European Journal of Anaesthesiology 1998, 15, 433–445 A meta-analysis of nausea and vomiting following maintenance of anaesthesia with propofol or inhalational agents J. R. Sneyd, A. Carr, W. D. Byrom* and A. J. T. Bilski* Department of Anaesthesia, Derriford Hospital, Derriford Road, Plymouth PL6 8DH, UK * Zeneca Pharmaceuticals, Mereside, Alderley Park, Macclesfield, Cheshire, UK Summary and log-odds ratios, calculated from the incidence data of each individual trial, were determined and A number of prospective randomized comparator combined using a fixed-effects meta-analysis ap- studies have suggested that there is a reduction in proach. Patients who received maintenance of an- post-operative nausea and vomiting following main- aesthesia with propofol had a significantly lower tenance of anaesthesia with propofol compared with incidence of post-operative nausea and vomiting in inhalational agents. We analysed these studies in more comparison with inhalational agents regardless of in- detail by examining the effects of induction agent, duction agent, choice of inhalation agent, presence/ choice of inhalation agent, presence/absence of ni- absence of nitrous oxide, age of patient or use of trous oxide, age of patient or use of opiate on the opiate. incidence of emesis. A search of the Zeneca database MEDLEY was undertaken and prospective randomized Keywords: , propofol; , comparator studies identified. These were examined vomiting; -. individually and independently by two of the authors Introduction with appropriate caveats, general conclusions may be realized. Post-operative nausea and vomiting (PONV) are a A previous report has described a meta-analysis relatively frequent cause of morbidity following gen- of PONV with propofol; however, its author relied eral anaesthesia. Reduction of PONV has attracted exclusively on a MEDLINE search and only identified considerable research, particularly because of the in- 28 studies covering 1116 patients [1]. These authors crease in day case surgery. Clinical anaesthesiologists concluded that the odds ratios for nausea and vomiting and authors of research publications have frequently following maintenance with propofol compared with commented on an apparent reduction in PONV fol- an inhalational agent were 0.29, 0.25 and 0.29 for lowing maintenance of anaesthesia with propofol, vomiting, nausea and ‘nausea and vomiting’, re- however, few studies have addressed this question spectively, indicating significant reductions with pro- directly. Despite the modest numbers of studies dir- pofol treatment. A second meta-analysis [2] of 6069 ectly investigating PONV, there exists a very large patients from 84 studies compared the effects of either number of publications describing the clinical use of induction or maintenance with propofol and other propofol and many of these were prospective ran- anaesthetic agents on PONV. Comparator agents used domized comparisons of maintenance of anaesthesia for maintenance included propanidid, midazolam and with propofol or inhalational agents. Meta-analysis alphaxalone. The study which divided PONV into early allows the combination of multiple independent stud- and late emetic events, suggested that the number- ies of low power to provide a pool of data from which, needed-to-treat [3] to prevent an early emetic event was 4.9. In contrast, prevention of a late event was Accepted January 1998 Correspondence: J. R. Sneyd. not significantly different from control. In our analysis, 1998 European Academy of Anaesthesiology 433