Efficacy and Safety of Midazolam Oral Solution for Sedative Hypnosis and Anti-anxiety in Children: A Systematic Review and Meta-Analysis

Background: Midazolam is recommended by health guidelines for sedation and hypnosis in children. Oral solution is a suitable dosage form for children. But there is no conclusive evidence for sedative-hypnosis and antianxiety effects by midazolam oral solution in children. Methods: Relevant studies were identified through searching PubMed, Embase, Cochrane Library, CINAHL, International Pharmaceuticals, four Chinese electronic databases, and relevant lists. Two reviewers independently selected trials, assessed trial quality, and extracted the data. Results: Eighty-nine randomized controlled trials (RCTs) comparing midazolam oral solution with placebo or blank (n = 33), dexmedetomidine (n = 15), ketamine (n = 11), different midazolam doses (n = 10), midazolam injection (n = 8), chloral hydrate (n = 7), diazepam (n = 5), N2O (n = 5), triclofos (n = 4), butorphanol (n = 2), fentanyl (n = 2), hydroxyzine (n = 1), and thiopental (n = 1) were identified. Meta-analysis showed no significant difference in the success rate and duration of sedation and hypnosis between midazolam oral and injectable solution (P > 0.05). The success rate of sedation and hypnosis of midazolam was higher than that of ketamine [risk ratio (RR) = 1.32, 95% CI (1.07, 1.62), I2 = 0%, P < 0.01]. No significant difference was found in the success rate of sedation and hypnosis, mask acceptance, and parental separation between midazolam oral solution and dexmedetomidine (P > 0.05), and the result of one cohort study was consistent. The results of RCTs and a prospective cohort study showed that the incidence of adverse drug reactions (ADR) was 19.57% (189/966). Incidence of adverse reactions between dose groups of (0.25, 0.5] and (0.5, 1.0] mg/kg was similar [Pf (95% CI) = 0.10 (0.04, 0.24) and Pf (95% CI) = 0.09 (0.02, 0.39), respectively], higher than that of the dose group of (0, 0.25] mg/kg [Pf (95% CI) = 0.01 (0.00, 0.19)]. Conclusions: Available evidence suggests that midazolam oral solution is as good as midazolam injection and dexmedetomidine and is better than ketamine. Based on efficacy and safety results, an oral midazolam solution dose of 0.5–1 mg/kg is recommended for children.


INTRODUCTION
Data from World Bank showed that children (aged 14 years and younger) accounted for 25.79% of the world's total population in 2018. The use of sedative-hypnotic drugs in assisting children to complete medical examinations and surgery has become more and more extensive (Chinese Medical Association, 2011).
Midazolam is an imidazole benzodiazepine that has an inhibitory effect on the central nervous system, which is used for examination, diagnosis, and pretreatment sedation. Midazolam is rapidly and completely absorbed after oral administration and will take effect within 10-30 min of intake. Hydroxylated by cytochrome P450 and CYP3A isoenzymes, midazolam has 1 ′ -hydroxymidazolam as a major oxidation product (de Wildt et al., 2002). According to (WHO, 2011) andEuropean Medicine Agency (2006) requirements for children's appropriate preparations, oral solutions are suitable for children.
Five guides (Mace et al., 2004; National Institute for Health Care Excellence, 2010, 2016; Editorial Board of Chinese Journal of Pediatrics, 2015; Chinese Medical Association Anesthesia Branch, 2017a) and three expert consensuses (Chinese Medical Association Pediatrics Branch Emergency Study Group, 2014;Chinese Medical Association Anesthesia Branch, 2017b,c) recommend midazolam for children to calm, hypnotize, and counter anxiety. The NICE guidelines recommend oral administration of midazolam for sedation before painful examination in children (National Institute for Health Care Excellence, 2010). Midazolam is included in the WHO, UK, and Indian children's formulas (World Health Organization Regional Office for South-East Asia New Delhi, 2011; Committee P F., 2017; WHO Expert Committee on the Selection Use of Essential Medicines, 2017). The UK Formulary recommends taking 0.5 mg/kg of midazolam (maximum dose 20 mg) orally in children 30-60 min before the test (Committee P F., 2017). Seven related systematic reviews were identified (Peng et al., 2014;Sun et al., 2014;Zhang et al., 2014;Guo, 2015;Pasin et al., 2015;Jun et al., 2017;Mataftsi et al., 2017). Midazolam has a lower success rate of sedative hypnosis compared to chloral hydrate, but there is no statistical difference in safety, and the quality of included studies is poor (Mataftsi et al., 2017). Dexmedetomidine is better than midazolam in children separated from parents in pre-anesthesia induction (Peng et al., 2014;Sun et al., 2014;Zhang et al., 2014;Guo, 2015;Pasin et al., 2015;Jun et al., 2017).
Another study suggests that preoperative sedative and anxiolytic effects of dexmedetomidine by nasal drip are comparable to those of midazolam (Guo, 2015). Dexmedetomidine slows heartbeat, lowers blood pressure, and prolongs sedation duration (Zhang et al., 2014).
There was no systematic review of the efficacy and safety of midazolam oral solutions for sedative hypnosis and antianxiety effects in children. This study systematically evaluated the efficacy and safety of midazolam oral solution based on original research evidence and compared the effectiveness of oral and injectable solutions of midazolam. The relationship between the dose of midazolam and its effectiveness and safety was evaluated.

