Haloperidol Versus 5-HT3 Receptor Antagonists for Postoperative Vomiting and QTc Prolongation: A Noninferiority Meta-Analysis and Trial Sequential Analysis of Randomized Controlled Trials.

Abstract

Haloperidol is an antipsychotic with well-known antiemetic potential. It is underutilized for postoperative nausea vomiting due to reported corrected QT interval (QTc) prolongation. This meta-analysis evaluates its safety and efficacy as an antiemetic in the perioperative period. Trials comparing haloperidol to 5-HT3 -receptor antagonists (5-HT3 -RA) for 24 postoperative vomiting incidences published up to May 2017 were searched in the medical database. Comparisons were made for antiemetic efficiency variables (vomiting incidence, rescue antiemetic need, and patients with complete response) during early (until 6 hours) and late postoperative phases. Eight randomized controlled double-blinded trials were included in the final analysis. Twenty-four-hour vomiting incidence was similar in groups (fixed effects, P = 0.52, I2 = 0%). Trial-sequential analysis confirmed noninferiority of haloperidol over 5-HT3 -RAs (α = 5%, β = 20%, δ = 10%), with "information size" being 859 (required > 812). Pooled results did not demonstrate superiority/inferiority of 5-HT3 -RAs over haloperidol in all other antiemetic efficacy variables (early and delayed). Negligible heterogeneity was found in all the comparisons made. Pooled Mantel Haenszel odds ratio for QTc prolongation was equivalent in both groups (fixed effects, P = 0.23, I2 = 0%). The mean dose of haloperidol used was 1.34 mg, and no trial reported extrapyramidal side effects. Trial-sequential analysis showed statistical equivalence (α = 5%, β = 20%, δ = 10%), with information size being 745 (required > 591). Publication bias was unlikely (Egger test, X-intercept = 2.07, P = 0.10). We conclude that haloperidol is equivalent to the well-established 5-HT3 -RAs in preventing vomiting during the first day after surgery. The incidence of QTc prolongation with haloperidol is statistically equivalent to 5-HT3 -RAs and thus should not be the factor that discourages its use for treatment/prophylaxis of postoperative nausea vomiting.