Chloral hydrate as a sedating agent for neurodiagnostic procedures in children

Choong Yi Fong; Chee Geap Tay; Lai Choo Ong; Nai Ming Lai

doi:10.1002/14651858.CD011786.pub2

. 2017 Nov 3;2017(11):CD011786. doi: 10.1002/14651858.CD011786.pub2

Chloral hydrate as a sedating agent for neurodiagnostic procedures in children

Choong Yi Fong ^1,^✉, Chee Geap Tay ¹, Lai Choo Ong ¹, Nai Ming Lai ²

Editor: Cochrane Epilepsy Group

PMCID: PMC6486182 PMID: 29099542

Abstract

Background

Paediatric neurodiagnostic investigations, including brain neuroimaging and electroencephalography (EEG), play an important role in the assessment of neurodevelopmental disorders. The use of an appropriate sedative agent is important to ensure the successful completion of the neurodiagnostic procedures, particularly in children, who are usually unable to remain still throughout the procedure.

Objectives

To assess the effectiveness and adverse effects of chloral hydrate as a sedative agent for non‐invasive neurodiagnostic procedures in children.

Search methods

We used the standard search strategy of the Cochrane Epilepsy Group. We searched MEDLINE (OVID SP) (1950 to July 2017), the Cochrane Central Register of Controlled Trials (CENTRAL) (the Cochrane Library, Issue 7, 2017), Embase (1980 to July 2017), and the Cochrane Epilepsy Group Specialized Register (via CENTRAL) using a combination of keywords and MeSH headings.

Selection criteria

We included randomised controlled trials that assessed chloral hydrate agent against other sedative agent(s), non‐drug agent(s), or placebo for children undergoing non‐invasive neurodiagnostic procedures.

Data collection and analysis

Two review authors independently assessed the studies for their eligibility, extracted data, and assessed risk of bias. Results were expressed in terms of risk ratio (RR) for dichotomous data, mean difference (MD) for continuous data, with 95% confidence intervals (CIs).

Main results

We included 13 studies with a total of 2390 children. The studies were all conducted in hospitals that provided neurodiagnostic services. Most studies assessed the proportion of sedation failure during the neurodiagnostic procedure, time for adequate sedation, and potential adverse effects associated with the sedative agent.

The methodological quality of the included studies was mixed, as reflected by a wide variation in their 'Risk of bias' profiles. Blinding of the participants and personnel was not achieved in most of the included studies, and three of the 13 studies had high risk of bias for selective reporting. Evaluation of the efficacy of the sedative agents was also underpowered, with all the comparisons performed in single small studies.

Children who received oral chloral hydrate had lower sedation failure when compared with oral promethazine (RR 0.11, 95% CI 0.01 to 0.82; 1 study, moderate‐quality evidence). Children who received oral chloral hydrate had a higher risk of sedation failure after one dose compared to those who received intravenous pentobarbital (RR 4.33, 95% CI 1.35 to 13.89; 1 study, low‐quality evidence), but after two doses there was no evidence of a significant difference between the two groups (RR 3.00, 95% CI 0.33 to 27.46; 1 study, very low‐quality evidence). Children who received oral chloral hydrate appeared to have more sedation failure when compared with music therapy, but the quality of evidence was very low for this outcome (RR 17.00, 95% CI 2.37 to 122.14; 1 study). Sedation failure rates were similar between oral chloral hydrate, oral dexmedetomidine, oral hydroxyzine hydrochloride, and oral midazolam.

Children who received oral chloral hydrate had a shorter time to achieve adequate sedation when compared with those who received oral dexmedetomidine (MD ‐3.86, 95% CI ‐5.12 to ‐2.6; 1 study, moderate‐quality evidence), oral hydroxyzine hydrochloride (MD ‐7.5, 95% CI ‐7.85 to ‐7.15; 1 study, moderate‐quality evidence), oral promethazine (MD ‐12.11, 95% CI ‐18.48 to ‐5.74; 1 study, moderate‐quality evidence), and rectal midazolam (MD ‐95.70, 95% CI ‐114.51 to ‐76.89; 1 study). However, children with oral chloral hydrate took longer to achieve adequate sedation when compared with intravenous pentobarbital (MD 19, 95% CI 16.61 to 21.39; 1 study, low‐quality evidence) and intranasal midazolam (MD 12.83, 95% CI 7.22 to 18.44; 1 study, moderate‐quality evidence).

No data were available to assess the proportion of children with successful completion of neurodiagnostic procedure without interruption by the child awakening. Most trials did not assess adequate sedation as measured by specific validated scales, except in the comparison of chloral hydrate versus intranasal midazolam and oral promethazine.

Compared to dexmedetomidine, chloral hydrate was associated with a higher risk of nausea and vomiting (RR 12.04 95% CI 1.58 to 91.96). No other adverse events were significantly associated with chloral hydrate (including behavioural change, oxygen desaturation) although there was an increased risk of adverse events overall (RR 7.66, 95% CI 1.78 to 32.91; 1 study, low‐quality evidence).

Authors' conclusions

The quality of evidence for the comparisons of oral chloral hydrate against several other methods of sedation was very variable. Oral chloral hydrate appears to have a lower sedation failure rate when compared with oral promethazine for children undergoing paediatric neurodiagnostic procedures. The sedation failure was similar for other comparisons such as oral dexmedetomidine, oral hydroxyzine hydrochloride, and oral midazolam. When compared with intravenous pentobarbital and music therapy, oral chloral hydrate had a higher sedation failure rate. However, it must be noted that the evidence for the outcomes for the comparisons of oral chloral hydrate against intravenous pentobarbital and music therapy was of very low to low quality, therefore the corresponding findings should be interpreted with caution.

Further research should determine the effects of oral chloral hydrate on major clinical outcomes such as successful completion of procedures, requirements for additional sedative agent, and degree of sedation measured using validated scales, which were rarely assessed in the studies included in this review. The safety profile of chloral hydrate should be studied further, especially the risk of major adverse effects such as bradycardia, hypotension, and oxygen desaturation.

Plain language summary

The effectiveness of chloral hydrate as a sedative agent for children undergoing neurodiagnostic procedures

Review question

In children undergoing non‐invasive neurodiagnostic procedures, is oral chloral hydrate more effective at producing adequate sedation and safer than other ways of achieving sedation?

Background

Neurodiagnostic procedures are non‐invasive neurological investigations important for children with suspected neurological disorders. These investigations include brain imaging and brain electrical activity testing. For these tests to be successfully performed, the child needs to remain still for at least 30 to 45 minutes during the investigation period. Sedative agents are required for children, who are usually unable to remain still for this period of time.

Search date

We performed a search in multiple medical databases in July 2017.

Study characteristics

Thirteen studies involving a total of 2390 children fit our inclusion criteria. These studies were all performed in hospitals that provided neurodiagnostic services. Most of the studies assessed three main outcome measures: i) proportion of children who were unsuccessfully sedated for the neurodiagnostic procedure, ii) length of time taken for adequate sedation, and iii) side effects associated with the sedative agent. The quality of the included studies was mixed, ranging from very low to high. The quality of the studies was affected mainly because those closely involved in the trials, such as the doctors giving the sedation or the parents of the child, were not masked from knowing which sedative agent was given to the child, which could have affected their recording or interpretation of the results.

Key results

We summarised the evidence of effectiveness and harms of oral chloral hydrate sedation when compared with other sedative medications. We included 13 studies with a total of 2390 children (age up to 18 years old). The studies were all conducted in hospitals that performed neurodiagnostic procedures. Our review suggests that oral chloral hydrate is just as effective a sedative agent with similar sedation failure rate when compared with oral dexmedetomidine, oral hydroxyzine hydrochloride, and oral midazolam; and probably a more effective sedative agent with lower sedation failure rate when compared with oral promethazine. While most of the included studies showed that chloral hydrate was safe with no increased side effects when compared to other sedative agents, one study reported an increased risk of adverse effects when compared with oral dexmedetomidine.

Quality of the evidence

The quality of most of the evidence was poor due to methodological flaws in the included studies and the small sample size of each study. Consequently our confidence in the results of the studies is reduced.The major factor affecting the quality of the evidence was lack of precision in the result estimates, as the calculated plausible range of the effects were wide.

Conclusions

Apart from intravenous pentobarbital and music therapy, oral chloral hydrate is either just as effective or more effective a sedative agent when compared to other sedative agents for children undergoing non‐invasive neurodiagnostic procedures. In view of the poor quality of the evidence, we could draw no clear conclusions on the effectiveness or safety of any paediatric sedative agent. The side effects profile of oral chloral hydrate when compared to other sedatives requires further study.

Summary of findings

Background

Description of the condition

Paediatric neurodevelopmental disorders are collectively a common problem that affect between 5% and 10% of all children (Horridge 2011; Shevell 2003; Shevell 2008). A structured clinical evaluation together with targeted neurodiagnostic investigations of children with possible neurodevelopmental disorders is important to establish the aetiology of the disorder, predict the prognosis, and develop management strategies (Horridge 2011; Shevell 2003; Shevell 2008; Silove 2013).

Paediatric neurodiagnostic (non‐interventional) investigations used in the assessment of children with possible neurodevelopmental disorders include brain neuroimaging and electroencephalography (EEG). Successful completion of neurodiagnostic procedures is challenging in the paediatric age group, as the child is required to be immobile during the procedure. In addition, routine paediatric EEG recording requires recording of a period of sleep. Sleep state EEG recording is important as it increases the yield in detecting interictal epileptiform abnormalities. Achieving spontaneous sleep is not possible in some children despite behavioural intervention (prior sleep deprivation). The use of an appropriate sedative agent in children who are unable to be immobile or do not naturally sleep (for EEG recording), or both, thus plays a pivotal role in ensuring successful completion of the procedure.

Selection of the most appropriate pharmacological and non‐pharmacological agents is an important factor for the success and safety of the neurodiagnostic procedure. Moreover, there are other pharmacological considerations when choosing a sedative agent for paediatric EEG because some sedative agents can suppress abnormal EEG activity or background activity (e.g. anaesthesia), and others can induce background activity changes that may obscure abnormalities (e.g. benzodiazepines and barbiturates).

Children with neurodevelopmental disorders can present with a variety of symptoms including epilepsy, developmental impairment, developmental regression, social communication deficits, and behavioural difficulties. These children will require a structured clinical evaluation with targeted investigations to enable an accurate diagnosis, provide appropriate intervention, and assist in predicting long‐term prognosis.

Neurodiagnostic investigations include EEG and brain neuroimaging. The duration of a standard awake and sleep paediatric EEG recording is approximately 30 to 60 minutes; a routine paediatric brain magnetic resonance imaging (MRI) scan takes approximately 45 to 60 minutes. Routine EEG is an important investigation tool in evaluating children with a clinical history that suggests epilepsy (NICE 2012). The EEG will assist in defining the specific epilepsy syndrome (Pillai 2006). Neuroimaging investigations (particularly brain MRI) are useful in detecting major or minor structural brain abnormalities that may assist in establishing the cause of the child's neurodevelopmental disorder (Battaglia 2003; Rodriguez 2007). Current guidelines advocate that neuroimaging investigations should be considered in those children with an abnormal antenatal or perinatal history, an abnormal neurological examination, a history of focal epilepsy, and in all epilepsies in children under two years of age (Battaglia 2003; NICE 2012; Rodriguez 2007).

Description of the intervention

Sedation is defined as a drug‐induced depression of consciousness, which is a continuum from wakefulness to anaesthesia (Starkey 2011). It assists in reducing anxiety, providing pain control, and minimising movement of the patient when undergoing a procedure. There are different levels of sedation, as defined by the American Society of Anesthesiologists. These include minimal sedation, moderation sedation, and deep sedation (ASA 2007). Moderate sedation involves the technique of administering sedatives to induce a state of depressed level of consciousness while allowing the patient to maintain independent control of their airway and oxygenation by preserving their protective airway reflexes (NICE 2010). Current guidelines recommend that moderate sedation is preferred over deep sedation and general anaesthesia for painless paediatric procedures (NICE 2010).

The ideal sedative agent to achieve moderate sedation should have rapid onset, a moderate duration of action, minimal or no adverse effects, and low cost. At present there is no ideal or first‐line sedative agent. The agents currently available for painless procedures include the following.

Chloral hydrate, a non‐opiate and non‐benzodiazepine sedative hypnotic drug.
Benzodiazepines, which act via the γ‐aminobutyric acid A receptors.
Hydroxyzine, a long‐acting first‐generation H₁ antagonist with central nervous system depressant activity and minimal circulatory and depressant activity.
Melatonin, a pineal hormone and natural sleep agent that can modulate the circadian rhythm of sleep through action on the suprachiasmatic nucleus in the hypothalamus.
Promethazine, an antiemetic and antisialagogue with sedative properties.
Dexmedetomidine, a selective alpha₂‐adrenoceptor agonist with sedative properties.

In addition, non‐pharmacological approaches such as music therapy have been shown to be beneficial in providing sedation during procedures in adult patients (Bechtold 2009), while there is some evidence that breastfeeding or breast milk helps infants tolerate painful or uncomfortable procedures better by reducing crying and improving infants' physiological responses during the procedures (Shah 2012).

How the intervention might work

Chloral hydrate is a central nervous system depressant and is one of the oldest sedatives (discovered in 1832). It is well absorbed orally as well as rectally and rapidly metabolised into the active metabolite trichloroethanol, which is responsible for its hypnotic and sedative effects. It is one of the most frequently used conscious sedative agents for children undergoing neurodiagnostic procedures including neuroimaging and EEG recording, as it has a good safety profile when used at sedative doses, with little effect on EEG background activity (Malviya 1997). Doses vary from 55 to 100 mg/kg in neonates and children younger than 12 years old, with a maximum dose of 2g (Starkey 2011). Chloral hydrate is absorbed from the gastrointestinal tract and starts to exert its effect within 60 minutes. Adverse reactions occur in 1.7% to 20% of children, with gastrointestinal side effects, particularly vomiting, being the most common (Starkey 2011). There have also been some concerns with regard to respiratory depression, long duration of action, and potential carcinogenicity (Haselkorn 2006; Kao 1999).

