The common and stressful adverse effects of postoperative nausea and vomiting (PONV) are a major concern in 30% to 70% of high-risk groups.1 Accordingly, numerous studies have assessed the efficacies of different drugs and treatment protocols in preventing PONV. Agents including anticholinergics, antidopaminergics, antihistamines, steroids, neuroleptics, and serotonin antagonists have shown variable effects.2–5 Hence, PONV remains a complication encountered frequently during the postoperative period.
Midazolam is a short-acting benzodiazepine that affects central benzodiazepine receptors, which contain a γ-aminobutyric acid receptor site and a chloride ion channel.6 Midazolam has a sedative effect and can be used for coinduction, because it reduces the requirements for propofol and thereby lowers the incidence of the cardiovascular effects of propofol.7 Midazolam is the most commonly used sedative agent in the ambulatory setting, most likely because it has fewer depressive cardiorespiratory effects than propofol.8
Numerous studies have reported that midazolam effectively prevents PONV.9–12 The use of midazolam not only may reduce the incidence of PONV but may also provide an anxiolytic effect. Moreover, midazolam may offer the benefits of lower cost and fewer side effects, such as headaches and extrapyramidal symptoms, that have been reported with other antiemetics.13 In several studies, such antiemetics did not provide superior benefits or outcomes compared with midazolam.14,15 However, the findings are variable, and the reported outcomes from several studies are conflicting, especially when midazolam has been used in combination with other antiemetics. Therefore, we aimed to determine the effectiveness of midazolam in preventing PONV by performing a systematic review and meta-analysis.
Study Protocol Registration
This systematic review was registered in PROSPERO (CRD42014009425) and was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines.16
We conducted a systematic review and meta-analysis of randomized controlled trials that investigated the effectiveness of midazolam in preventing PONV in comparison with controls. Studies combining midazolam and other drugs were included, provided the only difference between the groups was the use of midazolam. The MEDLINE®, Embase, and Cochrane Central Register of Controlled Trials (CENTRAL) databases were searched for all relevant articles written in English up to March 1, 2014 (inclusive). In addition, the reference lists of the full articles retrieved were searched manually. The search strategy, which included a combination of free text, Medical Subject Headings, and EMTREE terms, is described in the Appendix.
Inclusion and exclusion criteria were determined before the systematic search. Two authors independently scanned the titles and abstracts of the reports identified via the search strategies described earlier. If a report was determined eligible from the title or abstract, the full article was retrieved. Potentially relevant studies chosen by at least one author were retrieved, and full-text versions were evaluated. Articles that met the inclusion criteria were assessed separately by 2 authors, and any discrepancies were resolved through discussion. If no agreement could be reached, a third investigator was involved to provide a resolution.
Inclusion and Exclusion Criteria
We included randomized controlled trials that collected data on the effectiveness of midazolam as a prophylactic agent for PONV compared with a control. Control groups were compared with an experimental group, including subjects who did not receive a drug or who received another drug, with the only difference between the groups of midazolam administration (e.g., midazolam + ondansetron versus ondansetron). The studies also included adult patients undergoing general anesthesia involving IV administration of midazolam. Data from abstracts, posters, case reports, comments, or letters to the editor, reviews, and animal studies were excluded. Studies that were not published in English were likewise excluded (Fig. 1).
The primary end points were the incidence of postoperative nausea (PON), postoperative vomiting (POV), and PONV. The severity of PON was also assessed. The use of rescue antiemetics and the incidence of adverse effects, such as headache, dizziness, and sedation, were secondary outcomes in the systematic review.
Subgroup analysis was performed regardless of whether PON, POV, and PONV incidence data were collected as primary outcomes or secondary outcomes. Also, a subgroup analysis was performed to rule out the effects of propofol as an induction agent. In addition, administration times were divided into 2 phases representing the induction and conclusion of surgery. In studies with >1 midazolam group or control group, all midazolam groups or all control groups were combined to avoid a unit of analysis error.
The quality of eligible studies was assessed independently by 2 members of the review group using the “risk of bias” tool within the Review Manager software program (version 5.2, The Cochrane Collaboration, Oxford, UK). Quality was evaluated based on potential sources of bias, including random sequence generation, allocation concealment, blinding of the participants, blinding of outcome assessment, incomplete outcome data, selective reporting, and overall risk bias. The data were then cross-checked. The methodology of each trial was graded as “high,” “low,” or “unclear,” to reflect a high risk or a low risk of bias, and uncertainty regarding the risk of bias, respectively.17
All interrelated data from the included studies were independently extracted and entered into a spreadsheet by 2 authors, and then cross-checked. Any discrepancy was resolved through discussion. If an agreement could not be reached, the dispute was resolved with the aid of a third investigator. The spreadsheet included the following items: (1) title, (2) authors, (3) name of journal, (4) publication year, (5) study design, (6) registration of clinical trial, (7) competing interest, (8) country, (9) risk of bias, (10) number of patients in study, (11) dose of midazolam and control drugs, (12) sex, (13) age, (14) weight, (15) height of patients, (16) duration of anesthesia, (17) ASA physical status, (18) inclusion criteria, (19) exclusion criteria, (20) type of surgery, (21) airway device, (22) induction agent, (23) maintenance agent, (24) use of nitrous oxide, (25) use of an opioid during the perioperative period, (26) timing of administration of the experimental drug, (27) other drugs used during surgery, (28) timing of rescue antiemetics, (29) rescue analgesics, (30) definition of nausea, vomiting, and retching, (31) number of cases of PON, POV, and PONV overall and during the early and late postoperative phases, and (32) the need for rescue antiemetics.
