A meta-analysis of prevention of postoperative nausea and vomiting: randomised controlled trials by Fujii et al. compared with other authors

The population sampling in randomised controlled trials by Fujii et al. have been shown to exhibit unusual distributions. This systematic review analysed the effectiveness of prophylactic antiemetics in trials by Fujii et al. compared with other authors. Granisetron was more effective in trials by Fujii et al., relative risk ratios (RRR (95% CI)): nausea 0.53 (0.42-0.67), p=0.00021; vomiting 0.60 (0.50-0.73), p=0.00094. Ramosetron was also more effective in studies by Fujii et al.: vomiting 0.60 (0.39-0.91), p=0.02; nausea or vomiting 0.71 (0.56-0.91); p=0.006. In comparison with granisetron, droperidol was less effective in trials by Fujii et al. than others: nausea 2.41 (1.72-3.36), p=2.5×10(-7); vomiting 1.73 (1.26-2.38), p=6.4×10(-4). Postoperative nausea and vomiting was less likely to trigger rescue antiemesis after granisetron and metoclopramide in studies by Fujii et al., 0.40 (0.27-0.60), p=9.7×10(-6). Triggered rates of rescue were not different in studies by others for droperidol, granisetron and metoclopramide, but were less common after granisetron than droperidol and metoclopramide in studies by Fujii et al., 0.50 (0.38-0.66), p=1.7×10(-6) and 0.47 (0.34-0.64), p=2.6×10(-6), respectively. There was no synergism between antiemetics in trials by other authors. In contrast, in studies by Fujii et al., postoperative nausea and vomiting was more likely if granisetron was administered alone: nausea 4.20 (1.94-9.08), p=2.6×10(-4) ; vomiting 4.50 (2.55-7.97), p=2.3×10(-7); nausea or vomiting 5.00 (2.84-8.81), p=2.5×10(-8). Similarly, droperidol was less effective in studies by Fujii et al. if administered alone: vomiting 2.76 (1.25-6.11), p=0.01; nausea or vomiting 2.96 (1.46-6.00), p=2.7×10(-3). The conclusion is that if, as recommended, data with unusual distributions are removed from meta-analysis and articles by Fujii et al. excluded, then the antiemetic effects of granisetron and ramosetron are greatly reduced; further, there is no evidence of synergism between antiemetics and indeed, some evidence of antagonism between antiemetic agents.