.8 (10. 7 venlafaxine.0) 2.8) 2 and 3 P: 26.3) Response ratef A: 65% P: 0% A: 50% P: 0% A: 40% Study/country McDougle et al. 1 b35% improvement in Y-BOCS.8 (7.1) (4. 2 b25% improvement in Y-BOCS.5.3 (9.4 Table 1 Characteristics of ten trials of antipsychotic augmentation in treatment-resistant obsessive—compulsive disorder included in the meta-analysis N (% male) Duration (weeks) 4 6 Antidepressanta Antipsychotic Mean age (SDb) (A= antipsychotic P = placebo) 35 (12) A: 39.b)/ Netherlands Carey et al. 5 paroxetine. P.3.1) 300 1 and 4 1 and 3 6 1.5 (10.4 (5.4 (5. 2 fluoxetine.7 (8.2 (0. (1994)/USA McDougle et al.1) Mean daily dose mg (SDb) 6.2.6.0) A: 27.2 (6.3 1.6 Quetiapine A: 37.7) A: 29.2.4. 3 clomipramine.6.8) A: 44.1) 91.3) A: 26.2 (15.6) P: 31.6 Quetiapine 168.1 (4.5 (4.4) P: 6% A: 40% P: 47.7) 215 (124) 12 weeks (mostly at maximum dose) 2 A: 24.4) P: 43.8) A: 30.7) P: 26.3) 1 fluvoxamine.2.7) A: 36 (14) P: 34 (12) A: 33.1) 2 8 8 1. Definition of response was identical to the definition of refractoriness.4.8)e (4.8) P: 25.3.8) A: 38 (13.2) P: 34.1 23. Skapinakis et al.4) P: 37.2) A: 20. 4 sertraline.4% 0% 31% 6 2. (2005)/ South Africa and Canada Fineberg and Gale (2005)/ UK a b c d e f 34 (76%) 36 (58%) 16 (56%) 20 (53%) 26 (50%) 44 (41%) 27 (52%) 40 (25%) 41 (46%) 21 (57%) 1 1.8 Quetiapine Quetiapine A: 28.2.3.9 (10.2 (4.4 (6.7 Risperidone 2.1 24.1 (41.6 (8.3 (4.4) P: 27.3) A: 36. (2002)/Turkey Denys et al. (2004a.2 (3.2 (3.1) 36. 3 CGI-I b2.9)e P: A: P: A: 19.4.5.5 Haloperidol Risperidone 8 1.2 (4.4 (4.9) (4. (2000)/USA Hollander et al.8 28.2 (5.5.4 (11.4) P: 31.6) P: 24.1 (8.9 (4. 6 citalopram.7) 12 weeks (8 at maximum dose) 12 weeks (most on maximum dose) 12 weeks (mostly at maximum dose) 12 weeks (most at maximum dose) 8 weeks (at moderate doses) 12 weeks 16 weeks (2 SRIs 8 weeks each) 12 weeks (at least 6 at maximum dose) Duration of pre-randomization antidepressanttreatment Definition of refractorinessc 1 and 3 and 4 1 and 3 and 4 Mean baseline Y-BOCSd A: 25. 4 consensus of the authors.9) 2 and 3 ARTICLE IN PRESS 6 1 Risperidone 0.2.1 (2.2 (0.6% A: 27% P: 10% 16 4.5 (17.3) P: 38.5 1 and 3 P: 0% A: 50% P: 20% A: 46% P: 0% A: 41% P: A: P: A: 41% 71.7.4 (4. standard deviation of mean change from baseline reported.4.8) A: 24.3. . Numbers extracted approximately from a graph. (2003)/USA Erzegovesi et al.

