Secondary Logo

Journal Logo

Postoperative nausea and vomiting

Total intravenous anaesthesia versus single-drug pharmacological antiemetic prophylaxis in adults

A systematic review and meta-analysis

Schaefer, Maximilian S.; Kranke, Peter; Weibel, Stephanie; Kreysing, Robert; Kienbaum, Peter

Author Information
European Journal of Anaesthesiology: October 2016 - Volume 33 - Issue 10 - p 750-760
doi: 10.1097/EJA.0000000000000520



Postoperative nausea and vomiting (PONV) are frequent after general anaesthesia.1 Without antiemetic prophylaxis, several million patients undergoing general anaesthesia every year will develop PONV.2–4 Together with pain, PONV is ranked as the most undesirable event after surgery5 and significantly increases healthcare costs.6 Therefore, international consensus guidelines were recently published7 recommending the identification of patients at high risk of PONV8 and administration of a risk-adapted prophylaxis including, when general anaesthesia cannot be avoided, the patient's risk for PONV should be reduced by substitution of inhalational anaesthetics with propofol-based total intravenous anaesthesia (TIVA).7 However, in several situations, modern inhalational anaesthetics may be preferable to propofol. With inhalational anaesthetics, end-tidal anaesthetic concentrations can be used to guide anaesthetic depth9 and decrease the incidence of intraoperative awareness,10 as well as allowing the precise control of weaning (irrespective of a patient's weight)11 and they have favourable effects on the cardiovascular system. Thus, if inhalational anaesthetics are not to be substituted by TIVA, the question is raised as to whether the reduction in the patient's risk of PONV associated with TIVA can be compensated sufficiently with pharmacological prophylaxis. Interestingly, a large international multicentre trial (IMPACT) even suggested an inferior prophylactic effect of TIVA as compared with using one single antiemetic drug as prophylaxis; however, the published analyses did not include a direct comparison of these two strategies.1 Furthermore, TIVA and pharmacological prophylaxis might show differential effects on PONV in the early and late postoperative phase. Therefore, the relative potency of TIVA, as compared with pharmacological prophylaxis, needs to be established based on a direct comparison. Because recent guidelines recommend TIVA as an equivalent intervention to one single antiemetic drug for prevention of PONV,7 we restricted our analysis to studies comparing TIVA with single-drug pharmacological prophylaxis.

Thus, to compare the effectiveness of antiemetic prophylaxis with propofol-based TIVA with that of single-drug prophylaxis in association with inhalational anaesthetics, we conducted a systematic literature search and subsequent meta-analysis of all the available randomised controlled trials.

Materials and methods

Search strategy and selection criteria

We performed this systematic review and meta-analysis according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis statement.12 The study protocol, including the full written search strategy, was registered and published online at the publicly available internet database PROSPERO (, registration number CRD42015019571) before the search was conducted. All deviations from that protocol are specified in the Supplemental Digital Content, ‘Deviations from the PROSPERO protocol’. Our search combined a three-part strategy, including updated syntax for identification of studies including prophylactic drugs for PONV,13 with identification of studies investigating TIVA and the modified Robinson's highly sensitive search for controlled clinical trials.14 Searched databases included MEDLINE (via Pubmed), EMBASE (via Ovid) and the Cochrane Central Register of Controlled Trials (CENTRAL, via Wiley) without any restrictions on language or publication year.

We included all randomised, controlled trials with at least one study group receiving propofol-based anaesthesia (TIVA group) without further antiemetic prophylaxis, and at least one study group receiving inhalational anaesthesia with additional single-drug antiemetic prophylaxis. Our definition of TIVA implied the use of propofol as the primary hypnotic but no preconditions were applied regarding the use of nitrous oxide (N2O). Studies were required to report the incidences of PONV, postoperative nausea (PON) or postoperative vomiting (POV). Studies and study groups which included multiple antiemetic prophylactics [i.e. TIVA with pharmacological antiemetic prophylaxis or inhalational anaesthesia with pharmacological antiemetic prophylaxis (IA + AE) with more than one prophylactic drug] were excluded, as well as studies involving patients less than 18 years of age. Studies were independently screened for inclusion and exclusion criteria by two investigators (M.S.S., R.K.).

