Robert S Phillips, Amanda J Friend, Faith Gibson, Elizabeth Houghton, Shireen Gopaul, Jean V Craig, Barry Pizer

Nausea and vomiting remain a problem for children undergoing treatment for malignancies despite new antiemetic therapies. Optimising antiemetic regimens could improve quality of life by reducing nausea, vomiting, and associated clinical problems. This is an update of the original systematic review. To assess the effectiveness and adverse events of pharmacological interventions in controlling anticipatory, acute, and delayed nausea and vomiting in children and young people (aged less than 18 years) about to receive or receiving chemotherapy. Searches included the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, LILACS, PsycINFO, conference proceedings of the American Society of Clinical Oncology, International Society of Paediatric Oncology, Multinational Association of Supportive Care in Cancer, and ISI Science and Technology Proceedings Index from incept to December 16, 2014, and trial registries from their earliest records to December 2014. We examined references of systematic reviews and contacted trialists for information on further studies. We also screened the reference lists of included studies. Two review authors independently screened abstracts in order to identify randomised controlled trials (RCTs) that compared a pharmacological antiemetic, cannabinoid, or benzodiazepine with placebo or any alternative active intervention in children and young people (less than 18 years) with a diagnosis of cancer who were to receive chemotherapy. Two review authors independently extracted outcome and quality data from each RCT. When appropriate, we undertook meta-analysis. We included 34 studies that examined a range of different antiemetics, used different doses and comparators, and reported a variety of outcomes. The quality and quantity of included studies limited the exploration of heterogeneity to narrative approaches only.The majority of quantitative data related to the complete control of acute vomiting (27 studies). Adverse events were reported in 29 studies and nausea outcomes in 16 studies.Two studies assessed the addition of dexamethasone to 5-HT3 antagonists for complete control of vomiting (pooled risk ratio (RR) 2.03; 95% confidence interval (CI) 1.35 to 3.04). Three studies compared granisetron 20 mcg/kg with 40 mcg/kg for complete control of vomiting (pooled RR 0.93; 95% CI 0.80 to 1.07). Three studies compared granisetron with ondansetron for complete control of acute nausea (pooled RR 1.05; 95% CI 0.94 to 1.17; 2 studies), acute vomiting (pooled RR 2.26; 95% CI 2.04 to 2.51; 3 studies), delayed nausea (pooled RR 1.13; 95% CI 0.93 to 1.38; 2 studies), and delayed vomiting (pooled RR 1.13; 95% CI 0.98 to 1.29; 2 studies). No other pooled analyses were possible.Narrative synthesis suggests that 5-HT3 antagonists are more effective than older antiemetic agents, even when these agents are combined with a steroid. Cannabinoids are probably effective but produce frequent side effects. Our overall knowledge of the most effective antiemetics to prevent chemotherapy-induced nausea and vomiting in childhood is incomplete. Future research should be undertaken in consultation with children, young people, and families that have experienced chemotherapy and should make use of validated, age-appropriate measures. This review suggests that 5-HT3 antagonists are effective in patients who are to receive emetogenic chemotherapy, with granisetron or palonosetron possibly better than ondansetron. Adding dexamethasone improves control of vomiting, although the risk-benefit profile of adjunctive steroid remains uncertain.