The Effect of Dexmedetomidine on Emergence Agitation or Delirium in Children After Anesthesia—A Systematic Review and Meta-Analysis of Clinical Studies

Background: We conducted this systematic review and meta-analysis to investigate the clinical effect of dexmedetomidine in preventing pediatric emergence agitation (EA) or delirium (ED) following anesthesia compared with placebo or other sedatives. Methods: The databases of Pubmed, Embase, and Cochrane Library were searched until 8th January 2020. Inclusion criteria were participants with age<18 years and studies of comparison between dexmedetomidine and placebo or other sedatives. Exclusion criteria included adult studies; duplicate publications; management with dexmedetomidine alone; review or meta-analysis; basic research; article published as abstract, letter, case report, editorial, note, method, or protocol; and article presented in non-English language. Results: Fifty-eight randomized controlled trials (RCTs) and five case-control trials (CCTs) including 7,714 patients were included. The results showed that dexmedetomidine significantly decreased the incidence of post-anesthesia EA or ED compared with placebo [OR = 0.22, 95% CI: (0.16, 0.32), I2 = 75, P < 0.00001], midazolam [OR = 0.36, 95% CI: (0.21, 0.63), I2 = 57, P = 0.0003], and opioids [OR = 0.55, 95% CI: (0.33, 0.91), I2 = 0, P = 0.02], whereas the significant difference was not exhibited compared with propofol (or pentobarbital) [OR = 0.56, 95% CI: (0.15, 2.14), I2 = 58, P = 0.39], ketamine [OR = 0.43, 95% CI: (0.19, 1.00), I2 = 0, P = 0.05], clonidine [OR = 0.54, 95% CI: (0.20, 1.45), P = 0.22], chloral hydrate [OR = 0.98, 95% CI: (0.26, 3.78), P = 0.98], melatonin [OR = 1.0, 95% CI: (0.13, 7.72), P = 1.00], and ketofol [OR = 0.55, 95% CI: (0.16, 1.93), P = 0.35]. Conclusion: Compared with placebo, midazolam, and opioids, dexmedetomidine significantly decreased the incidence of post-anesthesia EA or ED in pediatric patients. However, dexmedetomidine did not exhibit this superiority compared with propofol and ketamine. With regard to clonidine, chloral hydrate, melatonin, and ketofol, the results needed to be further tested due to the fact that only one trial was included for each control drug.


INTRODUCTION
Emergence agitation (EA) or delirium (ED) manifests as a series of sudden complex psychomotor disorders, characterized by perceptual disturbances, delusions, and disorientation following sedation or general anesthesia (1). So far, the specific mechanism of EA or ED has not been clear. The preschool children undergoing ophthalmology or otorhinolaryngology procedures under inhalation agents are susceptible population (2). According to some studies, the incidence of EA or ED after general anesthesia in children ranges from 10 to 80% (3) and significantly increases the occurrence of other complications after anesthesia, like self-injury, prolonged post-anesthesia care unit (PACU) stay, poor satisfaction of parents and care providers, and so on (4). Therefore, it is necessary to find effective measures to prevent or treat EA or ED.
Some studies have reported the pharmacological strategies to prevent EA or ED, including midazolam, propofol, ketamine, opioids, and α 2 adrenergic receptor agonists (5)(6)(7)(8). Activation of an α 2 adrenergic receptor can contribute to pharmacological effects of sedation, analgesia, and anti-inflammation; thus, an α 2 adrenergic receptor may be a target for prevention and treatment of EA or ED (9,10). A study from Ydemann et al. (11) found that clonidine significantly decreased the incidence of postoperative agitation in children after sevoflurane anesthesia compared with placebo. Another commonly used α 2 adrenergic receptor agonist dexmedetomidine shows a higher ratio of specificity for α 2 receptor (α 2 :α 1 1600:1) compared with clonidine (α 2 :α 1 200:1) (12,13). Although dexmedetomidine is used as an offlabel drug in children, increasing studies about the effect of dexmedetomidine on EA and ED in pediatric patients have been completed. We conducted this meta-analysis for clinical trials to evaluate the effect of dexmedetomidine on EA or ED following sedation or general anesthesia in pediatric patients compared with placebo and other drugs.