Why it is important to do this review

Neurodiagnostic procedures (particularly neuroimaging) in children are important investigative tools and have been increasingly used over the last decade to assess and manage children with neurodevelopmental disorders. However, a significant proportion of children may be unable to complete the procedure due to sedation failure. Sedation failure is a great inconvenience to the child and their family and requires rescheduling of another hospital admission to complete the procedure, often under general anaesthesia (which carries additional associated risks).

Both the National Institute for Health and Care Excellence (NICE) 2010 guideline and the American College of Emergency Physicians 2008 guideline recommend chloral hydrate for moderate sedation during painless procedures in the paediatric population (Mace 2008; NICE 2010). However, neither guideline has suggested the superiority of chloral hydrate over other agents. Some studies have shown that chloral hydrate is unsuccessful in a significant proportion of children, resulting in failure to complete the procedure (Beebe 2000; Malviya 1997). There have been a few randomised controlled trials (RCTs) comparing the efficacy of chloral hydrate with other sedative agents and complementary therapy (e.g. music therapy). However, to date no meta‐analysis of the previously published RCTs has been performed in order to determine superiority among these agents.

The aim of this review was to allow clinicians to take an evidence‐based approach in deciding which sedating agent is best for children undergoing neurodiagnostic procedures. The care of children would improve if both sedation failure rates and adverse effects due to sedation were kept to a minimum.

Objectives

To assess the effectiveness and adverse effects of chloral hydrate as a sedative agent for non‐invasive neurodiagnostic procedures in children.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised or quasi‐randomised trials.

Types of participants

Children (from birth to 18 years old) who underwent non‐invasive neurodiagnostic procedures (including brain neuroimaging and sleep EEG) who required sedation before the procedure.

Types of interventions

Oral or rectal chloral hydrate.

Comparison: other sedative/sleep‐inducing agents (e.g. promethazine, hydroxyzine, melatonin, midazolam, pentobarbital, etomidate, sevoflurane) or complementary therapies (e.g. music therapy), or achieved sleep without a sedative agent, for instance after a milk feed.

Types of outcome measures

Primary outcomes

Proportion of children who successfully completed neurodiagnostic procedure without interruption by the child awakening.
Proportion of children who required a further dose of either the same sedative agent or the addition of a different sedative agent.
Time to adequate sedation (minutes or as measured by specific validated scales such as the Ramsay Sedation Score).

Secondary outcomes

Proportion of children who had sedation failure or inadequate level of sedation.
Sedation duration
Sleep onset latency (time interval for child to fall asleep) pre‐procedure.
Yield of EEG findings (expressed either as the rate of abnormal EEG findings or additional EEG artefact findings attributable to the sedative agent, which may prevent interpretation of the EEG) or yield of neuroimaging findings (expressed as the rate of non‐interpretable MRI attributable to motion artefact due to inadequate sedation).
Number of children with clinical adverse effects and severe drug side effects (e.g. hypotension, hypoxia, apnoea, laryngospasm, significant vomiting, refractory agitation, bradycardia).

Search methods for identification of studies

Electronic searches

We searched the following databases:

Cochrane Epilepsy Group Specialized Register (via Cochrane Central Register of Controlled Trials, searched on 19 July 2017)
Cochrane Central Register of Controlled Trials (CENTRAL) (Issue 7, 2017);
MEDLINE (PubMed (National Library of Medicine)) (1950 to 19 July 2017);
Embase (1980 to 19 July 2017).

We outlined detailed search strategies for the above databases in Appendix 1, Appendix 2, and Appendix 3.

We also searched for ongoing clinical trials and unpublished studies via the following sites:

ClinicalTrials.gov (www.clinicaltrials.gov);
World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp/en/).

We did not apply any language restrictions in our searches.

Searching other resources

We searched references cited in relevant studies, Cochrane Reviews, guidelines, review articles, and conference proceedings, including abstracts from the Society for Pediatric Anesthesia and the European Society for Paediatric Anaesthesiology. We also contacted experts in the field if necessary to identify further relevant studies.

Data collection and analysis

Selection of studies

Two review authors (CYF, CGT) independently performed the first round of searching for relevant studies. They then screened the identified studies for inclusion in the analysis with the help of an arbiter (LCO), using the predefined inclusion and exclusion criteria.

We collected information on study design and setting, participant characteristics (including age and the presence of additional co‐ morbidities such as cerebral palsy), study eligibility criteria, details of the intervention(s) given, and the outcomes assessed. We also collected study funding source and any conflicts of interest stated by the investigators.

We accepted published and unpublished studies, both in full article and abstract form, as long as adequate extraction of outcome data was possible. We contacted the authors of unpublished studies and studies available only as abstracts to request further information about the studies, including specific details of the methodologies employed as well as the full outcome data, so as to include them in our meta‐analysis. We only included final data from each study and not data from interim analyses.

Data extraction and management

Two review authors (CYF, CGT) independently extracted relevant data from each included study using a data collection form adapted for this review. We gathered the following information from each study.

Study design (RCT, quasi‐RCT, or cluster‐RCT, and the number of arms evaluated)
'Risk of bias' items, including the methods used and our 'Risk of bias' judgement (low risk, unclear risk, or high risk)
- Random sequence generation
- Allocation concealment
- Blinding of participants, care personnel, and evaluators
- Missing data (loss to follow‐up)
- Risk of selective outcome reporting
- Other possible sources of bias
Stratification factors
Participant factors
- Number (total per group)
- Setting
- Diagnostic criteria
- Country
- Age
- Sex
- Underlying diagnoses
- Indications for neurodiagnostic procedure: whether for diagnostic or monitoring purposes
Intervention data: route of administration, dosage, and duration of chloral hydrate and comparator medication, and whether chloral hydrate was administered alone or in combination with another medication
Follow‐up data
- Duration of follow‐up period
- Numbers lost to follow‐up
- Reasons for loss to follow‐up
Outcome data, grouped into dichotomous and continuous outcomes, including adverse events

For details of each outcome included in our review (names, definitions, and possible unit of measurement), please refer to Types of outcome measures.

We screened for multiple enrolment by matching the initial number of participants recruited against the total numbers at each step in the study. We contacted the authors of the study for additional information if we were unclear as to important aspects of the trial. Any disagreements were resolved by discussion, leading to a consensus, or by involving an arbiter (LCO) if necessary.

Assessment of risk of bias in included studies

Two authors (CGT, NML) independently assessed risk of bias for each of the included studies and cross‐checked the results of these assessments. A third review author (CYF) discussed and resolved any inconsistencies.

In accordance with Cochrane's tool for assessing risk of bias in RCTs (Higgins 2011), we assessed the following 'Risk of bias' domains:

sequence generation;
allocation concealment;
blinding;
incomplete data/attrition bias;
selective outcome reporting;
other.

Measures of treatment effect

We reported the outcome estimates for categorical data using risk ratios (RRs), risk differences (RDs), and the number needed to treat for an additional beneficial outcome (NNTB) or the number needed to treat for an additional harmful outcome (NNTH). For continuous data, which included time to adequate sedation and duration of sedation, we used mean differences (MDs) with their respective 95% confidence intervals (CIs). If pooled analyses were not possible for reasons such as major discrepancies in study characteristics or outcome reporting as detailed under Assessment of heterogeneity, we reported the results of the studies narratively.

Unit of analysis issues

No cluster‐RCTs were included in this review. Had we included cluster‐RCTs, we would have adopted the following strategy when dealing with such studies.

For cluster‐RCTs (e.g. trials in which the assignment to the intervention or control group was made at the level of the unit or hospital ward), we would have assessed whether adjustment had been made for the effects of clustering in order to account for non‐independence among the participants in a cluster via the use of an appropriate analysis model, such as the generalised estimating equation (GEE) model. If the unit of analysis was not stated in the study, we would have inspected the width of the standard error (SE) or 95% CI of the estimated treatment effects. Had we found an inappropriately small SE or a narrow 95% CI, we would have asked the authors of the study to provide information on the unit of analysis.

If no adjustment was made for the effects of clustering, we would have performed adjustment by multiplying the SEs of the final effect estimates by the square root of the 'design effect', represented by the formula '1 + (M ‐ 1) x ICC', where M is the average cluster size (number of children per cluster) and ICC is the intracluster correlation. We would determine the average cluster size (M) from each trial by dividing the total number of children by the total number of clusters. We would use a relatively large assumed ICC of 0.10, as we considered this to be a realistic estimate based on previous studies about implementation research (Campbell 2001). We would combine the adjusted final effect estimates from each trial with their SEs in meta‐analysis using generic inverse‐variance methods, as stated in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

If determination of the unit of analysis was not possible, we would have included the studies concerned in a meta‐analysis using whatever effect estimates were provided by the authors. We would then perform a sensitivity analysis to assess how the overall results were affected by the inclusion of these studies.

For multiple‐arm studies, we would adjust the data by following the methods stated in Chapter 16 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Specifically, if only two arms were relevant, we would only include the relevant arms. If there were more than two relevant arms (e.g. chloral hydrate of two different dosages versus placebo), we would set up separate pair‐wise comparisons, for example higher‐dose chloral hydrate versus placebo (Comparison 1) and lower‐dose chloral hydrate versus placebo (Comparison 2). In such cases, we would not total the number of children in the placebo group to avoid double‐counting.

Dealing with missing data

We determined the dropout rates from each study. We also assessed whether the authors analysed their data according to the intention‐to‐treat principle by comparing the number of children initially randomised and the total number analysed. We considered an absolute dropout rate of 20% or higher as important. We also adopted a 'worst‐case scenario' approach in judging the dropout rate: if we found that the direction of the effect estimate changed with the worst‐case scenario approach, we considered the dropout rate as significant.

If we found a significant dropout rate with no reasonable explanation, we classified the study as having a high risk of attrition bias for the domain incomplete outcome data. If we considered the missing data to be critical to the final estimates in our meta‐analysis, we contacted the authors of the individual studies to request further data.

We planned to perform sensitivity analyses to assess how the overall results were affected by the inclusion of studies with a high risk of attrition bias from incomplete outcome data.

Assessment of heterogeneity

We assessed clinical heterogeneity by comparing the distribution of important participant factors between trials (age, gender, seizure type, duration of epilepsy) and trial factors (clinical setting of the studies, the use of co‐interventions, 'Risk of bias' items including random sequence generation, allocation concealment, blinding, and losses to follow‐up).

For statistical heterogeneity, we first visually inspected the forest plots to see if there was any gross inconsistency in the trial results. We used the I² statistic to quantify the extent and importance of inconsistency in the results (Higgins 2011). We used the following cutoffs for the reporting of heterogeneity.

0% to 40%: heterogeneity might not be important
30% to 60%: may represent moderate heterogeneity
50% to 90%: may represent substantial heterogeneity
75% to 100%: considerable heterogeneity

If we found moderate, substantial, or considerable heterogeneity, we re‐examined the clinical and methodological characteristics of the studies, using the criteria listed above, to determine whether the degree of heterogeneity could be explained by differences in those characteristics, and whether a pooled analysis would be appropriate. If we considered a pooled analysis appropriate, we then performed this analysis using a random‐effects model.

Assessment of reporting biases

Had there been more than 10 studies available for an outcome to enable a meaningful assessment, we would have used a funnel plot to screen for publication bias. If we detected significant asymmetry in the funnel plot, which is suggestive of the possibility of publication bias, we would have further assessed it using Begg's rank correlation and Egger's test and would have included a statement in the Results with a corresponding note of caution in the Discussion.

Data synthesis

We performed meta‐analyses using Review Manager 5 software with a fixed‐effect model (RevMan 5.3), unless there was moderate or high heterogeneity, in which case we used the random‐effects model alongside an exploration of possible causes of heterogeneity, as described in the Assessment of heterogeneity section. Our primary data analyses followed the intention‐to‐treat principle; namely, we used the original number of participants allocated to each study arm as the denominator in subsequent analyses. We expressed our results as RRs, RDs, NNTB, NNTH, and MDs with their respective 95% CIs, as detailed in the Measures of treatment effect section.

Quality of evidence

We used the GRADE approach, as outlined in the GRADE Handbook (Schünemann 2013), to assess the quality of evidence for the most important outcomes across all comparisons, as listed below, whether or not there were data for the outcome.

Prespecified primary outcomes

Proportion of children who successfully completed neurodiagnostic procedure without interruption by the child awakening.
Proportion of children who required a further dose of either the same sedative agent or the addition of a different sedative agent.
Time to adequate sedation (minutes or as measured by specific validated scales such as the Ramsay Sedation Score).

Prespecified secondary outcomes

Proportion of children who had sedation failure or inadequate level of sedation.
Sedation duration (minutes).
Yield of EEG findings, including non‐interpretable EEG or sedative‐induced artefact.
Number of children with clinical adverse events (any).

Two review authors independently assessed the quality of the evidence for each of the outcomes listed above. We considered evidence from RCTs as high quality, but downgraded the evidence one level for serious (or two levels for very serious) limitations based upon the following: design (risk of bias), consistency across studies, directness of the evidence, precision of estimates, and presence of publication bias (Guyatt 2008; Schünemann 2013). We used the GRADEpro GDT platform (Gradepro GDT 2015) to create ‘Summary of findings’ tables to report the quality of the evidence.

The GRADE approach results in an assessment of the quality of a body of evidence in one of four grades:

High: we are very confident that the true effect lies close to that of the estimate of the effect.
Moderate: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.
Low: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.
Very low: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of the effect.

