Droperidol has been one of the most popular antiemetic drugs in the surgical setting for several decades. However, in 2001, the US Food and Drug Administration (FDA) strengthened its cautions regarding the use of droperidol by issuing a ‘black box’ warning, the most serious warning for a FDA-approved drug (http://www.fda.gov/safety/MedWatch/Safetyinformation/Safety). The FDA noted that the use of droperidol had been associated with QT segment prolongation and/or torsades de pointes, and, in some cases, had resulted in fatal cardiac arrhythmia. However, the impact of antiemetic doses of droperidol on cardiac arrhythmia has remained contentious and the FDA decision has been challenged ever since. There is support for this approach from two clinical studies that failed to show any significant difference in QT length with intravenous droperidol 0.625 and 1.25 mg compared to placebo.1,2 In two other studies, the effect of droperidol 0.75 mg on QT prolongation was similar to ondansetron 4 mg,3 and the combination of droperidol 1 mg and ondansetron 4 mg did not further prolong QT.4
In 2007, despite the FDA warning, an international consensus panel still recommended intravenous droperidol 0.625 and 1.25 mg as a first-line antiemetic, alone or in combination with other antiemetic drugs, for the prevention of postoperative nausea and vomiting (PONV).5 These recommendations were based on data from a large multicentre study comparing these two doses6 and a meta-analysis of studies that mainly tested doses of 1.25 mg or above.7 Unfortunately, the 2007 consensus panel did not specify whether the lower of the two recommended doses (0.625 mg) was as efficacious as, or less efficacious than, the higher dose (1.25 mg), and it remained unclear whether doses below 0.625 mg had any worthwhile antiemetic efficacy.
We may assume that the cardiac risk of droperidol is dose-dependent and that, therefore, anaesthetists increasingly tend to use minimal effective doses. To better understand the importance of droperidol use for PONV prophylaxis in Europe today, we contacted all 28 Council members of the European Society of Anaesthesiology and asked them about the use of droperidol as an antiemetic in their country (Fig. 1). In 19 of 25 European countries, representing an estimated number of 83 000 anaesthetists and a population of over 711 million, droperidol is regularly used. In some countries (for instance, Spain, United Kingdom), droperidol was only recently reapproved after an intermediate withdrawal due to the FDA black box warning. Reported doses vary between 0.5 and 2.5 mg. However, in only seven of 19 countries (37%) are doses of droperidol of ≤1 mg commonly used.
In this systematic review of randomised controlled trials (RCT), we set out to test the efficacy of low-dose droperidol (arbitrarily defined as doses ≤1 mg) for the prevention of PONV in adults.
We performed a comprehensive search for published reports of RCTs testing intravenous single-dose regimens of droperidol for the prevention of PONV after general anaesthesia in adults. We searched in Medline, Embase and Central using different search strategies with combinations of the key words ‘nausea’, ‘vomiting’, ‘emesis’, ‘droperidol’, ‘surgery’, ‘an(a)esthesia’, ‘postoperative’ and the LIMIT ‘Type of article: RCT’. The last electronic search was in June 2011. Additional trials were obtained from bibliographies of retrieved reports. There was no language restriction.
Inclusion and exclusion criteria
We considered full-published reports of placebo-controlled comparisons that tested intravenous single-dose droperidol regimens 1 mg or less or 15 μg kg−1 or less. Studies that were not published as full reports (for instance, abstracts or letters), review articles and animal studies, as well as trials without a placebo or ‘no treatment’ group, were excluded. Grey literature was not searched. We also excluded trials when droperidol was used for the treatment of established PONV, when droperidol was combined with another antiemetic or when it was used as an antiemetic in a patient-controlled analgesia device.
Quality scoring and data extraction
We screened the abstracts of all retrieved reports and excluded articles that clearly did not meet our inclusion criteria. We independently read the included reports and assessed their quality of data reporting using a modified 6-points 4-items Oxford scale.8 Discrepancies were resolved by discussion.
We extracted information about patients, surgery, droperidol regimen, efficacy endpoints and adverse effects from each included study. To facilitate comparisons between trials, we extrapolated all variable doses (μg kg−1) to fixed doses (mg) using average body weight of patients as reported in the original trials. When a trial did not report on average body weight of the study participants we used the median of all average body weights of all trials.