The combined rate ratios for the other drugs were not significantly different from unity. 3. The combined response rate ratio for the studies using higher doses was 12.. Fineberg et al.82 (0.4.5.ARTICLE IN PRESS Antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD (Erzegovesi et al. 5 studies were classified into the 35% reduction group (Atmaca et al.84) using a random effects model (Fig.37]).. 2000. 2003. 4 studies were classified into the standard or high dose group that could induce higher than 60% to 65% D2 occupancy: one haloperidol study (McDougle et al...45. Since this result was in the opposite direction of what we had hypothesized. 2004). 2005. 1994.31 (95% confidence interval 1. Definition of refractoriness and definition of response was not uniform across studies. 2005)..002. Using the percentage reduction in the Y-BOCS criterion. Hollander et al. Overall. Definition of refractoriness/response Fig. There is only one study with Haloperidol so conclusions cannot be made for this drug. 2005) and one olanzapine study (Shapira et al. Although the number of studies is limited within each class.....b.. 2004a.. Omitting this study from the analysis slightly reduced the combined response rate ratio (2. The combined rate ratio for the lower dose studies was 1. 2005) and two quetiapine studies (Atmaca et al.. 2005)..5.85).. Results are shown in Fig. Denys et al. Shapira et al. The duration of antidepressant treatment before randomization was generally 12 weeks but there was some variation on the duration of use of the maximum tolerated dose. 1994). 2005. b) standard/high dose minus tics: one olanzapine (Bystritsky et al. 2005. 6. .85—3... Five studies included patients with comorbid tic disorders (Carey et al. Stratification by tic disorders Contrary to what would have been expected based on the previous literature inclusion of tic disorders was associated with a smaller and non-significant response rate ratio (Fig. fluoxetine and clomipramine (Table 1).to 16-week durations) (Atmaca et al.20— 50.78 [1. Erzegovesi et al.21—6. with small between studies heterogeneity.. However. 2 also shows the analysis of different classes of antipsychotic drugs.. 2005). Bystritsky et al. The studies used a variety of antidepressant drugs but most often fluvoxamine. Shapira et al.3. it is worth noting that the results for risperidone are more consistent and there is no statistical heterogeneity.2. 2005. Carey et al. Fineberg et al.5. Sensitivity analysis Only one study scored 2 or less on the modified Jadad scale (Atmaca et al..75 (95% CI 3..4. one risperidone study (Erzegovesi et al.. 2002. 2005. There were 73 responses in the antipsychotic group (response rate 46%) and 24 responses in the placebo group (response rate 16%). and d) low dose plus tics: one olanzapine (Shapira et al. 40 to risperidone.91). 2004). Low versus standard/high dose Fig. 2003.. 2000) and the remaining 5 into the 25% reduction group. c) low dose minus tics: one risperidone (Erzegovesi et al. 2).. Fineberg et al. in all studies there was consistency between definition of refractoriness and response. 5 3. 2004. (2002).. 2005. The mean dose of the antipsychotic drugs was generally lower than the dose normally used for treating schizophrenia. The mean age of the participants was below 40 years old for the majority of the studies.. and some studies applied stricter criteria for refractoriness or response. Type of antipsychotic drug used Fig.b).. 5)... 3. Sub-group analysis 3. p = 0. two risperidone studies (Hollander et al. 2005. McDougle et al.5. higher doses were associated with a higher response rate but this was more pronounced in studies that included a) 3. This stratification significantly reduced the heterogeneity between studies.. Generally. 2004. Meta-analysis outcome For the 10 trials included in the meta-analysis response rate ratios were heterogeneous (X 2 = 26. 2002. the only outlier was the study of Atmaca et al.. 35 to olanzapine and 65 to quetiapine) and 148 patients were randomized to placebo. 2004a. 2004a. 4 shows the stratification of studies according to the dosing scheme used. 1994) and one risperidone study (McDougle et al.. Fineberg et al. 2004) and two quetiapine studies (Carey et al. 2004). 3 shows the results for the studies stratified according to the definition used for refractoriness.. McDougle et al. Only three studies had more female than male patients.. Carey et al. 2002..b. 157 patients were randomized to antipsychotic drug (17 were randomized to haloperidol.5. 1994. 2004). I 2 = 66%). According to our definitions. 2 olanzapine studies (6-week durations) (Bystritsky et al.3. 2004) and one risperidone study (Hollander et al. Denys et al. In the low dose group... This study also differed in that the patients randomized had less severe OCD symptoms at the point of randomization (YBOCS score was 19 or 20 for placebo or antipsychotic respectively). This stratification reduced considerably the heterogeneity. The combined response rate ratio for the 10 studies was 3. 2002). McDougle et al. 3. For this reason we classified the studies into four categories: standard/high dose plus tics: one haloperidol study (McDougle et al. McDougle et al... Six studies were classified into the low dose category: all four quetiapine studies (Atmaca et al.. 2000).1. Studies that used the 25% reduction in Y-BOCS scores criterion were not significantly different from placebo in the combined analysis. McDougle et al. 2004).40—7. and 4 quetiapine studies (6. we also examined the possibility of an interaction between the dose used and the inclusion or exclusion of patients with tic disorders.. McDougle et al. 3. 2000) and one olanzapine study (Bystritsky et al.. Denys et al. 2003). 2002. Only one study had a duration of 8 weeks and used moderate only doses of antidepressants in the prerandomization phase (Shapira et al. 2000).

