Midazolam for sedation before procedures

Aaron Conway; John Rolley; Joanna R Sutherland

doi:10.1002/14651858.CD009491.pub2

. 2016 May 20;2016(5):CD009491. doi: 10.1002/14651858.CD009491.pub2

Midazolam for sedation before procedures

Aaron Conway ^1,^2,^✉, John Rolley ³, Joanna R Sutherland ⁴

Editor: Cochrane Emergency and Critical Care Group

PMCID: PMC6517181 PMID: 27198122

Abstract

Background

Midazolam is used for sedation before diagnostic and therapeutic medical procedures. It is an imidazole benzodiazepine that has depressant effects on the central nervous system (CNS) with rapid onset of action and few adverse effects. The drug can be administered by several routes including oral, intravenous, intranasal and intramuscular.

Objectives

To determine the evidence on the effectiveness of midazolam for sedation when administered before a procedure (diagnostic or therapeutic).

Search methods

We searched the Cochrane Central Register of Controlled Trials (CENTRAL to January 2016), MEDLINE in Ovid (1966 to January 2016) and Ovid EMBASE (1980 to January 2016). We imposed no language restrictions.

Selection criteria

Randomized controlled trials in which midazolam, administered to participants of any age, by any route, at any dose or any time before any procedure (apart from dental procedures), was compared with placebo or other medications including sedatives and analgesics.

Data collection and analysis

Two authors extracted data and assessed risk of bias for each included study. We performed a separate analysis for each different drug comparison.

Main results

We included 30 trials (2319 participants) of midazolam for gastrointestinal endoscopy (16 trials), bronchoscopy (3), diagnostic imaging (5), cardioversion (1), minor plastic surgery (1), lumbar puncture (1), suturing (2) and Kirschner wire removal (1). Comparisons were: intravenous diazepam (14), placebo (5) etomidate (1) fentanyl (1), flunitrazepam (1) and propofol (1); oral chloral hydrate (4), diazepam (2), diazepam and clonidine (1); ketamine (1) and placebo (3); and intranasal placebo (2). There was a high risk of bias due to inadequate reporting about randomization (75% of trials). Effect estimates were imprecise due to small sample sizes. None of the trials reported on allergic or anaphylactoid reactions.

Intravenous midazolam versus diazepam (14 trials; 1069 participants)

There was no difference in anxiety (risk ratio (RR) 0.80, 95% confidence interval (CI) 0.39 to 1.62; 175 participants; 2 trials) or discomfort/pain (RR 0.60, 95% CI 0.24 to 1.49; 415 participants; 5 trials; I² = 67%). Midazolam produced greater anterograde amnesia (RR 0.45; 95% CI 0.30 to 0.66; 587 participants; 9 trials; low‐quality evidence).

Intravenous midazolam versus placebo (5 trials; 493 participants)

One trial reported that fewer participants who received midazolam were anxious (3/47 versus 15/35; low‐quality evidence). There was no difference in discomfort/pain identified in a further trial (3/85 in midazolam group; 4/82 in placebo group; P = 0.876; very low‐quality evidence).

Oral midazolam versus chloral hydrate (4 trials; 268 participants)

Midazolam increased the risk of incomplete procedures (RR 4.01; 95% CI 1.92 to 8.40; moderate‐quality evidence).

Oral midazolam versus placebo (3 trials; 176 participants)

Midazolam reduced pain (midazolam mean 2.56 (standard deviation (SD) 0.49); placebo mean 4.62 (SD 1.49); P < 0.005) and anxiety (midazolam mean 1.52 (SD 0.3); placebo mean 3.97 (SD 0.44); P < 0.0001) in one trial with 99 participants. Two other trials did not find a difference in numerical rating of anxiety (mean 1.7 (SD 2.4) for 20 participants randomized to midazolam; mean 2.6 (SD 2.9) for 22 participants randomized to placebo; P = 0.216; mean Spielberger's Trait Anxiety Inventory score 47.56 (SD 11.68) in the midazolam group; mean 52.78 (SD 9.61) in placebo group; P > 0.05).