We extracted cumulative incidences of nausea, and vomiting (including retching), within 6 h after surgery and within 24 h. Cumulative incidences during the two time periods were used as indicators of short-term or ‘early’ (0–6 h) and long-term or ‘late’ (0–24 h) antiemetic efficacy. Consequently, early events were included in late events.
We analysed dichotomous data by calculating relative risks (RRs) with 95% CIs.
Heterogeneity between studies was formally tested using the Cochrane χ2 test for heterogeneity. If the P-value for heterogeneity was more than 0.1, studies were considered homogeneous and the estimates were aggregated using a Mantel-Haenszel fixed effect model. If the P-value of the heterogeneity test was less than 0.1, studies were considered heterogeneous and reasons for heterogeneity were investigated. In particular, using a model of metaregression, we investigated whether differences in effects were due to differences in the doses of droperidol. If no source of heterogeneity was identified, we pooled the data using a random effects model.
We computed numbers needed to treat (NNT) to benefit one individual. Because NNT depend on the true underlying, or baseline, risk and because we were unable to identify that risk for all study groups, we used three hypothetical control event rates strata, 25, 50, and 75%, as surrogates of a low, medium and high-baseline risk, respectively. We then calculated for each baseline risk stratum the NNT with 95% CI using the aggregate (meta-analytical) RR and the hypothetical control event rate. Negative NNT were regarded as numbers needed to harm.
Data were processed and analysed using Excel 2007 (Microsoft, Redmond, USA), Revman 5.0.24 (The Nordic Cochrane Centre, Copenhagen, Denmark), and STATA (STATA release 11, StataCorp LP, Texas, USA).
Retrieved and analysed trials
We retrieved 86 reports; 62 were subsequently excluded (Fig. 2). One was a duplicate of a full report already published.6,9 We excluded the duplicate9 and analysed the original.6 We eventually analysed data from 25 trials published in 24 reports between 1984 and 2008 (Table 1).1,6,10–31 In total, 1525 patients received droperidol and 1432 received placebo. The median quality score of the trials was 3 (range, 1 to 6).
Most trials tested fixed doses and only a few tested weight-adjusted doses (Table 1). The median of all average body weights of patients in all trials was 67 kg. When all weight-adjusted doses were converted to fixed doses, the range was 0.25–1.0 mg (median, 0.625 mg).
Two large multicentre studies were published concomitantly in one report;6 they each tested droperidol 0.625 mg in 256 patients. In these large trials, the median control event rate (i.e. the incidence of nausea and vomiting in control patients) was 54% (range, 53 to 55%) for early outcomes and 70% (range, 63 to 82%) for late outcomes (Fig. 3).6 The other 23 trials were all single-centre; they tested droperidol 0.25–1.0 mg in 17–86 patients (median, 38). In these smaller trials, the median control event rate was 25% (range, 2 to 63%) for early outcomes and 40% (range, 7 to 96%) for late outcomes (Fig. 3).
Eight trials reported on prevention of early nausea.19,21,24,26,27,29–31 The tested dose-range was 0.25–0.94 mg. There was no evidence of statistical heterogeneity between study estimates (P = 0.783). Overall, the RR of early nausea comparing the risk in patients receiving droperidol with those receiving placebo was 0.45 (95% CI, 0.35 to 0.58) (Fig. 4, Supplemental Digital Content 1, http://links.lww.com/EJA/A26).
Ten trials (nine reports) reported on the prevention of early vomiting.6,19,21,24,26,28–31 The tested dose range was 0.25–1.0 mg. There was no evidence of statistical heterogeneity (P = 0.587). Overall, the RR of early vomiting comparing the risk in patients receiving droperidol with those receiving placebo was 0.65 (95% CI, 0.57 to 0.74). More than 80% of the weight was attributed to the two large multicentre trials that were documented in one report.6 Excluding these two trials changed the overall RR estimate to 0.53 (95% CI, 0.36 to 0.78) (Fig. 4, Supplemental Digital Content 2, http://links.lww.com/EJA/A26).
Twelve trials (11 reports) reported on the prevention of late nausea.1,6,10,14,15,19,23,25,28,30,31 Two of these provided the cumulative incidences of nausea or vomiting within 48 h.23,28 These data were combined with cumulative 24 h data. The tested dose range was 0.63–1.0 mg. The study estimates were heterogeneous (P = 0.002). However, there was no evidence of dose-responsiveness. Overall, the RR of late nausea in patients receiving droperidol compared with those receiving placebo was 0.74 (95% CI, 0.62 to 0.87). More than 40% of the weight was attributed to the two large multicentre trials.6 Excluding these two trials changed the RR estimate to 0.59 (95% CI, 0.41 to 0.83) (Fig. 4, Supplemental Digital Content 3, http://links.lww.com/EJA/A26).