6 ARTICLE IN PRESS P. . by type of drug and overall compared with Placebo. Skapinakis et al. Figure 2 Response rate ratios of antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD.

Antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD Figure 3 Response rate ratios of antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD. compared with Placebo. by definition of refractoriness (see Experimental procedures). ARTICLE IN PRESS 7 .

by dose of drug (see Experimental procedures). Skapinakis et al.8 ARTICLE IN PRESS Figure 4 Response rate ratios of antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD. . compared with Placebo. P.

Antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD Figure 5 Response rate ratios of antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD. ARTICLE IN PRESS 9 . compared with Placebo. by inclusion of comorbid tic disorders (see Experimental procedures).

. Skapinakis et al.10 ARTICLE IN PRESS P. compared with Placebo. Figure 6 Response rate ratios of antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD. by inclusion of tics and dose of drug (see Experimental procedures).

compared with Placebo.Antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD ARTICLE IN PRESS Figure 7 Response rate ratios of antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD. by duration of trial. 11 .

Interpretation of the main finding — clinical implications All antipsychotic drugs influence the dopaminergic pathways in areas of the brain that have been linked to the pathophysiology of OCD.52. Stratification by duration of the trial Results of this subgroup analysis are shown in Fig. Most common reason for dropout in the drug group was a side effect and for the placebo group was lack of effect.. A duration of trials of at least 8 weeks was generally associated with a better outcome. a combined weighted relative risk for dropouts of 1. Only quetiapine and clozapine show consistently lower D2 occupancy rates (Seeman and Tallerico. controlled. was associated with a short-term higher response rate in adults with treatment resistant OCD compared with placebo. There is now evidence that a dopamine dysfunction may be involved in the pathophysiology of OCD (Denys et al. as augmenting agents of serotonergic antidepressants. There were few studies within each class of antipsychotic drugs and future publication of more and larger trials may influence the results in either direction. and not in twelve hours as is normally done. and therefore it should not be assumed that the association reported refers to individual patients. Skapinakis et al. it is likely that some of the results would be different if more and larger trials had been published.5. definition of refractoriness and response differed between studies and this was found to be associated with the outcome.98). 3. sedation and weight gain in olanzapine and quetiapine). 2002). 2003) and this has been considered as indirect evidence of an increased dopaminergic activity in OCD. most of the patients treated with antipsychotic drugs reported mild side effects typical of the drug used (akathisia in haloperidol. Moreover we found that lower doses had a 4. There were 24 dropouts (out of 130 patients) in the antipsychotic group compared to 12 (out of 121) in the placebo group. This group however was quite heterogeneneous. Use of lower doses was not associated with a better outcome compared to placebo but further stratification by inclusion or exclusion of patients with tic disorders showed that this could be due to the studies that used a low dose in the presence of comorbid tic disorders. Issues of response (or non-response) have not been standardized in OCD research and it is likely that other definitions might yield slightly different results in either the combined or subgroup analysis. The results of this meta-analysis show that the use of antipsychotic drugs is associated with a higher response in patients who had not responded previously to a serotonergic antidepressant... Although Type II errors would be more likely given the small number of trials. Dopamine D2 receptor blockade is considered as a sufficient and necessary condition for antipsychotic action (Kapur and Seeman. No study reported any serious adverse events. 4.3. 2004b. 3. The same applies for our stratification of studies by dose and inclusion of tic disorders.. Neuroimaging studies have shown that patients with OCD have higher dopamine transporter densities and lower dopamine D2 binding ratios compared to controls (Kim et al. 2001). P. although the effect to other dopamine or serotonin receptors may also be important especially for extrapyramidal symptoms or other side effects (Kapur et al.2.ARTICLE IN PRESS 12 patients with tic disorders. Although in theory the advantage of randomization is that it can be assumed that all potential confounding variables have been . an imbalance in the two arms cannot be excluded given the small number of patients randomized. 2005) or because there were no dropouts in either group (McDougle et al. 1999). 2001).1. In this calculation we have excluded two studies either because they did not report dropouts in each group (Erzegovesi et al. It should also be noted that the variable binclusion of patients with tic disordersQ is at the study level. Main findings In this meta-analysis we found that the use of antipsychotic drugs. Second. Even though we carried out subgroup analyses based on predefined subgroups and also we made specific hypotheses after reviewing the previous literature.. Predefined subgroup analyses showed reduced heterogeneity and positive outcome for risperidone (among the three drugs with 2 or more studies) and for using standard/high doses of antipsychotic drugs. The results should therefore be interpreted with caution. 7. studies were generally small and only half of them had a total sample of more than 30 patients. such as the basal ganglia (Kim et al. Dropouts Although the primary aim of this analysis was to report on the efficacy..6. Limitations of the study Our results should be interpreted in the context of the following limitations: First. Studies with a (post-randomization) duration of at least 8 weeks showed a statistically significant result while this was not evident in the group of studies with less duration.5. in which case an ecological bias could take place. we cannot exclude the possibility of Type I errors for the reported positive results. their D2 occupancy rate is quite high at therapeutic doses (Kapur and Seeman. we also report on the safety and tolerability of the use of antipsychotic drugs in OCD patients. the total number of trials included in the analysis (ten) was small and this makes the interpretation of subgroup analyses difficult. Even though antipsychotic drugs differ in their D2 blockade potency. However. 4. Stein. but it has been suggested that these two drugs have a very fast dissociation from the D2 receptor and if their occupancy is measured in two hours after administration. Third. Discussion 4. However. 3. 1994). most are capable of blocking more than 65% of the D2 receptors if given in sufficient doses. We also found considerable heterogeneity between studies.43 (95% CI 0. The combined response rate ratio for the three studies that used a low dose and included patients with tics was not significantly different from unity. 2003). 1999). sedation and dry mouth in risperidone. one of the advantages of the metaanalysis is that it can take into account these differences through sub-group analyses or meta-regression techniques.