Data extraction and preparation

Our primary outcome variable was the incidence of PONV within 24 h after surgery (‘overall PONV’). Additionally, the following data were independently extracted by two investigators (M.S.S., R.K.): type of inhalational anaesthetics and type of intraoperative opioid; use and dose of N2O; timing, dose and type of pharmacological antiemetic prophylaxis; type of surgery and duration of anaesthesia; PON, POV and the need for rescue medication; incidences of early and late PONV, as defined by the authors; incidences of any reported adverse events with focus on but not limited to dizziness, drowsiness, shivering, headache, gastrointestinal symptoms, extrapyramidal symptoms, awareness, prolongation of the cardiac QT interval perioperative arrhythmia and quality of recovery. If available, supplemental material and appendices were also screened for the aforementioned data. Because POV without PON is extremely rare in adults, PON was taken as a surrogate for PONV, when only PON was reported.

Risk of bias assessment

Studies were scored for risk of bias using The Cochrane Collaboration ‘Risk Of Bias’ Assessment Tool, which judges risk of bias in six domains: random sequence generation and allocation concealment (selection bias), blinding of participants and study personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias) and selective reporting (reporting bias). Additionally, we estimated the overall risk of bias for each study. Assessments were made by two investigators (M.S.S. and S.W.) who were blind to each other's assessments. Upon disagreement, judgement by a third investigator (R.K.) was used as a referee's decision. A sensitivity analysis was performed for overall PONV, excluding studies with a high risk of overall bias. For outcomes that were reported by 10 or more included studies, publication bias was assessed using funnel plots. After visual inspection, funnel plot asymmetry was tested with the Arcsine test and a conventionally used alpha of of 0.1.15,16

Statistical analysis

All references were imported into Endnote X7.1 (Thomson Reuters, New York, USA). Extracted data were entered into Review Manager 5.3.5 (The Cochrane Collaboration, Copenhagen, Denmark), which was used for all analyses except for funnel plots. The latter were created and tested for asymmetry using the R software (version 3.2.2, The R Project) with the meta-package. Heterogeneity was assessed using the I2 statistic with χ2 test for significance.17,18 Subsequently, pooled estimates were calculated using inverse variance weighting with random-effects modelling and alpha set to 0.05. Because the funnel plot for the primary outcome overall PONV showed significant asymmetry with different results between smaller and larger studies, we conducted an additional explorative analysis; to reduce the influence of smaller studies in the model, fixed-effect modelling instead of random-effects modelling for calculation of risk ratios for the risk for overall PONV was used.

To explore heterogeneity, a subgroup analysis was performed on the main outcome (overall PONV) with stratification according to total sample size (number of included patients ≤50, ≤100, ≤200, >200). To avoid a type 1 error because of repetitive testing, we adjusted alpha for subgroup analyses using the Bonferroni approach: as we created a total of four strata, alpha was corrected to 0.013 (0.05 ÷ 4).16 Furthermore, because no preconditions had been specified regarding the use of N2O, we conducted an additional sensitivity analysis which excluded all studies with an uneven distribution of N2O administration between the two groups (i.e. where N2O was used with inhalational anaesthetics but not with TIVA, or vice versa). All results are reported as relative risks (RRs) with corresponding 95% confidence interval, with a RR less than 1 favouring TIVA and a RR greater than 1 favouring IA + AE. Overall, quality of evidence was rated using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach19 and incorporated in a Summary of Findings Table,20 which was created for the primary outcome.


Our database search was carried out on 29 April 2015 and yielded 3106 articles, 690 from MEDLINE, 511 from CENTRAL and 1905 from EMBASE. After elimination of duplicates, 2206 articles were screened for inclusion and exclusion criteria. Our search results included 21publications from two authors who have been discredited,21–24 and even though not all of these articles have been retracted, we decided to exclude all 21 before further screening. Figure 1 summarises reasons for the exclusion of the remaining 2171 articles. At the end of the screening process, 14 randomised controlled trials with a total of 2051 patients were included for final analysis.1,25–37 Study characteristics and anaesthesia details including specifics of antiemetic prophylaxis are summarised in Table 1.

Fig. 1
Fig. 1:
Flow chart indicating screening process with details of study exclusion.
Table 1
Table 1:
Overview and characteristics of included studies

Primary analysis

Overall PONV was reported by 12 out of 14 studies. The remaining two studies31,34 only reported PONV separately for early and late periods. Despite an effort to contact the corresponding authors, data on overall PONV were not available. Thus, data from these two studies were only included in analyses of early and late PONV. In most studies, the distinction between early and late PONV was made at 2 h after awakening from anaesthesia (Table 2).

