Twenty-five RCTs (2,853 adults) were included.
Most studies were underpowered and only 6 studies reported prior sample size calculations. One study accounted for almost half (47%)of the patients in the ondansetron versus metoclopramide meta-analysis. Thirteen studies had adequate concealment of randomisation, 7 studies had adequate sequence generation, 15 studies reported or gave the impression that no exclusions had taken place.
There was no clear benefit for ondansetron compared with metoclopramide or droperidol. The significant heterogeneity found at the 0.1 level (heterogeneity p=0.08) for ondansetron versus metoclopramide was explained by type of surgery and duration of outcome assessment. Outcome type, ondansetron dose level, and the validity criteria did not influence the outcome.
There was no evidence of statistical heterogeneity among studies comparing ondansetron and droperidol (p=0.11 for POV and p=0.07 for PONV).
The funnel plot suggested the possibility of publication bias in studies comparing ondansetron and metoclopramide, although it was estimated that a fail-safe N of 469 was required to increase the observed p-value for PONV to 0.01.
Ondansetron versus metoclopramide (16 RCTs involving 1,981 patients): The pooled RR of POV using a random-effects model was 0.74 (95% confidence interval, CI: 0.54, 1.00, p=0.051). The pooled risk reduction of PONV was 0.72 (95% CI: 0.57, 0.91, p=0.0056). The number-needed-to-treat, assuming the emesis risk with metoclopramide was 20%, was 12 to 56. Type of operation.
Minor gynaecological (9 RCTs involving 667 women) surgery: the pooled RR was 0.57 (95% CI: 0.38, 0.84, p=0.005); the heterogeneity p-value was 0.06.
Major gynaecological (4 RCTs involving 1,122 women) surgery: the pooled RR was 0.87 (95% CI: 0.68, 1.11, p=0.254); the heterogeneity p-value was 0.19.
Duration of outcome assessment.
For 0 to 6 hours (3 RCTs involving 283 patients): the pooled RR was 0.44 (95% CI: 0.22, 0.87, p=0.018); the heterogeneity p-value was 0.06.
For 0 to 24 hours (13 RCTs involving 1,698 patients): the pooled RR was 0.81 (95% CI: 0.65, 1.02, p=0.074); the heterogeneity p-value was 0.06.
Ondansetron versus droperidol (13 RCTs involving 872 patients): the pooled RRs were 0.91 (95% CI: 0.65, 1.29, p=0.61) and 0.94 (95% CI: 0.69, 1.28, p=0.61) for POV and PONV, respectively.