An alternative mechanism could be that the lower doses are sufficient to block the serotonin 5HT2 but not the dopamine D2 receptors and thus worsen OCD symptoms (Lykouras et al. Therefore. Bystritsky. we cannot exclude this possibility for both drugs. 2004. J. van Ameringen.J. previous research has shown that the co-administration of atypical antipsychotics and SSRIs may increase synergistically extracellular dopamine and/or noradrenaline. J. Tezcan. S. This would result in a marginally significant combined response rate ratio.pnpbp. Int. Quetiapine augmentation in patients with treatment resistant obsessive— compulsive disorder: a single-blind. 1990. Gecici... B. P. Kuloglu.. The one haloperidol study seems promising but it is difficult to make recommendations based on a single study. de Geus.003. Br. Results for other drugs do not seem to be sensitive to additional trials that may be published in the future. one cannot exclude the possibility that higher doses are associated with a greater non-specific anxiolytic effect which is reflected in patients’ measurements of OCD symptoms. randomized. Implications for research Future trials on antipsychotic drugs in treatment-resistant OCD should try to avoid using low doses of antipsychotic drugs. 2004) had an unusually high response rate for placebo. especially in the prefrontal cortex (Dennys et al. Trials that have included patients with tic disorders generally report that the antipsychotic drugs were not better in the subgroup of patients with tic disorders. Delgado.K.R. D. Price. Only 13 the first haloperidol study reported that the drug was effective in the tics subgroup. randomised. The duration of the trial was found to be significantly associated with a better outcome.. 2004. Risperidone is more likely effective in doses closer to 2 mg/day. it is likely that this has resulted in the negative result.. 2002. Westenberg. References Atmaca. there should be careful evaluation of the presence of comorbid tic disorders..M. 17. to comment on the relative efficacy of specific antipsychotic drugs. Psychiatry 157. M. Department of Hygiene and Epidemiology of the University of Ioannina. placebo-controlled study [ISRCTN83050762]... Henninger. there is a need for larger trials of longer duration in order to evaluate both efficacy and safety in the long-term.11.4. Hollander. Bartz.. placebo-controlled study. Stein.J. D. Trikalinos. 565 – 568.. Ackerman. In the case of olanzapine. Vythilingum. These empirical data support the hypothesis of an increased dopaminergic activity in treatment-resistant OCD patients. given the small number of trials. Haloperidol was effective at a mean dose of 6 mg/day but it may also be effective in doses of 2 to 3 mg/day that are capable of blocking more than 65% of D2 receptors (Kapur et al. Saxena. J. 762 – 765.. A. Alternative dosing schemes within a trial are also of interest. It has also been suggested by others (Fineberg et al.. as we suggest.. Alternatively. Notwithstanding these limitations.G. one should take into account that most of these studies also used low doses. Psychiatry. It is very difficult. D. Finally. 2000). Patients with tic disorders may constitute a difficult subgroup of OCD patients. H. Shapira et al. 2000)... even a study of 200 patients with a 60% response rate in the active arm versus 10% response rate in the placebo arm would not be capable of yielding a significant combined rate ratio for olanzapine. 