MATERIALS AND METHODS
This systematic review and meta-analysis was performed according to the guidelines of the 2009 PRISMA (Preferred Reporting Items for Systematic reviews and Meta-analyses) (Supplementary Table 1) (14).

Search Strategy and Study Selection
We searched the databases including "Pubmed, " "Embase, " and "Cochrane Library" through the PICOS (Population, Intervention, Comparison, Outcome, Study design) method until 8th January 2020. The entry words included "child" OR "children" OR "pediatric" AND "dexmedetomidine" OR "precedex" OR "MPV-1440" OR "MPV 1440" OR "Dexmedetomidine Hydrochloride" OR "Hydrochloride, Dexmedetomidine" AND "agitation" OR "delirium, " and the search scope was "all fields." Because all studies about the effect of dexmedetomidine vs. other drugs (placebo or other sedatives) on agitation or delirium in pediatric patients were eligible in this meta-analysis, we did not confine the search words of control drugs and study design. The inclusion criteria included the following: (1) participants with age<18 years; and (2) management with prophylactic dexmedetomidine and placebo or other sedatives. The exclusion criteria included the following: (1) participants with age≥18 years; (2) management with dexmedetomidine alone; (3) review or meta-analysis; (4) basic research; (5) article published as an abstract, letter, case report, editorial, note, method, or protocol; and (6) article presented in non-English language.

Data Analysis
The aim of this meta-analysis was to investigate whether dexmedetomidine had advantage in reducing the incidence of EA or ED following sedation or general anesthesia in pediatric patients compared with placebo or other sedatives.
Three authors were independently responsible for reviewing the titles, abstracts, or both and summarized the data of the included literatures. Another two authors were in charge of the data discrepancy adjustment.
Two authors were responsible for extracting the following information: (1) authors; (2) publication year; (3) number of the total participants in each study; (4) age range of all the participants; (5) country of publication; (6) procedures that the participants underwent; (7) time of dexmedetomidine or other sedative administration; (8) infusion speed or dosage of dexmedetomidine or other sedatives; and (9) number of patients with EA or ED following sedation or general anesthesia.
Two authors independently assessed the quality of included studies. The risk of bias of randomized controlled trials (RCTs) were assessed by the Cochrane Collaboration Risk of Bias Assessment tool including seven items: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and others (bias due to vested financial interest and academic bias). If a trial had one or more of the items to be judged as high or unclear risk of bias, this trial was classified as having high risk (15). The bias risk of case-control trials (CCTs) was assessed by the Newcastle-Otawa Quality Assessment Scale (NOS) comprising three domains: selection, comparability, and outcome for cohort studies. There were four stars in the selection domain, two stars in the comparability domain, and three stars in the exposure domain. Trials with cumulative seven stars or more were considered to be of high quality, those with six stars were considered to be of moderate quality, and those with less than six stars were considered to be of low quality (Supplementary Table 2) (16). If the two authors had different assessment results, they consulted the third or the fourth one. Eventually, the authors reached consensus. All included trials were grouped based on different control drugs.
RevMan Review Manager version 5.3 (Cochrane collaboration, Oxford, UK) and Stata version 12.0 (Stata Corp, College Station, TX, USA) were used to perform statistical analyses. The values of I 2 and the Mantel-Haenszel chisquare test (P-value for heterogeneity) were used to evaluate the heterogeneity of included studies. And the values of I 2 <40%, 40-60%, and >60% represented low, moderate, and high heterogeneity, respectively (17). If I 2 >50% or a P-value for heterogeneity <0.1 was identified, the method of random-effect model analysis was applied to analyze the data. Conversely, if I 2 < 50% or a P-value for heterogeneity ≥0.1 was presented, the method of a fixed-effect model was used (18). The dichotomous outcome was reported as odds ratios (OR) with 95% confidence interval (CI). The statistical tests were two-sided, and a P-value for overall effect < 0.05 was considered to have significant difference.
Sensitivity analysis was conducted to solve the problem of significant heterogeneity (I 2 > 40%) through the method of subgroup analysis or one-by-one literature removal. Metaregression was used to investigate the heterogeneity sources for the group with I 2 > 40% according to possible risk factors. A subgroup analysis proceeded based on the risk factor with P < 0.05 by meta-regression analysis; conversely, the method of oneby-one literature removal was used if P-values of all risk factors were 0.05 or more.
The random-effect model with OR was selected due to high I 2 in the groups of placebo (I 2 = 75%), midazolam (I 2 = 57%), and propofol (or pentobarbital) (I 2 = 58%), whereas the fixed-effect model with OR was selected because of low I 2 in the group of opioids (I 2 = 0%) and ketamine (I 2 = 0%).
With regard to other control sedatives or drug combination, no heterogenicity was presented because only one literature was retrieved for each group. The results did not demonstrate significant difference in the incidence of EA or ED after anesthesia when comparing dexmedetomidine with clonidine

