Fifty-two RCTs (n=12,649) were included.
Details of the included trials are available on the BMJ website. See Web Address at end of abstract. Most trials were short term (median follow-up 6.5 weeks) although 5 trials provided follow-up data for one year or more. Most trials compared the effectiveness of atypical antipsychotics with haloperidol. There were few data on quality of life, specific side-effects or cost effectiveness, so these outcomes have not been reported in the review.
Amisulpride: compared to conventional antipsychotics, the standardised weighted mean difference (WMD) was -0.35 (95% CI: -0.52, -0.18) in favour of amisulpride. Fixed-effect pooled OR for drop-out was 0.55 (95% CI: 0.38, 0.79) and the random-effects pooled OR was 0.54 (95% CI: 0.33, 0.85). The standardised WMD on the Simpson Angus Scale for extrapyramidal side-effects was -0.44 (95% CI: -0.26, -0.61).
Clozapine: the standardised WMD for the overall symptom score in short-term trials was -0.68 (95% CI: -0.82, -0.55) in favour of clozapine. Patients allocated to clozapine were less likely to drop out (using the fixed-effect model, OR 0.52, 95% CI: 0.40, 0.67) although the difference became non significant with the random-effects model (OR 0.69, 95% CI: 0.45, 1.19). Results of the 2 long-term trials were inconsistent.
Olanzapine: for overall symptom score, compared to conventional antipsychotics, the standardised WMD was -0.22 (95% CI: -0.30, -0.14) in favour of olanzapine. The pooled ORs for drop-out were 0.52 (95% CI: 0.44, 0.61) and 0.63 (95% CI: 0.40, 1.17) with the fixed-effect and random-effects models, respectively. There was a 4.8% (95% CI: 3.1, 6.5) reduction in dystonia and a 14.1% (95% CI: 11.0, 17.2) reduction in akathisia with olanzapine. Olanzapine was associated with a 12% (95% CI: 8, 15) increase in excessive appetite, compared with haloperidol.
Quetiapine: there was no difference in the overall symptom score between quetiapine and conventional antipsychotics (standardised WMD -0.03, 95% CI: -0.23, +0.18)), nor was there any reliable evidence of a reduced drop-out rate (OR 0.70, 95% CI: 0.46, 1.06).
Risperidone: in 6 short-term trials, the fixed-effect standardised WMD for overall symptom scales was -0.15 (95% CI: -0.27, -0.04) in favour of risperidone, and the random-effects standardised WMD was -0.16 (95% CI: -0.47, +0.16). Substantial heterogeneity was observed between trials; potential explanations being interactions between the use of risperidone and specific patient groups and variability in the comparators. The fixed-effect OR (10 trials) for drop-out from risperidone was 0.59 (95% CI: 0.46, 0.74) and the random-effects OR was 0.62 (95% CI: 0.31, 1.34).
Standardised WMD was -0.39 (95% CI: -0.51, -0.27) in favour of risperidone for the Parkinson scale, -0.26 (95% CI: -0.39, -0.12) for the dystonia scale and -0.16 (95% CI: -0.28, -0.04) for the dyskinesia scale.
In 2 long-term naturalistic trials, the pooled standardised WMD at 12 months for the overall positive and negative syndrome scale score was -0.40 (95% CI: -0.27, -0.54) in favour of risperidone.
Sertindole: results of trials are not provided in the review since the drug is not currently available in the UK.
No evidence was identified to suggest that any individual atypical antipsychotic had a specific effect on either positive or negative symptoms (benefits were evenly spread between positive and negative symptoms).
The dose of haloperidol significantly affected outcome in the 23 trials in which it was used. Within the range of mean doses of haloperidol reported (approximately 6 to 22.5 mg/day), meta-regression identified a significant advantage for atypical antipsychotics as the dose of haloperidol increased (regression coefficient -0.021, 95% CI: -0.003, -0.038). The observed advantage in favour of the atypical drug disappeared as the dose of haloperidol decreased. A similar effect was seen for chlorpromazine (7 trials). Further to this, trials using 12 mg of haloperidol or less were compared to those using a higher dose: in the 12 mg or less group, the random-effects standardised WMD was -0.09 (95% CI: -0.07, -0.26), whereas in the greater than 12 mg group, it was -0.28 (95% CI: -0.13, -0.44). There was no difference in drop-out rates between atypical antipsychotics and haloperidol in the 12 mg or less group (pooled risk difference -0.1%, 95% CI: -4.6, +4.4), but the pooled drop-out rate in the greater than 12 mg group was -8.3% (95% CI: -1.3, -15.2).
These results were unaffected by the removal of trials of treatment- resistant patients, those taking sertindole or long-term trials.
Two large trials indicated that reduced extrapyramidal side-effects remained apparent at lower doses of comparator drugs.
Olanzapine versus risperidone (2 trials): no difference in efficacy was found. Olanzapine was better tolerated, equivalent to a difference in drop out of approximately 7% (95% CI: 0.4, 13.6).
Risperidone versus clozapine: 2 trials were found (n=146) but data were not used due to insufficient power.
Indirect comparison in meta-regression models did not identify any individual atypical antipsychotic to be more or less effective when the dose of the comparator drug was taken into account.