2004b). Denys. while there was no effect in OCD patients without tics (McDougle et al. 2005. Goodman. Finally. F. Fluvoxamine/pimozide treatment of concurrent Tourette’s and obsessive—compulsive disorder. In addition. For example.. BMC Psychiatry 24 (5 (1)).. This may be an indirect indication of a previously inadequate treatment with the serotonergic antidepressant. Quetiapine augmentation of SRIs in treatment refractory obsessive—compulsive disorder: a doubleblind. 115 – 119.. G.1016/j. these patients may need larger doses of antipsychotic drugs. O.L. Augmentation of serotonin reuptake inhibitors in refractory obsessive—compulsive disorder using adjunctive olanzapine: a placebo-controlled trial..... Seedat.ARTICLE IN PRESS Antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD smaller (non-significant) effect compared to standard/high doses.L. Gorbis. L.. Chaney. H. E. K. Vapnik. Studies that used the 25% reduction on the Y-BOCS scale criterion for refractoriness/response were associated with a non-significant combined rate ratio.. In addition. R. However.. 5.. for his comments on an earlier draft of the manuscript. Clin. Clin.. 4. Is obsessive—compulsive disorder an anxiety disorder? Prog. Acknowledgements The authors would like to thank Dr. W.M. 2004a.2005. 1996). P. Ramasubbu et al. van Megen. E. 2005. Thomas A. D.. The use of more selective D2 antagonists could offer useful data. S. the results of this meta-analysis support the use of risperidone as a first choice and haloperidol as a second choice in the short-term management of treatmentresistant OCD.H. doi:10. 1994)... M. M. The haloperidol study largely confirms that but we need more studies to make a recommendation. Carey. This finding emphasizes the need for extending the period of observation for at least 8 weeks before one can conclude that the augmenting agent is or is not effective in reducing OCD symptoms. However. Psychopharmacol.. Rosen.S. especially if higher doses are given in the future.D. E. 2006). 2003.E. Maidment.. Zhang et al..A. A double-blind. In our combined analysis we noted that studies that included patients with tic disorders were associated with smaller response rates not significantly different from placebo. 2006.. Biol.. Muller.. It should also be noted that in patients without tics smaller doses may be tried. that two of these studies (Carey et al. Our analysis showed an association between the definition of refractoriness and the response rate ratio. Psychiatry 65. These cortical dopaminergic pathways are important in regulating the mesolimbic dopaminergic pathways that may be affected in OCD (Denys et al. for example comparing low doses versus standard doses with placebo. in the case of quetiapine we calculated that in order to achieve a combined response rate ratio similar to that of risperidone a trial of 70 patients (35 in each arm) would be needed with 25 responses in the quetiapine arm and 5 responses in the placebo arm (71% versus 14% response rate)... J. We noted previously that in the case of risperidone there was low between studies heterogeneity. T. placebo-controlled trial of quetiapine addition in patients with obsessive—compulsive disorder . Neuro-psychopharmacol.