Thirteen trials reported on the prevention of late vomiting.1,6,10,14–16,19,20,23,28,30,31 The tested dose range was 0.5–1.0 mg. The study estimates were heterogeneous (P < 0.001). However, there was no evidence of dose-responsiveness. Overall, the RR of late vomiting in patients receiving droperidol compared with those receiving placebo was 0.61 (95% CI, 0.47 to 0.80). About 40% of the weight was attributed to the two large multicentre trials.6 Excluding these two trials changed the RR estimate to 0.50 (95% CI, 0.32 to 0.79) (Fig. 4, Supplemental Digital Content 4, http://links.lww.com/EJA/A26).
Six trials (295 patients received droperidol 0.25–0.75 mg) reported on a composite nausea and vomiting endpoint only.11–13,17,18,22 These data were not analysed further.
Individual meta-analyses (Forrest plots) on early and late nausea and vomiting are freely accessible from the authors’ web page (http://anesthesiologie.hug-ge.ch/data.htm).
Numbers needed to treat for antiemetic efficacy
NNT consistently decreased (i.e. the antiemetic efficacy improved) with increasing control event rates (i.e. with increasing baseline risks; Table 2).
For early outcomes, and within each baseline risk stratum, NNT point estimates tended to be lower for nausea compared with vomiting, suggesting that antinausea efficacy was superior to antivomiting efficacy. For late outcomes, and within each baseline risk stratum, NNT point estimates tended to be lower for vomiting compared with nausea suggesting that antivomiting efficacy was superior to antinausea efficacy. Ninety-five percentage CI around the NNT for early nausea did not overlap with those for late nausea; however, 95% CI around the NNT for early vomiting overlapped with those for late vomiting.
Further beneficial and harmful effects
Significantly fewer patients treated with droperidol were reported to have a headache postoperatively; the NNT point estimate was 24 (Table 3, Supplemental Digital Content 5, http://links.lww.com/EJA/A26).1,6,13–15,17,21,22,24–26,30 Significantly more patients receiving droperidol were restless; the number needed to harm was nine (Table 3, Supplemental Digital Content 6, http://links.lww.com/EJA/A26). For both outcomes, there was no evidence of dose-responsiveness. There were no statistically significant differences in the incidence of sedation or dizziness. Six trials (301 patients received droperidol) reported on the presence or absence of extrapyramidal symptoms.13,15,16,20,21,31 In one trial, two of 74 patients receiving droperidol 0.625 mg presented such symptoms.21 There were no reports of QT prolongation or cardiac arrhythmia.
An international consensus panel recommended droperidol 0.625 or 1.25 mg for the prevention of PONV in adults undergoing general anaesthesia.5 Our analyses suggest that doses below 1 mg do have worthwhile antiemetic efficacy and that doses above 1 mg are not warranted.
The antinausea effect of low-dose droperidol was not maintained over the 24 h period. In contrast, the antivomiting efficacy was very similar at long-term (i.e. within 24 h) compared with short-term (within the first 6 h). This suggests that for the maintenance of an adequate antinausea effect during a 24 h period, one single prophylactic dose may not be sufficient.
Patients receiving droperidol complained of postoperative headache less often, something that has been described before.32 What is new knowledge is that protection against headache is apparent even at doses below 1 mg. Not unexpectedly, patients treated with droperidol were more often restless. However, for dizziness or sedation, aggregate data did not show a significant difference between those who received droperidol and controls. Finally, extrapyramidal symptoms happened but they appeared to be rare. No case of torsades de pointes or cardiac arrest was reported.
It may be argued that two previously published large-scale trials have already provided all the evidence on efficacy and harm of low-dose droperidol.6 However, these trials tested only one dose (0.625 mg) and they did not report on early nausea outcomes. To better understand the usefulness of an antiemetic drug, it is important to distinguish between its early and late efficacy, and also between its antinausea and antivomiting efficacy. Additionally, our enquiry among ESA council members revealed that, although droperidol is widely used in Europe, administered doses are often inappropriately high, including 2.5 mg in some countries. This suggests that dissemination and implementation of evidence from meta-analysis32 and large trials,6 and of recommendations from international expert panels,5 have been insufficient. Finally, had we not included all the smaller trials, we would have missed valuable data on adverse drug reactions and would not have been able to test for dose-responsiveness.