Table 2
Table 2:
Definitions of early and late postoperative nausea and vomiting, as defined by the respective authors

There was no difference in the risk of PONV between the groups that received TIVA or IA + AE when analysed using a random-effects model (Fig. 2). Similarly, there was no difference between the groups for the risk of POV, the need for rescue medication, or early PONV (Fig. 3). However, patients receiving TIVA had a significantly higher risk of experiencing PONV in the late postoperative phase [1.41 (1.10 to 1.79), P = 0.006, Fig. 3]. There was significant heterogeneity among studies in the analyses of overall PONV (I2 = 54%), need for rescue medication (I2 = 58%) and POV (I2 = 60%) (Figs. 2 and 3). When a fixed effect model was applied, TIVA was inferior to IA + AE in preventing overall PONV [1.16 (1.04 to 1.30), P = 0.01, Fig. 2]. GRADE rating for quality of evidence was ‘moderate’ for the main outcome of overall PONV (Table 1, Supplemental Digital Content shows the GRADE Table,

Fig. 2
Fig. 2:
Forest plot of the primary outcome overall PONV. Additional results using a fixed effect model are reported because of significant funnel plot asymmetry with smaller studies favouring TIVA. IA + AE, inhalational anaesthesia with pharmacological antiemetic prophylaxis; IV, inverse variance; PONV, postoperative nausea and vomiting; TIVA, total intravenous anaesthesia.
Fig. 3
Fig. 3:
Forest plots for POV, postoperative need for antiemetic rescue medication, PONV in the early and PONV in the late phase. IA + AE, inhalational anaesthesia with pharmacological antiemetic prophylaxis; IV, inverse variance; PONV, postoperative nausea and vomiting; POV, postoperative vomiting; TIVA, total intravenous anaesthesia.

Adverse events and quality of recovery

Postoperative adverse events were investigated by six studies but only four of these studies reported analysable data.25,30,31,35 Frequencies of postoperative shivering, dizziness and headache were investigated by two studies, whereas postoperative myalgia and drowsiness were each reported by only one study. No differences were observed between the two groups except for a tendency towards a decreased risk of postoperative shivering after TIVA [0.46 (0.19 to 1.11), P = 0.08, Supplemental Digital Content Figure 1,]. Two other studies provided only descriptive information with no incidence data: ‘the side-effect profile did not differ’ between the anaesthetic techniques29 and ‘there were no differences between the groups’ in drowsiness and adverse events.36 None of these six studies assessed intraoperative awareness, prolongation of the cardiac QT interval postoperative arrhythmias or gastrointestinal symptoms. Only one study reported on the quality of recovery,31 and it found no difference in the ‘Quality of Recovery Score 40’, median 155, interquartile range (149 to 159) versus 152 (150 to 156) for TIVA and IA + AE, respectively (P = 0.114).

Sensitivity and subgroup analyses

In three studies, the use of nitrous oxide was not evenly distributed between the two groups.33,35,36After excluding these studies from the different analyses, the overall risks were basically unchanged: PONV, 1.13 (0.86 to 1.47), P = 0.39, I2 = 52%; POV, 1.26 (0.73 to 2.20), P = 0.41, I2 = 61%; early PONV, 1.07 (0.88 to 1.30), P = 0.48, I2 = 0%; the need for rescue medication, 1.17 (0.52 to 2.65), P = 0.71, I2 = 73%. However, the risk of late PONV was even more pronounced in patients who received TIVA [1.51 (1.17 to 1.94), P = 0.002, I2 = 31%].

Subgroup analyses by sample size revealed a significantly higher risk of PONV with TIVA in studies involving 101 to 200, and greater than 200 patients: 1.44 (1.11 to 1.87), P = 0.006 and 1.24 (1.07 to 1.44), P = 0.005, respectively, (a Forest plot is provided in Supplemental Digital Content Figure 2, There was no statistically significant difference between groups in smaller studies involving 100 patients or less.