Shammi. Denys.. 397 – 401.. Hollander. Hill.. R. Kapur. A review of pharmacologic treatments for obsessive—compulsive disorder. Castle. 20. Moore. Seeman. S. 1996. Am.. Kapur. M.K. G. McQuay. and reliability. L.. Oya... Wasylink. Dopamine transporter density of basal ganglia assessed with [123I]IPT SPET in obsessive—compulsive disorder. C. B. 286 – 293. Psychopharmacol. C. R. 794 – 801.. Serv. Lewis. Price. Lapierre. 553 – 559. Westenberg. F..A.. Dennys..ARTICLE IN PRESS 14 refractory to serotonin reuptake inhibitors. L. Risperidone and haloperidol augmentation of serotonin reuptake inhibitors in refractory obsessive—compulsive disorder: a crossover study. Gale. W... 2002.. The Yale-Brown Obsessive Compulsive Scale. Psychiatry 46. 2003. Charney. Siliprandi. Psychiatry 66. S. Kapur. Matsui. 15 (21). and side effects: a double-blind PET study of firstepisode schizophrenia. Bebbington. 5. F. Gen.J. J. D..A. Jadad. Clin. Med. The epidemiology of obsessive—compulsive disorder in five US communities. Psychiatry 27. H... Synergistic dopamine increase in the rat prefrontal cortex with the combination of quetiapine and fluvoxamine. Gavaghan.. Association between the dopamine D(2) receptor TaqI A2 allele and low activity COMT allele with obsessive—compulsive disorder in males. Clin. J. risperidone. L. C... Nucl.R. V. L.B.K. Seeman.C... J.. Price..J.. 107 – 129.M. Price. Z. Baxter.. ECDEU Assessment Manual for Psychopharmacology. Burnam. 1998. 2003...N. C. Rabavilas. Am... C... Kapur.T. Lee. S. A. 1990.G.. 2005. J.. 195 – 203. 1999. Houle.. 2000a. Farrell. Biol. A. Gale. Ramasubbu.. J..J. McDougle.M. J.. A. J. Fineberg. Price. G. G. Disord.. K. G.P. Epperson. 2005. May.B.. C. Psychopharmacol.H. D.. Arch. K. 18. A.S. Gill..H. Denys. Lee... Stat. G. Jenkinson. Karno. Affect. A.. R.. Higgins. Psychiatry 57. G.S. Charney. Rasmussen. 11 – 17. Alevizos.. 1040 – 1048. Clin. Jones.A. clinical response. placebo-controlled study. 1539 – 1558. J.. 1994. 97 – 103. Pallanti.. D. Pokos.. Neuropsychopharmacol.. 2005. Eur. D.. North Am.1016/ j. A review of the reported cases. E. E. Erzegovesi.K. E.. Zipursky. Clinical and theoretical implications of 5-HT2 and D2 receptor occupancy of clozapine. Golding. Michalopoulou.A.. Gen. use.. 1999. 1997. 1637 – 1643. Rockville. Leckman. Krystal.P. Wilson. Obsessive—compulsive symptoms induced by atypical antipsychotics.S.R. J. R. Kim. L... Arch. J. Pelton. 775 – 789.. Remington.. Does fast dissociation from the dopamine d(2) receptor explain the action of atypical antipsychotics? A new hypothesis.H. Uppfeldt.. Psychiatry 45. P... H. Compr..L. P.F..M.. van Nieuwerburgh. Guglielmo.euroneuro. S. Int.. S. S.. 69 – 74..G. Sareen. 2005.. R... 652 – 654. Fleischmann. M.. Goodman.. 493 – 507. Nordstrom. The National Psychiatric Morbidity surveys of Great Britain—initial findings from the household survey. Psychiatry 51.. 1094 – 1099. 336 – 343.G. High levels of dopamine D2 receptor occupancy with low-dose haloperidol treatment: a PET study. 