Search Strategy
PubMed, Embase, Cochrane Library, CINAHL, International Pharmaceuticals, China National Knowledge Infrastructure (CNKI), Chinese Biomedical Literature Database (CBM), Wanfang Database, VIP Database for Chinese Technical Periodicals (VIP), the WHO Clinical Trials Registry Platform, Cochrane Central Registry of Controlled Trials, and ClinicalTrials.gov were searched from their inception to August 2018. The retrieval strategy was specific and different for each database, including a combination of medical subject headings and free text terms for ("midazolam" or "dormicum" or "versed") and ("child" or "newborn" or "infant" or "neonate" or "toddler" or "teenager" or "adolescent" or "pediatric"). We systematically searched the official website of the National Drug Administration and the Center for Adverse Reaction Monitoring for reports of midazolam adverse reactions in countries and regions around the world.

Inclusion/Exclusion Criteria
A report was selected for inclusion if (a) participants were children aged 0-18 years; (b) the intervention group only used midazolam oral solution, and the route of administration was oral; (c) for comparisons, the control group was blank control, placebo, midazolam injection (intravenous, intramuscular, and subcutaneous), or other sedative-hypnotic drugs, and the dose and course of treatment were not limited; (d) studies focused on the efficacy and safety outcome of midazolam; (e) studies were randomized controlled trials (RCTs), cohort studies, case-control studies, case series studies, case reports, and cross-sectional survey studies.
Studies were excluded if (a) they were repeated published studies; (b) they were non-Chinese and non-English studies; (c) their full text is not available; (d) they were comparative studies with different routes of administration of midazolam (except for oral vs. injectable).

Outcome Parameters
The primary outcome was the success rate of sedation and hypnosis (the ratio of the number of people who successfully completed an examination or surgery to the total number of people). Secondary outcomes were depth of sedation (sedative hypnosis depth scores), anxiety scores, duration of sedation and hypnosis (the time when children fell asleep to the time responding to command), the time of falling asleep (the time from the end of the medication to the state of falling into sleep), and the type and incidence of adverse reactions.

Data Extraction
Two reviewers (Cheng and Xu) independently screened the titles and abstracts of every record. Full articles were obtained when either information conformed to satisfy the selection criteria outlined previously or not enough to ascertain because of limited information. Data were independently extracted by each reviewer and entered into a standardized form. The data extraction form included general characteristics and outcome measurements. Discrepancies were resolved by a third researcher, Chen.

Data Analysis
Results for dichotomous outcomes were expressed as risk ratios (RR) with 95% confidence intervals (CIs), and for continuous outcomes, the mean difference (MD) with 95% CIs was accounted for. P ≤ 0.05 was considered statistically significant. Final outcomes of treatment vs. placebo or other medicines were used for the analysis, as recommended by the Cochrane Handbook for Systematic Review of Interventions, except where large pretreatment differences were identified; for these studies, the change from baseline was compared instead to prevent skewing of results. Where mean and/or standard deviation values were not reported, these were calculated based on reported CIs or P-values.
The meta-analyses were performed using Review Manager 5.3 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration) software. Descriptive analysis was employed for data that cannot be meta-analyzed. The incidence of adverse reactions was analyzed by meta-analysis of uncontrolled dichotomous data. Pf[95% CI] referred to the effect volume after correcting the relevant factors. Pf was equal to odds ratio (OR) divided by (1 + OR). With reference to MedDRA 20.1, the types of adverse reactions were divided into cardiovascular system, digestive system, nervous system, and so on. In combining the studies, the conservative random effects model was employed, since the underlying effects can differ across studies and populations that are not necessarily homogeneous (DerSimonian and Laird, 1986). Statistical heterogeneity was analyzed by χ 2 tests. Heterogeneity was quantified, where 25% = small, 50% = moderate, and 75% = high heterogeneity (Higgins, 2011). To cope with the potential heterogeneity across studies, subgroup analyses were conducted.
RCTs and cohort studies were included when analyzing the effectiveness of midazolam. Case-control studies and case reports were also included in the safety analysis.

