A Review of Antipsychotics in the Treatment of Obsessive Compulsive Disorder

Reports of improvements in most cases treated with adjunctive risperidone in four open case studies (McDougle et al., 1995a; Saxena et al., 1996; Stein et al., 1997; Pfanner et al., 2000) were followed up by McDougle et al. (2000) who reported the first double-blind, placebo-controlled study of risperidone augmentation in 36 patients unresponsive to 12 weeks treatment on an SRI. In their completer analysis of 33 patients, risperidone (2.2 mg) was superior to placebo in reducing Y-BOCS scores (mean post-baseline improvement of 31.8%) as well as anxiety and depressive symptoms and was well-tolerated. Of the risperidone completers, 9/18 (50%) were considered ‘marked responders’ (using the same criteria as the haloperidol study) or ‘partial responders’ (2 out of the 3 marked responder criteria), compared to none treated with placebo. There was no difference between those with and without comorbid tics or schizotypy. However, those who had two or more SRIs were less likely to respond to risperidone addition than those who had failed to respond to only one. A smaller double-blind study by Hollander et al. (2003a) examined patients failing to respond to at least 2 trials of SRIs using an intent-to-treat analysis. Four of ten (40%) randomized to risperidone (0.5–3 mg) were responders (CGI-I ≤ 2 and baseline Y-BOCS decrease of ≥25%) at the 8-week end-point, compared with none of the six patients randomized to placebo. However, the mean reduction in baseline Y-BOCS scores was only 19% and the difference was not significant, possibly owing to a small sample size. Less impaired insight at entry predicted a better response with risperidone, perhaps reflecting less severe illness. A third trial (Erzegovesi et al., 2005) enrolled 45 drug-naïve subjects in a 12-week open label fluvoxamine phase, after which ten treatment-non responders (Y-BOCS reduction <35% and CGI-I > 2) were randomized to 6 weeks low dose (0.5 mg) risperidone and ten to placebo augmentation. The study did not report a between group analysis, but five risperidone-treated cases responded (≥35% YBOCS improvement) compared to two on placebo.

Quetiapine, as an add-on to SRIs, has also been the subject of recent investigation. Contradictory results were obtained from open label studies, most of which showed benefits in up to 50% individuals treated (Denys et al., 2002; Sevincok and Topuz, 2003; Bogan et al., 2005) although another study showed little benefit (Mohr et al., 2002). The single-blind study by Atmaca et al. (2002) found a clinical response in 14/27 (64%) cases.

Three double-blind, placebo-controlled studies of quetiapine have been recently completed and these also show contradictory results. In the first (Fineberg et al., 2005) 21 patients with DSM-IV OCD without significant Axis I comorbidity, and who had failed to respond to at least 6 months SRI treatment, were treated for 16 weeks with SRI and quetiapine (≤400 mg) or SRI and placebo. Quetiapine was well tolerated with just one participant in each group dropping out. The quetiapine group's reduction in baseline Y-BOCS scores averaged 14% at week 16 compared to 6% under placebo. Considerable variation in treatment-response to quetiapine was seen, with most patients responding at the placebo-level. However 3/11 of the quetiapine group showed at least a partial treatment-response by Pallanti et al.'s (2002a) criteria, compared to just 1/10 under placebo. There was no relationship between treatment response and Yale Tic Severity Scale, Y-BOCS or MÅDRS scores, or previous length of illness or number/duration of failed treatments. Although the sample size was relatively small the study was powered at 90% to detect as significant a 4 point difference on the Y-BOCS.

In the study by Carey et al. (2005) 42 subjects who had responded inadequately (Y-BOCS reduction < 25% or CGI-I >2) to open-label treatment with an SRI for 12 weeks were randomized to either placebo or flexible doses of quetiapine for 6 weeks. It is noteworthy that cases received only 6 weeks treatment at maximum tolerated dose or manufacturers recommended maximum dose, which may have been too short to discriminate SRI resistance. Quetiapine was generally well tolerated and there was significant improvement from baseline to endpoint on the Y-BOCS in both the quetiapine and placebo groups (quetiapine, n = 20, p < 0.0001; placebo, n = 21, p = 0.001) with 40% (n = 8) of quetiapine and 47.6% (n = 10) of placebo treated individuals being classified as responders. Quetiapine did not demonstrate a significant benefit over placebo at the end of the 6-week treatment period. Similarly quetiapine failed to separate from placebo in the subgroup of subjects (n = 10) with co-morbid tics. The failure to demonstrate an advantage for the active drug has been attributed to the inclusion of cases of unstable OCD or partially SRI-resistant individuals who went on to respond further to extended treatment with the SRI.