Sensitivity Analysis
Meta-regression was performed to investigate the heterogeneity sources by assessing the potential factors including the year of publication, study methods, the country of authors, the time of drug administration, the type of surgery, routes of drug administration, the bias risk of the study, and the range of patients' age for the groups of placebo and midazolam. Unexpectedly, all P-values of these risk factors were over 0.05 (Supplementary Tables 4, 5). Afterward, the method of one-byone literature removal was used. Seven trials (20,23,39,44,46,49,54) were found to be the main sources of heterogeneity in the placebo group (I 2 dropped from 75 to 36%) and two trials (61,69) in the midazolam group (I 2 dropped from 57 to 28%). Due to a small number of included trials in the group of propofol (or pentobarbital), the method of one-by-one literature removal was directly used to lower the heterogeneity. When we removed the retrospective trial from Mason et al. (76), the value of I 2 in the propofol (or pentobarbital) group dropped from 58 to 13%, and

DISCUSSION
This meta-analysis included 58 RCTs and 5 CCTs that compared the prophylactic effect of dexmedetomidine vs. placebo or other sedatives on post-anesthesia EA or ED in pediatric patients undergoing medical procedures. The results showed that dexmedetomidine strikingly decreased the incidence of post-anesthesia EA or ED compared with placebo, midazolam, or opioids, whereas the significant difference was not exhibited compared with propofol (or    pentobarbital), ketamine, clonidine, chloral hydrate, melatonin, and ketofol, respectively.
Currently, the specific predisposing causes of EA or ED following medical procedures in children remain unclear. Children undergoing general anesthesia are prone to suffer post-anesthesia EA or ED due to their immature central nervous system, preoperative fear and anxiety about unfamiliar surroundings, and postoperative pain (84)(85)(86). In addition, the children undergoing inhalation anesthesia through sevoflurane, isoflurane, or desflurane may suffer from a high incidence of post-anesthesia EA or ED (87,88). Various medications have been used to prevent EA or ED in pediatric patients, like benzodiazepines, opioids, propofol, ketamine, clonidine, dexmedetomidine, and so on (11,(89)(90)(91)(92)(93).
Dexmedetomidine, as a highly selective α 2 adrenergic receptor agonist, can produce pharmacological effects of antianxiety, sedation, and analgesia without overt respiratory and circulatory inhibition in a routine dose (94,95). Meanwhile, dexmedetomidine can improve the cognitive function in children during recovery from general anesthesia (96) and contributes to dose-dependent inhibition of EA or ED after medical procedures (97). The optimal dose (ED 95 ) of dexmedetomidine for preventing EA was 0.30 µg/kg (95% CI: 0.21-1.00 µg/kg) (83). An animal experiment demonstrated that dexmedetomidine could enhance spatial learning and memory in neonatal rats under physiological conditions through promoting hippocampal neurogenesis (98). In this meta-analysis, nine trials had different dexmedetomidine groups according to different dosages (20,33,34,40,43,45,53,56) or administration routes of this drug (21). Patients in the control groups of these nine trials were treated with a placebo (20,21,33,34,40,43,45,53,56), and patients in another control group in the study from Pestieau et al. received fentanyl treatment (45). We chose the dexmedetomidine group with higher incidence of EA or ED. Therefore, the pooled results were more convincing in the powerful prophylactic effect of dexmedetomidine on the occurrence of EA or ED in children compared with placebo and opioids.
Dexmedetomidine can be administered in a variety of ways, like intravenous, transnasal, oral, inhalation, caudal or nerve block, and so on; thus, pediatric patients can easily accept it. The pooled results of 53 trials comparing dexmedetomidine with placebo and midazolam showed that dexmedetomidine could work in various ways and was superior to placebo or midazolam in inhibiting EA or ED in children. However, compared with propofol (or pentobarbital) or ketamine, dexmedetomidine did not demonstrate its superiority in reducing pediatric EA or ED following anesthesia. The possible explanations included the following: (1) the efficacy of propofol (or pentobarbital) or ketamine in suppressing EA or ED occurrence was no less than that of dexmedetomidine; and (2) the number of relevant prospective studies needed to be further increased. Because only one article was included, we could not perform meta-analysis for trials in the group of clonidine, chloral hydrate, melatonin, or ketofol.