Assessment of Bias
The risk of bias was assessed for all clinical trials included in the quantitative and qualitative analysis. Bias of RCTs was assessed by using the Cochrane Handbook for Systematic Review of Interventions. As per recommendations in the Cochrane Handbook for Systematic Review of Interventions version 5.3, bias was assessed based on the following six domains: (1) sequence generation, (2) allocation concealment, (3) blinding, (4) incomplete outcome data, (5) selective outcome, and (6) other biases. The Newcastle-Ottawa Scale (NOS) was used to assess the quality of cohort studies and case-control studies. Critical appraisal checklists of the Joanna Briggs Institute (JBI) were used to assess the bias of case reports and case series. Two researchers (Cheng and Xu) independently completed the quality evaluation, and discrepancies were resolved by the third researcher Chen.

Results of the Literature Search
The search yielded a total of 15,865 references (duplication = 3,977 references). We excluded 11,547 references after reviewing the title and abstract. A further 241 references were excluded after full-text reviews, because 87 involved combination therapy or did not concern midazolam oral solution, 79 did not concern children, and 75 were review or conference papers. In total, we identified 89 RCTs, eight cohort studies, 12 case series, and one case report that met our inclusion criteria (Figure 1).

Quality Assessment
Quality assessment for RCTs, 45% (40/89) of studies used an adequate method of random sequence generation. Thirteen percent (12/89) of studies implemented adequate allocation concealment. Fifty-two percent (46/89) used the methods of blinding to patients and researchers. Thirty-nine percent (35/89) used the methods of blinding to the outcome measurer. The mean score of risk of bias of eight cohort studies was 5.75. Results of quality assessment of 12 case series and one case report were in Supplementary Tables 2-5.
One RCT (Ghali et al., 2011) compared midazolam oral solution with dexmedetomidine, including 120 children, which showed that the anxiety scores of midazolam oral solution were higher and that the anxiolytic effect of midazolam oral solution was worse [MD = 12.76, 95% CI (11.25, 14.27), P < 0.01].

Comparison of Different Doses of Oral Midazolam
According to the British National Formulary for Children and the instructions of theUS midazolam syrup(Product name:Versed), the dose of midazolam oral solution was divided into three dose groups (0-0.25), mg/kg (0.25-0.5), mg/kg and (0.5-1.0) mg/kg.

