To assess whether (1) higher doses of neuroleptic drugs for maintenance treatment of psychosis are any more effective than lower doses, and (2) higher doses are safer.

MEDLINE was searched from 1966 to 1989. Review articles and reference lists from other papers were handsearched.

Study designs of evaluations included in the review

Randomised controlled trials (RCTs) of at least a 4-week duration.

Specific interventions included in the review

Maintenance treatment using different doses of neuroleptic drugs, for which a conversion factor to chlorpromazine was available. The drugs studied were: phenothiazines, chlorpromazine, butaperazine, fluphenazine hydrochloride, trifluopromazine, fluphenazine, thiotixene, fluphenazine enanthate, fluphenazine decanoate, loxapine, haloperidol, thioridazine, pimozide and flupenthixol decanoate.

Participants included in the review

Adult patients over 16 with chronic psychotic disorders. The majority of patients had a diagnosis of schizophrenia, which was treated in both in- and out-patient settings.

Outcomes assessed in the review

Clinical improvement was defined as improvement on a clinical global impression scale, or no psychotic relapse or no rehospitalisation. Any adverse reaction, distinguishing between general and neurological reactions.

How were decisions on the relevance of primary studies made?

The authors do not state how the papers were selected for the review, or how many of the authors performed the selection.

The quality of studies was summarised using a combination score with the following key criteria: selection description, therapeutic regimen definition, blinding of observers, blinding of randomisation process, prior estimate of sample size, reporting of withdrawals, handling of withdrawals, and reporting of side-effects. Two readers, blinded to everything but the methodology section of the study, assessed quality according to predefined criteria and determined a consensus estimate.

The authors do not state how the data were extracted for review, or how many of the authors performed the extraction. The authors attempted to undertake analysis based on intention to treat, so patients who dropped out for lack of symptomatic control were considered as unimproved, and those who dropped out because of adverse reactions were considered to suffer side-effects. Drug dosages were converted to a chlorpromazine equivalent to enable comparison, and were categorised into 4 quartile bands.

How were the studies combined?

The studies were combined using a random-effects meta-analysis with a covariate adjusting for dose. The models investigated changes in the proportion of people showing clinical improvement and the average number of side-effects per patient. The analysis was weighted by sample size.

How were differences between studies investigated?

The authors considered differences in the 4 categories of dose using a regression model.

Twenty-two RCTs comparing high and low dosage of drugs, of which 10 included information on side-effects. A total of 1,638 patients were considered; 904 patients were considered in the analysis of side-effects.

For doses beyond the range 166 to 375 mg, no significant increase in effect could be detected, compared to the highest dose group. The proportion of patients improving on doses less than 166 mg was 23% (95% confidence interval, CI: 7, 39) lower than those on doses greater than 830 mg. There was a non significant difference of 6% (95% CI: -4, 16) between those receiving 166 to 375 mg and the highest dose. Patients on lower doses (less than 375 mg) experienced significantly fewer side-effects. Comparing doses less than 375 mg to those greater than 830 mg, there was 18% (95% CI: 7, 29) less parkinsonism, on average 0.359 (95% CI: 0.16, 0.56) fewer neurological side-effects and 1.009 (95% CI: -1.21, -0.81) fewer side-effects of any kind.

There is no therapeutic advantage to be gained beyond the dose range of 166 to 375 mg equivalent of chlorpromazine. Adverse reactions increase significantly above these doses. On average, a dose up to 165 mg equivalent is significantly less effective, compared with a dose of more than 830 mg, and no significant therapeutic advantage is found above 375 mg equivalent of chlorpromazine. Both neurological and overall side-effects were significantly increased with higher doses.

This is a thorough and methodologically-rigorous review, although the results of the analysis are not clearly presented. The dose levels identified in the studies need to be considered in the context of current dosage recommendations: the daily maintenance dose recommended for treatment of chronic psychosis is 75 to 300 mg chlorpromazine (see Other Publications of Related Interest).

Moderate doses of neuroleptic drugs (approximately 165 to 375 mg equivalent of chlorpromazine) should be favoured for the maintenance treatment of chronic psychosis.

National Center for Health Services Research and Health Care Technology Assessment (USA).

British Medical Association, Royal Pharmaceutical Society of Great Britain. British National Formulary 1994 Sep;28.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.