Intranasal midazolam versus placebo (2 trials; 149 participants)

Midazolam induced sedation (midazolam mean 3.15 (SD 0.36); placebo mean 2.56 (SD 0.64); P < 0.001) and reduced the numerical rating of anxiety in one trial with 54 participants (midazolam mean 17.3 (SD 18.58); placebo mean 49.3 (SD 29.46); P < 0.001). There was no difference in meta‐analysis of results from both trials for risk of incomplete procedures (RR 0.14, 95% CI 0.02 to 1.12; downgraded to low‐quality evidence).

Authors' conclusions

We found no high‐quality evidence to determine if midazolam, when administered as the sole sedative agent prior to a procedure, produces more or less effective sedation than placebo or other medications. There is low‐quality evidence that intravenous midazolam reduced anxiety when compared with placebo. There is inconsistent evidence that oral midazolam decreased anxiety during procedures compared with placebo. Intranasal midazolam did not reduce the risk of incomplete procedures, although anxiolysis and sedation were observed. There is moderate‐quality evidence suggesting that oral midazolam produces less effective sedation than chloral hydrate for completion of procedures for children undergoing non‐invasive diagnostic procedures.

Plain language summary

Midazolam for sedation before procedures

Review question

We wanted to find out whether midazolam makes medical procedures more comfortable for children and adults, as well as whether it makes the procedure easier to perform.

Background

Children and adults can become anxious during medical procedures and the procedures can be painful. Pain and anxiety can sometimes make the procedure more difficult to perform for the medical staff, due to movement or a lack of co‐operation from the patient. Sedative medications, including midazolam, are used to reduce pain and anxiety. They can be injected directly into the bloodstream (with an almost immediate effect), injected into muscle tissue, given as a nasal spray, or swallowed as a tablet or solution.

Study characteristics

The evidence is up‐to‐date to January 2016. We included 30 trials involving 2319 participants. We looked at trials that compared midazolam with no active treatment ('dummy' treatment/placebo) or a different medication for sedation before a procedure. The trials involved children and adults having procedures to diagnose medical problems rather than procedures for treatment of a disease. We disregarded trials where people received a general anaesthetic or other medications for sedation or pain relief in addition to midazolam during their procedure.

Key results

Midazolam administered into the bloodstream compared with other medications did not seem to make the participants more drowsy, reduce anxiety or pain, or make the procedure easier to perform. This is based on the low‐quality evidence currently available. A potential benefit is that children and adults who received midazolam compared with no active treatment did not remember as much about the procedures. Midazolam made them drowsy, reduced anxiety and made it easier to perform a procedure. There is moderate‐quality evidence that a solution of midazolam given to children to drink before a procedure was not as effective as a different medication called chloral hydrate. A nasal spray of midazolam before a procedure made the participants drowsy and reduced their anxiety, but this did not make it easier to perform procedures on them. This review cannot be used to assess the harms of midazolam for sedation before a procedure.

Quality of the evidence

We rated the evidence, in the main, as being of low quality. Particularly concerning was that many trials did not explain how participants were randomized to either midazolam or to a different treatment, and that the results did not give us a very clearly defined answer.

Summary of findings

Background

Midazolam is used for sedation before diagnostic and therapeutic medical procedures. It is an imidazole benzodiazepine that has depressant effects on the central nervous system (CNS) with rapid onset of action, good effectiveness and few adverse effects. The drug can be administered by several routes including oral, intravenous, intranasal and intramuscular. We compare the effectiveness of midazolam versus placebo and other sedatives.

Description of the condition

The major goal of premedication is to provide sedation and anxiolysis in order to facilitate therapeutic and diagnostic interventions. The characteristics of midazolam are that it has a rapid onset of action, short duration of sedation and low toxicity.

Description of the intervention

Midazolam is one of the most commonly used sedative medications for surgical and non‐surgical procedures. It is currently indicated for sedation, anxiolysis and amnesia preoperatively and during procedures, including ventilation of critically‐ill patients; as a co‐induction agent; and as a supplement to nitrous oxide and oxygen.

