A comparative evaluation of dexmedetomidine and midazolam in pediatric sedation: A meta‐analysis of randomized controlled trials with trial sequential analysis

Abstract Background The present study with trial sequential analysis (TSA) was conducted to evaluate comprehensively the efficacy and safety of dexmedetomidine and midazolam in pediatric sedation, and to investigate whether the outcomes achieved the required information size to draw the conclusions. Methods PubMed, Embase, and Cochrane Library were searched from inception to October 2019. All randomized controlled trials used dexmedetomidine and midazolam in pediatric sedation were enrolled. Sedative efficacy, postoperative analgesic effect, and incidence of emergence agitation were considered as the co‐primary outcomes. Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system was applied to rate the quality of evidences. Results We acquired data from 34 studies involving 2281 pediatric patients. The results indicated that administration of dexmedetomidine was associated with less incidence of emergence agitation (RR = 0.78, with 95% CI [0.65, 0.92]) and more satisfactory sedation at parental separation (RR = 0.31, with 95% CI [0.24, 0.41]) compared to midazolam, and the current sample sizes were sufficient with unnecessary further trials. Two groups did not differ significantly in sedation level at mask induction (RR = 0.86, with 95% CI [0.74, 1.00]). And using of dexmedetomidine was associated with less incidence of postoperative analgesic rescue (RR = 0.57, with 95% CI [0.35, 0.93]), but the number of patients was too few to achieve the required information size and to draw reliable conclusions. Premedication of dexmedetomidine was associated with significant less value of SBP, heart rate, increased incidence of bradycardia, and a lower rate of shivering. And there were no differences about onset of sedation and recovery time between two groups. Conclusions Given that more satisfactory sedation at separation from parents and less incidence of emergence agitation, dexmedetomidine is preferred for pediatric sedation. However, compared with midazolam, the superiority of dexmedetomidine in providing adequate sedation at mask induction and postoperative analgesic effects has not yet been defined.


| INTRODUC TI ON
Anxiety and distress developed in pediatric patients during perioperative period bring the challenges for anesthesiologist and pediatric clinicians. Uncooperative physically resistance from children results in increased difficulties in separation from parents, mask application, and induction of anesthesia. 1,2 It is estimated that up to 60%-70% of children suffered anxiety, anguish, and fear throughout the perioperative period or diagnostic procedures. 3 Sedative premedications can help to reduce the anxiety, minimize the emotional discomforts, ease the parental separation, and smooth the induction of anesthesia. 4 Many premedicants via different routes have been tried in clinical practice.
Compared with other benzodiazepine, midazolam has rapid onset and high metabolic clearance, and its sedative efficacy in pediatric premedication has been demonstrated widely. However, untoward effects including negative postoperative behavioral changes, cognitive impairment, respiratory depression, and insufficient prevention of postoperative emergence agitation have been reported in children premedicated with midazolam, 5,6 which makes it a less-thanideal option in pediatric sedation.
As one highly selective α2-adrenoceptor agonist (selectivity ratio for α2-adrenoceptor:α1-adrenoceptor is 1600:1) with sedative and analgesic characteristics, dexmedetomidine provides cooperative and arousable sedation without clouded consciousness and respiratory depression. Owing to these beneficial effects, it has been demonstrated to be a useful pre-anesthesia medication in children. 7 In an effort to evaluate the influences of the two premedications on pediatric perioperative sedation, Pasin et al 8  were approximate, and authors also described that available data were still lacking. More evidences with large sample size were required to draw the reliable conclusions.
Therefore, on the basis of combining the latest evidences, the present updated meta-analysis of RCTs was conducted to evaluate comprehensively the effects of two premedicants in pediatric sedation at separation from parents and mask induction, hemodynamic status, and various adverse effects. And the trial sequential analysis (TSA) was also performed to determine whether the findings achieved the required information size to draw the conclusions.

| MATERIAL S AND ME THODS
The present meta-analysis was performed in accordance with the recommendations in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement 10 and the guidelines described in the Cochrane Handbook.

| Search strategy
Two independent reviewers (BL and YF) performed the literature search. The databases including PubMed, Embase, and Cochrane Library were searched systematically. The strategies used for searching were including infant, child, adolescent, dexmedetomidine, midazolam, and randomized controlled trial (Appendix S1). Only human studies were involved, and there were no restrictions of language.
The final literature search was performed on October 7, 2019.

| Eligibility criteria
The studies meeting the following criteria were selected for further analysis:

| Participants
The patients were the children (<18 years old) who experienced different surgical and diagnostic procedures.

| Intervention and comparison
Using dexmedetomidine vs midazolam as the premedication (regardless of the route and dose of administration).

| Outcome measures
Given that satisfactory separation from parents and satisfactory induction or facemask compliance with limited postoperative pain and agitation were considered as the ideal characteristics of pediatric sedatives, 9 the co-primary outcomes were as follows: (a) number of patients with satisfactory separation from parents, (b) number of patients with satisfactory induction or mask acceptance, (c) number in providing adequate sedation at mask induction and postoperative analgesic effects has not yet been defined.  ters, onset of sedation, and recovery time between two groups were considered as the secondary outcomes. The incidence of adverse events, including shivering, bradycardia, nausea, and vomiting, were also analyzed.

