A meta-analysis and critical review of the effects of conventional neuroleptic treatment on cognition in schizophrenia: opening a closed book
Mishara AL and Goldberg TE;
Commented by , 24 May 2004
Aim of the study
The introduction of the “atypical” antipsychotics has raised our expectations concerning the effects of antipsychotic drugs. A major hope is that the new compounds improve the cognitive functions of patients with schizophrenia. Recently a number of studies have therefore compared the cognitive effects of new generation and conventional antipsychotics. However, the effects of the comparators - conventional antipsychotics – compared to placebo on cognition had never been clear.
Many narrative reviews had concluded that conventional antipsychotics had either no effect on cognition or even deteriorated cognitive function. However, a systematic review and quantitative meta-analysis had never been available. The authors filled this gap by re-analysing the question of whether conventional antipsychotics have enhancing, neutral or deteriorating effects on cognition.
Method
The authors systematically searched the literature with the electronic databases Medline, Science Direct, Current Contents, PsychInfo, PsychLit and cross-referencing of the reference lists of identified trials. Included were trials in adult patients with schizophrenia that compared the effects of conventional antipsychotics with placebo or no medication on cognition. A standard measure, Cohen’s d, was used as an effect size estimator which corresponds to the difference between the means in cognitive parameters of patients taking conventional antipsychotics and placebo, divided by the pooled SD.
When a study reported several measures of cognition, the effect sizes of the different measures were averaged in order to avoid that a study was given undue weight simply because it used a number of cognitive outcome parameters. Studies with large numbers of participants were given more weight in the meta-analysis than studies with small numbers using an inverse of the variance method. The results of the individual trials were meta-analytically combined using a random-effects model that takes the variability of study methods into account.
Results
34 studies with a total of 208 effect sizes on cognition were included in the analysis. The pooled effect size summarizing all studies was d=0.22 with a 95% confidence interval of 0.10 to 0.34. This result demonstrates a statistically significant, positive effect on cognition. According to the authors it is of low moderate size. A number of potential moderator variables (especially drug dose, but also methodological quality of the trials, study designs, publication year etc.) were analysed as to whether they had an impact on the results, but no statistically significant effects were found.
The authors also calculated pooled effect sizes on individual neurocognitive domains. Conventional antipsychotics had statistically significant positive effects on most individual cognitive areas (effect sizes between 0.13 and 0.29) with the exception of motor function which was affected negatively (d = -0.11).
Discussion
This meta-analysis fills an important gap in our understanding of the effects of conventional antipsychotics. For a long time there had been a great insecurity on the cognitive effects of these compounds. There was a widely held belief that typical deteriorate cognitive function, but this assumption was rejected by the meta-analysis. This result seems plausible, because one of the core symptoms of acutely psychotic patients is formal thought disorder, i.e. very abnormal cognitive functioning, and this symptom is improved by conventional antipsychotic agents.
On the other hand the results of any meta-analysis depend on the quality of the individual studies. For example, a number of studies in the meta-analysis that did not show beneficial effects of conventional antipsychotics were relatively large, while many positive studies were small. It is difficult to explain these discrepancies from the presented data.
The authors therefore suggest that the impact of conventional medication on cognitive function should be re-evaluated rather than considering there meta-analysis to be a firm result. Furthermore, the question of course remains whether atypical antipsychotics improve cognitive functions even more than conventional compounds.
Although reviews and experts in the field support this view, a number of issues such as the influence of high doses of conventional antipsychotics used in the studies remain.