Eleven studies (n=1,023) were included in the review. Six studies evaluated ondansetron (n=745) (one study also evaluated dexamethasone and another also evaluated metoclopramide); these were randomised double-blind placebo-controlled trials; length of follow-up ranged from 24 hours to two weeks. Five studies evaluated other antiemetic agents (n=278); it was not reported what the study designs were; length of follow-up ranged from two hours to 24 hours.
The quality of the studies varied. The studies which evaluated ondansetron were of higher quality (Downs and Black scores 20 to 26, Delphi scores 7 to 9). Quality scores for the studies that evaluated the other agents were Downs and Black scores 8 to 18 and Delphi scores 2 to 7.
Ondansetron
Ondansetron was associated with a statistically significant decrease in the risk of hospital admission (Relative risk 0.52, 95% confidence interval: 0.27, 0.95. Number needed to treat was 14, 95% confidence interval: 9, 44. Five trials), intravenous fluid administration (Relative risk 0.41, 95% confidence interval: 0.28, 0.62. Number needed to treat was five with 95% CI: 4, 8. Four trials), cessation of vomiting in the emergency department (Relative risk 0.45, 95% confidence interval: 0.33, 0.62. Number needed to treat was five, 95% confidence interval: 4, 7. Four trials) when compared with placebo. There was no significant difference between ondansetron and placebo for return to outpatient care. It was reported that no significant heterogeneity or publication bias was detected. Sensitivity analyses did not significantly affect the results
The only adverse event reported was diarrhoea. Five studies reported on this outcome. Ondansetron may have been associated with an increase in diarrhoeal episodes up to 48 hours post drug administration, but no differences were found after this time.
Other antiemetic agents
The authors provided a narrative synthesis for the other antiemetic agents domperidone, metoclopramide, trimethobenzamide hydrochloride, pyrilamine-pentobarbital, promethazine hydrochloride and dexamethasone. The authors stated that these studies had low numbers, were of poor quality and produced inconsistent results.