NPC Archive Item: Aripiprazole in schizophrenia: how does it compare with olanzapine and risperidone▼?

NOTE – This is an archive post from the NPC and has not been updated since first publication. Therefore, some hyperlinks may no longer be working.
MeReC Rapid Review NPC Logo

22 February 2010

A Cochrane review suggests that aripiprazole may be slightly less efficacious than olanzapine and similar to risperidone▼ in the treatment of schizophrenia. However, the clinical significance of this difference is unclear. It appears to have a slightly different side-effect profile which may offer advantages for some patients.

Level of evidence:
Level 2 (limited quality patient-oriented evidence) according to the SORT criteria.

This study supports the NICE schizophrenia guideline and health professionals should continue to follow this. Difference in efficacy between the different antipsychotics (apart from clozapine) are small – choice of antipsychotic should be made in partnership with the service user (and carer if appropriate) taking into account the relative potential of individual antipsychotics to cause extrapyramidal side effects (such as akathisia), metabolic side effects (such as weight gain), and other side effects (including unpleasant subjective experiences).

Aripiprazole may be slightly less efficacious than olanzapine, although the clinical significance of any difference is unclear. It seems to have similar efficacy to that of risperidone. Like all antipsychotics, aripiprazole can cause side effects although its profile in this regard may suit some service users more than others. A patient decision aid is available on the schizophrenia floor of NPCi that includes a chart of the relative side effect profiles of different antipsychotics, including aripiprazole, which may be helpful when considering which antipsychotic is most appropriate for a patient to try.

What is the background to this?
Aripiprazole is a relatively new second generation antipsychotic (SGA or atypical). A previous Cochrane review considered the randomised controlled trial (RCT) evidence for the effectiveness of aripiprazole, primarily looking at prevention of relapse. This found that, although aripiprazole might be effective for the treatment of schizophrenia, as indicated from studies versus placebo, it did not differ greatly from first generation antipsychotics (FGA or typicals) and other SGAs with respect to treatment response, efficacy or tolerability. A subsequent Cochrane Review specifically considered the clinical effects of aripiprazole in comparison with FGAs. It found that there was little difference between aripiprazole and FGAs in terms of efficacy. However, it also found some advantages for aripiprazole in terms of tolerability, although attrition rates for both groups were high. Advantages were identified for aripiprazole with regard to extrapyramidal side effects (EPS), development of hyperprolactinaemia, sinus tachycardia, and blurred vision. However, aripiprazole was more likely to cause dizziness and nausea.

The present Cochrane review specifically considered the clinical effects of aripiprazole in comparison with other SGAs in schizophrenia. The review identified only four RCTs trials (two versus risperidone, and two versus olanzapine).

What does this study claim?
Aripiprazole was less efficacious than olanzapine with regard to changes in general mental state (2 RCTs, n=794, weighted mean difference [WMD] Positive And Negative Syndrome Scale [PANSS] total score 4.96, 95%CI 1.85 to 8.06). However it was associated with fewer side effects, such as cholesterol increase, weight gain, sedation and prolactin associated side effects. Compared with risperidone, there was no statistically significant difference in efficacy (2 RCTs, n=372, WMD in PANSS total score 1.50, 95%CI –2.96 to 5.96). Dystonia, QTc abnormalities, prolactin and cholesterol increase were less frequent, while tremor was more frequent with aripiprazole than with risperidone.

So what?
This study suggests that aripiprazole may have slightly worse efficacy than olanzapine, and similar efficacy to risperidone, in the treatment of schizophrenia, but may be better tolerated with regard to certain side effects. However, there are many limitations which need to be considered when interpreting the results of these studies.

Although the study identified changes in general mental state (PANSS) favouring olanzapine over aripiprazole, no clinically important difference in response (as defined in original studies) or global clinical improvement (Clinical Global Impression Scale [CGI]) was found. The clinical significance of any differences are, therefore, difficult to ascertain. Three of the studies were short-term (four to six weeks) and one was medium term (26 weeks). The withdrawal rates for all treatment groups were high (overall 38.5%, and as high as 72% in the 26-week study). Although all trials were randomised and double-blind, details were rarely presented and it is unclear whether this was done appropriately. Also, in view of the different side-effect profiles, it is not certain whether or not blinding was effective, at least for subjective outcomes. There was also a high risk of bias from selective reporting; in three trials only adverse events that occurred in at least 5–10% of patients were reported, so rare but important adverse events may have been missed. All trials were sponsored by the manufacturer of aripiprazole.

The authors suggested the need for more long-term studies (i.e. at least six months) with a wider range of SGAs. CGI was suggested as the primary outcome, but also relapse, leaving the study early (for any reason), adverse events, service outcomes (e.g. hospitalisation), employment, family and patient satisfaction, as well as changes in mental state.