948 – 950. X.L. Neuropsychopharmacol. D.12..M. Sood. H. D. H. G. Neuropsychopharmacol.. DaSilva.. DaSilva. Remington. Lee. S.. Gen. 15. LH.H. 167 – 174. Am. Gale. C. placebo-controlled study in patients with and without tics. Kirshner. 1976. Obsessive—compulsive disorder with poor insight. Remington.. J.. Li. Reynolds. Deforce.S. 1006 – 1011. A. Delgado. The role of dopamine in obsessive—compulsive disorder: preclinical and clinical evidence. 82 – 83. Y.. Jenkins. S. T. Prog. Y. Am. Sivakumaran.R. A review of antipsychotics in the treatment of obsessive compulsive disorder.. Reiss. L. N. M. Eur.A. 2002. Seeman. H.. Kapur. Houle. L.M. 1996... Y. C.. Zohar. Zipursky. Meltzer. D. W.. Pharmacopsychiatry 33. Bellodi.. Psychiatry 65 (suppl 14). Jackson.. Zipursky. 2003. Kapur. Psychiatry 65. Hoolander.. M. G.M. Central D2-dopamine receptor occupancy in relation to antipsychotic drug effects — a doubleblind PET study of schizophrenic patients.. T. D. Kjernisted. W. Development... D.. Jones. J. N. H.. K. I. Psychiatry 158.. Koshimune. 1998. Psychol. 360 – 369... R. N. Cheon. Goodman. 2000. N. Klompmakers. Lander... S. D. Lowdose risperidone augmentation of fluvoxamine treatment in obsessive—compulsive disorder: a double-blind. C. 2005. C. Clin. Med. 1989. 2004b... H. T. Bienstock. Haloperidol addition in fluvoxaminerefractory obsessive—compulsive disorder. G. 3rd ed. Clin.H.. Pauli. Horwath. The relationship between risperidone plasma levels and dopamine D2 occupancy: a positron emission tomographic study. Zohar.J.. J. Heninger.A. Psychiatry 147.H. Remington.. Leckman. W. 1993. C. N.. J. Carroll.J. Mazure. 2000. Roberts. Baldini Rossi. S. Tallerico..S. Arch. S.. G. Guy.D. Wilson. 2003. C. 8. Mol.. A. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin. J. Thompson. 2001.. G.... S. Weerasundera. J. Med. Psychiatry 33. J..J.D. 2004.. P. Psychiatry 157. Stein. N. Brugha. Skapinakis et al. 1 – 12. D. S. Psychiatry 39.A..A.. Neuroleptic addition in fluvoxamine-refractory obsessive—compulsive disorder. 150 – 157.. 1988. Sivakumaran... L. Gen. 27. 2000.. Jones. R. Arch. B.. J. 5-HT2 and D2 receptor occupancy of olanzapine in schizophrenia: a PET investigation. J. 23.. Matsunaga.L. 302 – 308.C. 6. Antipsychotic treatment in obsessive—compulsive disorder: a literature review. Miyata. Adding quetiapine to SRI in treatment-resistant obsessive—compulsive disorder: a randomized controlled treatment study. A.. J. E.M.B. Psychiatry 155. Int. J. F. 54. Psychiatry 153.. Jones. Psychiatry 156. Int.A. Kapur. J.. 333 – 346. R. 2006. Aust.. P. Int. E. Ravindran. Stein. Neuropsychopharmacol... Zipursky.. T... L. 223 – 226. Risperidone augmentation in treatment-resistant obsessive— compulsive disorder: a double-blind.S. Halldin.. Remington.. S. Neuropsychopharmacol.. McDougle. J. D. D. Quantifying heterogeneity in a meta-analysis.. Biol. T.. S. Zipursky.. and olanzapine in schizophrenia. Zipursky.R. 2002. Rapid release of antipsychotic drugs from dopamine D2 receptors: an explanation for low receptor . Lykouras. The epidemiology and crossnational presentation of obsessive—compulsive disorder.L. Psychiatr. Am.. Relationship between dopamine D(2) occupancy.J. 227 – 235. Psychopharmacology (Berl) 176. E. Remington. McDougle... Tolbert. Ryu. Heninger..2005.. T. Eleff. 