Antiemetic efficiency clearly depends on the underlying, or baseline risk. In a high-risk population, less effort is needed to prevent one bad outcome. As expected, NNT decreased with increasing control event rates, confirming that prevention is only worthwhile when the baseline risk is high. In the hypothetical high-risk setting, NNT to prevent nausea and vomiting were between about 3.5 and 5. These numbers suggested improved efficacy. However, it may be argued that even in a high-risk setting, low-dose droperidol alone should not be regarded as a universal prophylaxis. Combinations of antiemetic drugs for optimal efficacy have been advocated before.5
As with other old antiemetic drugs,33 droperidol is not exempt from central adverse reactions. However, in the dose range analysed, adverse effects did not seem to be a major limiting factor. Not unexpectedly, the risk of restlessness was increased; about one in nine patients receiving droperidol 0.25–1.0 mg will be restless postoperatively, who would not be if they received a placebo. It remained unclear whether this adverse effect increased postoperative morbidity or interfered with patient satisfaction. Sedation or dizziness was not an issue and this was probably due to the low-dose range that was tested. Extrapyramidal symptoms are typical adverse effects of butyrophenones and there is evidence that these may even occur with doses as low as 0.625 mg.21 Depending on the chosen denominator, between 0.7 (two of 301) and 2.7% (two of 74) patients receiving droperidol 0.625 mg were reported to have such symptoms postoperatively. We do not know whether the two reported cases needed treatment. There was no report of arrhythmia including QT prolongation and no case of torsades de pointes or cardiac arrest. Because none of these trials was designed to study the cardiac toxicity of droperidol, it remains unknown whether such cases occurred but were missed, or whether they were not reported, although this is unlikely given the severity of this adverse effect. Clinical studies have suggested that with droperidol doses below 1 mg, QT prolongation was no different from placebo1,2 or ondansetron 4 mg,3,4 but we still cannot completely rule out an increased risk of cardiac arrhythmia even with doses below 1 mg. In our analysis, 1525 patients had received droperidol 0.25–1 mg and there was no report of cardiac death; therefore, we can be 95% confident that the chance of this major adverse event happening is at most three in 1525 (or 0.2%).34
We cannot be certain that our analysis does not overestimate the beneficial effect of low-dose droperidol. The literature search was comprehensive but selective publication of positive results cannot be ruled out. The quality score of studies ranged from 1–6 (median, 3) and low-quality trials tend to overestimate treatment efficacy. However, our methodological quality score appreciates the quality of data reporting of randomised trials, which, strictly speaking, does not reveal how the trials were actually performed. Authors sometimes perform adequate randomisation but fail to report on the necessary methodological details.35
Small trials reported large variability in control event rates, indicating that in small trials anything may happen by random chance.36 It has been argued that smaller trials should, therefore, be excluded from meta-analysis.37 The two largest trials reported extraordinary high control event rates due to selective inclusion of high-risk patients, mainly women with a positive history of PONV or motion sickness.6 However, despite high control event rates, the large trials reported a degree of antiemetic efficacy of droperidol that was lower than in smaller trials.
We were unable to show evidence of dose-responsiveness. This may mean that there is none in the dose range between 0.25 and 1.0 mg, or that our analyses were lacking power to show a dose-response relationship. For late outcomes, the tested dose range was narrow, ranging from 0.5 to 1.0 mg only. This may be yet another explanation for why we were unable to find dose-responsiveness. The lack of late data for very low doses (i.e. ≤0.5 mg) may reflect the presence of publication bias; it cannot be ruled out that trials testing very low doses failed to report on late data because these were negative.
It has been suggested that the antiemetic effect of droperidol may be sex-dependent.38 Although 13 trials included mixed sex, only one reported on the number of men and women needing rescue treatment separately but without providing any analyses of these data.21 This meta-analysis does not further our understanding on sex differences of the antiemetic effect of droperidol.