The data were initially extracted from tables or text. In cases involving missing or incomplete data, an attempt was made to contact the study authors to obtain the relevant information. If the author did not respond or did not have current information, the data were extracted from available figures.
The review and meta-analysis were conducted using Comprehensive meta-analysis software (version 2.0, Biostat, Englewood, NJ) and Review Manager software (version 5.2, The Cochrane Collaboration, Oxford, UK). For dichotomous data, a pooled risk ratio (RR) and 95% confidence intervals (CIs) were calculated. If the 95% CI included a value of 1, the difference was not considered statistically significant. We calculated the mean difference (MD) for continuous data and reported the 95% CI. If the 95% CI included a value of 0, the difference was not considered statistically significant.
We used the χ2 test and the I 2 test for heterogeneity. If the P value was <0.10 or an I 2 value was >50%, the heterogeneity was considered significant. A fixed-effects model was selected if P value for χ2 test >0.10 and the I 2 value was <50%. In cases with I 2 value >50%, random-effects models were used. Because all the number of combined studies that showed substantial heterogeneity were <10, t statistics (Hartung-Knapp-Sidik-Jonkman method) were used instead of Z test in all random-effects analysis to lower the error rate. A subgroup analysis was performed based on the timing of the PON, POV, and PONV assessments; the time of midazolam administration; the use of perioperative opioids; the use of propofol as an induction agent; and whether the aim of the study included PONV as a primary or secondary outcome. If heterogeneity was >50%, sensitivity analysis was likewise performed by excluding one study at a time to evaluate the influence of a single study on the overall effect estimate.
We calculated the number needed to treat (NNT) using a 95% CI based on the absolute risk reduction as an estimate of the overall clinical impact of the intervention.18 Publication bias was identified by the Egger linear regression test, and a P value of <0.05 was used to identify the presence of a publication bias, or funnel plots for each data set were visually assessed for asymmetry. If a publication bias was present, a trim-and-fill analysis was performed to evaluate the effect of the bias.
Literature Search and Study Characteristics
A search of the MEDLINE, Embase, and CENTRAL databases returned 1296 studies that were initially evaluated. After excluding duplicates, 1197 studies remained. Another 1040 studies were excluded after the review of titles and abstracts. Studies that were not published in English were likewise excluded. Of the 151 studies that remained, 83 additional studies were excluded because 39 of the studies involved children and 44 studies involved administration of experimental drugs. Forty-two studies were excluded because they were performed under monitored or regional anesthesia care and not under general anesthesia. The full texts of the remaining 26 studies were reviewed in detail, and 11 additional studies were excluded because 8 did not include a control group,15,19–25 2 administered midazolam for therapeutic management,14,26 and 1 did not report findings for PON, POV, or PONV.27 Manual review of the references of the included studies identified 13 additional studies for assessment. However, only one met the inclusion criteria for the present review.28 Collectively, 16 studies met the inclusion criteria and were included in this systematic review and meta-analysis.6,8–10,12,14,26,28–38
The characteristics of the 16 studies that met the inclusion criteria are summarized in Tables 1, 2, and 3. Midazolam was administered at induction in 11 studies8–10,12,29,30,32,33,35,37,38 and at the end of the surgery in 3 studies.6,28,31 Two studies compared the timing of administration,34,36 2 studies included a nausea visual analog scale for assessing the severity of PON,10,32 and 3 studies did not involve administration of perioperative opioids.12,37,38 Twelve of the 16 studies6,10,12,28,29,31–34,36–38 were aimed at comparing the incidence of PONV, and 4 studies involved the collection of PONV incidence as a secondary end point.8,9,30,35
In the studies by Park et al.34 and Safavi and Honarmand,36 subjects were divided into groups based on drug administration at the induction or conclusion of surgery. The group receiving midazolam or placebo administration at induction was designated Park1 and Safavi1, respectively, and the groups receiving the drugs at the end of surgery were designated Park2 and Safavi2, respectively. To avoid duplicate counting, the groups were divided in both studies.
In the study by Honarmand et al.,6 4 separate cohorts received haloperidol, midazolam, haloperidol plus midazolam, or saline. Similarly, 4 separate cohorts in the study conducted by Makhdoom and Farid33 received dexamethasone, midazolam, dexamethasone plus midazolam, or saline. Therefore, comparison of midazolam with a placebo was designated Honarmand1 and Makhdoom1, respectively, and the comparison of midazolam plus another drug with the second drug was designated Honarmand2 and Makhdoom2, respectively.