| Study design
Randomized controlled trials with no language limitations.

| Grading the quality of evidence
The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) methodology was used to assess the quality of evidence and strength of recommendations. The quality of all primary and secondary outcomes was independently assessed by two reviewers (BL and YF). On the basis of risk of bias, inconsistency, indirectness, imprecision, and publication bias, the quality was classified as high, moderate, low, or very low. 12 And GRADE profiler (version 3.6) software was used.

| Statistical analysis
Statistical analyses were done with Review Manager 5.0 software (The Cochrane Collaboration). The risk ratio (RR) with 95% confidence interval (CI) and the Mantel-Haenszel method (fixed or random models) were used to analyze dichotomous data. For continuous data, standardized mean difference (SMD) was chosen for the estimation. The I-squared (I 2 ) test was chosen to weigh the impact of heterogeneity on the results. If significant heterogeneity (present at I 2 > 50%) emerged, the sensitivity analysis was performed by omitting each study individually, and the random effects model was chosen; otherwise, the fixed-effects model was applied. Publication bias was evaluated by using Begg's test when approximate ten studies or more were included in meta-analysis. A P value <.05 was considered statistically significant.

| Trial sequential analysis
The

| Literature search results
A total of 440 relevant items were identified initially. One hundred eighty-four of them were excluded by duplicate removal, and 165 were excluded by reviewing the title and abstract. In these 165 excluded items, 69 were the protocols or registered trials (still recruiting or not), two were animal researches, 42 were studies performed in adult patients, 33 were unrelated reviews or meeting abstracts, 16 were studies with irrelevant topics, and three were similar systematic reviews published in 2014 and in 2015. A total of 57 items were excluded by full-text reviewing, 54 of them were owing to the inappropriate comparisons, two of them reported the uncorrelated outcomes or the outcomes with inappropriate format, and the full text of the rest one cannot be gained after contacting the authors.
Finally, 34 studies were selected in the consequent analysis.  The identification procedure of eligible items is described in Figure S1.

| Basic characteristics of enrolled studies
These included studies were published from 2004 to 2019 (33 in English and one in Chinese) and were enrolled a total of 2281 pediatric patients (ages ranged from 6 months to 18 years). The primary outcomes "the number of patients with satisfactory separation from parents" and "the number of patients with satisfactory induction or mask acceptance" were reported separately in 18 studies and in 20 studies. And the primary adverse events "the incidence of postoperative pain needed analgesics rescue" and "the incidence of emergence agitation" were mentioned in eight studies and in 14 studies.
The secondary outcomes including general hemodynamic parameters (systolic blood pressure, diastolic blood pressure, mean arterial pressure, and heart rate), onset of sedation, recovery time, and the incidences of various adverse events (shivering, bradycardia, nausea, and vomiting) were also reported in different studies. The main characteristics of these enrolled studies were summarized in Table 1.

| Risk of bias assessment
In accordance with the Cochrane Collaboration tool for assessing risk of bias, we evaluated the mentioned-above items. A total of 65% (22/34) studies performed an adequate method of random sequence generation, and 12 studies reported allocation concealment with detailed descriptions (using opaque, sealed envelopes). Twenty-five studies described the blinding procedure of participants and personnel, and 25 studies mentioned the blinding procedure of outcome assessment.
A total of eight studies were high-quality studies with low risk of bias in all items. The detail of risk of bias assessment was shown in Figure S2.  Figure 1A). Although cumulative Z curves did not reach the required information size, the results of TSA indicated that the curves crossed both the conventional boundary and the trial sequential monitoring boundary. The level of evidence about the intervention effect was sufficient with unnecessary further trials ( Figure 3A). Publication bias was detected in analysis by using of Begg's test (P = .006; Figure 4A). Therefore, in order to estimate and adjust for the number and outcomes of missing studies, we performed Duval's trim and fill method. 48 And the results from sensitivity analyses of trim and fill method (no new studies added) revealed that the result was reliable.

| Primary outcome 2: the number of patients with satisfactory induction or mask acceptance
A total of twenty studies 17,18,21,23,[25][26][27][28][29][30]33,34,[39][40][41][42][43][45][46][47] with 1398 patients were analyzed. The I 2 of 76% demonstrated that significant heterogeneity was existed. However, in sensitivity analysis, all attempts to reduce the value of I 2 to below 50% by excluding one single study were not successful. Therefore, random effects model was used. The using of dexmedetomidine was associated with higher rate of satisfactory induction or satisfactory mask acceptance compared to midazolam, but no significant differences were observed between two groups (71.11% vs 61.88%, RR = 0.86, with 95% CI [0.74, 1.00], P = .06, I 2 = 76%; Figure 1B). The TSA indicated that cumulative Z curves did not cross any of the boundaries, and the current number of patients was too few to achieve the required information size (2112 patients). The further evidences with large sample size are required ( Figure 3B). Begg's (P = .381) test suggested that publication bias was not found ( Figure 4B).

