Thirty RCTs (n=1,719 participants) were included in the review. Most trials were cross-over design. Half of the trials had a sample size of less than 50 patients. Six trials were rated A (low risk of bias) and 24 trials were rated B (moderate risk of bias).
Anti-emetic efficacy: Compared with neuroleptics, dronabinol had a statistically significantly lower risk of nausea and vomiting (RR 0.67, 95% CI 0.47 to 0.96, I2=79%; five trials). The number needed to treat with dronabinol to prevent one episode of vomiting was 3.4 patients. Compared with placebo, there was no statistically significant difference in risk of nausea and vomiting with dronabinol (RR 0.47, 95% CI 0.19 to 1.16, I2=91%; two trials). Compared with neuroleptics, there was no statistically significant difference in risk of nausea and vomiting with nabilone (RR 0.88, 95% CI 0.72 to 1.08, I2=64%; five trials) and levonantradol (RR 0.94, 95% CI 0.75 to 1.18, I2=37%; four trials).
Preference for anti-emetic therapy: Pooling 18 trials of anti-emetics demonstrated a statistically significant preference for cannabis-based treatment compared with control (RR 0.33, 95% CI 0.24 to 0.44, I2=65%).
Funnel plot assessment showed symmetry, which indicated a low risk of publication bias.