The postoperative period was divided into the overall, early, and late phases. The early phase was defined as 0 to 6 hours postoperatively, and the late phase as 6 to 24 hours postoperatively. The overall phase was included to capture the maximal number of studies that contained PON, POV, and PONV data with a variable data collection period and was defined as the first period of data collection (Table 4). Because several studies defined the early and late phase differently, we combined the earliest period of data collection with early incidence of PON, POV, PONV, and severity of PON. For the early phase, 3 studies included postanesthesia care unit data,8,28,30 4 studies included data collected at 0 to 2 hours,32,34,36,37 and 1 study collected data at 0 to 60 minutes.35 In addition, the widest overlapping period of data collection for late incidence of PON, POV, PONV, and late PON severity was combined. For the late phase, one study included data from 8 to 14 hours,10 a second study contained 2- to 24-hour data,34 3 studies reported 6- to 12-hour data,32,36,37 and 1 additional study included 6- to 24-hour data.28
Risk of Bias
Fifteen studies referenced the random sequence generation method,6,8–10,12,29–38 and only 2 studies described allocation concealment.32,34 In every study, outcome assessors were blinded, and no incomplete data were reported. The overall risks of bias are shown in Table 5.
Overall, Early, and Late PON
Eleven studies compared the effectiveness of midazolam in preventing overall PON with those of a placebo or another drug.8,9,12,29–31,33–35,37,38 Five studies compared its effectiveness in reducing the incidence of early PON,8,30,34,35,37 and 2 studies compared its effects on late PON.34,37 Analysis of the combined results revealed that midazolam was associated with a statistically significant reduction in the overall incidence of PON (risk ratio [RR], 0.51; 95% CI, 0.40–0.65; I 2 = 35%;
= 0.10; NNT = 6; number of included studies [n] = 11) compared with the control groups. However, the effect of midazolam on the incidence of early and late PON compared with the control groups is unknown (RR, 0.56; 95% CI, 0.17–1.82;
= 0.069; I 2 = 51%; NNT = 7; n = 5; and RR, 0.74; 95% CI, 0.42–1.31;
= 0.14; I 2 = 49.7%; NNT = 11; n = 2, respectively).
Effect of Administration Timing on Overall PON
Nine studies8,12,29,30,33–35,37,38 analyzed the effectiveness of midazolam administered at induction of surgery in preventing overall PON, whereas 2 studies31,34 assessed its effectiveness after administration at the conclusion of surgery. The combined results of both analyses indicated that midazolam caused a statistically significant reduction in the incidence of overall PON, regardless of the timing of administration (RR, 0.53; 95% CI, 0.39–0.71; I 2 = 31%;
= 0.16; NNT = 7; n = 9; and RR, 0.30; 95% CI, 0.14–0.66; I 2 = 0%;
= 0.61; NNT = 4; n = 2, respectively; Fig. 2).
Severity of PON
Two studies10,32 compared the effects of midazolam on the severity of PON using a visual analog scale score. The postoperative periods of the 2 studies were divided into the early phase (0–6 hours after surgery) and the late phase (6–24 hours after surgery). However, we were not able to clarify the effect on the severity of PON after administration of midazolam in the early or late postoperative phases when compared with a placebo (MD, −0.38; 95% CI, −3.39 to 2.64; I 2 = 77.2%;
= 0.036; n = 2; and MD, −0.04; 95% CI, −0.71 to 0.62; I 2 = 0%;
= 0.39; n = 2, respectively).
Effect of Midazolam in Conjunction with Propofol as an Induction Agent on Overall PON
Seven studies8,9,12,30,31,35,38 evaluated the effectiveness of midazolam in reducing overall PON with the use of propofol as an induction agent, whereas 3 studies33,34,37 assessed its effectiveness without propofol. The combined results of both analyses revealed that midazolam caused a statistically significant reduction in the incidence of overall PON, regardless of propofol administration (RR, 0.64; 95%, CI 0.47–0.87; I 2 = 45%;
= 0.09; NNT = 9; n = 7; and RR, 0.39; 95% CI, 0.23–0.65; I 2 = 35%;
= 0.19; NNT = 5; n = 3, respectively).
Data Collected as Primary or Secondary Outcomes
PON incidence was collected as a primary end point in 7 studies,12,29,31,33,34,37,38 whereas 4 studies8,9,30,35 collected PON incidence as a secondary end point. The combined results of both analyses regarding primary and secondary outcomes revealed that midazolam caused a statistically significant reduction in the overall incidence of PON, regardless of the aim of the study (RR, 0.46; 95% CI, 0.34–0.63; I 2 = 5%;
= 0.39; NNT = 5; n = 7; and RR, 0.64; 95% CI, 0.44–0.93; I 2 = 42%;
= 0.16; NNT = 8; n = 4, respectively).
Overall, Early, and Late POV
Ten studies assessed the effects of midazolam on the overall POV,10,12,29–34,37,38 whereas 4 studies compared the effects on the incidence of early POV,30,32,34,37 and 3 studies evaluated its effects on the incidence of late POV.32,34,37 The combined results revealed that midazolam caused a statistically significant reduction in the incidence of overall POV (RR, 0.46; 95% CI, 0.33–0.65; I 2 = 0%;
= 0.87; NNT = 8; n = 10; Fig. 3). Also, the results suggest the beneficial effect of midazolam on the incidence of early or late POV when compared with a placebo, but these did not reach the statistical significance (RR, 0.48; 95% CI, 0.22–1.01; I 2 = 0%;
= 0.54; NNT = 16; n = 4; and RR, 0.40; 95% CI, 0.13–1.27; I 2 = 0%;
= 0.85; NNT = 23; n = 3, respectively).