Our analyses may serve as a rational basis for future research. Knowing that low doses of droperidol are antiemetic in the surgical setting, albeit to a limited degree, begs the question as to how low-dose droperidol performs in combination with other antiemetics, for instance, 5HT3 receptor antagonists or dexamethasone. Also, it seemed that the antinausea effect was not maintained to the same degree as the antivomiting effect. It may be worthwhile testing, how often a prophylactic low-dose droperidol regimen needs to be repeated to maintain adequate 24 h antinausea efficacy.39
In conclusion, a survey suggests that droperidol is still a popular antiemetic among European anaesthetists. Data from randomised trials provide evidence that low-dose droperidol (≤1 mg or ≤15 μg kg−1) is clearly efficacious for the prevention of PONV. An additional, but weak, beneficial effect is the prevention of postoperative headache. Cardiac adverse events were not reported. However, even at low doses, potentially serious central adverse effects such as extrapyramidal symptoms may occur. In view of the fact that droperidol at doses of 1 mg or less is antiemetic, and that adverse drug reactions are likely to be dose-dependent, there is an argument to stop using doses above 1 mg.
This work was supported by the Evidence-based Critical Care, Anaesthesia and Pain Treatment (EBCAP) Institute (Geneva, Switzerland) (salary of N.E.), and institutional funds from the Department of Anaesthesiology, Critical Care & Clinical Pharmacology, Geneva University Hospitals. From 2005 to 2007, the Evidence-based Critical Care, Anaesthesia and Pain Treatment (EBCAP) Institute (Geneva, Switzerland) received an unrestricted grant from SINTETICA SA, CH-6850 Mendrisio, Switzerland. SINTETICA SA is a manufacturer of droperidol. None of the authors reports any conflict of interest in relation to this study.
1. White PF, Song D, Abrao J, et al. Effect of low-dose droperidol
on the QT interval during and after general anesthesia: a placebo-controlled study. Anesthesiology
2. Nakao S, Hatano K, Sumi C, et al. Sevoflurane causes greater QTc interval prolongation in elderly patients than in younger patients. Anesth Analg
3. Charbit B, Albaladejo P, Funck-Brentano C, et al. Prolongation of QTc interval after postoperative nausea
treatment by droperidol
or ondansetron. Anesthesiology
4. Charbit B, Alvarez JC, Dasque E, et al. Droperidol
and ondansetron-induced QT interval prolongation: a clinical drug interaction study. Anesthesiology
5. Gan TJ, Meyer TA, Apfel CC, et al. Society for Ambulatory Anesthesia guidelines for the management of postoperative nausea
. Anesth Analg
6. Fortney JT, Gan TJ, Graczyk S, et al. S3A-409 and S3A-410 Study GroupsA comparison of the efficacy, safety, and patient satisfaction of ondansetron versus droperidol
as antiemetics for elective outpatient surgical procedures. Anesth Analg
7. Domino KB, Anderson EA, Polissar NL, Posner KL. Comparative efficacy and safety of ondansetron, droperidol
, and metoclopramide for preventing postoperative nausea
: a meta-analysis. Anesth Analg
8. Elia N, Tramèr MR. Ketamine and postoperative pain: a quantitative systematic review of randomised trials. Pain
9. Hill RP, Lubarsky DA, Phillips-Bute B, et al. Cost-effectiveness of prophylactic antiemetic therapy with ondansetron, droperidol
, or placebo. Anesthesiology
10. Badaoui R, Yagoubi A, Pouilly A, et al. Comparaison de l’ondansétron et du dropéridol pour la prévention des nausées et des vomissements après cholécystectomie par voie laparoscopique. Cahier d’Anesthésiologie
11. Beattie WS, Lindblad T, Buckley DN, Forrest JB. Menstruation increases the risk of nausea
after laparoscopy. A prospective randomized study. Anesthesiology
12. Contreras-Dominguez V, Carbonell-Bellolio P. Prophylactic antiemetic therapy for acute abdominal surgery. A comparative study of droperidol
, metoclopramide, tropisetron, granisetron and dexamethasone. Rev Bras Anestesiol
13. Cozanitis D, Asantila R, Eklund P, Paloheimo M. A comparison of ranitidine, droperidol
or placebo in the prevention of nausea
