Combination Antipsychotics: What is the Evidence?


Objective: To identify clinical situations where combination antipsychotic drug therapies are used, and to review the evidence supporting the use of  such combinations.

Methods: MEDLINE (1966-April 2000), Cochrane Library and EMBASE (1980-April 2000) searches were performed to identify literature pertinent to the practice of combining antipsychotics.  Articles selected included clinical trials, anecdotal reports and discussion papers pertaining to  antipsychotic augmentation or combination therapy.  Key words used in the search were antipsychotic, neuroleptic, combination and augmentation.

Results: Examples where combination antipsychotics are used include: the augmentation of clozapine for partially or non-responsive patients; the addition of a traditional antipsychotic to an atypical agent; the short-term co-administration of traditional agents when initiating therapy with a novel antipsychotic; and as a result of an incomplete switchover from one antipsychotic to another.  Despite the commonality of such practices, only one randomized controlled trial could be identified which involved the addition of sulpiride (a potent D2 antagonist) to clozapine.  The remaining indexed literature was comprised of two open, prospective trials, a single retrospective review and four separate anecdotal reports.

Conclusion: Clinicians are advised to evaluate the rationale for combining antipsychotics before doing so as there is a paucity of published evidence to support such practices.  Combination antipsychotic therapy can potentially lead to more complex, less well-tolerated treatment regimens that may reduce patient compliance and increase treatment costs.


Combination antipsychotic use or the practice of augmenting the use of one antipsychotic with another is a common clinical scenario.  The potential disadvantages of combining antipsychotics include: increased risk of adverse effects; increased risk of noncompliance; and increased drug costs. The  frequent use of combination antipsychotic therapy  may stem from the perceived slow onset of action of the novel agents and the need for quick recoveries in a system under pressure to shorten or avoid hospital inpatient stays.  Presumed enhanced efficacy when using agents with dissimilar pharmacologies may also lead some clinicians to trial  combined regimens for  patients who are partially or completely unresponsive to monotherapy.  Because of the potential disadvantages of combining ant psychotics, evidence for improved efficacy of combination therapy over single agents is required. Accordingly, the  description of clinical situations where combinations of antipsychotics may be used and assessment of   published evidence to support such practices was sought out.  Thus, the objective of this paper is to identify clinical situations where combination antipsychotic drug therapies are used, and to review the evidence supporting the use of  such combinations.


A literature search using MEDLINE (1966-April 2000), Cochrane Library and EMBASE (1980-April 2000) was performed to identify literature pertinent to the practice of combining antipsychotics. Articles selected included clinical trials, anecdotal reports and discussion papers pertaining to  antipsychotic augmentation or combination therapy.  Key words used in the search were antipsychotic, neuroleptic, combination and augmentation.


Only one randomized controlled trial could be identified while the remaining indexed literature was comprised of two open, prospective trials, a single retrospective review and four separate anecdotal reports (Table 1).

Clozapine Augmentation: A Controlled Clinical Trial

At present, the bulk of published data on combining antipsychotics involves clozapine. Typically,  clozapine is relegated to last resort therapy due to the risk of agranulocytosis and the requirement for regular blood monitoring. Approximately 30-60% of patients resistant to previous antipsychotic trials respond to clozapine monotherapy. While there is no clear consensus on how to manage patients whose symptoms persist despite an adequate course of clozapine, the results of a study by Shiloh et al suggest a possible role for sulpiride, a selective D2 antagonist. In this double-blind, placebo-controlled study, 28 patients with schizophrenia who had partial or unsatisfactory responses to clozapine at adequate doses for at least 12 weeks were randomized to receive additional therapy with either sulpiride or placebo for a period of 10 weeks.  The investigators reported that Brief Psychiatric Rating Scale (BPRS) scores were reduced by 8.7 points (20.7%) in the sulpiride group versus 2.3 points (5.4%) in the placebo group (p<0.05).  The between group differences in positive and negative symptom score reductions were reported to be 5.8 points (12.3%) and 6.8 points (10.4%), respectively, in favour of sulpiride (p<0.05).  There was a trend toward younger patients exhibiting a greater than 20% reduction in BPRS scores (35.3 versus 45.4 years),  suggesting that a subgroup of patients with chronic schizophrenia may obtain significant benefits from this combination.  Overall, the combination of clozapine and sulpiride was well tolerated with the exception of hyperprolactinemia.  There was a 4- to 7-fold increase  noted in the serum prolactin levels of the sulpiride-treated subjects (p<0.05) while no changes were observed in the clozapine+placebo group.  Despite this, none of the 5 female patients with hyperprolactinemia complained of galactorrhea and 3 of the 5 were amenorrheic prior to enrollment in the trial and remained so for the duration of the study. This side effect did not result in  an alteration in treatment.