In this meta-analysis, high heterogenicity was detected in trials comparing dexmedetomidine with placebo (I 2 = 75%), midazolam (I 2 = 57%), and propofol (or pentobarbital) (I 2 = 58%), respectively. Subgroup analysis is an effective method to solve large heterogenicity among studies (99). We suggested some possible risk factors associated with overt heterogenicity including the year of publication, study methods, the country of authors, the time of drug administration, the type of surgery, routes of drug administration, the bias risk of the study, and the range of patients' age. Meta-regression was used to identify heterogenicity sources. If the P-value of meta-regression was <0.05 through analyzing one risk factor, the subgroup analysis was performed based on this risk factor (99,100). However, in this meta-analysis, all P-values of meta-regression were more than 0.05 through analyzing all possible risk factors in the placebo and midazolam groups. Hence, we considered that significant heterogeneity may be the result of a combination of multiple factors. The meta-analysis by a random-effect model can decrease the effect of significant heterogeneity on the results, although this method does not solve heterogeneity (101). In addition, the method of trial exclusion is also an effective method to solve large heterogenicity for meta-analysis (102). When we excluded seven trials (20,23,39,44,46,49,54) in the placebo group, two trials (61,69) in the midazolam group, and one trial (76) in the propofol (or pentobarbital) group, all values of I 2 dropped to below 40%. Interestingly, the pooled results were consistent with those prior to sensitivity analysis.
It is necessary to elaborate the strengths and limitations of our meta-analysis. Firstly, this meta-analysis presented a comprehensive and up-to-date analysis of dexmedetomidine vs. placebo or other sedatives in pediatric patients. Sixty-three included trials with unlimited study methods (RCTs and CCTs) and various administration routes and dosages were grouped according to control drugs; thus, the pooled outcomes revealed the effect of dexmedetomidine on pediatric EA or ED more comprehensively. Secondly, sensitivity analysis was conducted in groups with high heterogeneity to remove the influence of heterogeneity on the overall results. Thirdly, this meta-analysis provided several directions for future clinical studies about the effect of dexmedetomidine on EA or ED in children. In addition, some limitations should be taken into account in this metaanalysis. Foremost, 39 RCTs and 2 CCTs in 63 included trials were assessed to be high bias risk, and so many trials with high-risk bias would affect the results. Additionally, the age gap of participants in 9 trials (46,52,54,58,64,71,72,75,79) was over 10 years, and a large age gap might be an important risk factor associated with the unreliability of outcomes. Lastly, non-uniform definitions of EA or ED were an additional limitation of this meta-analysis. There were five strategies diagnosing EA or ED in included trials, i.e., three-point scale, four-point scale, five-point scale, pediatric Anesthesia Emergence Delirium (PAED) scale, and the Confusion Assessment Method for the ICU.

CONCLUSION
In conclusion, compared with placebo, midazolam, and opioids, dexmedetomidine significantly decreased the incidence of post-anesthesia EA or ED in pediatric patients. However, dexmedetomidine did not exhibit this superiority when compared with propofol and ketamine. With regard to clonidine, chloral hydrate, melatonin, or ketofol, the results needed to be further tested due to the fact that there was only one trial in each study.

DATA AVAILABILITY STATEMENT
All datasets presented in this study are included in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
XW and JL designed this meta-analysis and supervised the acquisition and analysis of the data. YR, RZ, and XJ were independently responsible for reviewing the titles, abstracts, or both and summarized the data of the included literatures. RZ and XJ conducted statistical analysis of the data. YR wrote the manuscript. All authors contributed to the article and approved the submitted version.