Risk of bias

An overview of the risk of bias assessment is presented in Table 3. Low risk of bias was present in four, moderate risk in seven and high risk in three studies. Excluding high risk of bias studies did not change the result for the main analysis on overall PONV [1.11 (0.89 to 1.37), P = 0.35], but it reduced heterogeneity (I2 40%, P = 0.10). Data from at least 10 studies were only available for overall PONV and early PONV. Therefore, funnel plot analysis was only performed for these two outcomes. Visual inspection of the funnel plot for overall PONV suggested asymmetry favouring TIVA, which was confirmed by Arcsine analysis for funnel plot asymmetry (P = 0.073, Fig. 4). No funnel plot asymmetry was detected for early PONV (P = 0.27, Supplemental Digital Content Figure 3,

Table 3
Table 3:
Risk of bias assessment for each trial
Fig. 4
Fig. 4:
Funnel plot for the primary outcome, overall postoperative nausea and vomiting with weighted regression. The filled circles represent estimated treatment effects (RR) and its precision (standard error) for each individual study. Also, the random-effects estimate (vertical dotted line) as well as the fixed-effect estimate (vertical dashed line) with 95% confidence interval limits (diagonal dashed lines) are shown in the figure. In the absence of small-study effects, the treatment effects (filled circles) would scatter symmetrically around a common average treatment effect resulting in a symmetrical triangle. However, the figure depicts three studies with high standard errors clustering asymmetrically to the left (favouring TIVA) with no corresponding studies on the right side of the figure (favouring IA + AE). IA + AE, inhalational anaesthesia with pharmacological antiemetic prophylaxis; RR, relative risk; TIVA, total intravenous anaesthesia.


The meta-analysis summarises data from 14 prospective randomised trials and a total of more than 2000 patients. In our primary analysis, there was no difference in the overall risk of PONV, POV, need for rescue medication and early PONV in patients receiving TIVA compared with patients receiving IA + AE. However, patients receiving TIVA were more likely to suffer from PONV in the late postoperative phase than patients receiving IA + AE. Also, we found evidence of a publication bias with smaller trials favouring TIVA.

As PONV has generally been attributed to inhalational anaesthetics,1 recommendations to reduce the patient's risk include substituting TIVA for the inhalational anaesthetics. This strategy is supported by the notion that propofol itself provides antiemetic properties.38 As, in different clinical settings, some advantages (or disadvantages) of a type of anaesthetic are judged to be less (or more) relevant than others, factors apart from the risk of PONV may guide a physician's choice of anaesthetic.9 In obese patients, time to awakening from anaesthesia is more precise after desflurane compared with propofol.10 Furthermore, propofol depresses sympathetic outflow and causes arterial hypotension particularly in patients at risk of experiencing adverse cardiovascular events.39 In contrast, sympathetic reflexes and arterial pressure are better maintained with inhalational anaesthetics.39 Moreover, in animal models and highly selected clinical situations, potent organ protective effects have been demonstrated for all inhalational anaesthetics.40,41 On the other hand, as well as decreasing PONV,1 propofol significantly reduces the incidence of postoperative shivering42 and is associated with a decreased incidence of emergence agitation.43 In this context, our analysis shows that, if a physician chooses not to substitute TIVA for an inhalational anaesthetic, then the administration of single-drug antiemetic prophylaxis is sufficient to compensate for the missing reduction in a patient's risk of PONV. Because inhalational anaesthetics increase the incidence of PONV, particularly in the early postoperative phase,44 to investigate differential effects on early and late PONV, we also evaluated PONV by the time of its occurrence. We did not find any difference between the two groups with respect to early PONV. In contrast, our analysis shows that patients receiving TIVA had a higher risk of experiencing PONV in the late postoperative phase, starting 2 to 6 h after surgery. As propofol has antiemetic properties,38 it may be assumed that early protection is caused by residual propofol at the effect sites after cessation of its administration. However, when propofol has been eliminated from the body, patients will be unprotected if additional pharmacological prophylaxis has not been initiated. These findings are of particular relevance for ambulatory surgery when antiemetic rescue medication to treat PONV beginning 2 to 6 h after surgery may not be available.45 The clinical relevance of late PONV following TIVA is supported by the observation that insufficient control of PONV is the leading cause of unexpected hospital admission after ambulatory surgery.46 Thus, the risk of late PONV and postdischarge nausea and vomiting45 needs to be recognised and administration of sufficient prophylaxis initiated.