How does this compare with other studies?
Not considered in this review is a 28-week, double blind, RCT comparing aripiprazole with olanzapine in schizophrenia, sponsored by the manufacturer of olanzapine, which was not published when the Cochrane Review was carried out. This study (n=566) did not find a statistically significant difference in time to all-cause discontinuation (P=0.067) or all-cause discontinuation rate (olanzapine 42.7% vs. aripiprazole 50.2%, P=0.053). Consistent with the Cochrane review, olanzapine did appear to have advantages with regard to efficacy-related outcomes (e.g. discontinuation rate due to efficacy-related reasons: olanzapine 8.9% vs. aripiprazole 16.8%, P=0.006), and disadvantages with regard to some side effects.  For example, increased appetite and weight increase were more frequent with olanzapine: mean weight increase olanzapine 3.4kg vs. aripiprazole 0.3kg for aripiprazole-treated patients, P<0.001; at least 7% body weight gain 40.3% vs. 16.4%, respectively, number needed to harm (NNH 4), P<0.001. Other treatment-emergent side effects (occurring in more than 5% of patients in olanzapine group) that were statistically significantly more frequent with olanzapine than aripiprazole were somnolence, sedation and depression. However, more patients experienced insomnia and upper abdominal pain with aripiprazole than with olanzapine. Olanzapine was also associated with greater decreases in prolactin levels and greater increases in fasting glucose levels. Whereas there were decreases in mean fasting total cholesterol and triglyceride levels with aripiprazole, there were increases in the mean values for these parameters with olanzapine (all comparisons P<0.05).

NICE published a clinical guideline for the management of schizophrenia in March 2009. NICE found that there was little evidence of clinically significant differences between oral antipsychotics in terms of efficacy outcomes from RCTs or overall effectiveness as measured in pragmatic studies such as CATIE and CUTLASS. The exception to this is clozapine which does appear to have some effectiveness advantages. NICE recommended that choice of antipsychotic should be made in taking into account the relative potential of individual antipsychotics to cause EPS (such as akathisia), metabolic side effects (such as weight gain), and other side effects (including unpleasant subjective experiences). The present review and subsequent study, did not identify any significant differences between aripiprazole and either olanzapine or risperidone with respect to EPS, but did identify some advantages for aripiprazole with regard to metabolic side effects (notably weight gain versus olanzapine).

Study details
Komossa K, Rummel-Kluge C, Schmid F, et al. Aripiprazole versus other atypical antipsychotics for schizophrenia. Cochrane Database of Systematic Reviews 2009, Issue 4. Art. No.: CD006569. DOI: 10.1002/14651858.CD006569.pub3

Systematic review and meta-analysis of clinical trials of aripiprazole compared with SGAs.

Studies included:
Two RCTs vs. olanzapine (n=703, 6 weeks; n=317, 26 weeks) and two RCTs vs. risperidone (n=83, 4 weeks; n=404, 4 weeks)

People with a diagnosis of schizophrenia or schizophrenia-like psychoses

Flexible dose (both groups) vs. olanzapine; fixed dose (both groups) vs. risperidone
Outcomes and results:

Aripiprazole vs. olanzapine:
There was no statistically significant difference in response (as defined by the original studies), the proportion leaving the study early, and global state (CGI). There was a statistically significantly greater change in PANSS in favour of olanzapine (n=794, WMD total score 4.96, 95%CI 1.85 to 8.06).

There was no statistically significant difference in mean change in various EPS, change in QTc interval, or blood glucose levels, but with olanzapine there were significant increases in the proportion of patients with increased cholesterol levels (relative risk [RR] 0.32, 95%CI 0.19 to 0.54, NNH 4), with increased prolactin levels (RR 0.27, 95%CI 0.12 to 0.60, NNH 8), sedation (RR 0.33, 95%CI 0.18 to 0.62, NNH 7) and with more than a 7% increase in body weight (RR 0.37, 95%CI 0.24 to 0.58, NNH 4).

Aripiprazole vs. risperidone
There was no statistically significant difference in response (as defined by the original studies), the proportion leaving the study early, global state (CGI), or in the change in PANSS (n=372, WMD total score 1.50, 95%CI –2.56 to 5.96).

There was no statistically significant difference in the number of people reporting at least one adverse event, QTc prolongation, blood glucose levels, weight gain, or mean scores on EPS rating scales for dyskinesia, akasthisia, or general EPS. Significantly fewer patients receiving aripiprazole had QTc abnormalities, as indicated by a smaller change of the QTc interval in ms from baseline (WMD –7.19, 95%CI –2.19 to –12.19). Dystonia occurred more frequently with risperidone (RR 0.14, 95%CI 0.05 to 0.41, NNH 8), whereas tremor occurred more frequently with aripiprazole (RR 4.66 95%CI 1.11 to 19.59, NNH 14). There was a significantly greater proportion of patients with increased prolactin levels in the risperidone group (RR 0.04, 95%CI 0.02 to 0.08, NNH 1). Changes in cholesterol levels were also greater in the risperidone group (WMD 22.3mg/dL, 95%CI 4.9 to 39.7)

Sponsorship: The review was sponsored by the Psychiatrische Klinik, Klinikum rechts der Isar, and Bundesministerium für Bildung und Forschung, Germany. All included studies were sponsored by the manufacturer of aripiprazole.

More information on the management of schizophrenia can be found on the schizophrenia section of NPC.

Please comment on this blog in the NPC discussion rooms, or using our feedback form.

Make sure you are signed up to NPC Email updates — the free email alerting system that keeps you up to date with the NPC news and outputs relevant to you