| Primary outcome 3: the number of patients requiring postoperative analgesics rescue
It was reported in eight studies with 640 patients. 16,20,23,26,28,35,40,44 Patients who received dexmedetomidine experienced significantly lower incidence of postoperative analgesics rescue than patients who severe pain compared to midazolam. The further trials to achieve the firm evidences are necessary ( Figure 3C). Begg's test (P = .711) indicated that publication bias was not found in the analysis ( Figure 4C).

| Primary outcome 4: the incidence of emergence agitation
The emergence agitation was mentioned in 14 studies with 969 pati ents. 19 Figure 3D). After Begg's test (P = .827), no publication bias was found in the analysis ( Figure 4D).

| Secondary outcomes
All secondary outcomes involving hemodynamic parameters, onset of sedation, recovery time, and incidence of different adverse effects were clarified in Table 2. The details about general hemodynamic parameters including systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP), and heart rate (HR) were reported separately in six studies, 14,20,25,31,33,34 three studies, 14,31,33 two studies, 15,16 and eight studies. [14][15][16]20,25,31,33,34 The results indicated that the using of dexmedetomidine was associated with significant less value of SBP (SMD = 0.99, with 95% CI , which was inconsistent with the previous outcome from the similar systematic review. 9 It might be resulted from the different sample sizes. And we also found two groups did not differ significantly in recovery time.
The analysis of various adverse events exhibited that the patients in dexmedetomidine group suffered increased incidence of bradycardia, and experienced lower rate of shivering. The incidence of postoperative nauseas or vomiting was not different between two groups. However, the reports about the incidence of the adverse events were relatively scarce.

| Quality of the evidence
We used GRADE approach to grading the level of each outcome in present study. Although the results from risk of bias assessment part indicated that the quality of trials design was reasonable, the GRADE summary of findings table demonstrated that the overall level of current evidence in our meta-analysis was moderate or low, which might be resulted from the inconsistency issue and, particularly, the limited number of events (Table S1).

| D ISCUSS I ON
Pediatric sedation is always served as one of conundrums during diagnostic and surgical procedures, such area changes rapid and engenders several debates among the anesthesiologists and pediatric specialists. 49 For instance, the optimal premedication between dex- Our study updated the previous conclusion about the superiority of dexmedetomidine in onset of sedation. 9 The combination of studies with expanded sample size demonstrated that the difference between dexmedetomidine and midazolam in onset time was not significant. Moreover, the summary of new evidences also indicated that no difference was found in recovery time between two groups.
Noteworthily, in evaluation of hemodynamic parameters and adverse events between two groups, the using of dexmedetomidine exhibited great reduction in systolic blood pressure and heart rate, and was associated with high incidence of bradycardia.
These phenomena might be derived from the biphasic effects of F I G U R E 4 Funnel plots of effect estimates for the co-primary outcomes. A, the number of patients with satisfactory separation from parents; B, the number of patients with satisfactory induction or mask acceptance; C, the number of patients requiring postoperative analgesics rescue; D, the incidence of emergence agitation. RR, risk ratio; SE, standard error α2-adrenoceptor. It enhanced the blood pressure temporarily as the transient vasoconstrictive effects in peripheral vasculature and then lowered the arterial pressure with decreasing sympathetic outflow. 51 Even though, some researchers still regarded the dexmedetomidine as one appropriate sedative option for pediatric patients, in consideration of great hemodynamic changes could be resolved by decelerating the rate of administration. 52,53 The secondary outcomes about different adverse events suggested that incidence of shivering was lower in patients received dexmedetomidine compared to midazolam. And no difference was found in occurrence rate of postoperative nausea and vomiting between two groups. However, owing to extremely limited sample size, the above data were not enough to draw a definitive and reliable conclusion. This was one of the limitations in present study. Hence, the focus in future should be moved on the evaluation of safety in using dexmedetomidine and midazolam as premedication in children.
Furthermore, the widespread moderate or low quality in outcomes evaluated by GRADE approach resulted from inconsistency (high heterogeneity) and imprecision (lack of events number).
Heterogeneity might be originated from different types of procedures, administration routes, and premedication doses. The sensitivity analysis performed by us discovered one trial which brought the significant heterogeneity in evaluation of patients requiring postoperative analgesics rescue, and then, we verified the reliability of conclusion by omitting it. And the other significant heterogeneity among studies led us to use random effects models for meta-analysis.

| CON CLUS ION
In conclusion, the current evidences suggest that dexmedetomidine is the preferred choice for pediatric patients than midazolam owing to its more satisfactory sedation at separation from parents and less incidence of emergence agitation. However, the superiority of using dexmedetomidine as premedication in providing adequate sedation at mask induction and postoperative analgesic effects compared to midazolam has not yet been defined. Additionally, to obtain firm evidences about the effects of dexmedetomidine vs midazolam on hemodynamic parameters and the safety of two premedicants, more high-quality trials are required. And the exploration of the optimal dose range and ideal route of using dexmedetomidine should also be considered in future.

ACK N OWLED G M ENTS
The authors thank the support from Group of People with Highest Risk of Drug Exposure of International Network for the Rational Use of Drugs, China.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest. .002* *Significant difference between groups (P < .05).