Effect of Administration Timing on Overall POV
Eight studies10,12,29,30,32–34,37 assessed the effects of midazolam on the overall POV when administered at induction, and 2 studies31,34 assessed the effects of midazolam administered at the conclusion of surgery. The combined results revealed that midazolam administered at induction caused a statistically significant reduction in the incidence of overall POV (RR, 0.45; 95% CI, 0.31–0.65; I 2 = 0%;
= 0.61; I 2 = 0%; NNT = 7; n = 8). Also, the effect of midazolam administered at the end of surgery showed the tendency of reduction in the incidence of overall POV (RR, 0.49; 95% CI, 0.20–1.19; I 2 = 0%;
= 0.56; NNT = 8; n = 2).
Effect of Midazolam in Conjunction with Propofol as an Induction Agent on Overall POV
Four studies12,30,31,38 assessed the effectiveness of midazolam administered with propofol as an induction agent in preventing overall PON, whereas 5 studies14,32–34,37 assessed its effectiveness without propofol. The combined results of both analyses revealed that midazolam caused a statistically significant reduction in the incidence of overall PON, regardless of propofol administration (RR, 0.37; 95% CI, 0.20–0.68; I 2 = 0%;
= 0.42; NNT = 8; n = 4; and RR, 0.54; 95% CI, 0.35–0.84; I 2 = 0%;
= 0.87; NNT = 8; n = 5, respectively).
Data Collected as Primary or Secondary Outcomes
Nine studies10,12,29,31–34,37,38 collected the incidence of POV as a primary end point, whereas only 1 study30 collected POV incidence as a secondary end point. The results of the combined analysis revealed that midazolam caused a statistically significant reduction in the overall incidence of POV as a primary outcome (RR, 0.48; 95% CI, 0.34–0.68; I 2 = 0%;
= 0.90; NNT = 8; n = 9). However, the analysis of POV incidence as a secondary end point could not be performed because of the small number of studies included.
Postoperative Nausea and Vomiting
Overall, Early, and Late PONV
Seven studies assessed the effects of midazolam on overall PONV,6,12,31,33,36–38 4 investigations compared the effects on the incidence of early PONV,6,32,34,36 and 2 studies evaluated the effects on the incidence of late PONV.28,36 The combined results revealed that midazolam caused a statistically significant reduction in the incidence of overall, early, and late PONV (RR, 0.45; 95% CI, 0.36–0.57; I 2 = 31%;
= 0.16; NNT = 3; n = 7; RR, 0.50; 95% CI, 0.32–0.79; I 2 = 30%;
= 0.22; NNT = 6; n = 4; and RR, 0.26; 95% CI, 0.12–0.59; I 2 = 43%;
= 0.17; NNT = 5; n = 2, respectively; Fig. 4).
Effect of Administration Timing on Overall PONV
Five studies12,33,36–38 assessed the effects of midazolam on overall PONV after administration at surgical induction, and 3 investigations6,28,36 assessed its effectiveness at the end of surgery. The combined results indicated that midazolam administration at both the induction and the conclusion of surgery caused statistically significant reductions in the incidence of overall PONV (RR, 0.44; 95% CI, 0.33–0.59; I 2 = 24%;
= 0.25; NNT = 3; n = 5; and RR, 0.50; 95% CI, 0.31–0.78; I 2 = 48%;
= 0.12; NNT = 4; n = 3, respectively).
Effects on Overall PONV With or Without Perioperative Opioids
Five studies6,28,31,33,36 compared the effectiveness of midazolam in reducing the overall PONV in combination with perioperative opioids, and 3 studies12,37,38 compared its effectiveness without perioperative opioids. The combined results revealed that midazolam administered with perioperative opioids caused a statistically significant reduction in the incidence of overall PONV (RR, 0.43; 95% CI, 0.32–0.59; I 2 = 1%;
= 0.42; NNT = 3; n = 5). Also, midazolam administered without perioperative opioids showed beneficial effect on the incidence of overall PONV, but this did not reach the statistical significance (RR, 0.46; 95% CI, 0.10–2.12; I 2 = 66.4%;
= 0.05; NNT = 3; n = 3).
Effect of Midazolam in Conjunction with Propofol as an Induction Agent on Overall PONV
Four studies12,31,36,38 analyzed the effectiveness of midazolam in reducing the overall PONV when used in conjunction with propofol as an induction agent, whereas 4 separate studies6,28,33,38 assessed its effectiveness without the administration of propofol as an induction agent. The combined results of both analyses revealed that midazolam caused a statistically significant reduction in the incidence of overall PON independent of propofol administration (RR, 0.48; 95% CI, 0.36–0.65; I 2 = 9%;
= 0.35; NNT = 3; n = 4; and RR, 0.41; 95% CI, 0.29–0.57; I 2 = 42%;
= 0.12; NNT = 3; n = 4, respectively).
Eleven studies9,12,28,29,31–36,38 compared the requirement for rescue antiemetics among midazolam recipients and controls. The combined results revealed statistically significant differences in the requirement for rescue antiemetics, but with substantial heterogeneity (RR, 0.52; 95% CI, 0.37–0.74; I 2 = 53%;
= 0.009; NNT = 5; n = 11). Exclusion of the study by Gilliland et al.9 before a sensitivity analysis revealed a decrease in heterogeneity and a statistically significant reduction in the requirement for rescue antiemetics (RR, 0.44; 95% CI, 0.34–0.56; I 2 = 0%;
= 0.67; NNT = 4; n = 10). In addition, the same outcome was observed when the study by Yeo et al.38 was excluded (RR, 0.57; 95% CI, 0.45–0.71; I 2 = 47%;
= 0.03; NNT = 5; n = 10).