By contrast, another double-blind, placebo-controlled study (Denys et al., 2004a) has shown significant efficacy for quetiapine (<300 mg) augmentation in a sample of 40 patients who had previously failed to respond to at least two SRIs. An intent-to-treat analysis for the 20 quetiapine-treated cases showed a significant advantage over placebo from 4 weeks on the Y-BOCS and a mean decrease of 31% on baseline Y-BOCS, compared to 6% in the placebo group at the 8-week endpoint. Decreases in Y-BOCS scores correlated positively with severity of obsessions at entry. Eight out of 20 (40%) quetiapine patients showed meaningful clinical responses by Pallanti et al.'s (2002a) criteria compared to just 2/20 (10%) under placebo.

The potentially encouraging results of open-label olanzapine studies (Weiss et al., 1999; Bogetto et al., 2000; Koran et al., 2000; Francobandiera, 2001; Crocq et al., 2002; D'Amico et al., 2003) were not supported by a double-blind, placebo-controlled study run by Shapira et al. (2004). This study included 44 partial-or non-responders to 8 weeks open-label fluoxetine; both the fluoxetine plus olanzapine (5–10 mg), and fluoxetine plus placebo groups improved over a 6-week study period, and there was no additional advantage for olanzapine augmentation. Nine of 22 (41%) showed at least a 25% post-baseline reduction in Y-BOCS scores in both the olanzapine and placebo groups. Differences in entry criteria between this study, which included ‘partial responders’ and which used short (8-week) pre-trial SRI treatment, and other studies that included more stable, treatment-refractory cases, may partly explain the differences in the results.

By contrast, Bystritsky et al. (2004) recruited a more treatment-refractory group of 26 OCD patients who had been unresponsive to at least two 12-week trials of SRIs and one trial of behaviour therapy. Eleven of 13 olanzapine treated patients (mean dose 11.2 mg) and seven of 13 placebo treated patients completed the study. Olanzapine was well-tolerated and, by intent-to-treat analysis, was significantly superior to placebo at the 6-week endpoint. The mean reduction of baseline Y-BOCS scores was just 16% but the placebo-group showed no improvement at all with only seven completing the study. Six (46%) of the olanzapine group were responders (25% improvement on Y-BOCS). Because of the high placebo dropout rate, the authors also ran a completer analysis which failed to reach significance so a larger study was recommended. Given the inconsistency of the data collected so far, the efficacy of olanzapine for treatment resistant OCD remains equivocal.

A single positive open-label study of the more selective dopamine antagonist amisulpride (200–600 mg) used as an adjunct to SRI in 20 cases of resistant OCD (Metin et al., 2003) merits replication under controlled conditions. However, the results for clozapine, given as a monotherapy to 12 adults with resistant OCD in an open-label design showed no effect (McDougle et al., 1995b). Aripiprazole (10–30 g) was also studied as a monotherapy using an open-label design in a mixture of drug-naive and resistant cases (Connor et al., 2005). Two out of seven cases withdrew because of side effects, and there was no significant change from baseline on Y-BOCS scores for the group, although two patients were judged to be responders.

Some authors have reported emergent obsessions during treatment of schizophrenia with atypical antipsychotics, which may be related to their mixed receptor antagonist properties (Sareen et al., 2004). Obsessive-compulsive symptoms have, however, long been recognized to be common among patients with schizophrenia (Poyurovsky and Koran, 2005) and may well be linked to the schizophrenic disorder as opposed to the antipsychotic treatment.

The small size of all studies to date limits the extent to which we can identify predictors of antipsychotic response, though presence of comorbid tics, insightfulness and true SRI-resistance may predict a better outcome. A FDG-PET study by Buchsbaum and colleagues (2002) has investigated SRI-responsive cases before and after risperidone and was reported to have found a superior response in cases with lower metabolic activity in the striatum. This echoes the PET finding by Saxena et al. (2003) who found that lower metabolic activity in the right caudate nucleus was associated with SRI-refractoriness in untreated OCD. It also hints at a radiological endophenotype representing SRI-resistant antipsychotic responsive OCD.

In the treatment-resistant OCD population, the use of first generation antipsychotics produced dose-dependent extrapyramidal side effects (EPSEs) such as akathisia to the extent that in some of the studies (e.g., McDougle, 1990) up to half the patients experienced significant impairment requiring coadministration of beta blockers and/or anticholinergics. The second generation antipsychotics appear to offer an advantage in this respect and are generally well-tolerated in the context of short-term efficacy trials. EPSEs are very uncommon in all the trials so far and dropout rates from side effects have been low. Weight gain has been reported in short-term trials of olanzapine (Bystritsy et al., 2004; Shapira et al., 2004) and drowsiness in trials of olanzapine (Bystritsky et al., 2004) and quetiapine (Fineberg et al., 2005). Longer-term studies are needed to examine this area with more confidence.