Subgroup analysis and investigation of heterogeneity

We planned to conduct the following subgroup analyses if relevant data were available.

Different neurodiagnostic procedures (e.g. brain MRI versus brain computed tomography (CT) versus EEG).
Age group of children (e.g. neonates versus older children).

Sensitivity analysis

We planned to perform sensitivity analyses for the primary outcomes and any secondary outcomes for which sufficient numbers of studies were available by evaluating the change in the effect estimates following the exclusion of studies with a high risk of:

selection bias (high risk for random sequence generation or allocation concealment, or both);
attrition bias (high risk for incomplete outcome data).

Results

Description of studies

Results of the search

We identified 279 records from our search of CENTRAL, MEDLINE and Embase. We identified a further 17 records that appeared to be relevant from our search of ClinicalTrials.gov and the WHO ICTRP. After removing duplicates, 218 records remained, of which 60 articles appeared to be relevant after we inspected the titles. We evaluated the abstracts or full text of the articles, or both, and excluded 45 of the 60 articles. Of the remaining 15 articles, we excluded one, as it was a duplicate publication of an included study, and another article for which we have requested additional information from the authors is awaiting further assessment of its eligibility. We included a total of 13 articles in the review. The flow diagram of the studies from the initial search to the meta‐analysis is shown in Figure 1. Descriptions of all the included studies are provided in the Characteristics of included studies table, and the excluded studies with the reasons for exclusion are given in the Characteristics of excluded studies table.

Included studies

We included 13 RCTs that were conducted in six countries, including Iran (4 studies), Turkey (3 studies), USA (3 studies), and Israel, Chile, and Spain (1 study each). All 13 trials were single‐centre RCTs. The number of children recruited ranged from 40, in D'Agostino 2000, to 582, in Thompson 1982. All studies were conducted on paediatric patients (age up to 18 years old). All studies included children of both sexes.

Five of these trials were sedation trials conducted on neuroimaging studies (2 studies on brain CT, 2 studies on brain MRI, 1 study on brain MRI or CT) (D'Agostino 2000; Fallah 2013; Malviya 2004; Marti‐Bonmati 1995; Thompson 1982); the remaining eight were sedation trials conducted on EEG studies (Ashrafi 2010; Ashrafi 2013; Bektas 2014; Gumus 2015; Loewy 2005; Lopez 1995; Razieh 2013; Sezer 2013).

There were 10 comparisons:

oral chloral hydrate (50 mg/kg or 100 mg/kg) versus oral dexmedetomidine (Gumus 2015);
oral chloral hydrate (75 mg/kg) versus intravenous pentobarbital (Malviya 2004);
oral chloral hydrate (75 mg/kg or 100 mg/kg) versus midazolam (given either orally or intranasally) (Ashrafi 2013; D'Agostino 2000; Fallah 2013);
oral chloral hydrate (50 mg/kg) versus oral melatonin (Ashrafi 2010);
oral chloral hydrate (60 mg/kg) versus music therapy (Loewy 2005);
oral chloral hydrate (50 mg/kg) versus oral hydroxyzine hydrochloride (Sezer 2013);
oral chloral hydrate (70 mg/kg) versus oral promethazine (Razieh 2013);
rectal chloral hydrate (50 mg/kg) versus rectal midazolam (Lopez 1995);
high‐dose oral chloral hydrate (100 mg/kg) versus low‐dose oral chloral hydrate (70 mg/kg) (Marti‐Bonmati 1995);
high‐dose oral chloral hydrate (100 mg/kg) versus low‐dose oral chloral hydrate (50 mg/kg) (Gumus 2015).

The initial sedation dose of chloral hydrate used in the studies ranged from 25 mg/kg to 100 mg/kg oral and 50 mg/kg rectal. The type and doses of the sedation agents used varied among the included studies: oral dexmedetomidine 2 or 3 mg/kg (Gumus 2015); intravenous pentobarbital 5 mg/kg (Malviya 2004); midazolam either 0.5 mg/kg orally (Ashrafi 2013; D'Agostino 2000), 0.2 mg/kg intranasally (Fallah 2013), or rectal midazolam 1 mg/kg (Lopez 1995); oral hydroxyzine hydrochloride 1 mg/kg (Bektas 2014; Sezer 2013); oral promethazine 1 mg/kg (Razieh 2013); and intramuscular atropine/meperidine/promethazine/secobarbital (AMPS) cocktail 0.08 mL/kg (Thompson 1982). Two studies compared sedation of high‐dose (100 mg/kg) chloral hydrate versus low‐dose (50 mg/kg or 70 mg/kg) chloral hydrate orally (Gumus 2015; Marti‐Bonmati 1995). No dose of melatonin was stated (Ashrafi 2010), and dosing/intensity was not stated for music therapy (Loewy 2005).

In general, the included studies assessed outcomes of success of sedation based on the following.

Time to onset to adequate sedation in minutes in eight studies (Fallah 2013; Gumus 2015; Loewy 2005; Lopez 1995; Malviya 2004; Marti‐Bonmati 1995; Razieh 2013; Sezer 2013).
Failure of sedation preventing successful completion of neurodiagnostic investigation in eight studies (D'Agostino 2000; Fallah 2013; Gumus 2015; Loewy 2005; Malviya 2004; Marti‐Bonmati 1995; Razieh 2013; Sezer 2013).
Total sedation/sleep duration in six studies (D'Agostino 2000; Gumus 2015; Loewy 2005; Lopez 1995; Marti‐Bonmati 1995; Sezer 2013).
Sedative adverse events in nine studies (Ashrafi 2010; Ashrafi 2013; D'Agostino 2000; Fallah 2013; Gumus 2015; Malviya 2004; Marti‐Bonmati 1995; Razieh 2013; Sezer 2013).
Uninterpretable neurodiagnostic investigation due to either excessive motion artefact for neuroimaging studies or excessive fast background activity on EEG studies in five studies (Ashrafi 2010; Ashrafi 2013; Lopez 1995; Malviya 2004; Sezer 2013).

Other outcomes assessed included caregiver's satisfaction scale (Razieh 2013), hospital stay post‐sedation (Razieh 2013), and return to normal baseline activity (Malviya 2004).

Excluded studies

We excluded a total of 45 studies based on one or more of the following.

Study design or article type (33 studies): the studies were either retrospective or prospective cohort studies, cross‐over study, prospective non‐randomised intervention studies, literature review articles, a questionnaire study, studies with research questions or outcomes that did not match our review, or commentaries.
Population (5 studies): the participants in the studies were either children undergoing dental procedures, auditory brainstem response tests, or ophthalmic examination.
Intervention (18 studies): the studies either assessed efficacy of chloral hydrate alone or in combination with other sedative agents or efficacy of other sedative agents that were not the intervention of our interest.
Outcome (2 studies): the studies reported effects of sedation on EEG results and did not assess the success of sedation.

A description of each study is provided in the Characteristics of excluded studies table.

Risk of bias in included studies

Risk of bias in the included studies varied widely. There was at least one high‐risk domain in eight out of the 13 included studies (Ashrafi 2010; Ashrafi 2013; Bektas 2014; Gumus 2015; Loewy 2005; Lopez 1995; Malviya 2004; Thompson 1982). We judged all of these eight studies to be at high risk for blinding of participants, except Bektas 2014, which we judged to be at unclear risk. Three studies were at low risk in all domains (Fallah 2013; Lopez 1995; Marti‐Bonmati 1995). The proportions of included studies at low, high, and unclear risk of bias in each domain is illustrated in Figure 2; the 'Risk of bias' judgement of each included study in each domain is depicted in Figure 3. We have also provided a detailed description of the risk of bias of each study in the Characteristics of included studies table. Our 'Risk of bias' assessments under each domain are summarised below.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

For random sequence generation, we judged seven of the 13 included studies to have low risk of bias (D'Agostino 2000; Fallah 2013; Gumus 2015; Lopez 1995; Malviya 2004; Marti‐Bonmati 1995; Razieh 2013). For allocation concealment, four studies had low risk of bias (D'Agostino 2000; Lopez 1995; Marti‐Bonmati 1995; Razieh 2013). In these studies, the authors clearly stated the method of sequence generation, which involved some form of random number scheme using either computers, random number table, or the toss of a coin. There were also clear statements in the methods that reassured the readers of the independence between sequence generation and allocation. We judged two studies to be at high risk of bias for sequence generation as well as allocation concealment (Loewy 2005; Thompson 1982), as children were allocated following a predictable sequence based on the day of the week they were admitted or via alternating and rotational assignments. Four of the 13 included studies had unclear risk of bias in one or both domains due to insufficient information provided in the articles (Ashrafi 2010; Ashrafi 2013; Bektas 2014; Sezer 2013).

Blinding

For blinding of participants, we judged seven studies to have high risk of bias (Ashrafi 2010; Ashrafi 2013; Bektas 2014; Gumus 2015; Loewy 2005; Malviya 2004; Thompson 1982). In three studies (D'Agostino 2000; Fallah 2013; Marti‐Bonmati 1995), it was stated clearly that blinding of participants was achieved, while in the other three studies (Lopez 1995; Razieh 2013; Sezer 2013), no statements were made regarding the blinding status of the participants. We considered blinding for these three studies to be unlikely because they compared chloral hydrate with another solution of a different appearance/preparation and taste.

Seven studies did not report blinding of outcome assessors (Ashrafi 2010; Ashrafi 2013; Bektas 2014; Gumus 2015; Malviya 2004; Sezer 2013; Thompson 1982), and we judged one study to be at high risk of bias, as the modality of sedation intervention given in this study (oral chloral hydrate versus music therapy via music therapist) was clearly different, making blinding of the outcome assessors not possible (Loewy 2005). In the remaining five studies (D'Agostino 2000; Fallah 2013; Lopez 1995; Marti‐Bonmati 1995; Razieh 2013), it was clearly stated that the outcome assessors were blinded.

Incomplete outcome data

In all studies except two (Lopez 1995; Thompson 1982), all children enrolled appeared to be included in the analysis. Both studies with missing data had a dropout rate of over 20% (22% in Lopez 1995 and 42.3% in Thompson 1982), resulting in a judgement of high risk of bias for this domain. We judged a third study as having high risk in this domain, as the authors did not follow an intention‐to‐treat analysis (Bektas 2014).

Selective reporting

We judged 10 studies as at low risk of bias for this domain (Ashrafi 2010; Ashrafi 2013; D'Agostino 2000; Fallah 2013; Gumus 2015; Lopez 1995; Malviya 2004; Marti‐Bonmati 1995; Razieh 2013; Sezer 2013), and three studies at high risk (Bektas 2014; Loewy 2005; Thompson 1982). The studies at high risk of reporting bias used a different categorical outcome sedation score between both comparison groups (reported sleep score 4 for chloral hydrate group and sleep score 3 for music therapy), making it unsuitable for direct comparison (Loewy 2005), and reported outcomes without sufficient details for a meta‐analysis, for example reporting means without standard deviations (Thompson 1982).

Other potential sources of bias

We did not identify other potential major sources of bias in the included studies.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5; Table 6; Table 7

Summary of findings for the main comparison. Chloral hydrate oral (50 mg/kg or 100 mg/kg) compared to dexmedetomidine oral (2 mg/kg or 3 mg/kg) as sedating agents for neurodiagnostic procedures in children.