Effect of Midazolam in Conjunction with Propofol as an Induction Agent in the Requirement for Rescue Antiemetics
The effectiveness of midazolam with propofol as an induction agent in cases requiring rescue antiemetics was analyzed in 6 studies.9,12,31,35,36,38 Four studies28,32–34 assessed the effectiveness of midazolam without the use of propofol as an induction agent. The combined results of combined analyses indicated that midazolam with propofol as an induction agent showed no evidence of difference in the use of rescue antiemetics (RR, 0.54; 95% CI, 0.024–1.09;
= 0.002; I 2 = 72%; NNT = 7; n = 6). Conversely, the effects of midazolam on the requirement for rescue antiemetics without propofol revealed statistically significant reductions without heterogeneity (RR, 0.50; 95% CI, 0.31–0.79; I 2 = 0%;
= 0.84; NNT = 7; n = 4).
Data Collected as Primary or Secondary Outcomes
Seven studies6,12,31,33,36–38 collected the incidence of PONV as a primary end point, but there were no studies that collected PONV incidence as a secondary end point. The results of the combined analysis revealed that midazolam caused a statistically significant reduction in the overall incidence of PONV as a primary outcome (RR, 0.45; 95% CI, 0.36–0.57; I 2 = 31%;
= 0.16; NNT = 3; n = 7).
The incidence of headaches among midazolam recipients was compared with controls in 4 studies.6,12,34,38 The analysis of the combined findings showed a neutral result (RR, 1.18; 95% CI, 0.56–2.49; I 2 = 0%;
= 0.86; NNT = 1045; n = 4).
The overall effect of midazolam on the incidence of dizziness was assessed in 4 studies,12,31,34,38 but there was no statistically significant effect of midazolam relative to the placebo (RR, 0.90; 95% CI, 0.46–1.77; I 2 = 0;
= 0.79; NNT = 53; n = 4).
The overall effect of midazolam on the incidence of postoperative sedation was assessed in 4 studies,9,12,31,34 but there were no statistically significant effect (RR, 1.33; 95% CI, 0.65–2.71; I 2 = 58%;
= 0.04; NNT = 15; n = 4).
Evidence of a publication bias was detected by the Egger linear regression test (P < 0.05; Table 4). Similarly, the funnel plot analyses for these values demonstrated asymmetry (Fig. 5). However, the funnel plots became symmetrical after adjustment using the trim-and-fill method (Fig. 5). The results before and after elimination of the effect of publication bias are shown in Table 6. The significance of the RR with 95% CIs did not change after the elimination of the publication bias.
The results of the current study suggested that midazolam reduced the overall incidence of PON, POV, and PONV, with an NNT of 3 for overall PONV. Furthermore, the results indicated that midazolam treatment can prevent nausea and vomiting in approximately 1 in 3 patients who would otherwise continue to suffer from PONV if administered a placebo.18 In addition, the need for rescue antiemetics was reduced among patients in the midazolam groups compared with control groups, and no evidence found that midazolam was associated with adverse effects, such as headache, dizziness, or sedation.
Midazolam was shown to have a preventive effect on PONV in the early, late, and overall recovery period and was also effective in preventing overall PON and POV. However, we were not able to clarify the effects of midazolam on the early and late recovery period of PON or POV. The combined results of early and late PON were considerably heterogeneous, otherwise there was no substantial heterogeneity in POV. It seems that subjective judgments of feeling nausea lead to a large variation among collected data, unlike vomiting.
The combined results of 2 studies indicated that midazolam appeared to be effective in preventing late PONV. However, 1 of the 2 studies28 administered midazolam as a continuous patient-controlled analgesia infusion, which might have adversely affected the results. Therefore, further studies are needed to evaluate the duration of the antiemetic effect. In addition, many patients currently receive antiemetics as opposed to no treatment. Therefore, further studies or a systematic review comparing midazolam with other antiemetics would be needed to confirm the potency of the antiemetic effects of midazolam.
Preoperative anxiety may increase gastric acid secretion, lower gastric pH, and delay gastric emptying, which leads to an increase in PONV.29 However, several studies reported that anxiety was not related to the incidence of PONV.39–41 In addition, the consensus guidelines for the management of PONV42 state that in adults, anxiety is not a proven risk factor and may be of limited clinical relevance. In the study conducted by Safavi and Honarmand,36 the incidence of PONV was significantly lower among subjects who received midazolam at induction compared with those who received the agent at the end of surgery. However, Park et al.,34 reported that no significant differences were observed between administration of midazolam at induction or at the end of surgery. In this meta-analysis, a statistically significant reduction in the incidence of PON, POV, and PONV relative to controls was observed after midazolam administration at induction. However, only the incidence of PON was significantly reduced after administration of midazolam at the end of surgery when compared with the control group. This would imply that the anxiolytic effect of midazolam may be important in its PONV efficacy, which is not in accord with the consensus guidelines.