2005. 181 – 191. placebo-controlled study. placebo-controlled study of risperidone addition in serotonin reuptake inhibitor-refractory obsessive—compulsive disorder.. C.R. Weissman. A. Iwasaki... N. A double-blind. Psychiatry 57. M. R. doi:10. Flournoy... Am. M. Farde. K. 921 – 928.001.H.. G. P. R. Marazziti. J. J. Trials 17. Koo... Westenberg.. Imaging 30. Forslund.J. 82. Fineberg. Koran. Kaplan. C. 2000b. J. M.. Psychiatr. Keuneman. Heninger. R. Arch. S. Sorenson. 1111 – 1118. Eur. Neuropsychopharmacol. R. 236 – 238..K. Psychopharmacol. 2004.. Fineberg. Pallanti. Kiriike. 514 – 520.. Clin. Wiesel.. T. A.A.. J. Goodman. Westenberg.J.. Treatment non-response in OCD: methodological issues and operational definitions. J. A double-blind. Psychiatry 43. P. National Institute of Mental Health Ed. S.. J. Houle. 736 – 743. T. Hollander. Serotonin and dopamine antagonism in obsessive—compulsive disorder: effect of atypical antipsychotic drugs. C.. Gen. J. M. Pato.. quetiapine in schizophrenia: a preliminary finding of an antipsychotic effect with only transiently high dopamine D2 receptor occupancy. K.. W. Do antipsychotics ameliorate or exacerbate obsessive compulsive disorder symptoms? A systematic review. S. 20. Evidence-based pharmacotherapy of obsessive—compulsive disorder. A positron emission tomography study of P.... A. R.

M.K.J. 2004.. Bymaster.2 for Windows.A. 250 – 262. The Cochrane Collaboration (2003). Perry. Psychiatry 55. T.T.. 121 – 127.W.. Skoog. D. I. W. placebocontrolled trial of olanzapine addition in fluoxetine-refractory obsessive—compulsive disorder.. J. Neuropsychopharmacology 23. Shapira. Arch..D. Can. Psychiatry 42. M. 1997. N. Disability and family burden in obsessive— compulsive disorder. Stein. D. Ward. Potts. G. Blier. Obsessive—compulsive disorder. G. K. F.ARTICLE IN PRESS Antipsychotic augmentation of serotonergic antidepressants in treatment-resistant OCD occupancy and early clinical relapse upon withdrawal of clozapine or quetiapine..K... J. Biol. A 40-year follow up of patients with obsessive—compulsive disorder. J. 919 – 928. 876 – 884. 553 – 555...C..P. Review Manager (RevMan) [Computer program]. P... Bao. 397 – 405. Murphy. W.. . G.D. H. Oxford. A double-blind.E. Gen. Psychiatry 156. Yang.. Wong. Goodman. England. Version 4. 2002. Synergistic effect of olanzapine and other antipsychotic agents in combination with fluoxetine on norepinephrine and dopamine release in rat prefrontal cortex. Zhang. 2000. 15 Steketee. Lancet 360. Psychiatry 56. B... Am. Tollefson. 1999.. Skoog. Mandoki.