When considering PONV prophylaxis with antiemetics, potential side-effects warrant discussion: high dosages of most available antiemetics prolong cardiac QT interval.47 For these reasons, the antiemetic droperidol was labelled with a black box warning by the American Food and Drug Administration in 2001. This was followed by an intense debate, with the result that now the dose administered for prophylaxis and treatment of PONV is considered to be well tolerated as it is some 10-fold lower than the doses given historically for neurolept anaesthesia.48–50 Other side-effects are characteristic of specific drugs,7 e.g. headache after 5-hydroxytryptamine (5-HT3) antagonists or sedation after droperidol or dimenhydrinate.51 Although slight increases in blood glucose concentration may be observed after dexamethasone administration in nondiabetic patients,52 low doses of dexamethasone are proven not to increase the risk of surgical site infection.53,54 To add to this discussion, we identified six studies which investigated perioperative adverse events, including dizziness, headache, myalgia, drowsiness and shivering. Except for a trend towards increased postoperative shivering following IA + AE, which has been known for decades to be associated with inhalational anaesthetics,42 we did not find any difference between the two groups. We acknowledge that for the individual side-effects, the data came from only one or two studies and we could not include two of the six studies in the analysis as no data on adverse events were given. Nevertheless, despite the small number of studies, we feel that our findings do not contradict the general notion that the use of low-dose antiemetics for PONV prophylaxis is well tolerated.

Our funnel plot analysis showed significant asymmetry, suggesting a publication bias favouring TIVA. This observation is supported by our subgroup analysis stratifying studies by study size: studies involving more than 100 patients favoured IA + AE, whereas smaller studies tended to favour TIVA. Although other reasons for funnel plot asymmetry exist,16 we decided to extended our analysis to assess the implications for our findings: when heterogeneity is high and random-effects modelling is applied then smaller studies are weighted disproportionately higher than larger studies, so we re-examined our primary analysis using a fixed effect model.16 With fixed effect modelling, TIVA was actually inferior to IA + AE in preventing overall PONV (Fig. 2). We would like to emphasise that this analysis was conducted a posteriori, and its results are merely explorative. To exclude any further bias because of poor study quality, individual risk of bias for each study was assessed, revealing an overall moderate risk. Excluding studies with high risk of bias did not change the results of our primary analysis. Finally, in the face of significant heterogeneity in our primary analysis, we estimated the overall quality of evidence19 to be ‘moderate’.


Even with antiemetic prophylaxis, PONV still occurred in a high proportion (39%) of patients across all studies. Thus, irrespective of their relative effectiveness, neither strategy in isolation reduced patients’ risk of PONV sufficiently. Accordingly, we would emphasise that only a multimodal approach, including administration of multiple prophylaxes, reduces the incidence of PONV satisfactorily in high-risk patients undergoing general anaesthesia. Thus, adding one or more additional antiemetic- to both strategies will further decrease the incidence of PONV. However, the aim of our analysis was to establish the relative effect of single pharmacological prophylaxis compared with TIVA, which is why we excluded studies investigating multiple prophylaxes.

Secondly, it has been shown that titration of inhalational anaesthesia to higher levels of bispectral index (BIS) with avoidance of unnecessary deep anaesthesia reduces the incidence of postoperative vomiting.55 Unfortunately, only one study controlled for anaesthesia depth,31 making a sensitivity analysis impossible. In the IMPACT trial, the BIS level was one of the randomised factors,56 but these data were not published.1 Therefore, it is possible that the relative effectiveness of pharmacological prophylaxis might be even more pronounced when anaesthetic depth is controlled; however, our data do not allow us to make any assumption about this effect.

Third, we pooled groups with different antiemetic classes and subclasses in our main analysis. Therefore, our primary findings are limited to the comparison of TIVA versus ‘uncontrolled’ pharmacological antiemetic prophylaxis and the results might differ between antiemetic classes. As most of the included studies investigated 5-HT3 antagonists, it might be concluded that our findings are more or less restricted to that class. However, as has been shown in the IMPACT trial,1 as well as many others, the prophylactic effect of antiemetics from different classes is comparable when adequate doses are used.7

Finally, although N2O is known to trigger PONV,57 we did not make any preconditions regarding its use. Consequently, administration of N2O varies among studies and has also been administered to ‘TIVA’ groups in four trials. In each of these, the corresponding IA + AE groups received the same dose of N2O. Because there is no specific interaction between N2O and propofol or pharmacological antiemetic prophylaxis,1 we tend to assume that N2O independently increased the risk for PONV in both groups to the same amount. In contrast, N2O was administered only in the IA + AE groups in three studies. Accordingly, we performed a sensitivity analysis excluding these studies which did not change any of our main findings. However, without these studies, patients receiving TIVA were even more likely to suffer from late PONV (risk ratio 1.51, as compared with 1.41).

