Fifteen trials were included (n=1,337, sample sizes ranged from 20 to 120 patients). The median quality score was 4 out of a maximum of 5. All studies were published in or after 1999. The funnel plot of the primary outcome was not symmetrical around the mean, so publication bias could not be ruled out.
5-HT3 antagonists significantly reduced the incidence of pruritus from 78 per cent to 66 per cent overall (OR 0.44, 95% CI: 0.3, 0.7, p=0.0002), representing a number needed to treat of six (95% CI: 4, 14). Subgroup analysis found that this benefit was in patients receiving morphine opioids and not with lipid-soluble opioids. 5-HT3 antagonists also significantly reduced the treatment request for pruritus (OR 0.58, 95% CI: 0.4, 0.8, p=0.0003) and intensity of pruritus (WMD -0.35, 95% CI: -0.6, -0.1, p=0.007). 5-HT3 antagonists significantly reduced the incidence of postoperative nausea and vomiting (Peto OR 0.43, 95% CI: 0.3, 0.6, p<0.00001), intensity of postoperative nausea and vomiting (WMD -0.12, 95% CI -0.2, 0.0, p=0.05) and need of rescue treatment (OR 0.42, 95% CI: 0.2, 0.9, p=0.02).