Treatment of postoperative nausea and vomiting

BMJ 2003; 327 doi: http://dx.doi.org/10.1136/bmj.327.7418.762 (Published 02 October 2003) Cite this as: BMJ 2003;327:762
  1. Martin R Tramèr (martin.tramer{at}hcuge.ch), consultant anaesthetist
  1. Division of Anaesthesiology, Department of Anaethesiology, Pharmacology, and Surgical Intensive Care, Geneva University Hospitals, CH-1211 Geneva 14, Switzerland

    Better data, improved control have been achieved during recent years

    Postoperative nausea and vomiting is a nuisance. The anaesthetist is usually blamed, despite evidence that postoperative nausea and vomiting results from several factors, some related to anaesthesia, others to surgery, and some to the patients themselves. The importance of postoperative nausea and vomiting is generally underestimated because it is self limiting, never becomes chronic, and almost never kills. However, its impact on the cost of health care is not negligible. Ten per cent of the population undergo general anaesthesia every year,1 and about 30% of them are affected by postoperative nausea and vomiting.2 This amounts to about two million people in the United Kingdom every year. About 1% of patients undergoing ambulatory surgery are admitted overnight because of uncontrolled postoperative nausea and vomiting.2

    Surgical patients prefer to suffer pain rather than postoperative nausea and vomiting3 and would be willing to pay considerable amounts of money for an effective antiemetic.4 However, successful control of postoperative nausea and vomiting has proved elusive. A major obstacle to the development of an effective treatment has been the lack of a valid animal model for postoperative nausea and vomiting. New insights into pathways for emesis and efficacy of antiemetics have come from animal research with highly emetogenic chemotherapy. Extrapolation of these data to postoperative nausea and vomiting has been of limited value. Anaesthetists therefore have to rely on the results of a myriad of clinical trials, most of small size and some of doubtful validity. Data on an almost infinite number of potentially useful antiemetic interventions have been published during the last 40 years. Despite this large body of literature fundamental data on dose responsiveness or profiles of adverse effects have remained unclear for most antiemetics, and no agreement has been reached on what constitutes a gold standard. As a consequence, anaesthetists have been using antiemetics irrationally.

    The good news is that notable progress towards improved control of postoperative nausea and vomiting has been achieved during recent years. The first landmark was the advent of several sponsored, high quality, dose finding studies of a 5-hydroxytryptamine3 receptor antagonist, ondansetron, in the early 1990s. For the first time in the history of research into postoperative nausea and vomiting, a manufacturer had launched a large scale trial to evaluate an antiemetic. Initial enthusiasm was subsequently tempered because the manufacturer did not prevent authors from flooding the anaesthetic literature with covert duplicate reports that led to an overoptimistic view of the drug's efficacy and safety.5

    Secondly, large amounts of the literature on postoperative nausea and vomiting have been systematically reviewed, critically appraised, and quantitatively synthesised.2 Today we understand the relative efficacy and harm of most antiemetic interventions. Droperidol, for example, a butyrophenone that has been withdrawn in some countries for reasons of safety, has a pronounced antinausea effect at doses that are so incredibly low that the occurrence of any relevant adverse effect becomes highly unlikely. Ondansetron, which was thought to represent the first universally effective antiemetic for postoperative nausea and vomiting, was shown to have a limited effect just on vomiting. Metoclopramide, one of the most popular antiemetics for decades, showed no worthwhile efficacy. Perhaps the most important nugget from these systematic reviews was that none of the drugs tested could be regarded as a gold standard, and none was good enough to be used on its own; at best, they achieved a number needed to treat to prevent postoperative nausea and vomiting of about five compared with placebo.2 However, systematic review also confirmed improved efficacy with combined treatment through an additive or synergistic effect—for example, by combining a 5-hydroxytryptamine3 receptor antagonist with droperidol or with dexamethasone.2

    Thirdly, investigators set out to verify predictive factors for postoperative nausea and vomiting.6 They mainly confirmed what clinicians have known for a long time, that female sex, a positive history of postoperative nausea and vomiting, opioids, and certain types of surgery (for example, gynaecological or urological) increased the risk of postoperative nausea and vomiting. An intriguing observation was that smokers are less likely to suffer from it.7 Predictive factors may be used as a tool to target prophylaxis towards patients who actually need it; for all others, a “wait and see” strategy may be chosen.

    Finally, anaesthetists have become increasingly aware of the emetogenic potency of anaesthetic techniques. For a patient who must not vomit (she may be a non-smoker with a history of postoperative nausea and vomiting, and she has wired jaws after maxillofacial surgery) the anaesthetist may choose the intravenous anaesthetic propofol and a mixture of air and oxygen and may try to avoid emetogenic drugs, such as physostigmine and opioids. These modalities are of little interest to non-anaesthetists; some may be of limited or short lived efficacy, and not all are feasible in every circumstance. However, such simple measures may contribute to a low baseline risk of postoperative nausea and vomiting.

    Modern anaesthesia tries to take advantage of this knowledge; a multimodal approach has been recommended recently.8 In a perfect world anaesthetists would be able to identify reliably the patient at high risk preoperatively, use an anaesthesia technique that is less likely to cause vomiting, and administer an antiemetic cocktail before emergence. The question then is, who merits such a prophylactic multimodal approach? Cost and adverse effects may be issues here. Forty years ago it was known that not all patients vomited after surgery, and of those who did, most did so only once or twice.9 This may be an argument in favour of the “wait and see” approach. Interestingly with the expensive 5-hydroxytryptamine3 receptor antagonists, treatment of established postoperative nausea and vomiting is efficacious at much lower doses than are necessary for successful prophylaxis.10 If 30% of surgical patients suffer from postoperative nausea and vomiting,2 and we assume that in about half of those symptoms are persisting, then the target population for prophylaxis is about 15%. These patients are suffering unnecessarily, they want their opioid analgesia to be stopped, and they may need overnight admission due to intractable vomiting. For those, further investigations are warranted. Valid data are needed on old molecules that are still widely used in clinical practice (for example, haloperidol or hyoscine).

    Hopefully, new compounds that block yet another receptor system of the emesis pathways and that have shown promising results in animal models will further improve the treatment of postoperative nausea and vomiting.11


    • Competing interests MT has been a consultant to Pfizer, UK; to Sintetica, Switzerland, a manufacturer of droperidol; and to Merck, US, manufacturer of aprepitant. He has been invited to an international consensus meeting on postoperative nausea and vomiting (sponsored by Aventis, US), has received fees from lectures from MSD Switzerland and Pfizer Norway, and is a recipient of a PROSPER grant from the Swiss National Science Foundation (No 3233–051939.97/2).


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.