Chloral hydrate orally (50 mg/kg or 100 mg/kg) compared to dexmedetomidine orally (2 mg/kg or 3 mg/kg) as sedating agents for neurodiagnostic procedures in children
Patient or population: children undergoing neurodiagnostic procedures  Setting: paediatric hospital or outpatient  Intervention: chloral hydrate oral (50 mg/kg or 100 mg/kg)  Comparison: dexmedetomidine oral (2 mg/kg or 3 mg/kg)
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with dexmedetomidine	Risk with chloral hydrate
Proportion of children who successfully completed neurodiagnostic procedure without interruption by the child awakening	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Proportion of children who required a further dose of either the same sedative agent or the addition of a different sedative agent	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Time to adequate sedation (minutes or as measured by specific validated scales such as the Ramsay Sedation Score)	The mean EEG time onset for adequate sedation (minutes) was 35.2.	The mean EEG time onset for adequate sedation (minutes) in the intervention group was 3.86 minutes shorter (5.12 shorter to 2.6 shorter).	‐	160  (1 RCT)	⊕⊕⊕⊝  MODERATE ¹
Proportion of children who had sedation failure or inadequate level of sedation	Study population		RR 1.14  (0.51 to 2.53)	160  (1 RCT)	⊕⊕⊕⊝  MODERATE ²
	119 per 1000	136 per 1000  (61 to 301)
Sedation duration (minutes)	The mean duration of sedation/sleep was 112.1 minutes³	The mean EEG sedation/sleep duration in the intervention group was 16.31 minutes longer (9.15 to 23.46 minutes longer)	‐	160  (1 RCT)	⊕⊕⊕⊝  MODERATE ⁴
Number of children with clinical adverse events (any)	Study population		RR 7.66  (1.78 to 32.91)	160  (1 RCT)	⊕⊕⊝⊝  LOW ⁵
	24 per 1000	182 per 1000  (42 to 784)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: confidence interval; EEG: electroencephalogram; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence   High quality: We are very confident that the true effect lies close to that of the estimate of the effect.  Moderate quality: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low quality: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low quality: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Chloral hydrate orally (75 mg/kg) compared to pentobarbital intravenously (5 mg/kg) as sedating agents for neurodiagnostic procedures in children
Patient or population: children undergoing neurodiagnostic procedures  Setting: paediatric hospital or outpatient  Intervention: chloral hydrate oral (75 mg/kg)  Comparison: pentobarbital intravenous (5 mg/kg)
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with pentobarbital	Risk with chloral hydrate
Proportion of children who successfully completed neurodiagnostic procedure without interruption by the child awakening	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Proportion of children who required a further dose of either the same sedative agent or the addition of a different sedative agent	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Time to adequate sedation (minutes or as measured by specific validated scales such as the Ramsay Sedation Score)	The mean time to adequate sedation was 9 minutes.	The mean time to adequate sedation in the intervention group was 19 minutes longer (16.61 to 21.39 minutes longer).	‐	70  (1 RCT)	⊕⊕⊝⊝  LOW ^{1 2}
Proportion of children who had sedation failure after 1 dose of sedative agent	Study population		RR 4.33  (1.35 to 13.89)	70  (1 RCT)	⊕⊕⊝⊝  LOW ^{1 3}
	86 per 1000	371 per 1000  (116 to 1000)
Proportion of children who had sedation failure after 2 doses of sedative agent (same or different)	Study population		RR 3.00  (0.33 to 27.46)	70  (1 RCT)	⊕⊝⊝⊝  VERY LOW ^{1 3}
	29 per 1000	86 per 1000  (9 to 785)
Non‐interpretable neuroimaging finding	Study population		RR 0.23  (0.03 to 1.94)	54  (1 RCT)	⊕⊝⊝⊝  VERY LOW ^{1 4}
	154 per 1000	35 per 1000  (5 to 298)
Number of children with clinical adverse event: oxygen desaturation	Study population		RR 0.67  (0.21 to 2.16)	70  (1 RCT)	⊕⊝⊝⊝  VERY LOW ^{1 5}
	171 per 1000	115 per 1000  (36 to 370)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence   High quality: We are very confident that the true effect lies close to that of the estimate of the effect.  Moderate quality: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low quality: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low quality: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Chloral hydrate orally (100 mg/kg or 75 mg/kg) compared to midazolam (0.2 mg/kg intranasally or 0.5 mg/kg orally) as sedating agents for neurodiagnostic procedures in children
Patient or population: children undergoing neurodiagnostic procedures  Setting: paediatric hospital or outpatient  Intervention: chloral hydrate oral (100 mg/kg or 75 mg/kg)  Comparison: midazolam (intranasal 0.2 mg/kg or oral 0.5 mg/kg)
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with midazolam	Risk with chloral hydrate
Proportion of children who successfully completed neurodiagnostic procedure without interruption by the child awakening	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Proportion of children who required a further dose of either the same sedative agent or the addition of a different sedative agent	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Time to adequate sedation (minutes or as measured by specific validated scales such as the Ramsay Sedation Score)	The mean time to adequate sedation was 10.92 minutes (intranasal midazolam).	The mean time to adequate sedation in the intervention group was 12.83 minutes longer (7.22 to 18.44 minutes longer).	‐	60  (1 RCT)	⊕⊕⊕⊝  MODERATE ¹
Proportion of children with inadequate level of sedation (Ramsay score below 4)	Study population		RR 0.11  (0.03 to 0.44)	60  (1 RCT)	⊕⊕⊕⊝  MODERATE ¹
	600 per 1000	66 per 1000  (18 to 264)
Proportion of children who had sedation failure after 1 dose of sedative agent	Study population		RR 0.17  (0.02 to 1.12)	33  (1 RCT)	⊕⊕⊝⊝  LOW ²
	545 per 1000	93 per 1000  (11 to 611)
Sedation duration (minutes)	The mean duration of sedation or sleep was 76 minutes (oral midazolam).	The mean duration of sedation or sleep in the intervention group was 19 minutes longer (3.4 minutes shorter to 41.4 minutes longer).	‐	33  (1 RCT)	⊕⊕⊝⊝  LOW ³
EEG sedative‐induced artefact	Study population		RR 0.58  (0.44 to 0.76)	198  (1 RCT)	⊕⊕⊕⊕  HIGH
	700 per 1000	406 per 1000 (308 to 532)
Number of children with clinical adverse events (any)	Study population		RR 0.20  (0.01 to 4.20)	198  (1 RCT)	⊕⊕⊝⊝  LOW ²
	20 per 1000	4 per 1000  (0 to 84)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: confidence interval; EEG: electroencephalogram; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence   High quality: We are very confident that the true effect lies close to that of the estimate of the effect.  Moderate quality: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low quality: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low quality: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Chloral hydrate orally (50 mg/kg) compared to melatonin orally as sedating agents for neurodiagnostic procedures in children
Patient or population: children undergoing neurodiagnostic procedures  Setting: paediatric hospital or outpatient  Intervention: chloral hydrate oral (50 mg/kg)  Comparison: melatonin oral
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with melatonin	Risk with chloral hydrate
Proportion of children who successfully completed neurodiagnostic procedure without interruption by the child awakening	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Proportion of children who required a further dose of either the same sedative agent or the addition of a different sedative agent	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Time to adequate sedation (minutes or as measured by specific validated scales such as the Ramsay Sedation Score)	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Proportion of children who had sedation failure or inadequate level of sedation	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
EEG sedative‐induced artefact	Study population		RR 0.33  (0.14 to 0.82)	348  (1 RCT)	⊕⊕⊝⊝  LOW ^{1 2}
	103 per 1000	34 per 1000  (14 to 85)
Number of children with clinical adverse events (any)	Study population		RR 1.00  (0.25 to 3.93)	348  (1 RCT)	⊕⊕⊝⊝  LOW ^{1 2}
	23 per 1000	23 per 1000  (6 to 90)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: confidence interval; EEG: electroencephalogram; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence   High quality: We are very confident that the true effect lies close to that of the estimate of the effect.  Moderate quality: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low quality: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low quality: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Chloral hydrate orally (50 mg/kg + 50 mg/kg) compared to hydroxyzine hydrochloride orally (1 mg/kg + 1 mg/kg) as sedating agents for neurodiagnostic procedures in children
Patient or population: children undergoing neurodiagnostic procedures  Setting: paediatric hospital or outpatient  Intervention: chloral hydrate oral (50 mg/kg + 50 mg/kg)  Comparison: hydroxyzine hydrochloride oral (1 mg/kg + 1 mg/kg)
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with hydroxyzine hydrochloride	Risk with chloral hydrate
Proportion of children who successfully completed neurodiagnostic procedure without interruption by the child awakening	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Proportion of children who required a further dose of either the same sedative agent or the addition of a different sedative agent	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
TIme to adequate sedation (minutes or as measured by specific validated scales such as the Ramsay Sedation Score)	The time onset to adequate sedation was 23.7 minutes.	The mean time to adequate sedation in the intervention group was 7.5 minutes shorter (7.85 to 7.15 minutes shorter)	‐	282  (1 RCT)	⊕⊕⊕⊝  MODERATE ¹
Proportion of children who had sedation failure or inadequate level of sedation	Study population		RR 0.33  (0.11 to 1.01)	282  (1 RCT)	⊕⊕⊝⊝  LOW ^{1 2}
	85 per 1000	28 per 1000  (9 to 86)
Sedation duration (minutes)	The mean duration of sedation or sleep was 85.1 minutes.	The mean duration of sedation or sleep in the intervention group was 3.1 minutes longer (2.23 to 3.97 minutes longer)	‐	282  (1 RCT)	⊕⊕⊕⊝  MODERATE ¹
Number of children with clinical adverse event: behavioural change	Study population		RR 1.17  (0.40 to 3.38)	282  (1 RCT)	⊕⊕⊝⊝  LOW ^{1 2}
	43 per 1000	50 per 1000  (17 to 144)
Number of children with clinical adverse event: nausea or vomiting	Study population		RR 1.25  (0.34 to 4.56)	282  (1 RCT)	⊕⊕⊝⊝  LOW ^{1 2}
	28 per 1000	35 per 1000  (10 to 129)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence   High quality: We are very confident that the true effect lies close to that of the estimate of the effect.  Moderate quality: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low quality: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low quality: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Chloral hydrate orally (70 mg/kg) compared to promethazine orally (1 mg/kg) as sedating agents for neurodiagnostic procedures in children
Patient or population: children undergoing neurodiagnostic procedures  Setting: paediatric hospital or outpatient  Intervention: chloral hydrate oral (70 mg/kg)  Comparison: promethazine oral (1 mg/kg)
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with promethazine	Risk with chloral hydrate
Proportion of children who successfully completed neurodiagnostic procedure without interruption by the child awakening	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Proportion of children who required a further dose of either the same sedative agent or the addition of a different sedative agent	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Time to adequate sedation (minutes or as measured by specific validated scales such as the Ramsay Sedation Score)	The mean time to adequate sedation was 33.84 minutes.	The mean time to adequate sedation in the intervention group was 12.11 minutes shorter (18.48 to 5.74 minutes shorter).	‐	60  (1 RCT)	⊕⊕⊕⊝  MODERATE ¹
Proportion of children with inadequate level of sedation (Ramsay score below 4)	Study population		RR 0.03  (0.00 to 0.45)	60  (1 RCT)	⊕⊕⊝⊝  LOW ²
	567 per 1000	17 per 1000  (0 to 255)
Proportion of children who had sedation failure or inadequate level of sedation	Study population		RR 0.11  (0.01 to 0.82)	60  (1 RCT)	⊕⊕⊕⊝  MODERATE ¹
	300 per 1000	33 per 1000  (3 to 246)
Number of children with clinical adverse event: behavioural change	Study population		RR 0.20  (0.01 to 4.00)	60  (1 RCT)	⊕⊕⊕⊝  MODERATE ¹
	67 per 1000	13 per 1000  (1 to 267)
Number of children with clinical adverse event: vomiting or nausea	Study population		RR 13.00  (0.76 to 220.96)	60  (1 RCT)	⊕⊕⊝⊝  LOW ³
	0 per 1000	0 per 1000  (0 to 0)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence   High quality: We are very confident that the true effect lies close to that of the estimate of the effect.  Moderate quality: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low quality: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low quality: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Chloral hydrate oral (60 mg/kg) compared to music therapy as sedating agents for neurodiagnostic procedures in children
Patient or population: children undergoing neurodiagnostic procedures  Setting: paediatric inpatient or outpatient  Intervention: chloral hydrate oral (60 mg/kg)  Comparison: music therapy
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with music therapy	Risk with chloral hydrate	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Proportion of children who successfully completed neurodiagnostic procedure without interruption by the child awakening	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Proportion of children who required a further dose of either the same sedative agent or the addition of a different sedative agent	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
Time to adequate sedation (minutes or as measured by specific validated scales such as the Ramsay Sedation Score)	The mean EEG time onset for adequate sedation was 23 minutes.	The mean EEG time onset for adequate sedation in the intervention group was 9 minutes more (2.15 fewer to 20.15 more).	‐	58  (1 RCT)	⊕⊝⊝⊝  VERY LOW ^{1 3}
Proportion of children who had sedation failure or inadequate level of sedation	Study population		RR 17.00  (2.37 to 122.14)	58  (1 RCT)	⊕⊝⊝⊝  VERY LOW ^{1 2}
	29 per 1000	500 per 1000  (70 to 1000)	RR 17.00  (2.37 to 122.14)	58  (1 RCT)	⊕⊝⊝⊝  VERY LOW ^{1 2}
Sedation duration (minutes)	The mean EEG sedation/sleep duration was 66 minutes.	The mean EEG sedation/sleep duration in the intervention group was 160 minutes more (121.07 more to 198.93 more).	‐	58  (1 RCT)	⊕⊝⊝⊝  VERY LOW ^{1 3}
Number of children with clinical adverse events (any)	‐	‐	‐	‐	‐	No study for this comparison assessed this outcome.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: confidence interval; EEG: electroencephalogram; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence   High quality: We are very confident that the true effect lies close to that of the estimate of the effect.  Moderate quality: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.  Low quality: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.  Very low quality: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Comparison	Outcome	Study ID	Effect estimate	P value
Chloral hydrate versus melatonin	Sleep onset latency	Ashrafi 2010	Median (range) in minutes Chloral hydrate group (35 (10 to 150)), melatonin group (45 (5 to 210))	0.113
	Sleep duration	Ashrafi 2010	Median (range) in minutes Chloral hydrate group (60 (15 to 240)), melatonin group (30 (15 to 240))	0.0001
	Drowsiness time	Ashrafi 2010	Median (range) in minutes Chloral hydrate group (60 (0 to 300)), melatonin group (20 (0 to 300))	0.0001
Chloral hydrate versus midazolam	Sleep onset latency	Ashrafi 2013	Median (range) in minutes Chloral hydrate group (32 (20 to 95)), midazolam group (20 (0 to 300))	< 0.001
	Sleep duration	Ashrafi 2013	Median (range) in minutes Chloral hydrate group (66.5 (56 to 98)), midazolam group (25.5 (12 to 38))	< 0.001
	Drowsiness time	Ashrafi 2013	Median (range) in minutes Chloral hydrate group (32 (21 to 36)), midazolam group (6 (2 to 9))	< 0.001
Chloral hydrate versus hydroxyzine	Time of sleep	Bektas 2014	Median (range) in minutes Chloral hydrate group (20 (5 to 120)), hydroxyzine group (30 (5 to 240))	0.309
	Excessive fast activity	Bektas 2014	Percentage Chloral hydrate (24.5%) vs hydroxyzine (11.6%)	0.02
	Success in falling asleep	Bektas 2014	Percentage Chloral hydrate (90.7%) vs hydroxyzine (89.6%)	0.841
Chloral hydrate versus intramuscular AMPS	Time to sedation	Thompson 1982	Mean (minutes) Chloral hydrate (55) vs AMPS (53)	N/A
Chloral hydrate versus intramuscular AMPS	Imaging failure	Thompson 1982	Percentage Chloral hydrate (15%) vs AMPS (12%)	N/A