A limitation of this study was a statistically significant difference in previous total duration of hospitalization between the 2 groups, with patients in the placebo arm (70.8 +/- 56.1 months) experiencing a mean pre-trial hospitalization period that was twice the duration of that reported for the study group (32.8 +/- 30.3 months)  (p<0.05).  These investigators address the implications of this limitation; however, they point out that this difference was not reflected in the clinical status of the patients at baseline.  Other limitations of this study included a small sample size, the exclusion of complete nonresponders to clozapine and limited study duration of only 10 weeks.  Continued clinical improvement with clozapine therapy may be seen after  12-24 months of treatment. Thus, it is not clear whether  the patients would have improved with clozapine alone or if sulpiride contributed to the overall improvement. Albeit encouraging, validation of these results by other investigators and with other potent D2 antagonists is needed to support this treatment approach.

Clozapine Augmentation: Uncontrolled Clinical Trials

In 1997, Friedman published a case series involving augmentation of clozapine with pimozide, another potent D2 antagonist. Although this was not a controlled trial, the results were also positive and consistent with the sulpiride trial described above. Pimozide, at doses ranging from 2-8 mg/day, was added to clozapine in 7 patients previously diagnosed with either schizophrenia or schizoaffective disorder.  These patients had been receiving treatment with clozapine for at least 10 months prior to the addition of pimozide, and were considered to be only partial responders.  The mean change in BPRS score was 24 points (24%) (p=0.003) over an average treatment period of 32 days.  Unfortunately, the authors did not describe how well the combination was tolerated by these patients and it is unclear whether any of these patients continued with both clozapine and pimozide, as a consequence of the favourable results.  While the limitations of this report are quite apparent, the case series does provide an impetus for future studies.

Possibly more compelling are “open trial publications”  on the augmentation of clozapine with risperidone and loxapine both describing favourable effects. The risperidone group consisted of 12 patients with either schizophrenia or schizoaffective disorder who had persistent psychotic symptoms after at least 12 months of clozapine therapy.  Patients received 4 weeks of treatment with risperidone  up to a maximum dose of 6 mg/day.  The mean reduction in total BPRS scores was 11.9 points (28%) (p=0.0002) and 10 of the 12 patients showed a greater than 20% reduction in such scores.  Risperidone did not seem to have significant effects on clozapine disposition as evidenced by clozapine levels drawn at baseline and at the end of the 4 weeks.

Loxapine was added to clozapine in an open 18-50 week trial with 7 patients diagnosed with schizophrenia or schizoaffective disorder, who had previous treatment with clozapine for at least 9 months.  All patients were concurrently attending a structured day hospital program, which may have contributed to their improvement and confounded the results.  A 19-38 point reduction in BPRS scores was reported, with 2 patients described as being “rapid responders”, with  improvement demonstrated within 3 weeks of initiating loxapine.  Clozapine serum levels were not significantly altered by the addition of loxapine as monitored in 4 of the 7 patients.  Unfortunately, little information was provided regarding reported adverse effects with this group other than the authors declaration that  no serious effects were observed.

Clozapine Augmentation: Anecdotal Reports

Finally, two anecdotal reports were identified describing  augmentation of clozapine with either olanzapine or risperidone. Gupta et al. report success with the addition of olanzapine to clozapine in 2 patients; however, the length of clozapine trial prior to the addition of olanzapine was not indicated.  One patient was reported to have a 35% improvement in symptoms and the second a 13-point reduction in BPRS score. The only adverse effect identified was a case of drooling in one patient.   The investigators stated that neither patient had significant extrapyramidal side effects or orthostatic hypotension with this combination.  In a report by McCarthy and Terkelsen, the augmentation of clozapine with risperidone in 2 patients whose clozapine dose was limited due to intolerable adverse effects, was described as successful. Again, the length of clozapine trial prior to the addition of risperidone was unclear and tolerance to the combination was not specified.  The benefits of these regimens remain inconclusive because of small sample sizes, lack of a control group and the likelihood of publication bias (i.e. lack of publication of negative trials of the above combinations).

Other Antipsychotic Combinations

Possibly more prevalent is the practice of augmenting antipsychotics other than clozapine.  Anecdotal reports of adding a conventional antipsychotic to a novel agent exist however, scarce.  Waring and colleagues describe success with using this strategy in 31 refractory schizophrenic patients over a 9-month review period,  and proposed that this approach could be an alternative to using clozapine. These authors report using low dose haloperidol, trifluoperazine or fluphenazine in conjunction with risperidone, olanzapine or  quetiapine.  Unfortunately, details of the  specific combinations used were not reported.  Two-thirds of these patients were reported to have improved and were discharged from hospital.  Little information is provided by the authors regarding patient tolerance to the combinations  other than noting the absence of serious adverse effects.

Takhar reported improvement after 7 weeks of pimozide 2 mg/day augmentation of olanzapine in a patient exhibiting residual psychotic symptoms despite 6 months of treatment with olanzapine 20 mg/day. The patient was noted to have had a BPRS score reduction of 18 points (32%).  Tolerability of this regimen is not discussed other than the patient did not display evidence of tardive dyskinesia. Although, provocative, these case reports suffer the same limitations as the anecdotal reports involving cocaine combinations.

