Trifluoperazine is an inexpensive accessible 'high potency' antipsychotic drug, widely used to treat schizophrenia or related psychoses. To estimate the effects of trifluoperazine compared with placebo and other drugs. Searches of the Cochrane Schizophrenia Group's register of trials (March 2002), supplemented with hand searching, reference searching, personal communication and contact with industry. All clinical randomised trials involving people with schizophrenia and comparing trifluoperazine with any other treatment. Studies were reliably selected and quality rated and data was extracted. For dichotomous data, relative risks (RR) were estimated, with 95% confidence intervals (CI). Where possible, we undertook intention-to-treat analyses. For statistically significant results, the number needed to treat (NNT) was calculated. We estimated heterogeneity (I-square technique) and publication bias. 1162 people from 13 studies were randomised to trifluoperazine or placebo. For global improvement, small short-term studies favoured trifluoperazine (n=95, 3 RCTs, RR 0.62 CI 0.49 to 0.78 NNT 3 CI 2 to 4). Loss to follow up was about 12% in both groups (n=280, 7 RCTs, RR 0.99 CI 0.62 to 1.57) and more people allocated trifluoperazine used antiparkinson drugs to alleviate movements disorders compared with placebo (n=195, 4 RCTs, RR 5.06 CI 2.49 to 10.27, NNH 4 CI 2 to 9). 2230 people from 49 studies were randomised to trifluoperazine or another older generation antipsychotic. Trifluoperazine was not clearly different in terms of 'no substantial improvement' (n=1016, 27 RCTs, RR 1.06 CI 0.98 to 1.14) or leaving the study early (n=930, 22 RCTs, RR 1.15 CI 0.83 to 1.58). Almost identical numbers of people reported at least one adverse event (60%) in each group (n=585, 14 RCTs, RR 0.99 CI 0.87 to 1.13), although trifluoperazine was more likely to cause extrapyramidal adverse effects overall when compared to low potency antipsychotics such as chlorpromazine (n=130, 3 RCTs, RR 1.66 CI 1.03 to 2.67, NNH 6 CI 3 to 121). One small study (n=38) found no clear differences between trifluoperazine and the atypical drug, sulpiride. Although there are shortcomings and gaps in the data, there appears to be enough consistency over different outcomes and periods to confirm that trifluoperazine is an antipsychotic of similar efficacy to other commonly used neuroleptics for people with schizophrenia. Its adverse events profile is similar to that of other drugs. It has been claimed that trifluoperazine is effective at low doses for patients with schizophrenia but this does not appear to be based on good quality trial based evidence.