Summary of Main Findings
Certain examinations and treatments for children require sedation and hypnosis. The success rate of sedation and hypnosis and the duration of sedative hypnosis are widely used outcome measures. Both its therapeutic effects and adverse reactions are due to its neuronal inhibitory pathways by affecting the gammaaminobutyric acid (GABA) receptor (Jacqz-Aigrain and Burtin, 1996). We systematically reviewed the efficacy and safety of midazolam oral solution for sedative hypnosis and antianxiety in children.
No statistically significant difference in the efficacy and adverse effects of midazolam for oral solution and injection was found. Neither was there a statistically significant difference in the incidence of adverse reactions between the midazolam oral solution and the blank or placebo group. There was inconsistent evidence that oral midazolam decreased anxiety during procedures compared with placebo. Oral solution has the advantages of convenience, non-invasiveness, safety, and economy (European Medicine Agency, 2006). Midazolam was more effective than other benzodiazepines (hydroxyzine and diazepam). Midazolam is preferred over other benzodiazepines because of its water solubility and rapid clearance.
The systematic review showed that the midazolam oral solution had a lower sedation success rate than chloral hydrate, consistent with the results of the previous systematic review (Mataftsi et al., 2017), but the evidence quality was low, and the chloral hydrate group dose was higher than the regular clinical dose. The overall incidence of adverse effects was comparable. Three RCTs showed that midazolam oral solutions had a higher incidence of mental adverse events [RR = 5.67, 95% CI (3.54, 9.09), I 2 = 0%, P < 0.05] and a lower incidence in the digestive system [RR = 0.24, 95% CI (0.07, 0.78), I 2 = 34%, P = 0.02]. One cohort study showed no statistically significant difference in each system, but the ratio of experimental groups compared with the control group was approximately 1:20, which could lead to false-negative results (Schmidt et al., 2018). Chloral hydrate is now commonly used in clinical practice for sedation and hypnosis for diagnosis or treatment. It is strangely odorous, so the child's compliance of oral solution is poor, which makes it easy to cough and even suffocate. The rectal administration of chloral hydrate is easy to stimulate the intestinal wall, and the administration process is troublesome. The active metabolites of chloral hydrate may cause long-term sedation and a narrow therapeutic index (Abbas et al., 1996). Both chloral hydrate and midazolam are recommended by the American Institute for Safe Medication Practices as a high-risk drug (Institute for Safe Medication Practices, 2019). The total adverse reaction rate of midazolam oral solution was not statistically different from that of chloral hydrate. Therefore, clinical applications of chloral hydrate and midazolam oral solution should be done with caution. The sedative and hypnotic success rate of oral solution of midazolam was equivalent to that of dexmedetomidine. The incidence of neuropsychiatric adverse reactions in midazolam oral solution is higher [RR = 12.00, 95% CI (2.93, 49.23), I 2 = 0%, P < 0.05]. A systematic review (Zhang et al., 2014) showed that dexmedetomidine reduced the risk of agitation or paralysis, chills, systolic blood pressure, and heart rate in children compared with midazolam. The bioavailability of dexmedetomidine oral sedation is poor, about 16%, and its nasal mucosa absorption is more stable (Uusalo et al., 2019). The pharmacokinetics of dexmedetomidine is highly characterized by individual differences, especially in the intensive care unit population (Weerink et al., 2017).
Our evidence suggests that clinical midazolam oral solutions are used in children for sedative hypnosis at doses ranging from 0.25 to 1.0 mg/kg. Our review showed that the time for children to fall asleep was longer in the dose of (0.25-0.5) mg/kg compared to (0.5-1.0) mg/kg, and the two groups were equivalent in relieving anxiety and parental success rate. And the incidence of adverse reactions was similar [10% for (0.25-0.5) mg/kg and 9% (0.5-1.0) mg/kg]. These results suggest that the oral solution of 0.25-1.0 mg/kg of midazolam is effective and safe, and the safe and effective dose of midazolam in the pediatric surgery outdoor anesthesia/sedation expert consensus (2017) is 0.50-0.75 mg/kg, so they are consistent to some extent (Chinese Medical Association, 2011).
We performed a subgroup analysis comparing midazolam with other drugs, based on different doses of midazolam. The doses of midazolam oral solution were divided into three dose groups: (0-0.25), (0.25-0.5), and (0.5-1.0) mg/kg. The results did not differ from those before grouping. This may be the result of the insufficient numbers of included studies.

Limitations
This systematic review had some limitations. First, the literature included in this study is limited to Chinese and English, and thus, there may be language bias. Second, 77 of the 89 RCTs did not clearly describe the allocation concealment. If researchers and authors document their experimental methods in future clinical trials and publications in detail, readers and reviewers can better understand the true content of the study. Third, of these 89 RCTs, the smallest sample size is 10 (median 60). A small sample size results in a less authentic result. We look forward to a larger population and long-term data to fully assess the efficacy and risk.

CONCLUSION
Limited evidence suggests that midazolam is effective and safe prior to a diagnosis or treatment procedure for sedation and hypnosis in children. Available evidence suggests that midazolam oral solution is as good as midazolam injection and dexmedetomidine and is better than ketamine. The success rate of sedation and hypnosis of midazolam oral solution was lower than that of chloral hydrate. Neuropsychiatric adverse reactions of midazolam are higher than those of chloral hydrate and dexmedetomidine. Digestive system adverse reactions of midazolam are lower than those of chloral hydrate.

AUTHOR CONTRIBUTIONS
LZh, ZC, LZe, LH, DY, XC, and PX conceived and designed the review. XC, PX, FQ, XJ, YW, and ML reviewed the literature and extracted data. XC and ZC wrote the manuscript.