Adverse Drug Events Related to Ziprasidone: A Meta-analysis of Randomized, Placebo-Controlled Trials.

Study Objective: To assess the drug-related risk of adverse events associated with ziprasidone.

Design: Meta-analysis of 19 randomized, placebo-controlled trials.

Patients: A total of 4132 adults taking oral ziprasidone who had adverse-event data reported in the studies identified.

Measurements And Main Results: A systematic review (January 1996-October 2010) was conducted by using the EMBASE and MEDLINE databases to identify Cochrane reviews, controlled clinical trials, meta-analyses, randomized controlled trials, and systematic reviews; studies were limited to those published in English and those conducted in humans. The www.ClinicalTrials.gov Web site was also searched for ziprasidone studies. A total of 887 citations were reviewed; 31 articles met the criteria for inclusion, of which 19 were included in the final analysis. Data were combined for the meta-analysis by using the Mantel-Haenszel method, random-effects model at 95% confidence. The overall rate of treatment-emergent adverse events for ziprasidone was 73% compared with 60% for patients receiving placebo (p<0.0001). Adverse events with the greatest frequency included somnolence (21%), extrapyramidal symptoms (13%), headache (13%), insomnia (11%) and respiratory disorders (10%). Adverse events with highest risk, evaluated by using the risk difference (RD) summary statistic (adverse events due to the drug itself and not the placebo effect), were somnolence (RD 14, 95% confidence interval [CI] 7-21), extrapyramidal symptoms (RD 6, 95% CI 1-10), asthenia (RD 5, 95% CI 1-8), weight gain (RD 4, 95% CI 2-7), dizziness (RD 4, 95% CI 2-6), and dyspepsia (RD 4, 95% CI 1-6). Adverse events reported but likely not caused by ziprasidone included headache (RD 0, 95% CI -2-3), QTc interval greater than 480 msec (RD 0, 95% CI -1-1), diarrhea (RD 0, 95% CI -2-2, and abdominal pain (RD 0, CI -2-2).

Conclusion: Ziprasidone use increased the risk of 18 specific adverse events when compared with placebo. Small reductions in risk for insomnia, pain, and agitation are likely among patients with schizophrenia but not those with bipolar disorder. The results of the study are limited by the concomitant use of other drugs allowed during the trials, underreporting of adverse events in the clinical trials, and the short length of the trials (most 3-6 wks).