Background: Levomepromazine is an 'older' typical antipsychotic medication licensed for use in schizophrenia but sparingly prescribed in the United Kingdom.
Objectives: To determine the clinical effects and safety of levomepromazine compared with placebo or antipsychotic medications for schizophrenia and schizophreniform psychoses.
Search strategy: We searched the Cochrane Schizophrenia Group's Register (December 2008) which is based on regular searches of, amongst others, BIOSIS, CENTRAL CINAHL, EMBASE, MEDLINE and PsycINFO. References of all identified studies were inspected for further trials. We also contacted relevant pharmaceutical companies for additional information.
Selection criteria: All randomised trials comparing levomepromazine with placebo or other antipsychotics for schizophrenia and schizophreniform psychoses were included.
Data collection and analysis: Data were extracted independently. For dichotomous outcomes, we calculated relative risk (RR) (random-effects model), 95% confidence intervals (CI) and, where appropriate, number needed to treat (NNT) was calculated. We avoided the use of number needed to harm (NNH), as an alternative we used number needed to treat for an additional beneficial outcome (NNTB) and number needed to treat for an additional harmful outcome (NNTH) to indicate the direction of effect. For continuous outcomes, we calculated weighted mean differences (WMD).
Main results: The review currently includes 4 RCTs with 192 participants. For our primary outcome of leaving the study early, levomepromazine was not significantly different compared with other antipsychotics. The levomepromazine arm was significantly better on CGI severity compared with chlorpromazine (n=38, 1 RCT, WMD -0.80 CI -1.51 to -0.09). Risperidone was better for CGI endpoint scores (n=42, 1 RCT, RR 2.33 CI 1.11 to 4.89, NNT 3 CI 2 to 10) compared with levomepromazine. Recipients given levomepromazine had a better BPRS endpoint score (n=38, 1 RCT, WMD -9.00, CI -17.46 to -0.54) and PANSS total score (n=38, 1 RCT, WMD -15.90, CI -30.30 to -1.50) than chlorpromazine. Risperidone recipients noticed a significant difference for the outcome 'at least 20% reduction' on BPRS endpoint score (n=42, 1 RCT, RR 3.33 CI 1.07 to 10.42, NNT 3 CI 2 to 14) compared with levomepromazine. Levomepromazine caused less tremor (n=41, 1 RCT RR 0.12 CI 0.02 to 0.87 NNTB 3 CI 2 to 8), less antiparkinsonian medication administration (n=79, 2 RCTs, RR 0.39 CI 0.17 to 0.90, NNTB 5, CI 2 to 21) compared with haloperidol. Levomepromazine caused less akathisia compared with chlorpromazine, but more hypotension compared with risperidone (n=42, 1 RCT, RR 2.50 CI 1.21 to 5.18, NNTH 3, CI 2 to 7). Dizziness was common with levomepromazine compared with other antipsychotic medications.
Authors' conclusions: Available data does not enable us to confidently comment on the effectiveness of levomepromazine for schizophrenia. Larger, more robust, studies comparing levomepromazine with other antipsychotics including clozapine are much needed.