Other Combination Antipsychotic Use

A related scenario to the aforementioned is the short-term administration of conventional agents when initiating therapy with a second-generation antipsychotic.  This has become commonplace at some centres because of the belief that the novel agents are not as effective for treating agitation and aggression that often accompany acute psychosis. Unpublished data from a retrospective analysis of two acute admitting units revealed that concomitant haloperidol use was highest with quetiapine and lowest with risperidone, supporting the impression that clinicians lack confidence in the new agents for inpatient treatment or possibly the inefficacy of quetiapine. The pharmacologic hypothesis appears to be that potent D2 blockade is required for resolution of acute psychotic episodes. Thus, quetiapine, a weak D2 blocker, would require augmentation with haloperidol, unlike risperidone, a potent D2 blocker. If this is employed, it is important to regularly reassess the need for both medications in order to minimize extrapyramidal side effects and ineffective polypharmacy.

Other cases of combination antipsychotic use occur when patients are being switched from one agent to another because of poor response or tolerance.  Most often the two therapies are overlapped and if a patient shows improvement during the cross-titration, there may be reluctance to make subsequent changes.  This practice is generally discouraged due to the potential for increased costs, noncompliance and reduced tolerability, and is not supported by published evidence.

Table 1: Clinical trials and anecdotal reports regarding the use of combination antipsychotics.

Trial Design/N Treatment Arms/ (N/N) Results (Mean Reduction in BPRS* Scores) Adverse Effects
Shiloh et al. 1997 RCT**/28 Clozapine + Sulpiride (16)

Clozapine + Placebo (12)

8.7 points (20.7%) vs. 2.3 points (5.4%) in favour of sulpiride addition (p<0.05) 4-7-fold increase in serum prolactin levels.
Friedman et  al. 1997 Retrospective review/7 Clozapine + Pimozide 24 points (47%) (p=0.003) Not reported.
Henderson et al. 1996 Open prospective trial/12 Clozapine + Risperidone 11.9 points (28%) Mild akathisia reported in 4 patients and return of hypersalivation in 1 patient.
Mowerman et al. 1996 Open prospective trial/7 Clozapine + Loxapine 19-38 points No serious adverse effects reported.
Gupta et al. 1998 Case reports/2 Clozapine + Olanzapine 13 point reduction for 1 patient and a  35% improvement in the second (not specified) Drooling described in 1 patient.
McCarthy et al. 1995 Case reports/2 Clozapine + Risperidone Not specified. Authors report improvement in both patients. Not reported.
Waring et al. 1999 Case reports/31 Conventional + Novel (haloperidol, trifluoperazine, and fluphenazine with risperidone, olanzapine and quetiapine) Not specified. Authors report clinical improvement in 2/3 of patients. No serious adverse effects and a few “unpleasant subjective” side effects reported.
Case report/1 Olanzapine + Pimozide 18 points (32%) No discussion of adverse


*BPRS= Brief Psychiatric Rating Scale
**RCT=randomized controlled trial


Not addressed by these reports is an evaluation of the trade-offs among benefits, safety and costs when combining antipsychotics.  Early experience with the novel agents indicated a more appealing adverse effect profile, primarily due to a reduction in the risk for extrapyramidal side effects and tardive dyskinesia.  However, as experience accumulates new risks associated with the novel agents have attracted concern, including obesity, diabetes (including diabetic ketoacidosis), hepatic dysfunction, and most recently cardiomyopathy. When traditional and novel antipsychotics are combined, patients are put at risk simultaneously for both the  old and new adverse effects. The question remains, and will continue until long-term, controlled data becomes available, whether the benefits of combination antipsychotics are worth the risks to health and budget.

Before considering combining antipsychotics, clinicians should first examine the reasons for doing so. The addition of a second antipsychotic to clozapine-resistant patients is considered an appropriate intervention, as there is yet no other treatment shown to be as or more effective than clozapine for this patient population.  The strongest evidence for such an approach, however limited at this time, is for sulpiride. Unfortunately, sulpiride’s availability is, at present, restricted and in Canada can only be obtained through  Health Canada’s special access program. The United States does not have a similar program for sulpiride and as such is not available.  Thus, further controlled studies are required to determine whether or not the same gains can be made by substituting other potent antipsychotics for sulpiride (e.g., pimozide, haloperidol).  The augmentation of the second-generation agents with a conventional agent is generally discouraged due to lack of evidence and risk for increased adverse effects, cost and noncompliance.  Patients should be given adequate treatment trials and if poor response is observed, clozapine should be considered.  The practice of using “prn” doses of conventional antipsychotics during initiation of a second-generation agent is thought to be a rational approach for acute psychotic episodes when agitation or aggression is present.  Clinicians should; however, assess whether benzodiazepines would be a more appropriate choice in these cases.  Finally, combination antipsychotic use as a result of incomplete cross-titration is not recommended. Instead, switching should always be completed to allow for proper assessment of the new agent.

Latest Posts