In summary, our meta-analysis shows that when substitution of TIVA for inhalational anaesthetics to reduce a patient's risk of PONV is not an option, then the PONV risk can be compensated to an equivalent degree by the use of a single-drug pharmacological prophylaxis. Finally, as neither strategy alone decreases PONV sufficiently in high-risk patients, the use of multiple prophylactic methods should be instituted to minimise the incidence of PONV and improve surgical outcome.

Acknowledgements relating to this article

Assistance with the study: none.

Financial support and sponsorship: this work was supported by departmental funding of the Department of Anaesthesiology, University Hospital Düsseldorf, Germany and the Department of Anaesthesiology and Critical Care, University Hospital Würzburg, Germany.

Conflicts of interest: PeKr has conducted funded Research by Acacia Ltd. and has received fees for lectures by Fresenius Kabi, Baxter GmbH, Germany and ProStrakan and is co-author of the cited consensus guideline on the management of PONV (SAMBA-Guideline), published by TJ Gan et al. PeKi has been consulting for Baxter GmbH Germany and Air Liquide Medical GmbH Germany, has received lecture fees and travelling expenses from both companies and is an associated editor of BMC Anesthesiology.

Presentation: preliminary data for this study have been presented at the European Society of Anaesthesiology meeting, 28 to 30 May 2016, London.

Comment from the editor: PeKr is an associate editor of the European Journal of Anaesthesiology.