Flumazenil is a benzodiazepine antagonist that can be used to rapidly reverse the sedative and other CNS effects of midazolam.

How the intervention might work

Midazolam has a fast recovery time and is used as premedication for many procedures including colonoscopy (Lazaraki 2007); gastrointestinal endoscopy (Fakheri 2010); magnetic resonance imaging (Hollenhorst 2001); and flexible bronchoscopy (Rolo 2012). The anterograde amnesic property of midazolam may be useful for premedication before a procedure, to reduce any associated unpleasant memories (Riss 2008). The disadvantages of midazolam include drug interactions, tolerance, and withdrawal syndrome, as well as adverse events including cognitive impairment (Riss 2008). There is also the possibility that midazolam‐induced anterograde amnesia may be viewed by some patients as undesirable.

Why it is important to do this review

Midazolam is used for sedation before procedures in a wide range of medical specialties. These include gastroenterology, respiratory medicine, gynaecology, cardiology and radiology. As midazolam is often not the only drug used for procedural sedation, it is important to determine if there are any relevant clinical differences between the effects of midazolam and other drugs. It may also be important to determine whether clinicians choose different medications according to patient characteristics (age, weight, gender), or if medication choice is influenced by the type of procedure or context. To our knowledge, no systematic review has been done on the use of midazolam for sedation prior to procedures. Our review aims to fill this gap.

Objectives

To determine the evidence on the effectiveness of midazolam for sedation when administered before a procedure (diagnostic or therapeutic).

Methods

Criteria for considering studies for this review

Types of studies

We included randomized controlled trials (RCTs) in which midazolam was used for sedation before a procedure (diagnostic or therapeutic). We included studies irrespective of language and publication status. We excluded prospective cohort studies and quasi‐randomized studies.

Types of participants

We included participants of any age (adults and children) who were undergoing a procedure preceded by sedation. We excluded any participants undergoing dental procedures, because a Cochrane review has been published about sedation with midazolam specifically in that setting (Lourenço‐Matharu 2012).

Types of interventions

We included midazolam by any route, at any dose or time, administered before a procedure. Participants who received a placebo before a procedure constituted the control group. We also included studies that compared midazolam with another drug for sedation before a procedure. We performed a separate analysis for each different drug comparison (for example, midazolam versus sedative A; midazolam versus sedative B).

We excluded studies that simultaneously compared different drugs and different routes (for example, intranasal midazolam plus intravenous sedative A versus intranasal sedative A plus intravenous midazolam; intravenous midazolam versus intranasal sedative A). We excluded studies where dexmedetomidine was the comparator, as there is another Cochrane review about sedation for this medication (Shailaja 2013).

Types of outcome measures

Primary outcomes

Effective sedation corresponding to adequate sedation level, anxiolysis, ability to complete proposed procedure. This was evaluated by:

1.1. Level of sedation on a sedation assessment scale;

1.2. Numeric rating scale of anxiety or number of participants rated as anxious;

1.3. Vital signs (heart rate, blood pressure, respiratory rate, and oxygen saturation);

1.4. Tolerance of procedure or participant co‐operation (as defined/measured by the authors of the trial);

1.5. Participant or proceduralist satisfaction (as defined/measured by the authors of the trial);

1.6. Proportion of incomplete procedures or where there was difficulty performing the procedures.

Secondary outcomes

2.1. Duration of sedation;

2.2. Onset time of sedation;

2.3. Offset time of sedation;

2.4. Anterograde amnesia (defined by number of participants who recalled the procedure);

2.5. Oversedation (as defined/measured by the authors of the trial);

2.6. Disinhibition or excitation;

2.7. Quality of recovery (as defined/measured by the authors of the trial);

2.8. Discomfort/pain (as defined/measured by the authors of the trial);

2.9. Allergic or anaphylactoid reactions;

2.10. Sedation reversal.

Search methods for identification of studies

Electronic searches

We searched the Cochrane Central Register of Controlled Trials (CENTRAL) to January 2016, MEDLINE in Ovid (1966 to January 2016) and Ovid EMBASE (1980 to January 2016).

