GENERAL ARTICLES Efficacy and Adverse Effects of Prophylactic Antiemetics During Patient-Controlled Analgesia Therapy: A Quantitative Systematic Review Martin R. Trame`r, MD, DPhil*, and Bernhard Walder, MD² Divisions of *Anaesthesiology and ²Anaesthesiological Investigations, Department APSIC, Geneva University Hospital, Geneva, Switzerland Nausea and vomiting are frequent adverse effects of patient-controlled analgesia (PCA) with opioids. To identify the optimal prophylactic antiemetic interven- tion in this setting, we performed a systematic search for randomized trials (MEDLINE, EMBASE, Cochrane library, reference lists, hand-searching, no language re- striction) published up to May 1998 that compared pro- phylactic antiemetic interventions with placebo or no treatment in the postoperative PCA-setting with opi- oids. Fourteen placebo-controlled trials (1117 patients) with different regimens of droperidol, ondansetron, hyoscine TTS, tropisetron, metoclopramide, propofol, and promethazine were analyzed. One PCA was with tramadol, all others were with morphine. At 24 h, the cumulative incidence of nausea and vomiting without antiemetics was approximately 50%. Droperidol 0.017– 0.17 mg/mg of morphine (0.5–11 mg/d droperidol) was statistically significantly more effective than pla- cebo without evidence of dose-responsiveness; the number needed to treat to prevent nausea compared with placebo was 2.7 (95% confidence interval 1.8 –5.2), and that to prevent vomiting was 3.1 (2.3– 4.8). Com- pared with placebo, the incidence of minor adverse ef- fects with droperidol was increased with doses .4 mg/d. Implications: Of 100 patients treated with droperidol added in a patient-controlled analgesia pump with morphine, 30 who would have vomited or been nauseated had they not received droperidol will not suffer these effects. There is no evidence of dose- responsiveness for efficacy with droperidol, but the risk of adverse effects is dose-dependent. There is a lack of evidence for other antiemetics. (Anesth Analg 1999;88:1354 –61) P atient-controlled analgesia (PCA) is a routine technique of pain treatment (1). Opioids are the mainstay of postoperative analgesia with PCA (2). However, postoperative pain control with opioid PCA is often perceived to be highly emetogenic (3). Although there is no evidence that the risk of nausea and vomiting with PCA is more frequent than that with IM opioids (4), patients may refuse to continue PCA treatment because of these side effects. The best antiemetic prophylaxis for PCA-related emesis is still not known (5). We systematically searched relevant and valid data on antiemetics that are concomitantly used with postoperative opioid PCA. The aim was to search the strongest evidence for the relative efficacy and harm of interventions that are used prophylactically to reduce the incidence of opi- oid PCA related nausea and vomiting. Methods A systematic search for relevant randomized con- trolled trials was conducted. The search was per- formed with MEDLINE (from 1966), EMBASE (from 1980), and Cochrane library (1998, issue I) using the free text keywords “nausea,” “vomiting,” “emesis,” “patient-controlled analgesia,” and their combination. We performed the last electronic search on April 3, 1998. Reference lists of retrieved reports, relevant review articles, and meta-analyses (1,2,4,5) were checked. Locally available anesthesia journals were hand-searched. No language restriction was applied. We contacted authors by letter when there was uncer- tainty about the data (for instance, when efficacy data were reported as scores or number of emesis episodes but not in dichotomous form) and requested informa- tion on additional unpublished data. This work was supported by a PROSPER research grant from the Swiss National Science Foundation (Grant No 3233– 051939.97). Accepted for publication October 6, 1998. Address correspondence and reprint requests to Martin Trame`r, Division of Anaesthesiology, Department APSIC, Geneva Univer- sity Hospital, CH-1211 Geneva 14, Switzerland. Address e-mail to martin.tramer@hcuge.ch. ©1999 by the International Anesthesia Research Society 1354 Anesth Analg 1999;88:1354–61 0003-2999/99