1. Apfel CC, Korttila K, Abdalla M, et al. A factorial trial of six interventions for the prevention of postoperative nausea and vomiting. N Engl J Med 2004; 350:2441–2451.
2. Skolnik A, Gan TJ. Update on the management of postoperative nausea and vomiting. Curr Opin Anaesthesiol 2014; 27:605–609.
3. Gan TJ. Postoperative nausea and vomiting: can it be eliminated? JAMA 2002; 287:1233–1236.
4. Weiser TG, Regenbogen SE, Thompson KD, et al. An estimation of the global volume of surgery: a modelling strategy based on available data. Lancet 2008; 372:139–144.
5. Macario A, Weinger M, Carney S, et al. Which clinical anesthesia outcomes are important to avoid? The perspective of patients. Anesth Analg 1999; 89:652–658.
6. Hill RP, Lubarsky DA, Phillips-Bute B, et al. Cost-effectiveness of prophylactic antiemetic therapy with ondansetron, droperidol, or placebo. Anesthesiology 2000; 92:958–967.
7. Gan TJ, Diemunsch P, Habib AS, et al. Consensus guidelines for the management of postoperative nausea and vomiting. Anesth Analg 2014; 118:85–113.
8. Apfel CC, Greim CA, Haubitz I, et al. A risk score to predict the probability of postoperative vomiting in adults. Acta Anaesthesiol Scand 1998; 42:495–501.
9. Jakobsson J. Desflurane: a clinical update of a third-generation inhaled anaesthetic. Acta Anaesthesiol Scand 2012; 56:420–432.
10. Xu L, Wu A-S, Yue Y. The incidence of intra-operative awareness during general anesthesia in China: a multicenter observational study. Acta Anaesthesiol Scand 2009; 53:873–882.
11. Juvin P, Vadam C, Malek L, et al. Postoperative recovery after desflurane, propofol, or isoflurane anesthesia among morbidly obese patients: a prospective, randomized study. Anesth Analg 2000; 91:714–719.
12. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009; 339:b2700.
13. Carlisle JB, Stevenson CA. Drugs for preventing postoperative nausea and vomiting. Cochrane Database Syst Rev 2006; 3:CD004125.
14. Biondi-Zoccai GGL, Agostoni P, Abbate A, et al. A simple hint to improve Robinson and Dickersin's highly sensitive PubMed search strategy for controlled clinical trials. Int J Epidemiol 2005; 34:224–225.
15. Rücker G, Schwarzer G, Carpenter J. Arcsine test for publication bias in meta-analyses with binary outcomes. Stat Med 2008; 27:746–763.
16. Sterne JAC, Sutton AJ, Ioannidis JPA, et al. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ 2011; 343:d4002.
17. Higgins JPT, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ 2003; 327:557–560.
18. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 21:1539–1558.
19. Guyatt GH, Oxman AD, Sultan S, et al. GRADE guidelines: 9. Rating up the quality of evidence. J Clin Epidemiol 2011; 64:1311–1316.
20. Guyatt GH, Oxman AD, Santesso N, et al. GRADE guidelines: 12. Preparing summary of findings tables-binary outcomes. J Clin Epidemiol 2013; 66:158–172.
21. Kranke P, Apfel CC, Roewer N, et al. Reported data on granisetron and postoperative nausea and vomiting by Fujii et al. Are incredibly nice!. Anesth Analg 2000; 90:1004–1007.
22. Carlisle JB. The analysis of 168 randomised controlled trials to test data integrity. Anaesthesia 2012; 67:521–537.
23. Tramèr MR. The Boldt debacle. Eur J Anaesthesiol 2011; 28:393–395.
24. Shafer SL. Editor's note: notice of retraction. Anesth Analg 2014; 119:1225.
25. Eberhart LH, Bernert S, Wulf H, et al. [Pharmacoeconomical model for cost calculation using a study on prophylaxis of nausea and vomiting in the postoperative phase as an example. Cost effectiveness analysis of a tropisetron supplemented desflurane anaesthesia in comparison to a propofol total intravenous anaesthesia (TIVA)]. Anaesthesist 2002; 51:475–481.
26. Gan TJ, Ginsberg B, Grant AP, et al. Double-blind, randomized comparison of ondansetron and intraoperative propofol to prevent postoperative nausea and vomiting. Anesthesiology 1996; 85:1036–1042.
27. Heinke W, Frank T, Meier P, et al. [Postoperative vomiting after pars plana vitrectomy]. Anaesthesiol Reanim 1996; 21:47–50.
28. Jellish WS, Leonetti JP, Murdoch JR, et al. Propofol-based anesthesia as compared with standard anesthetic techniques for middle ear surgery. J Clin Anesth 1995; 7:292–296.
29. Jokela RM, Kangas-Saarela TA, Valanne JV, et al. Postoperative nausea and vomiting after sevoflurane with or without ondansetron compared with propofol in female patients undergoing breast surgery. Anesth Analg 2000; 91:1062–1065.
30. Khan P. Comparative study between granisetron ondansetron and propofol for the prevention of emesis after gynaecological laproscopy. J Postgrad Med Inst 2005; 19:135–143.
31. Mei W, Li M, Yu Y, et al. Tropisetron alleviate early postoperative pain after gynecological laparoscopy in sevoflurane based general anaesthesia: a randomized, parallel-group, factorial study. Eur J Pain 2014; 18:238–248.
32. Özünlü O, Karaman S, Firat V. Anaesthetic technique and postoperative nausea, vomiting and cognitive functions in outpatients undergoing gynecologic laparoscopic surgery. Türk Anest Rean Der Dergisi 2005; 33:209–216.