#1	Search *child[Text Word]**
#2	Search children[MeSH Terms]
#3	Search *infant[Text Word]**
#4	Search *neonat[Text Word]**
#5	Search neonates[MeSH Terms]
#6	Search newborn[Text Word]
#7	Search infant, newborn[MeSH Terms]
#8	Search *paediatric[Text Word]**
#9	Search toddler[Text Word]
#10	Search adolescent[Text Word]
#11	Search adolescent[MeSH Terms]
#12	Search teenager[Text Word]
#13	Search teenager[MeSH Terms]
#14	Search (#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11 OR #12 OR #13)
#15	Search "non‐invasive"[Text Word]
#16	Search "non‐interventional"[Text Word]
#17	Search *investigation[Text Word]**
#18	Search *assessment[Text Word]**
#19	Search *evaluation[Text Word]**
#20	Search (#15 OR #16 OR #17 OR #18 OR #19)
#21	Search *neurodiagnos[Text Word]**
#22	Search brain[Text Word]
#23	Search brain[MeSH Terms]
#24	Search cerebral[Text Word]
#25	Search cerebral[MeSH Terms]
#26	Search neuroimaging[Text Word]
#27	Search neuroimaging[MeSH Terms]
#28	Search nuclear medicine[Text Word]
#29	Search nuclear medicine[MeSH Terms]
#30	Search electroencephalography[Text Word]
#31	Search electroencephalography[MeSH Terms]
#32	Search "EEG"[Text Word]
#33	Search EEG[MeSH Terms]
#34	Search magnetic resonance imaging[Text Word]
#35	Search magnetic resonance imaging[MeSH Terms]
#36	Search "MRI"[Text Word]
#37	Search MRI[MeSH Terms]
#38	Search Computed tomography[Text Word]
#39	Search Computed tomography[MeSH Terms]
#40	Search Single‐photon emission computed tomography[Text Word]
#41	Search Single‐photon emission computed tomography[MeSH Terms]
#42	Search "SPECT"[Text Word]
#43	Search SPECT[MeSH Terms]
#44	Search Positron emission tomography[Text Word]
#45	Search Positron emission tomography[MeSH Terms]
#46	Search nerve conduction study[Text Word]
#47	Search nerve conduction[MeSH Terms]
#48	Search (#21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29 OR #30 OR #31 OR #32 OR #33 OR #34 OR #35 OR #36 OR #37 or #38 OR #39 OR #40 OR #41 OR #42 OR #43 OR #44 OR #45 OR #46 OR #47)
#49	Search Chloral hydrate[Text Word]
#50	Search Chloral hydrate[MeSH Terms]
#51	Search Trichloroacetaldehyde monohydrate[Text Word]
#52	Search Trichloroacetaldehyde monohydrate[MeSH Terms]
#53	Search Noctec[Text Word]
#54	Search Somnos[Text Word]
#55	Search (#49 OR #50 OR #51 OR #52 OR #53 OR #54)
#56	Search (#14 AND (#20 OR #48) AND #55)
#57	Search "randomized controlled trial"[Publication Type]
#58	Search "controlled clinical trial"[Publication Type]
#59	Search (randomised[Title/Abstract]) OR randomized[Title/Abstract]
#60	Search placebo[Title/Abstract]
#61	Search clinical trials[MeSH Major Topic]
#62	Search randomly[Title/Abstract]
#63	Search trial[Title]
#64	Search (#57 OR #58 OR #59 OR #60 OR #61 OR #62 OR #63)
#65	Search (animals [mh] NOT humans [mh])
#66	Search (#64 NOT #65)
#67	Search (#56 AND #66)

#1	child*:ti,ab,kw
#2	MeSH descriptor: [Child] explode all trees
#3	infant*:ti,ab,kw
#4	neonat*:ti,ab,kw
#5	MeSH descriptor: [Infant, Newborn] explode all trees
#6	"newborn":ti,ab,kw
#7	paediatric*:ti,ab,kw
#8	"toddler":ti,ab,kw
#9	adolescent:ti,ab,kw
#10	MeSH descriptor: [Adolescent] explode all trees
#11	"teenager":ti,ab,kw
#12	(#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11)
#13	"non‐invasive":ti,ab,kw
#14	"non‐interventional":ti,ab,kw
#15	investigation*:ti,ab,kw
#16	assessment*:ti,ab,kw
#17	evaluation*:ti,ab,kw
#18	#13 Or #14 OR #15 OR #16 OR #17
#19	neurodiagnos*:ti,ab,kw
#20	"brain":ti,ab,kw
#21	MeSH descriptor: [Brain] explode all trees
#22	cerebral:ti,ab,kw
#23	neuroimaging:ti,ab,kw
#24	MeSH descriptor: [Neuroimaging] explode all trees
#25	"nuclear medicine":ti,ab,kw
#26	MeSH descriptor: [Nuclear Medicine] explode all trees
#27	electroencephalography:ti,ab,kw
#28	MeSH descriptor: [Electroencephalography] explode all trees
#29	EEG:ti,ab,kw
#30	"magnetic resonance imaging":ti,ab,kw
#31	MeSH descriptor: [Magnetic Resonance Imaging] explode all trees
#32	MRI:ti,ab,kw
#33	"computed tomography":ti,ab,kw
#34	MeSH descriptor: [Tomography, X‐Ray Computed] explode all trees
#35	Single‐photon emission computed tomography:ti,ab,kw
#36	MeSH descriptor: [Tomography, Emission‐Computed, Single‐Photon] explode all trees
#37	SPECT:ti,ab,kw
#38	Positron emission tomography:ti,ab,kw
#39	MeSH descriptor: [Positron‐Emission Tomography] explode all trees
#40	nerve conduction study:ti,ab,kw
#41	MeSH descriptor: [Neural Conduction] explode all trees
#42	#19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29 OR #30 OR #31 OR #32 OR #33 OR #34 OR #35 OR #36 OR #37 or #38 OR #39 OR #40 OR #41
#43	Chloral hydrate:ti,ab,kw
#44	MeSH descriptor: [Chloral Hydrate] explode all trees
#45	Trichloroacetaldehyde monohydrate:ti,ab,kw
#46	Noctec:ti,ab,kw
#47	Somnos:ti,ab,kw
#48	#43 OR #44 OR #45 OR #46 OR #47
#49	#12 and (#18 or #42) and #48 in trials

#1	child*:ab,ti
#2	"children"/exp
#3	infant*:ab,ti
#4	neonat*:ab,ti
#5	newborn:ab,ti
#6	"infant, newborn"/exp
#7	paediatric*:ab,ti
#8	toddler: ab,ti
#9	adolescent:ab,ti
#10	"adolescent"/exp
#11	teenager:ab,ti
#12	#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11
#13	'non‐invasive':ab,ti
#14	'non‐interventional':ab,ti
#15	investigation*:ab,ti
#16	assessment*:ab,ti
#17	evaluation*:ab,ti
#18	#13 OR #14 OR #15 OR #16 OR #17
#19	neurodiagnos*:ab,ti
#20	brain:ab,ti
#21	"brain"/exp
#22	cerebral:ab,ti
#23	"cerebral"/exp
#24	neuroimaging:ab,ti
#25	"neuroimaging"/exp
#26	nuclear medicine:ab,ti
#27	"nuclear medicine"/exp
#28	electroencephalography:ab,ti
#29	"electroencephalography"/exp
#30	"EEG":ab,ti
#31	magnetic resonance imaging:ab,ti
#32	"magnetic resonance imaging"/exp
#33	"MRI":ab,ti
#34	computed tomography:ab,ti
#35	"computed tomography"/exp
#36	Single‐photon emission computed tomography:ab,ti
#37	"Single‐photon emission computed tomography"/exp
#38	"SPECT":ab,ti
#39	Positron emission tomography:ab,ti
#40	"Positron emission tomography"/exp
#41	nerve conduction study:ab,ti
#42	"nerve conduction"/exp
#43	#19 OR #20 Or #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 OR #29 OR #30 OR #31 OR #32 OR #33 OR #34 OR #35 OR #36 OR #37 or #38 OR #39 OR #40 OR #41 OR #42
#44	'Chloral hydrate':ab,ti
#45	"Chloral hydrate"/exp
#46	'Trichloroacetaldehyde monohydrate':ab,ti
#47	'Noctec':ab,ti
#48	'Somnos':ab,ti
#49	#44 OR #45 OR #46 OR #47 OR #48
#50	'randomized controlled trial'/exp
#51	'randomization'/exp
#52	'controlled study'/exp
#53	'multicenter study'/exp
#54	'double blind procedure'/exp
#55	'single blind procedure'/exp
#56	random* OR cross?over* OR factorial* OR placebo* OR volunteer*:ab,ti
#57	(singl* OR doubl* OR trebl* OR tripl) NEAR (blind:ab,ti OR mask*:ab,ti)
#58	#50 OR #51 OR #52 OR #53 OR #54 OR #55 OR #56 OR #57
#59	#12 AND (#18 OR #43) AND #49 AND #58