In the study by Splinter et al.,43 a 50 to 75 μg/kg midazolam dose was recommended for prophylactic antiemetic use. In the current meta-analysis, the doses of midazolam in the reviewed studies ranged from 35 to 75 μg/kg. However, the 35 to 75 μg/kg dosing range might be an excess of quantity for preventing PONV with the risk of sedation. Therefore, when using midazolam, which is a hypnotic agent, the sedative effects, respiratory depression, and prolonged recovery time might be a major safety concern. However, no significant differences were apparent between midazolam and control groups after a safety analysis, but the different sedation scoring scales in the studies limited the combination of data. Fifteen of the 16 relevant studies reported that there were no significant differences between midazolam and control groups. Among 3 studies in which midazolam was administered continuously,9,28,31 only 1 study31 reported that the incidence of mild sedation in the midazolam group was higher than levels in the control group. However, the incidence of moderate to severe sedation or respiratory depression was not observed in any group.
A subgroup analysis was performed to evaluate the effects of midazolam administration on PONV when used in conjunction with propofol as an induction agent. The results of the analysis indicated that midazolam had a statistically significant effect on overall PON, POV, and PONV whether propofol was administered or not. However, the requirement for rescue antiemetics showed no evidence when propofol was used in conjunction with midazolam as an induction agent. Conversely, the requirement for rescue antiemetics revealed a significant reduction compared with the control group when only midazolam was used as an induction agent without propofol. We suggest that the use of propofol, which has been reported to decrease the incidence of early PONV,44,45 might have affected the requirement for rescue antiemetics.
Nonetheless, an apparent publication bias was evident in several comparisons of midazolam and controls in the current study and may have influenced the results. Nevertheless, the significance of the study findings was not altered after consideration of the apparent publication bias. Consequently, we had confidence in our findings, despite the omission of data from unpublished studies.
To include maximum data regarding the antiemetic effects of midazolam, several studies contained PONV data as an inevitable secondary end point, which might have increased the likelihood that confounding variables influenced the results. Therefore, we performed a subgroup analysis based on whether PON, POV, and PONV incidence data were collected as primary or secondary outcomes. The combined results of the analyses revealed that midazolam caused a statistically significant reduction in the incidence of overall PON, regardless of the aim of the studies.
The current study has several limitations. First, only published studies were included in our meta-analysis. Second, only 2 studies included the late effectiveness of midazolam on PONV, one of which included the continuous infusion of midazolam using a patient-controlled analgesia device. Third, many studies either did not assess side effects or adverse events or approached them in different ways. Hence, the late effectiveness of midazolam on PONV and the results of safety analyses should be interpreted cautiously. Despite its limitations, the present meta-analysis demonstrated strength through the application of rigorous methodology and provided the first systematic review assessing the prophylactic effect of midazolam on the incidence of PONV. Moreover, added strengths of our study included the large sample size, the thorough identification of articles, the data extraction, and the resolution of all discrepancies by 3 independent investigators. Furthermore, our subgroup and sensitivity analyses yielded stable and robust findings.
In summary, the prophylactic administration of midazolam reduced the incidence of overall PON, POV, and PONV. Furthermore, the effects were more evident during the early postoperative phase than during the late phase, and the reductions in the incidence of PON, POV, and PONV were greater when midazolam was administered at the induction of surgery.
- randomized controlled trial.pt
- randomized controlled trial$.mp
- controlled clinical trial.pt
- controlled clinical trial$.mp
- random allocation.mp
- exp double-blind method/
- 8.exp single-blind method/
- clinical trial.pt
- clinical trial$.mp
- exp clinical trial/
- (clin$ adj25 trial$).mp
- ((singl$ or doubl$ or tripl$ or trebl$) adj25 (blind$ or mask$)).mp
- exp research design/
- research design.mp
- 10 or 19
- Case report.tw.
- Historical article.pt.
- 20 not 25
- exp postoperative nausea and vomiting/
- vomit* OR nausea* OR PONV.mp
- (postoperative adj6 nausea adj6 vomiting).mp
- 26 and 30
- exp midazolam/
- 31 and 34
- randomi?ed controlled trial$.mp.
- ‘controlled clinical trial’/exp
- controlled AND clinical AND trials
- controlled clinical trial$.mp.
- ‘random allocation’/exp
- random allocation.mp.
- #1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9
- ‘clinical trial (topic)’/exp
- clinical AND trial$.mp.
- #11 OR #12 OR #13 OR #14
- #10 OR #15
- ‘case study’/exp
- ‘case report’/exp
- ‘abstract report’/exp
- #17 OR #18 OR #19 OR #20
- #16 NOT #21
- ‘postoperative nausea and vomiting’/exp
- ‘postoperative nausea’/exp
- ‘postoperative vomiting’/exp
- Nausea or vomiting or PONV.mp.
- #23 OR #24 OR #25 OR #26
- #22 AND #27
- #29 OR #30
- #28 AND #31
Name: Eun Jin Ahn, MD.
Contribution: This author helped conduct the study, analyze the data, and write the manuscript.
Attestation: Eun Jin Ahn has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Hyun Kang, MD, PhD, MPH.
Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.
Attestation: Hyun Kang has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Geun Joo Choi, MD.
Contribution: This author helped analyze the data and write the manuscript.
Attestation: Geun Joo Choi has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Chong Wha Baek, MD, PhD.
