Perphenazine for schizophrenia.

Perphenazine is an old phenothiazine antipsychotic with a potency similar to haloperidol. It has been used for many years and is popular in the northern European countries and Japan. To examine the clinical effects and safety of perphenazine for those with schizophrenia and schizophrenia-like psychoses. We searched the Cochrane Schizophrenia Group's register (June 2001), references of all included studies and contacted pharmaceutical companies and authors of included studies in order to identify further trials. We included all randomised controlled trials that compared perphenazine with other treatments for people with schizophrenia and/or schizophrenia-like psychoses. We excluded trials of depot formulations of perphenazine. Two reviewers independently inspected citations and, where possible, abstracts. We ordered papers, inspected and quality assessed them. We extracted data, again working independently. If loss to follow up was greater than 50% we considered results as 'prone to bias'. For dichotomous data we calculated the relative risk (RR), the 95% confidence interval (CI) and, where appropriate, the number needed to treat/harm (NNT/H) on an intention-to-treat basis. For continuous data, we calculated weighted mean differences (WMD). The review currently includes 25 studies with 2478 patients, 2285 of whom had been randomised to interventions that were relevant for the review such as perphenazine, other antipsychotic drugs or placebo. The trials were carried out between 1961 and 1993. All but one trial were short term with a duration of between ten days and 12 weeks. Descriptions of allocation and blinding were usually incomplete. Six studies (n=300) compared perphenazine with placebo. Perphenazine was associated with fewer participants leaving the trials early due to relapse or worsening of symptoms (n=84, RR 0.1 CI 0.03 - 0.4, NNT 2 CI 1 to 20). Twenty studies compared perphenazine (n=738) with other antipsychotics (n=1278). Perphenazine seemed as effective as other antipsychotics ('global state unimproved or worse' n=1327, RR 1.0 CI 0.9 to 1.2). We found no clear differences in terms of specific aspects of efficacy, behaviour or tolerability. However, interpretation of findings of the review was limited by poor reporting and the use of 24 different comparator antipsychotics in the 20 trials. Although perphenazine has been randomised for more than 40 years, incomplete reporting and the variety of comparators used make it impossible to draw clear conclusions. At best we can say that perphenazine showed similar effects and adverse events as several of the other pooled antipsychotic drugs. Since perphenazine is a relatively inexpensive and frequently used compound, further trials are justified to clarify the properties of this classical antipsychotic drug.