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 EEG time onset for adequate sedation (minutes)	1	160	Mean Difference (IV, Fixed, 95% CI)	‐3.86 [‐5.12, ‐2.60]
1.1 high dose (100 mg/kg chloral hydrate vs 3ug/kg dexmetodimidine	1	82	Mean Difference (IV, Fixed, 95% CI)	‐5.60 [‐7.33, ‐3.87]
1.2 low dose (50mg/kg chloral hydrate vs 2ug/kg dexmetodimidine	1	78	Mean Difference (IV, Fixed, 95% CI)	‐1.90 [‐3.74, ‐0.06]
2 EEG sedation failure	1	160	Risk Ratio (M‐H, Fixed, 95% CI)	1.14 [0.51, 2.53]
2.1 high dose	1	82	Risk Ratio (M‐H, Fixed, 95% CI)	0.70 [0.12, 3.97]
2.2 low dose	1	78	Risk Ratio (M‐H, Fixed, 95% CI)	1.33 [0.54, 3.32]
3 EEG sedation / sleep duration (minutes)	1	164	Mean Difference (IV, Fixed, 95% CI)	16.31 [9.15, 23.46]
3.1 high dose	1	82	Mean Difference (IV, Fixed, 95% CI)	21.70 [11.76, 31.64]
3.2 low dose	1	82	Mean Difference (IV, Fixed, 95% CI)	10.5 [0.19, 20.81]
4 EEG sedation adverse event: total	1	160	Risk Ratio (M‐H, Fixed, 95% CI)	7.66 [1.78, 32.91]
4.1 high dose	1	82	Risk Ratio (M‐H, Fixed, 95% CI)	10.5 [1.41, 78.33]
4.2 low dose	1	78	Risk Ratio (M‐H, Fixed, 95% CI)	4.67 [0.55, 39.89]
5 EEG sedation adverse event: hypotension	1	160	Risk Ratio (M‐H, Fixed, 95% CI)	0.35 [0.01, 8.34]
5.1 high dose	1	82	Risk Ratio (M‐H, Fixed, 95% CI)	0.35 [0.01, 8.34]
5.2 low dose	1	78	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 EEG sedation adverse event: bradycardia	1	160	Risk Ratio (M‐H, Fixed, 95% CI)	0.39 [0.02, 9.23]
6.1 high dose	1	82	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.2 low dose	1	78	Risk Ratio (M‐H, Fixed, 95% CI)	0.39 [0.02, 9.23]
7 EEG sedation adverse event: behavioural change	1	160	Risk Ratio (M‐H, Fixed, 95% CI)	5.24 [0.26, 105.97]
7.1 high dose	1	82	Risk Ratio (M‐H, Fixed, 95% CI)	5.24 [0.26, 105.97]
7.2 low dose	1	78	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
8 EEG sedation adverse event: nausea or vomiting	1	160	Risk Ratio (M‐H, Fixed, 95% CI)	12.04 [1.58, 91.96]
8.1 high dose	1	82	Risk Ratio (M‐H, Fixed, 95% CI)	15.73 [0.93, 266.73]
8.2 low dose	1	78	Risk Ratio (M‐H, Fixed, 95% CI)	8.14 [0.43, 152.41]
9 EEG sedation adverse event: oxygen desaturation	1	160	Risk Ratio (M‐H, Fixed, 95% CI)	3.31 [0.35, 31.16]
9.1 high dose	1	82	Risk Ratio (M‐H, Fixed, 95% CI)	3.15 [0.13, 75.05]
9.2 low dose	1	78	Risk Ratio (M‐H, Fixed, 95% CI)	3.49 [0.15, 83.03]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Neuroimaging time onset for adequate sedation (minutes)	1	70	Mean Difference (IV, Fixed, 95% CI)	19.0 [16.61, 21.39]
2 Neuroimaging sedation failure after 2 administrations of sedative agent (same or different)	1	70	Risk Ratio (M‐H, Fixed, 95% CI)	3.0 [0.33, 27.46]
3 Neuroimaging sedation failure after 1 administration of sedative agent	1	70	Risk Ratio (M‐H, Fixed, 95% CI)	4.33 [1.35, 13.89]
4 Neuroimaging uninterpretable	1	54	Risk Ratio (M‐H, Fixed, 95% CI)	0.23 [0.03, 1.94]
5 Neuroimaging sedation adverse event: oxygen desaturation	1	70	Risk Ratio (M‐H, Fixed, 95% CI)	0.67 [0.21, 2.16]
6 Neuroimaging sedation adverse event: nausea or vomiting	1	70	Risk Ratio (M‐H, Fixed, 95% CI)	6.0 [0.76, 47.29]
7 Neuroimaging sedation adverse event: paradoxical reaction	1	70	Risk Ratio (M‐H, Fixed, 95% CI)	0.09 [0.01, 1.58]
8 Neuroimaging sedation adverse event: return to baseline activity postdischarge	1	70	Mean Difference (IV, Fixed, 95% CI)	‐6.0 [‐11.43, ‐0.57]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 EEG time onset for adequate sedation (minutes)	1	58	Mean Difference (IV, Fixed, 95% CI)	9.0 [‐2.15, 20.15]
2 EEG sedation failure	1	58	Risk Ratio (M‐H, Fixed, 95% CI)	17.0 [2.37, 122.14]
3 EEG sedation / sleep duration (minutes)	1	58	Mean Difference (IV, Fixed, 95% CI)	160.0 [121.07, 198.93]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 EEG time onset for adequate sedation (minutes)	1	59	Mean Difference (IV, Fixed, 95% CI)	‐95.7 [‐114.51, ‐76.89]
2 EEG sedation/ sleep duration (minutes)	1	59	Mean Difference (IV, Fixed, 95% CI)	15.1 [3.35, 26.85]
3 EEG sedative‐induced artefact	1	53	Risk Ratio (M‐H, Fixed, 95% CI)	1.25 [0.73, 2.12]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Neuroimaging time onset for adequate sedation (minutes)	1	60	Mean Difference (IV, Fixed, 95% CI)	12.83 [7.22, 18.44]
2 Neuroimaging inadequate level of sedation achieved (Ramsay score 4)	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.11 [0.03, 0.44]
3 Neuroimaging sedation failure after 1 administration of sedative agent	1	33	Risk Ratio (M‐H, Fixed, 95% CI)	0.17 [0.02, 1.12]
4 Neuroimaging sedation / sleep duration (minutes)	1	33	Mean Difference (IV, Fixed, 95% CI)	19.0 [‐3.40, 41.40]
5 EEG sedative‐induced artefact	1	198	Risk Ratio (M‐H, Fixed, 95% CI)	0.58 [0.44, 0.76]
6 EEG sedation adverse event: total	1	198	Risk Ratio (M‐H, Fixed, 95% CI)	0.20 [0.01, 4.20]
7 Neuroimaging adverse event: behavioural change	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 7.87]
8 Neuroimaging adverse event: vomiting	2	93	Risk Ratio (M‐H, Fixed, 95% CI)	5.29 [0.84, 33.14]
9 Neuroimaging sedation failure with intranasal midazolam	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.39 [0.19, 0.79]
10 Neuroimaging sedation failure with oral midazolam	1	33	Risk Ratio (M‐H, Fixed, 95% CI)	0.08 [0.01, 1.30]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 EEG sedative‐induced artefact	1	348	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.14, 0.82]
2 EEG sedation adverse event: total	1	348	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.25, 3.93]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 EEG time onset for adequate sedation (minutes)	1	282	Mean Difference (IV, Fixed, 95% CI)	‐7.5 [‐7.85, ‐7.15]
2 EEG sedation failure	1	282	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.11, 1.01]
3 EEG sedation / sleep duration (minutes)	1	282	Mean Difference (IV, Fixed, 95% CI)	3.10 [2.23, 3.97]
4 EEG sedative‐induced artefact	1	282	Risk Ratio (M‐H, Fixed, 95% CI)	1.33 [0.47, 3.74]
5 EEG sedation adverse event: behavioural change	1	282	Risk Ratio (M‐H, Fixed, 95% CI)	1.17 [0.40, 3.38]
6 EEG sedation adverse event: nausea or vomiting	1	282	Risk Ratio (M‐H, Fixed, 95% CI)	1.25 [0.34, 4.56]
7 EEG failure after 1 administration of sedative agent	1	282	Risk Ratio (M‐H, Fixed, 95% CI)	0.5 [0.18, 1.43]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 EEG time for adequate sedation (minutes)	1	60	Mean Difference (IV, Fixed, 95% CI)	‐12.11 [‐18.48, ‐5.74]
2 EEG inadequate level of EEG sedation achieved (Ramsay score 4)	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.03 [0.00, 0.45]
3 EEG sedation failure	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.11 [0.01, 0.82]
4 EEG sedation adverse event: behavioural change	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.2 [0.01, 4.00]
5 EEG sedation adverse event: vomiting or nausea	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	13.0 [0.76, 220.96]
6 EEG Ramsay Sedation Score after 1 administration of sedative agent	1	60	Mean Difference (IV, Fixed, 95% CI)	1.53 [1.00, 2.06]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Neuroimaging time onset for adequate sedation (minutes)	1	97	Mean Difference (IV, Fixed, 95% CI)	‐7.0 [‐7.62, ‐6.38]
2 Neuroimaging sedation failure after 1 administration of sedative agent	1	97	Risk Ratio (M‐H, Fixed, 95% CI)	0.46 [0.19, 1.09]
3 Neuroimaging sedation / sleep duration (minutes)	1	97	Mean Difference (IV, Fixed, 95% CI)	8.0 [5.81, 10.19]
4 Neuroimaging sedation adverse event: total	1	97	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.49, 2.32]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 EEG time onset for adequate sedation (minutes)	1	76	Mean Difference (IV, Fixed, 95% CI)	‐5.10 [‐7.05, ‐3.15]
2 EEG sedation failure	1	76	Risk Ratio (M‐H, Fixed, 95% CI)	0.23 [0.05, 0.99]
3 EEG sedation/sleep duration (minutes)	1	76	Mean Difference (IV, Fixed, 95% CI)	17.80 [8.50, 27.10]
4 EEG sedation adverse event: total	1	76	Risk Ratio (M‐H, Fixed, 95% CI)	2.25 [0.77, 6.55]
5 EEG sedation adverse event: behavioural change	1	76	Risk Ratio (M‐H, Fixed, 95% CI)	4.51 [0.22, 90.96]
6 EEG sedation adverse event: nausea or vomiting	1	76	Risk Ratio (M‐H, Fixed, 95% CI)	2.1 [0.59, 7.52]
7 EEG sedation adverse event: oxygen desaturation	1	76	Risk Ratio (M‐H, Fixed, 95% CI)	0.9 [0.06, 13.87]

Methods	Single‐centre RCT (Iran)
Participants	All patients aged 1 to 72 months that were unco‐operative with the EEG setup or referred to our electrodiagnostic department for sleep EEG recording were enrolled. A total of 348 children (male‐to‐female ratio of 1.3:1) were enrolled, 174 children in each group of chloral hydrate (1 to 72 months of age) and melatonin (2 to 64 months of age).
Interventions	2‐arm comparison: chloral hydrate 5% (1 mL/kg; 174 children) melatonin (2 to 6 mg orally; 174 children) 0.5 to 1 h before EEG performance
Outcomes	sleep onset latency sleep duration drowsiness time adverse events
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Materials and methods: “Patients were randomly divided in two groups of melatonin and chloral hydrate for sedation.” The authors stated that patients were randomised but no further detail on the methods of randomisation was described.
Allocation concealment (selection bias)	Unclear risk	As above.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	It was not stated whether the participants and personnel were blinded to the allocation. However blinding appeared very unlikely, as melatonin and chloral hydrate differed in appearance and taste. As the data collected included neurological diagnosis, sleep onset latency, sleep duration, drowsiness time and adverse drug events, which included outcomes that required subjective assessment, non‐blinding of the personnel could have influenced the care of the participants and the outcomes.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	It was unclear whether the 2 neurologists who interpreted the EEG and the EEG technicians who recorded the rest of the outcome data were blinded to the allocation.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Although the number of withdrawals or participants with missing data were not directly stated, it appeared that all 348 children (174 in each group) who were initially randomised were analysed, as calculated from the results section based on the number of EEGs obtained.
Selective reporting (reporting bias)	Low risk	The pre‐specified outcomes of sleep onset latency, sleep onset latency, sleep duration, drowsiness time and adverse drug events were reported in the results. An additional outcome of EEG yield, or the number of abnormal EEGs, which was specified in our review, were also reported. However, the data for the continuous outcomes of sleep onset latency, sleep duration and drowsiness time were skewed, and they were reported in median and range and were unsuitable to be included in our meta‐analysis.
Other bias	Low risk	None identified.

Methods	Single‐centre RCT (Turkey)
Participants	341 children (mean age: 60.92 ± 53.81 months, 194 male and 147 female) that were unco‐operative with the EEG setup or referred for sleep EEG were enrolled. They were randomly divided in 2 groups of hydroxyzine and chloral hydrate.
Interventions	2‐arm comparison: chloral hydrate administered orally as a suspension (mean dosage 26.38 ± 14.73 mg/kg; 147 children) hydroxyzine administered orally as a solution (mean dosage 1.43 ± 0.74 mg/kg; 112 children) If first drug failed, the other drug (hydroxyzine or chloral hydrate) was given (28 children received a combination of chloral hydrate and hydroxyzine).
Outcomes	frequency of EEG background rhythm, sleep onset latency the amplitude of the background rhythm, epileptic abnormalities (focal epileptic discharges, generalised epileptic discharges, multifocal epileptic discharges), the presence of fast activity, amplitude and frequency of sleep spindles duration of sleep Number of children with successful sedation (spontaneous sleep) side effects
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Methods: "The patients who could not sleep spontaneously were randomly divided in two groups of hydroxyzine and chloral hydrate taking into account age, diagnosis, and mental retardation." The authors stated that patients were randomised but no further detail on the methods of randomisation was described.
Allocation concealment (selection bias)	Unclear risk	As above.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	It was not stated whether the participants and personnel were blinded to the allocation. However blinding appeared very unlikely, as hydroxyzine and chloral hydrate differed in appearance and taste. Furthermore, those that failed the first drug were given the second. As the data collected included neurological diagnosis, sleep onset latency, sleep duration, drowsiness time and adverse drug events, which included outcomes that required subjective assessment, non‐blinding of the personnel could have influenced the care of the participants and the outcomes.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	It was unclear whether the EEG technicians who recorded the side effects were blinded to the allocation. It was also unclear who recorded the number of successful sedation and time interval to go to sleep.
Incomplete outcome data (attrition bias)   All outcomes	High risk	Although the number of withdrawals or participants with missing data were not directly stated, it appeared that all children who were initially randomised were analysed, as calculated from the results section based on the number of EEGs obtained. However, the authors did not follow intention to treat analysis as they put those 28 in a separate group (who failed Chloral hydrate or hydroxyzine and received 2nd sedative agent were grouped into chloral hydrate and hydroxyzine group).
Selective reporting (reporting bias)	High risk	The pre‐specified outcomes of frequency, sleep onset latency, EEG changes ( the amplitude of the background rhythm, epileptic abnormalities), sleep duration and adverse drug events were reported in the results. An additional outcome of EEG yield, or the number of abnormal EEGs, which was specified in our review, were also reported. However, the data for the continuous outcomes of time of sleep were skewed, and they were reported in median and range and were unsuitable to be included in our meta‐analysis.
Other bias	Low risk	None identified

Methods	Double‐blinded, single‐centre RCT (USA)
Participants	Children were enrolled in an outpatient neuroimaging study. Eligible were children between 2 months and 8 years of age. 40 children enrolled in the study, 33 completed the protocol.
Interventions	2‐arm comparison chloral hydrate (75 mg/kg, maximum 2 g; 11 children) midazolam (0.5 mg/kg, maximum 10 mg; 22 children) Identically appearing, cherry‐flavoured liquids If inadequately sedated after 30 min, received a supplementary dose of the same medication at 50% of the original dosage.
Outcomes	the principal outcome measurement was the ability to induce sufficient sedation to perform the intended neuroimaging study a secondary efficacy outcome was the proportion of children who required supplementary medication side effects
Notes	Due to an unexpectedly high sedation failure rate, an interim analysis was performed after the first 40 children were enrolled, and the study was terminated as a result of the findings.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Methods: "A randomized sequence of 100 total subjects with 50 in each group was generated using a random number table." Randomisation method was explained and valid.
Allocation concealment (selection bias)	Unclear risk	As above, no further description on the person performing the randomisation to enable an assessment on whether random sequence was generated independently from allocation.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Methods: "Children were administered freshly prepared, identically appearing, cherry flavored liquids in body weight equivalent volumes...... Neither the patient nor any of the investigators were aware of the active component given to individual patients." Blinding of participants were well described.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Neither the patient nor any of the investigators were aware of the active component given to individual patients.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Results: “Randomized children who did not complete the protocol included one with respiratory distress, one who ate a full meal prior to intended drug administration, one who fell asleep after intravenous line placement and four patients who cancelled their appointments after randomization.” Due to an unexpectedly high sedation failure rate, an interim analysis was performed after the first 40 patients were enrolled. The study was terminated early as a result of the findings showing high failure rate in one arm.
Selective reporting (reporting bias)	Low risk	The pre‐specified outcomes of the ability to induce sufficient sedation , duration of sedation, maximum change in anxiety scores, proportion of patients who required supplementary medication and side effect were reported in the results.
Other bias	Low risk	None identified

Methods	Single‐blinded, single‐centre RCT (Iran)
Participants	Children aged 1 to 10 years, referred to CT centre for elective brain CT scan. These children were in ASA class 1 or 2. 60 children were recruited.
Interventions	2‐arm comparison: 100 mg/kg oral chloral hydrate with 1 mL of intranasal normal saline as placebo (30 children) 0.2 mg/kg intranasal midazolam with oral normal saline as placebo (30 children)
Outcomes	The primary outcomes were efficacy in adequate sedation and completing of CT scan.  Secondary outcomes included clinical side effects.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The trial used computer generated equal randomization and allocation ratio was 1:1 for the two groups. Randomisation and blinding was done by an investigator with no clinical involvement in the trial. Data collectors, outcome assessors and data analysts were all kept blinded to the allocation.” The stated method of randomisation was use of a computer‐generated equal randomisation for the 2 groups.
Allocation concealment (selection bias)	Low risk	“The trial used computer generated equal randomization and allocation ratio was 1:1 for the two groups. Randomisation and blinding was done by an investigator with no clinical involvement in the trial. Data collectors, outcome assessors and data analysts were all kept blinded to the allocation.”
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	“Randomisation and blinding was done by an investigator with no clinical involvement in the trial. Data collectors, outcome assessors and data analysts were all kept blinded to the allocation.” “The children were randomized to receive either single dose of 100 mg/kg oral chloral hydrate with one millilitre of intranasal normal saline as placebo (Group I) or 0.2 mg/kg intranasal midazolam with oral normal saline as placebo (Group II).” The participants and personnel were blinded to the appearance of the medications, as both were served (either one is placebo). Even though intranasal normal saline and oral normal saline (as placebo) taste differently from the medications, children would not be able to tell what the actual medications they received were. The investigators were not involved in the trial.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	“The trial used computer generated equal randomization and allocation ratio was 1:1 for the two groups. Randomisation and blinding was done by an investigator with no clinical involvement in the trial. Data collectors, outcome assessors and data analysts were all kept blinded to the allocation.” All the personnel involved in the assessment were blinded.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Even though the author did not report any withdrawals or missing data, it appeared that all of the 60 children who were initially randomised were analysed, as calculated from the results section based on the number of CT brain obtained.
Selective reporting (reporting bias)	Low risk	"The primary outcomes were efficacy in adequate sedation and completing of CT scan. Secondary outcomes included clinical side effects, serious adverse events.” The author reported data on rate of successful CT brain and Ramsay Sedation Score for the primary outcomes. In addition, time from drug administration to adequately sedated, time after taking the drug to completing CT scan, caregiver's satisfaction scale, and total stay time in CT centre were also reported.
Other bias	Low risk	None identified.