The mechanism by which antipsychotics might enhance SRI-mediated antiobessional activity remains speculative, owing to their different receptor-blocking profiles. Moreover, the mechanism underpinning action of even SRIs in OCD still remains unclear. While it is assumed that increased serotonergic activation in relevant brain areas, such as prefrontal cortex, thalamus and striatum play a role, the failure of acute tryptophan depletion (Barr et al., 1994) to reverse the treatment effect casts some doubt on this hypothesis, although the patients may not have been tested under optimal conditions with a challenge or an exposure strategy.

All antipsychotics found to be effective in treating OCD so far share affinity at 5HT_2A, D2 and alpha 1 or 2 receptors. Some authors have considered differential blockade and activation of 5-HT_2A and non 5-HT_2A receptors as important (Marek et al., 2003). Second generation antipsychotics such as risperidone, olanzapine and, possibly to a lesser extent quetiapine, are distinguished by their potent antagonism at 5HT_2A sites. Support for this hypothesis comes from the findings of an uncontrolled study that mirtazapine, which also antagonizes, inter-alia, 5HT_2A receptors, might enhance the anti-obsessional effect of SRIs in non-resistant cases (Pallanti et al., 2004). However, the robust efficacy seen in the studies of resistant OCD with haloperidol and pimozide, which are substantially more potent at blocking dopamine than serotonin receptors, suggest a dopaminergic contribution (Denys et al., 2004b), although it should be noted that these studies included a high proportion of cases with comorbid tic disorders, which themselves respond to antipsychotic drugs. It remains unclear as to how effective these drugs might be for OCD cases without comorbid tics. Positive results from open studies of amisulpride, if confirmed in RCTs, would support a dopamine hypothesis for SRI-resistant OCD. A recent microdialysis study in rats by Denys et al. (2004c) looking at extracellular monoamines demonstrated that whereas monotherapy with either fluvoxamine or quetiapine increased striatal and prefrontal serotonin levels, only a combination of the drugs resulted in a major increase in prefrontal levels of extracellular dopamine, which they hypothesized might be necessary as well as enhanced prefrontal serotonin for treatment augmentation. Pharmacological dissection in controlled studies using more selective antagonists, such as amisulpride will be valuable. Investigation of cognitive enhancing agents such as the norepinephrine transporter inhibitor atomoxetine, which has also been reported to raise dopamine levels in prefrontal cortex but not in dependence-promoting areas such as ventral tegmental area and nucleus accumbens (Bymaster et al., 2002), may have potential for this area.

Evidence for alternatives to antipsychotics as add-on treatments in OCD is sparse. Some uncontrolled case studies suggest that increasing SRI doses above formulary limits can procure a better effect (Byerly et al., 1996; Bejerot and Bodland, 1998). Intravenous clomipramine has been shown to be more effective than placebo in a single double-blind study (Fallon et al., 1998) and a positive open study of intravenous citalopram has been reported (Pallanti et al., 2002b). However, no controlled studies of the co-administration of different SRIs have been published. Positive results have, nonetheless, been reported from small, uncontrolled case series combining clomipramine with an SRI (Szegedi et al., 1996; Pallanti et al., 1999) although this strategy should be approached with caution since the pharmacokinetic interactions of these drugs at hepatic cytochrome P450 systems could be hazardous. In this respect, citalopram is least likely to interact and may be the safest SRI to combine with clomipramine (Pallanti et al., 1999). Controlled studies have confirmed the lack of efficacy of lithium augmentation in OCD (McDougle et al., 1991; Pigott et al., 1991). Similarly, double-blind placebo controlled studies have demonstrated that combination of buspirone with an SRI is not helpful (Pigott et al., 1992a; Grady et al., 1993; McDougle et al., 1993). Clonazepam as a monotherapy fails to impact on the core symptoms of OCD (Hollander et al., 2003b) though it may help with associated anxiety. Pigott et al. (1992b) reported limited efficacy for clonazepam given together with an SRI in a double-blind placebo-controlled study. Dannon et al. (2000) demonstrated efficacy for pindolol when combined with paroxetine in a double-blind placebo-controlled study, but another study combining it with fluvoxamine did not (Mundo et al., 1998). Blier and Bergeron (1996) found a beneficial effect for pindolol only when 1-tryptophan was openly added to the combination. Barr et al. (1997) investigated the addition of desipramine to SRI in a double-blind placebo-controlled study, and found no added benefit.