Contribution: This author helped conduct the study and analyze the data.
Attestation: Chong Wha Baek has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Yong Hun Jung, MD, PhD.
Contribution: This author helped design the study and write the manuscript.
Attestation: Yong Hun Jung has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
Name: Young Choel Woo, MD, PhD.
Contribution: This author helped conduct the study and analyze the data.
Attestation: Young Choel Woo has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.
This manuscript was handled by: Tong J. Gan, MD, MHS, FRCA.
1. Gan TJ. Postoperative nausea and vomiting—can it be eliminated? JAMA. 2002;287:1233–6
2. Tramèr MR, Reynolds DJ, Moore RA, McQuay HJ. Efficacy, dose-response, and safety of ondansetron in prevention of postoperative nausea and vomiting: a quantitative systematic review of randomized placebo-controlled trials. Anesthesiology. 1997;87:1277–89
3. Henzi I, Walder B, Tramèr MR. Dexamethasone for the prevention of postoperative nausea and vomiting: a quantitative systematic review. Anesth Analg. 2000;90:186–94
4. Khalil S, Philbrook L, Rabb M, Wells L, Aves T, Villanueva G, Amhan M, Chuang AZ, Lemak NA. Ondansetron/promethazine combination or promethazine alone reduces nausea and vomiting after middle ear surgery. J Clin Anesth. 1999;11:596–600
5. Henzi I, Walder B, Tramèr MR. Metoclopramide in the prevention of postoperative nausea and vomiting: a quantitative systematic review of randomized, placebo-controlled studies. Br J Anaesth. 1999;83:761–71
6. Honarmand A, Safavi M, Khalili G, Mohammadnejad F. Prophylactic administration of haloperidol plus midazolam reduces postoperative nausea and vomiting better than using each drug alone in patients undergoing middle ear surgery. Saudi J Anaesth. 2012;6:145–51
7. Yoon SH. Concerns of the anesthesiologist: anesthetic induction in severe sepsis or septic shock patients. Korean J Anesthesiol. 2012;63:3–10
8. Godsiff L, Magee L, Park GR. Propofol versus propofol with midazolam for laryngeal mask insertion. Eur J Anaesthesiol Suppl. 1995;12:35–40
9. Gilliland HE, Prasad BK, Mirakhur RK, Fee JP. An investigation of the potential morphine sparing effect of midazolam. Anaesthesia. 1996;51:808–11
10. Heidari SM, Saryazdi H, Saghaei M. Effect of intravenous midazolam premedication on postoperative nausea and vomiting after cholecystectomy. Acta Anaesthesiol Taiwan. 2004;42:77–80
11. Hvarfner A, Hammas B, Thörn SE, Wattwil M. The influence of propofol on vomiting induced by apomorphine. Anesth Analg. 1995;80:967–9
12. Jung JS, Park JS, Kim SO, Lim DG, Park SS, Kwak KH, Cho JD, Jeon YH. Prophylactic antiemetic effect of midazolam after middle ear surgery. Otolaryngol Head Neck Surg. 2007;137:753–6
13. Kranke P, Eberhart LH. Possibilities and limitations in the pharmacological management of postoperative nausea and vomiting. Eur J Anaesthesiol. 2011;28:758–65
14. Heidari SM, Rahimi M, Saeb F, Shafa A. Comparison of the effect of intravenous midazolam, metoclopramide and their combination on the treatment of postoperative nausea and vomiting: a double blinded randomized clinical trial. Pak J Med Sci. 2011;27:847–50
15. Lee Y, Wang JJ, Yang YL, Chen A, Lai HY. Midazolam vs ondansetron for preventing postoperative nausea and vomiting: a randomised controlled trial. Anaesthesia. 2007;62:18–22
16. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JP, Clarke M, Devereaux PJ, Kleijnen J, Moher D. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. J Clin Epidemiol. 2009;62:e1–34
17. Wu SJ, Xiong XZ, Lin YX, Cheng NS. Comparison of the efficacy of ondansetron and granisetron to prevent postoperative nausea and vomiting after laparoscopic cholecystectomy: a systematic review and meta-analysis. Surg Laparosc Endosc Percutan Tech. 2013;23:79–87
18. Naing C, Aung K, Mak JW. Reporting ‘number needed to treat’ in meta-analyses: a cross-sectional study. J Evid Based Med. 2012;5:232–7
19. Clyburn P, Kay NH, McKenzie PJ. Effects of diazepam and midazolam on recovery from anaesthesia in outpatients. Br J Anaesth. 1986;58:872–5
20. Crawford ME, Carl P, Bach V, Ravlo O, Mikkelsen BO, Werner M. A randomized comparison between midazolam and thiopental for elective cesarean section anesthesia. I. Mothers. Anesth Analg. 1989;68:229–33
21. El-Deeb A, Ali Y, Rashdy H. Evaluation of combination antiemetic prophylaxis in high risk emetogenic patients undergoing thyroid surgery: a randomized double-blind study. Egypt J Anaesth. 2011;27:203–6
22. Fragen RJ, Caldwell NJ. Awakening characteristics following anesthesia induction with midazolam for short surgical procedures. Arzneimittelforschung. 1981;31:2261–3