PERMALINK

Chloral hydrate as a sedating agent for neurodiagnostic procedures in children

Choong Yi Fong

Chee Geap Tay

Lai Choo Ong

Nai Ming Lai

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Quality of evidence

Prespecified primary outcomes

Prespecified secondary outcomes

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

1.

Included studies

Excluded studies

Risk of bias in included studies

2.

3.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Summary of findings for the main comparison. Chloral hydrate oral (50 mg/kg or 100 mg/kg) compared to dexmedetomidine oral (2 mg/kg or 3 mg/kg) as sedating agents for neurodiagnostic procedures in children.

Summary of findings 2. Chloral hydrate oral (75 mg/kg) compared to pentobarbital intravenous (5 mg/kg) as sedating agents for neurodiagnostic procedures in children.

Summary of findings 3. Chloral hydrate oral (100 mg/kg or 75 mg/kg) compared to midazolam (intranasal 0.2 mg/kg or oral 0.5 mg/kg) as sedating agents for neurodiagnostic procedures in children.

Summary of findings 4. Chloral hydrate oral (50 mg/kg) compared to melatonin oral as sedating agents for neurodiagnostic procedures in children.

Summary of findings 5. Chloral hydrate oral (50 mg/kg + 50 mg/kg) compared to hydroxyzine hydrochloride oral (1 mg/kg + 1 mg/kg) as sedating agents for neurodiagnostic procedures in children.

Summary of findings 6. Chloral hydrate oral (70 mg/kg) compared to promethazine oral (1 mg/kg) as sedating agents for neurodiagnostic procedures in children.

Summary of findings 7. Chloral hydrate oral (60 mg/kg) compared to music therapy as sedating agents for neurodiagnostic procedures in children.

1. Outcomes with skewed data as reported by individual studies.

Comparison 1: Chloral hydrate versus dexmedetomidine

Primary outcomes

Proportion of children who successfully completed neurodiagnostic procedure without interruption by the child awakening

Proportion of children who required a further dose of either the same sedative agent or the addition of a different sedative agent

Time to adequate sedation (minutes or as measured by specific validated scales such as the Ramsay Sedation Score)

1.1. Analysis.

Secondary outcomes

Proportion of children who had sedation failure or inadequate level of sedation

1.2. Analysis.

Methods	Single‐centre RCT (Israel)
Participants	58 children from a paediatric inpatient unit who underwent EEG procedure over a 4‐year period
Interventions	2‐arm comparison of success of sedation between: chloral hydrate administered orally at 60 mg/kg (up to a maximum of 1.5 g; 34 children) music therapy whereby child was played and sung soothing music according to caregiver's request (24 children)
Outcomes	child's level of sleep/sedation during procedure using Beth Israel Medical Center sedation scale time to achieve sleep/sedation from the onset of the intervention length of sleep/sedation from beginning of the EEG until awakening
Notes	Quasi‐RCT whereby the children were identified and assigned to 1 of 2 treatment groups, chloral hydrate or music therapy, based on the day of the week they were admitted. Children recruited on Mondays received chloral hydrate, children recruited on Tuesdays received music therapy.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Method: “The subjects were identified and assigned to one of 2 treatment groups, chloral hydrate or music therapy, based on the day of the week they were admitted. Subjects recruited on Mondays received chloral hydrate, and subjects recruited on Tuesdays received music therapy.” Quasi‐randomisation was performed according to the day children were recruited, which is predictable. In addition, the allocation of the 58 children to these 2 groups was not balanced, with 34 in the music therapy group and 24 in the chloral hydrate group. Data were also skewed with the mean age of children assigned to music therapy group (mean 2.44) being lower than the mean age assigned to the chloral hydrate group (mean 3.21), the difference showing a tendency towards statistical significance (P = 0.53).
Allocation concealment (selection bias)	High risk	Method: “The subjects were identified and assigned to one of 2 treatment groups, chloral hydrate or music therapy, based on the day of the week they were admitted. Subjects recruited on Mondays received chloral hydrate, and subjects recruited on Tuesdays received music therapy.” As stated above; allocation follows a predictable sequence (according to day child was recruited).
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Blinding was not possible, as participants received completely different modality of intervention (either chloral hydrate or music therapy).
Blinding of outcome assessment (detection bias)   All outcomes	High risk	The authors did not state whether the outcome assessors were blinded to the allocation, however blinding seems unlikely, as the modality of intervention was completely different (chloral hydrate given orally or music therapy with music therapist). As the outcomes collected included child’s level of sleep/sedation during procedure, time to achieve sleep/sedation, and length of sleep/sedation, some of which required a subjective assessment, non‐blinding of personnel could have influenced the care of the participants and outcomes. Medical residents or music therapy interns who recorded the outcome data were not blinded to the allocation, as they collected the data before, during, and after the EEG procedure, which would have included witnessing the music therapy taking place.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Results: “Of the total, 2 children (1 from each group: music therapy and chloral hydrate) were not able to go through the EEG on the day of the test and were not included in the analyses. One patient was using a medication that interacted with chloral hydrate and the other patient cancelled and rescheduled due to the parent’s request to be present during the test.” The missing outcome data were equal across both groups and a small number (1 = 1.7% for each group). However, although not stated, it is possible that the 1 participant who cancelled and rescheduled could be related to the type of sedation assigned for the EEG procedure. It is unlikely that the missing data meaningfully changed the outcome of the study.
Selective reporting (reporting bias)	High risk	The prespecified outcomes of length of sleep/sedation, time to achieve sleep/sedation, and level of sleep/sedation were reported in the results. The level of sleep/sedation was reported as a categorical outcome (scale of score 0 to 5) in medication, however the outcome was reported inappropriately whereby the author reported sleep score 4 for chloral hydrate and sleep score 3 for music therapy, making it unsuitable for direct comparison.
Other bias	Low risk	None identified.

Methods	Randomised study (Chile)
Participants	Children aged 1 to 5 years, sent to “Servicio de Neuropsiquiatrfa Infantil, Hospital Clinico San Borja‐Arriara'n” for EEG from June to December 1993, excluding those who were treated with barbiturates or benzodiazepines. 92 children were recruited.
Interventions	Sedation group assignment was based on a tossed coin to receive either rectal chloral hydrate (50 mg/kg; 32 children) or rectal midazolam (1 mg/kg; 27 children), with another control group of 33 children.
Outcomes	time to achieve sleep/sedation from the onset of the intervention sleep duration EEG artefact due to sedative agent
Notes	Article was in Spanish, and assessment was performed with English translation.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The assignment was based on coin tossing for sedation group (rectal chloral hydrate vs rectal midazolam).
Allocation concealment (selection bias)	Unclear risk	“To the children who were not sleepy or sleeping at the time of launch the examination nor had contraindications to the sedation, were administered 50 mg /kg of chloral hydrate in solution 5% or midazolam at 5 mg/ml dose parenteral solution 1 mg/kg, both rectally. Due to the small volume of midazolam it was diluted with 3 ml of solution of NaCl 0.9%, to avoid that it stays in the probe.” It is unclear whether allocation was concealed.
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	“the EEG tracings were analyzed by two medical electro‐encephalographers (EM, LT) who did not know what sedatives were administered and who were asked to review possible base alterations attributable to a sedative (impregnation), and to what medication it was attributed to” It is unclear whether or not participants and personnel were blinded.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	“the EEG tracings were analyzed by two medical electro‐encephalographers (EM, LT) who did not know what sedatives were administered and who were asked to review possible base alterations attributable to a sedative (impregnation), and to what medication it was attributed to” 2 independent electroencephalographers interpreted the EEG result.
Incomplete outcome data (attrition bias)   All outcomes	High risk	Participants dropped out of the midazolam group, as EEG outcome data were only reported for 21 children (originally 27 children were recruited in this arm). The dropout rate of 22% was significant.
Selective reporting (reporting bias)	Low risk	Data were skewed in the chloral hydrate group (sleep latency 21.8 ± 17.5 min; and duration of sleep 61 ± 31.2 min).
Other bias	Low risk	None identified.

Methods	Single‐centre RCT (USA)
Participants	70 children who were undergoing sedation for MRI over a 1‐year period
Interventions	2‐arm comparison of the efficacy and adverse events of sedation between: chloral hydrate (75 mg/kg orally up to maximum dose of 2 g; 35 children) pentobarbital (incremental 2 mg/kg intravenous doses titrated to a maximum of 5 mg/kg or 150 mg; 35 children)
Outcomes	The primary outcome was success of sedation using validated University Michigan Sedation Scale (scale of 0 to 4 with score 4 being unrousable, i.e. sedation successful). Secondary outcome measures included: quality of MRI scans (score 1 to 3; with score 3 major motion artefact with scan incomplete); parents' overall satisfaction with sedation experience (score 1 to 4; score 4 = very satisfied); procedural adverse events.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Method: "Using random number tables, children were randomized to one of two study groups. Group 1 received incremental 2 mg/kg) intravenous (i.v.) doses of PB, titrated to a maximum of 5 mg/kg or 150 mg, administered approximately 10 min prior to MRI. Group 2 received 75 mg/kg of CH orally (maximum dose 2 g) in a single dose approximately 20 min prior to the procedure."
Allocation concealment (selection bias)	Unclear risk	There was no description to enable an assessment of whether random sequence was generated independently from allocation.
Blinding of participants and personnel (performance bias)   All outcomes	High risk	It was not stated whether participants and personnel were blinded to the allocation. However, blinding seems unlikely due to the different routes of administration of the interventions.
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	It was not stated whether the outcome assessors were blinded to the allocation.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No withdrawals were reported. Although the children with missing data were not specifically reported, it appears that all the 70 children who were initially randomised were analysed as calculated from the results section.
Selective reporting (reporting bias)	Low risk	The prespecified outcomes of success of sedation using University Michigan Sedation Scale, time interval to readiness of procedure, quality of MRI scans (score 1 to 3; with score 3 major motion artefact with scan incomplete), parents' overall satisfaction with sedation experience (score 1 to 4; score 4 = very satisfied), and procedural adverse events were reported.
Other bias	Low risk	None identified.

Methods	Double‐blinded, single‐centre RCT (Spain)
Participants	97 consecutive children receiving sedation for MRI
Interventions	2‐arm comparison oral chloral hydrate (70 mg/kg; 50 children) oral chloral hydrate (100 mg/kg; 47 children)
Outcomes	Primary outcome: successful sedation and completion of scan Secondary outcome: adverse reactions
Notes	Comparison is chloral hydrate itself.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Materials and methods: “We entered 97 consecutive children receiving sedation for MRI in a prospective, controlled, double‐blind, randomized trial” and “The children were randomly allocated, by means of a computer generated chart, to oral chloral hydrate 70 mg/kg (group A, n = 50) or 100 mg/kg (group B, n = 47).”
Allocation concealment (selection bias)	Low risk	Materials and methods: “Two strawberry‐flavoured chloral hydrate syrups containing 70 or 100mg/ml were prepared by the pharmacy department.” As described above,allocation concealment occurred as chloral hydrate medication of 2 different concentrations of the same flavour were made by pharmacy.
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	As described above, blinding occurred for participants and personnel as chloral hydrate medication of 2 different concentrations of the same flavour and volume were made by pharmacy.
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	A nurse performed the outcome assessment. Although not stated, it is likely that blinding of outcome assessment occurred for the reasons stated above.
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No withdrawals were reported and no missing data were directly stated. It appeared all the 97 children who were initially randomised were analysed as calculated from the results section.
Selective reporting (reporting bias)	Low risk	The pre‐specified outcomes of mean time to onset of sedation, mean time to spontaneous awakening, effectiveness of sedation and adverse reactions were reported in the results.
Other bias	Low risk	None identified.