after hysterectomy. Can J Anaesth
14. Eberhart LH, Lindenthal M, Seeling W, et al. Dolasetron, droperidol
and a combination of both in prevention of postoperative nausea
after extracapsular cataract extraction under general anesthesia. Anasthesiol Intensivmed Notfallmed Schmerzther
15. Eberhart LH, Morin AM, Hoerle S, et al. Droperidol
and dolasetron alone or in combination for prevention of postoperative nausea
after vitrectomy. Ophthalmology
16. Fassoulaki A, Galanaki E. The antiemetic effect of droperidol
: is it dose dependent? Acta Anaesthesiol Belg
17. Foster PN, Stickle BR, Laurence AS. Akathisia following low-dose droperidol
for antiemesis in day-case patients. Anaesthesia
18. Jorgensen NH, Coyle JP. Intravenous droperidol
after alfentanil anesthesia without increasing recovery time. J Clin Anesth
19. Joris JL, Poth NJ, Djamadar AM, et al. Supplemental oxygen does not reduce postoperative nausea
after thyroidectomy. Br J Anaesth
20. Klahsen AJ, O’Reilly D, McBride J, et al. Reduction of postoperative nausea
with the combination of morphine and droperidol
in patient-controlled analgesia. Can J Anaesth
21. Kreisler NS, Spiekermann BF, Ascari CM, et al. Small-dose droperidol
effectively reduces nausea
in a general surgical adult patient population. Anesth Analg
22. Lim KS, Lim BL, Tee CS, Vengadasalam D. Nausea
after termination of pregnancy as day surgery cases: comparison of 3 different doses of droperidol
and metoclopramide as antiemetic prophylaxis. Singapore Med J
23. Martin J, Williams D, Weis FR Jr. Comparison of three anesthetic techniques on emetic symptoms using sufentanil for outpatient surgery. AANA J
24. Millar JM, Hall PJ. Nausea
after prostaglandins in day case termination of pregnancy. The efficacy of low dose droperidol
25. Morin AM, Bezler T, Eberhart LH, et al. [The effect of low droperidol
dosages on postoperative anxiety, internal tension, general mood and PONV]. Anaesthesist
26. O’Donovan N, Shaw J. Nausea
in day-case dental anaesthesia. The use of low-dose droperidol
27. Pandit SK, Kothary SP, Pandit UA, et al. Dose-response study of droperidol
and metoclopramide as antiemetics for outpatient anesthesia. Anesth Analg
28. Paxton LD, McKay AC, Mirakhur RK. Prevention of nausea
after day case gynaecological laparoscopy. A comparison of ondansetron, droperidol
, metoclopramide and placebo. Anaesthesia
29. Tang J, Watcha MF, White PF. A comparison of costs and efficacy of ondansetron and droperidol
as prophylactic antiemetic therapy for elective outpatient gynecologic procedures. Anesth Analg
30. Valanne J, Korttila K. Effect of a small dose of droperidol
and recovery after outpatient enflurane anaesthesia. Acta Anaesthesiol Scand
31. Wang TF, Liu YH, Chu CC, et al. Low-dose haloperidol prevents postoperative nausea
after ambulatory laparoscopic surgery. Acta Anaesthesiol Scand
32. Henzi I, Sonderegger J, Tramèr MR. Efficacy, dose-response, and adverse effects of droperidol
for prevention of postoperative nausea
. Can J Anaesth
33. Schnabel A, Eberhart LH, Muellenbach R, et al. Efficacy of perphenazine to prevent postoperative nausea
: a quantitative systematic review. Eur J Anaesthesiol
34. Hanley JA, Lippman-Hand A. If nothing goes wrong, is everything all right? Interpreting zero numerators. JAMA
35. Devereaux PJ, Choi PT, El-Dika S, et al. An observational study found that authors of randomized controlled trials frequently use concealment of randomization and blinding, despite the failure to report these methods. J Clin Epidemiol
36. Moore RA, Tramèr MR, Carroll D, et al. Quantitative systematic review of topically applied nonsteroidal anti-inflammatory drugs. BMJ
37. Nuesch E, Trelle S, Reichenbach S, et al.
Small study effects in meta-analyses of osteoarthritis trials: meta-epidemiological study. BMJ 2010; 341:c3515.
38. Apfel CC, Korttila K, Abdalla M, et al. A factorial trial of six interventions for the prevention of postoperative nausea
. N Engl J Med
39. Kooij FO, Klok T, Hollmann MW, Kal JE. Automated reminders increase adherence to guidelines for administration of prophylaxis for postoperative nausea
. Eur J Anaesthesiol