23. Puri GD, Palvannan S, Chari P, Singh H, Gujral JS. Midazolam as an induction agent in mitral stenosis patients for closed mitral commissurotomy. Indian J Med Res. 1993;98:174–7
24. Quario Rondo L, Thompson C. Efficacy of propofol compared to midazolam as an intravenous premedication agent. Minerva Anestesiol. 2008;74:173–9
25. Sanjay OP, Tauro DI. Midazolam: an effective antiemetic after cardiac surgery—a clinical trial. Anesth Analg. 2004;99:339–43
26. Di Florio T, Goucke CR. The effect of midazolam on persistent postoperative nausea and vomiting. Anaesth Intensive Care. 1999;27:38–40
27. Lohakare R, Sangawar AV, Ghosh AA. Influence of intravenous fentanyl and / or midazolam on induction of anaesthesia with thiopentone. J Anaesthesiol Clin Pharmacol. 2004;20:273–8
28. Kim DS, Koo GH, Kang H, Baek CW, Jung YH, Woo YC, Kim JY, Park SG. The antiemetic effect of midazolam or/and ondansetron added to intravenous patient controlled analgesia in patients of pelviscopic surgery. Korean J Anesthesiol. 2012;62:343–9
29. Bauer KP, Dom PM, Ramirez AM, O’Flaherty JE. Preoperative intravenous midazolam: benefits beyond anxiolysis. J Clin Anesth. 2004;16:177–83
30. Hasani A, Maloku H, Sallahu F, Gashi V, Ozgen SU. Preemptive analgesia with midazolam and diclofenac for hernia repair pain. Hernia. 2011;15:267–72
31. Huh BK, Jung S, White W, Jeon Y. Anti-emetic effect of midazolam added to morphine patient-controlled analgesia after total abdominal hysterectomy. Anaesth Intensive Care. 2010;38:481–5
32. Kim WJ, Kang H, Shin HY, Baek CW, Jung YH, Woo YC, Kim JY, Koo GH. Ramosetron, midazolam, and combination of ramosetron and midazolam for prevention of postoperative nausea and vomiting: a prospective, randomized, double-blind study. J Int Med Res. 2013;41:1203–13
33. Makhdoom NK, Farid MF. Prophylactic antiemetic effects of midazolam, dexamethasone, and its combination after middle ear surgery. Saudi Med J. 2009;30:504–8
34. Park EY, Lee SK, Kang MH, Lim KJ, Kim YS, Choi E, Park YH. Comparison of ramosetron with combined ramosetron and midazolam for preventing postoperative nausea and vomiting in patients at high risk following laparoscopic gynaecological surgery. J Int Med Res. 2013;41:654–63
35. Richardson MG, Wu CL, Hussain A. Midazolam premedication increases sedation but does not prolong discharge times after brief outpatient general anesthesia for laparoscopic tubal sterilization. Anesth Analg. 1997;85:301–5
36. Safavi MR, Honarmand A. Low dose intravenous midazolam for prevention of PONV, in lower abdominal surgery—preoperative vs intraoperative administration. Middle East J Anaesthesiol. 2009;20:75–81
37. Shirdashtzadeh N, Eshraghi N, Eshraghi A. Comparison of parenteral promethazine versus midazolam effect as a preoperative medication on postoperative nausea and vomiting after appendectomy. Caspian J Intern Med. 2011;2:270–3
38. Yeo J, Jung J, Ryu T, Jeon YH, Kim S, Baek W. Antiemetic efficacy of dexamethasone combined with midazolam after middle ear surgery. Otolaryngol Head Neck Surg. 2009;141:684–8
39. Haavik PE, Søreide E, Hofstad B, Steen PA. Does preoperative anxiety influence gastric fluid volume and acidity? Anesth Analg. 1992;75:91–4
40. Lydon A, McGinley J, Cooke T, Duggan PF, Shorten GD. Effect of anxiety on the rate of gastric emptying of liquids. Br J Anaesth. 1998;81:522–5
41. Wang SM, Kain ZN. Preoperative anxiety and postoperative nausea and vomiting in children: is there an association? Anesth Analg. 2000;90:571–5
42. Gan TJ, Diemunsch P, Habib AS, Kovac A, Kranke P, Meyer TA, Watcha M, Chung F, Angus S, Apfel CC, Bergese SD, Candiotti KA, Chan MT, Davis PJ, Hooper VD, Lagoo-Deenadayalan S, Myles P, Nezat G, Philip BK, Tramèr MRSociety for Ambulatory Anesthesia. . Consensus guidelines for the management of postoperative nausea and vomiting. Anesth Analg. 2014;118:85–113
43. Splinter WM, MacNeill HB, Menard EA, Rhine EJ, Roberts DJ, Gould MH. Midazolam reduces vomiting after tonsillectomy in children. Can J Anaesth. 1995;42:201–3
44. Tramèr M, Moore A, McQuay H. Propofol anaesthesia and postoperative nausea and vomiting: quantitative systematic review of randomized controlled studies. Br J Anaesth. 1997;78:247–55
© 2016 International Anesthesia Research Society
45. Visser K, Hassink EA, Bonsel GJ, Moen J, Kalkman CJ. Randomized controlled trial of total intravenous anesthesia with propofol versus inhalation anesthesia with isoflurane-nitrous oxide: postoperative nausea with vomiting and economic analysis. Anesthesiology. 2001;95:616–26