Evid Based Mental Health 15:47 doi:10.1136/ebmental-2012-100637
  • Therapeutics

Review: extrapyramidal side effects vary between different second-generation antipsychotic drugs



Are the second-generation antipsychotic drugs (SGAs) different in their capacity to induce extrapyramidal side effects (EPS)?


Primary outcome: use of antiparkinson medication at least once (used as a global measure for EPS).



Systematic review and meta-analysis.

Data sources

The register of the Cochrane Schizophrenia Group was searched up to May 2007, and MEDLINE was searched up to July 2009, with no language restrictions. The search terms used were all 36 possible combinations of the names of the SGAs, including trade names. The SGAs in question were amisulpride, aripiprazole, clozapine, olanzapine, quetiapine, risperidone, sertindole, ziprasidone and zotepine. Manufacturers were contacted for details of published and unpublished studies, and study authors were contacted for missing data in their studies.

Study selection and analysis

Blinded randomised controlled trials (RCTs) directly comparing different orally administered SGAs for the treatment of schizophrenia or related disorders. Only studies meeting the quality criteria A (adequate randomisation) and B (primarily studies stated to be randomised without further details) of the Cochrane Handbook 2005 were included. At least three reviewers independently extracted data from the studies. Study results were pooled using random effects methods. Heterogeneity was assessed using the χ2 test and I2 statistic. Analyses were carried out using Review Manager Version 5.0 and STATA Version 7.

Main results

Fifty-four trials (116 relevant arms) met inclusion criteria. Statistically significant differences in use of antiparkinson medication were observed between many of the SGAs (see Webextra table). Risperidone increased use of antiparkinson medication compared with olanzapine, ziprasidone, quetiapine and clozapine. Quetiapine reduced use of antiparkinson medication compared with ziprasidone, olanzapine and risperidone. Amisulpride did not differ from the drugs to which it was compared (olanzapine, risperidone and ziprasidone). Aripiprazole increased use of antiparkinson medication compared with olanzapine, and was not significantly different to risperidone. Clozapine reduced use of antiparkinson medication compared with risperidone and zotepine, and was not significantly different to olanzapine and ziprasidone. Olanzapine had mixed results; it increased use of antiparkinson medication compared with quetiapine, but reduced use compared with aripiprazole, risperidone and ziprasidone. Findings are summarised in the Webextra table.


There are differences between second-generation antipsychotics in their capacity to induce extrapyramidal side effects requiring antiparkinson medication.

Abstracted from

[Abstract/FREE Full text]


  • Sources of funding Technische Universitat München; German Federal Ministry of Education and Research; National Institute of Mental Health, Advanced Center for Intervention and Services Research Center; Maryland Psychiatric Research Center; EliLilly; Sanofi-Aventis.


  1. Stanley N Caroff
  1. Department of Psychiatry, VAMC/Univ Pennsylvania, Philadelphia, Pennsylvania, USA

Rummel-Kluge et al provide important evidence on the RR of extrapyramidal side effects (EPS) among second-generation antipsychotic drugs (SGAs) based on a meta-analysis of clinical trials. EPS remain clinically relevant because they can be mistaken for or worsen psychotic symptoms. They are sometimes lethal, irreversible, necessitate additional burdensome side effects from antiparkinsonian agents, potentially stigmatising, and can influence treatment adherence and outcome.

Building on previous work showing that older high-potency antipsychotics are significantly more likely to cause EPS,1 the authors tested whether there are differences in liability among SGAs. Although they found statistically significant differences in the need for antiparkinsonian drugs, the authors note that the effect sizes were small and diminished when controlled for dose, consistent with previous data that failed to find differences in the risk of EPS between SGAs and placebo.2 Although these differences in EPS may be less important for drug choice compared with older, high-potency drugs, they are essential to be considered in high-risk groups, for example, in patients with histories of EPS or basal ganglia disease.

The authors discuss the limitations of antiparkinsonian medication as a proxy measure of EPS, especially carryover or withdrawal effects in patients who entered trials with EPS and switched from prebaseline treatment. In the CATIE trial, similar differences were found among SGAs in administration of antiparkinsonian medications, but these differences resolved when a broader analysis was applied, including adverse event reports, rating scale criteria and treatment discontinuation due to emergent EPS.3

Nevertheless, the small but significant differences between SGAs in antiparkinsonian medication underscore the need to prescribe antipsychotics based on careful assessment of individual patient psychopathology, treatment history and vulnerabilities to adverse effects, matched with the specific pharmacologic profile of each drug.


  • Competing interests None.