Study	Reason for exclusion
Badalaty 1990	Study compared a high and low dose of diazepam with chloral hydrate in the sedation of young children for dental treatment; The outcome measure is dental procedure. Basis of exclusion: type of population and outcome measure.
Bluemke 2000	Study performed based on a sedation database to evaluate the successful sedation procedures, adverse reactions, and cost‐effectiveness of various sedative agents (chloral hydrate, pentobarbital sodium, diazepam, alprazolam). This is not an RCT. Basis of exclusion: study design.
Casillas 1995	Non‐randomised single‐group study, only oral chloral hydrate was being assessed with additional second dose if first dose failed for MRI sedation; outcomes were successful sedation and adverse reaction. Basis of exclusion: study design and intervention.
Cortellazzi 2007	Retrospective study on the efficacy of chloral hydrate sedation and supplementation with sevoflurane, intramuscular or intravenous ketamine, and intravenous pentobarbital and midazolam if failed chloral hydrate. Not an RCT. Basis of exclusion: study design.
Cutler 2007	A retrospective review of all sedations administered by a paediatric sedation service to children (ages 0 to 18 years) undergoing imaging procedures in the radiology department. This was not an RCT. Basis of exclusion: study design.
Dacher 1996	This article was in the French language and was not an RCT. The participants received both rectal chloral hydrate and oral hydroxyzine. Basis of exclusion: study design and intervention.
Dallman 2001	This was an RCT, however, it compared the safety, efficacy, and recovery time of intranasal midazolam spray administered using an atomiser to orally administered chloral hydrate and promethazine for the sedation of paediatric dental patients. Did not meet eligibility criteria for types of outcome (dental procedure) and intervention (chloral hydrate + promethazine vs intranasal midazolam). Basis of exclusion: population type and intervention.
Dearlove 2007	Letter to editor (commentary on adverse reaction). Not an RCT. Basis of exclusion: type of article.
Dirani 2017	A non‐randomised study that compared sequential administration of melatonin, hydroxyzine (if needed), and chloral hydrate (if needed) and chloral hydrate alone in children who underwent EEG. Children in the 2 groups being compared were recruited at different periods. Basis of exclusion: study design.
Edwards 2011	Review evaluating the advantages and disadvantage of sedation and anaesthesia for MRI and alternatives including neonatal comforting techniques, sleep manipulation, and appropriate adaptation of the physical environment. Several factors that would influence the choice of imaging preparation were also discussed. Not an RCT (a review article). Basis of exclusion: type of article.
Eich 2011	Comparative observational study aiming to evaluate 2 institutional anaesthetic protocols for children undergoing elective MRI: propofol only vs propofol plus S‐ketamine. Not an RCT (a comparative observational study) and did not meet eligibility criteria for types of intervention (propofol and ketamine). Basis of exclusion: study design and intervention.
Fallah 2014	This was an RCT. Children aged 1 to 7 years who did not naturally sleep and were unco‐operative for EEG were recruited. Children were in ASA class 1 (healthy persons) or 2. 90 children (39 girls and 51 boys) aged 3.34 ± 1.47 years were investigated. 3‐arm comparison: chloral hydrate 40 mg/kg chloral hydrate 40 mg/kg and promethazine 1 mg/kg chloral hydrate 40 mg/kg and hydroxyzine 2 mg/kg All intervention and comparison groups received chloral hydrate, making them not the intervention of interest. Basic of exclusion: type of intervention.
Fallah 2014a	This was an RCT. Children aged 1 to 7 years in ASA class 1 or 2 were recruited to assess the efficacy of sedative agents inducing deep sedation and completion of MRI examination. Secondary outcomes included clinical side effects. 2 arm comparison: chloral hydrate 40 mg/kg and hydroxyzine 2 mg/kg chloral hydrate 40 mg/kg and midazolam 0.5 mg/kg Both intervention and comparison groups received chloral hydrate, making them not the intervention of interest. Basic of exclusion: type of intervention.
Finnemore 2014	This was a retrospective cohort study of all infants sedated for clinical or research MRI scanning. The aim of this study was to look for clinically significant adverse effects of chloral hydrate used in a large cohort of infants sedated for MRI. Not an RCT (retrospective study), did not meet eligibility criteria for outcome measures (only assessed adverse reaction of chloral hydrate), and did not compare chloral hydrate with another sedation agent. Basis of exclusion: study design.
Funk 2000	Review article discussing various factors that influenced MRI/CT: general anaesthesia or sedation, anaesthesia and image quality, and technical developments. Not an RCT (review article). Basis of exclusion: type of article.
Gan 2016	RCT comparing different dosages of intranasal dexmedetomidine as rescue medication in paediatric ophthalmic examination after chloral hydrate failed. Basis of exclusion: population and intervention.
Greenberg 1991	Prospective non‐randomised study in which high‐dose oral chloral hydrate (80 to 100 mg/kg) and low‐dose oral chloral hydrate (40 to 75 mg/kg), with a maximum total dose of 2 g, was administered to children for CT examinations (did not state the specified body parts and including abdomen). Success rate and adverse reactions were evaluated. Not an RCT (single‐group study: high‐dose and low‐dose chloral hydrate without comparison). Basis of exclusion: study design.
Greenberg 1994	Prospective non‐randomised study evaluating the safety and efficacy of thioridazine as an adjunct to chloral hydrate sedation in children undergoing MR imaging who are difficult to sedate. All children in the study had a history of unsuccessful sedation with chloral hydrate alone or were mentally retarded. Not an RCT and did not meet eligibility criteria for types of intervention (no comparison). Basis of exclusion: study design and intervention.
Gupta 2010	This was an oral presentation abstract to determine if currently available evidence supports use of melatonin for EEG sedation. Not an RCT (abstract). Basis of exclusion: type of article.
Hare 2012	Review article evaluating 4 randomised trials and comparing the efficacy and safety of chloral hydrate versus midazolam for use in paediatric sedation for painless imaging including echocardiography. Not an RCT (review article). Basis of exclusion: type of article.
Hoffman 2002	This article evaluated the risk reduction in paediatric procedural sedation. Not an RCT. Basis of exclusion: type of article and study design.
Hollman 1996	Letter to editor and commentary on chloral hydrate vs midazolam sedation for neuroimaging studies. Not an RCT (letter to editor and commentary). Basis of exclusion: type of article.
Hubbard 1992	Non‐comparative retrospective study evaluating the safety and efficacy of oral chloral hydrate for infants and intravenous pentobarbital for older children. Not an RCT (non‐comparative retrospective study). Basis of exclusion: study design.
Kannikeswaran 2009	Retrospective study of children 1 to 18 years who required sedation for an elective brain MRI. Children < 2 years of age were sedated with oral chloral hydrate, while children 1 to 7 years of age were sedated with intravenous pentobarbital. Additional doses of pentobarbital or fentanyl were administered if failed sedation. Children older than 8 years of age were given midazolam intravenously or orally for sedation. Not an RCT (retrospective study) and did not meet eligibility criteria for types of comparison (pentobarbital vs fentanyl). Basis of exclusion: study design and intervention.
Keeter 1990	Questionnaire study that aimed to document current sedation practices in CT examination of children in the USA. A questionnaire was sent to a random sample of 2000 hospitals with CT scanners. Not an RCT (questionnaire survey). Basis of exclusion: study design.
Keidan 2004	Retrospective study conducted in 2 large, urban hospitals in Israel. The study population consisted of 200 infants who underwent auditory brainstem response examination and were sedated with chloral hydrate (no comparison). This study was not an RCT (non‐comparative retrospective study) and did not meet eligibility of types of study design, population, intervention. Basis of exclusion: study design, population, and intervention.
Lee 2012	This retrospective study aimed to evaluate sedation success in children given chloral hydrate at 2 dosing regimens for brain MRI scan. This was not an RCT and did not met eligibility criteria for types of comparison used (no direct comparison with another sedative agent). Basis of exclusion: study design and intervention.
Li 2014	This prospective randomised study aimed to evaluate sedation success in children for brain CT, auditory brainstem responses, and visual evoked potentials who failed chloral hydrate sedation. Child was then randomly assigned to receive intranasal dexmedetomidine at various doses. The study did not meet eligibility criteria for types of intervention (only used intranasal dexmedetomidine for those who failed chloral hydrate sedation) and types of comparison used. Basis of exclusion: type of intervention.
Low 2008	This retrospective study aimed to evaluate the success and safety of chloral hydrate sedation protocol for children undergoing brain MRI. Not an RCT (a retrospective study evaluating efficacy of chloral hydrate) and did not meet eligibility criteria for types of comparison used. Basis of exclusion: study design and type of intervention.
Marchi 2004	Prospective observational study of children undergoing deep sedation (using either chloral hydrate or propofol) for brain MRI. Not an RCT as the children were not randomised. Basis of exclusion: study design.
Mason 2004	This retrospective study aimed to evaluate the success of sedation using chloral hydrate (patients sedated between 1997 and 1999) and pentobarbital (patients sedated between 2000 and 2002) for brain imaging. Not an RCT (retrospective review comparing chloral hydrate with pentobarbital). Basis of exclusion: study design.
Mathew 2014	Prospective RCT of children undergoing auditory brainstem response testing randomised for sedation with either midazolam nasal spray with oral placebo or syrup chloral hydrate with placebo nasal spray. Did not meet eligibility criteria for type of participant (sedation used for auditory test not neurodiagnostic procedure). Basis of exclusion: population type.
McCarver‐May 1996	Cross‐over study. Term newborn infants who had both CT and single‐photon emission computed tomography scanning after extracorporeal membrane oxygenation bypass were re‐emitted for the study. The order of neuroimaging studies was randomised. For the first study, chloral hydrate was given orally. After 48 hours, midazolam was given intravenously for the second study. Not an RCT (cross‐over study). Basis of exclusion: study design.
Mehta 2004	Prospective observation study of efficacy of clonidine as sedating agent in children with autism undergoing EEG. Not an RCT and did not meet eligibility criteria of type of intervention (did not use chloral hydrate for sedation). Basis of exclusion: study design and type of intervention.
Nichols 2005	Retrospecitive review of children who failed sedation using chloral hydrate or midazolam for diagnostic brain imaging. Not an RCT. Basis of exclusion: study design.
Reynolds 2016	Double‐blinded RCTs comparing efficacy of intranasal dexmedetomidine and oral chloral hydrate for auditory brainstem response in children. Did not meet eligibility criteria for type of participant (sedation was used for auditory test not neurodiagnostic procedure). Basis of exclusion: population type.
Ronchera‐Oms 1994	Retrospective review of efficacy of chloral hydrate sedation for children undergoing brain MRI scan. Not an RCT and did not meet eligibility criteria for types of comparison (no comparison with other sedative agent). Basic of exclusion: study design and type of intervention.
Rooks 2003	A prospective, non‐randomised observational study assessing the sedation effects of oral pentobarbital sodium against oral chloral hydrate in 2 separate groups of children who were undergoing radiologic imaging. Although the method of allocation was not stated, it is unlikely that the participants were randomised. Basis of exclusion: study design.
Rues 2002	Retrospective review and prospective observational study assessing the efficacy of sedation using a pre‐existing sedation protocol (for under 2 years old: oral chloral hydrate +/‐ oral diphenhydramine/hydroxyzine followed by IV midazolam; for over 2 years old: IV pentobarbital +/‐ IV midazolam) for brain MRI or CT. Not an RCT and did not meet eligibility criteria for type of intervention (children over 2 years old were given IV medication not chloral hydrate). Also did not meet eligibility criteria for type of comparison (looked at effectiveness of sedation protocol without comparing with other sedative agents). Basis of exclusion: study design and type of intervention.
Sury 2006	RCT of children who received additional second‐line sedation (either melatonin or placebo) after failure of first‐line chloral hydrate sedation for brain MRI. Basis of exclusion: type of intervention.
Takasaka 1999	Retrospective review of effect of sedation on EEG recording. Not an RCT and did not meet eligibility criteria for type of outcome measure (assessed whether sedatives alter EEG recording, did not assess success of sedation). Basis of exclusion: study design and type of outcome measure.
Treluyer 2004	Retrospective study assessing efficacy of chloral hydrate sedation for brain CT or MRI. Not an RCT and did not meet eligibility criteria for type of comparison (only assessed efficacy of chloral hydrate). Basis of exclusion: study design and type of intervention.
Wang 2005	RCT assessing effect of sedation (chloral hydrate) versus sleep deprivation on brain EEG results. Did not meet eligibility criteria for type of comparison (as no other sedative agent was used) and type of outcome measure (assessed effect of sedation on EEG results, did not assess success of sedation). Basis of exclusion: type of intervention and type of outcome measure.
Yuen 2017	Retrospective study comparing melatonin versus chloral hydrate as sedating agent in children undergoing EEG. Basis of exclusion: study design.
Zhang 2016	RCT assessing the effectiveness of intranasal dexmedetomidine as a rescue sedative agent as compared to a second dose of chloral hydrate in children who had 1 dose of chloral hydrate while undergoing non‐invasive diagnostic procedures. Basis of exclusion: intervention.

Methods	Prospective, double‐blind, randomised study Randomisation of the study drugs was performed by an independent pharmacist using a computer‐generated random number program and was concealed from the study investigators. Neither the child, nor any of the investigators, nor the healthcare providers knew the active component of the study medication.
Participants	All paediatric patients ≤ 12 years of age who were judged to need sedation for diagnostic or therapeutic procedures in Day Care Unit, King Abdulaziz Medical City, Riyadh, Saudi Arabia
Interventions	2‐arm comparison chloral hydrate (75 mg/kg, maximum dose of 2 g) prepared midazolam (0.5 mg/kg, maximum dose of 10 mg) The second dose was 30 mg/kg for the chloral hydrate group and 0.25 mg/kg for the midazolam group if failed sedation in the first dose. This study also involves non‐neurodiagnostic procedures, eye exam and others besides MRI and CT scan. However, no specific parts of body were mentioned for MRI and CT scan sessions. No separate analysis for these groups (neurodiagnostic procedure) were available.
Outcomes	Primary outcome: successful sedation Secondary outcome: adverse reaction
Notes	Wrote and emailed author on 4 November 2015 requesting the separate number of MRI and CT brain performed and the analyses.