33. Paech MJ, Lee BH, Evans SF. The effect of anaesthetic technique on postoperative nausea and vomiting after day-case gynaecological laparoscopy. Anaesth Intensive Care 2002; 30:153–159.
34. Park SH, Lee HG, Jeong CY, et al. Postoperative nausea and vomiting after total thyroidectomy: sevoflurane combined with prophylactic ramosetron vs. propofol-based total intravenous anesthesia. Korean J Anesth 2014; 66:216–221.
35. Park SK, Cho EJ. A randomized controlled trial of two different interventions for the prevention of postoperative nausea and vomiting: total intravenous anaesthesia using propofol and remifentanil versus prophylactic palonosetron with inhalational anaesthesia using sevoflurane-nitrous oxide. J Int Med Res 2011; 39:1808–1815.
36. Purhonen S, Koski EM, Niskanen M, et al. Efficacy and costs of 3 anesthetic regimens in the prevention of postoperative nausea and vomiting. J Clin Anesth 2006; 18:41–45.
37. White H, Black RJ, Jones M, et al. Randomized comparison of two antiemetic strategies in high-risk patients undergoing day-case gynaecological surgery. Br J Anaesth 2007; 98:470–476.
38. Ewalenko P, Janny S, Dejonckheere M, et al. Antiemetic effect of subhypnotic doses of propofol after thyroidectomy. Br J Anaesth 1996; 77:463–467.
39. Neukirchen M, Kienbaum P. Sympathetic nervous system: evaluation and importance for clinical general anesthesia. Anesthesiology 2008; 109:1113–1131.
40. Landoni G, Greco T, Biondi-Zoccai G, et al. Anaesthetic drugs and survival: a Bayesian network meta-analysis of randomized trials in cardiac surgery. Br J Anaesth 2013; 111:886–896.
41. Li F, Yuan Y. Meta-analysis of the cardioprotective effect of sevoflurane versus propofol during cardiac surgery. BMC Anesthesiol 2015; 15:128.
42. Crossley AW. Six months of shivering in a district general hospital. Anaesthesia 1992; 47:845–848.
43. Costi D, Cyna AM, Ahmed S, et al. Effects of sevoflurane versus other general anaesthesia on emergence agitation in children. Cochrane Database Syst Rev 2014; 9:CD007084.
44. Apfel CC, Kranke P, Katz MH, et al. Volatile anaesthetics may be the main cause of early but not delayed postoperative vomiting: a randomized controlled trial of factorial design. Br J Anaesth 2002; 88:659–668.
45. Apfel CC, Philip BK, Cakmakkaya OS, et al. Who is at risk for postdischarge nausea and vomiting after ambulatory surgery? Anesthesiology 2012; 117:475–486.
46. Gold BS, Kitz DS, Lecky JH, et al. Unanticipated admission to the hospital following ambulatory surgery. JAMA 1989; 262:3008–3010.
47. Staikou C, Stamelos M, Stavroulakis E. Impact of anaesthetic drugs and adjuvants on ECG markers of torsadogenicity. Br J Anaesth 2013; 112:217–230.
48. Nuttall GA, Eckerman KM, Jacob KA, et al. Does low-dose droperidol administration increase the risk of drug-induced QT prolongation and torsade de pointes in the general surgical population? Anesthesiology 2007; 107:531–536.
49. Habib AS, Gan TJ. Food and drug administration black box warning on the perioperative use of droperidol: a review of the cases. Anesth Analg 2003; 96:1377–1379.
50. Gan TJ, White PF, Scuderi PE, et al. FDA ‘black box’ warning regarding use of droperidol for postoperative nausea and vomiting: is it justified? Anesthesiology 2002; 97:287.
51. Schaub I, Lysakowski C, Elia N, et al. Low-dose droperidol (≤1 mg or ≤15 (g kg-1) for the prevention of postoperative nausea and vomiting in adults: quantitative systematic review of randomised controlled trials. Eur J Anaesthesiol 2012; 29:286–294.
52. Abdelmalak BB, Bonilla AM, Yang D, et al. The hyperglycemic response to major noncardiac surgery and the added effect of steroid administration in patients with and without diabetes. Anesth Analg 2013; 116:1116–1122.
53. Kurz A, Fleischmann E, Sessler DI, et al. Effects of supplemental oxygen and dexamethasone on surgical site infection: a factorial randomized trial. Br J Anaesth 2015; 115:434–443.
54. Bolac CS, Wallace AH, Broadwater G, et al. The impact of postoperative nausea and vomiting prophylaxis with dexamethasone on postoperative wound complications in patients undergoing laparotomy for endometrial cancer. Anesth Analg 2013; 116:1041–1047.
55. Nelskylä KA, Yli-Hankala AM, Puro PH, et al. Sevoflurane titration using bispectral index decreases postoperative vomiting in phase II recovery after ambulatory surgery. Anesth Analg 2001; 93:1165–1169.
56. Apfel CC, Korttila K, Abdalla M, et al. An international multicenter protocol to assess the single and combined benefits of antiemetic interventions in a controlled clinical trial of a 2x2x2x2x2x2 factorial design (IMPACT). Control Clin Trials 2003; 24:736–751.
57. Myles PS, Leslie K, Chan MTV, et al. The safety of addition of nitrous oxide to general anaesthesia in at-risk patients having major noncardiac surgery (ENIGMA-II): a randomised, single-blind trial. Lancet 2014; 384:1446–1454.

Supplemental Digital Content

© 2016 European Society of Anaesthesiology