We combined the sensitive strategies described in Section 6.4 of theCochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) to search for RCTs in MEDLINE and EMBASE.

We searched CENTRAL using the terms given in Appendix 1. We adapted our MEDLINE search strategy (Appendix 2) to reflect the subject headings found in the thesauri used by EMBASE (Appendix 3). We used the free‐text terms in all databases and in combination with subject headings when thesauri are a component of a database.

We imposed no language restrictions.

Searching other resources

For ongoing trials, we searched the following databases on 13th July 2015: metaRegister of Controlled Trials (www.controlled‐trials.com/mrct) and Clinical Trials (clinicaltrials.gov). We also screened the reference lists of all eligible trials and reviews.

Data collection and analysis

Selection of studies

We screened all titles and abstracts for eligibility. Two authors (AC and JR) independently performed this screening (see Appendix 4 for a copy of the study selection form). We resolved disagreements by discussion to decide on trial inclusion. In the case of insufficient published information to make a decision about inclusion, we attempted to contact the first author of the relevant trial. We compiled a list of eligible trials, each with a unique identifier on a 'Form for eligible trials' (see Appendix 5).

Data extraction and management

Two authors (AC and JR) independently extracted data onto a paper form. A copy of this paper form is in Appendix 6. We resolved discrepancies by discussion. AC attempted to contact an author of the relevant trial if we required additional information.

Assessment of risk of bias in included studies

Two authors (AC and JR) independently assessed the methodological quality of the eligible trials. We resolved disagreements by discussion.

We performed 'Risk of bias' assessment using the 'Risk of bias' tool as described in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) and by Jüni 2001. A copy of the form we used for this is in Appendix 7.

We assessed each trial according to the quality domains of random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and any other potential threats to validity.

We considered a trial as having a low risk of bias if we assessed all domains as adequate. We considered a trial as having a high risk of bias if we assessed one or more domains as inadequate or unclear. We planned to conduct sensitivity analyses to determine whether excluding studies at high risk of bias affected the results of the meta‐analysis.

We reported the 'Risk of bias' table as part of the Characteristics of included studies table, and present 'Risk of bias' summary figures that detail all of the judgements made for all included studies in the review.

Measures of treatment effect

For dichotomous variables, we calculated the risk ratio (RR). For continuous variables, we calculated the mean difference (MD) when studies reported their results through the same variables measured with the same instruments (same units of measurement). When continuous data were related to the same aspect in the participants but were measured with different instruments (and did not have an interchangeable unit of measurement) we pooled them using the standardized mean difference (SMD). We calculated the 95% confidence interval (CI) as the measure of variance for all statistical methods.

Unit of analysis issues

To avoid unit of analysis issues, we planned to consider repeated observations as separate outcomes and group them accordingly for analysis (Morão 2011). However, the trials included in the review reported the change in time‐separated observations (such as the change in oxygen saturation from before to after the administration of sedation), so we were unable to do this (Differences between protocol and review). We sought pre‐cross‐over data for trials that used a cross‐over design.

Dealing with missing data

If trials did not report withdrawals, we assumed there were none. We used an available‐case analysis as the default for meta‐analysis and we also considered sensitivity analysis using best‐case (all participants who withdrew did not experience the event) and worst‐case (all participants who withdrew did experience the event) scenarios for any missing data. No outcomes measured with continuous variables had missing data that needed to be included in the meta‐analyses.

Assessment of heterogeneity

We assessed the clinical heterogeneity of included trials as:

clinical diversity (e.g. different types of procedures, different forms of midazolam administration, participants' ages, etc.);
methodological diversity ('Risk of bias' assessment);
statistical heterogeneity (a manifestation of clinical or methodological diversity, or both, among the trials).

We assessed statistical heterogeneity with the I² statistic, thereby estimating the percentage of total variance across studies due to heterogeneity rather than chance (Higgins 2002). We considered an I² statistic value greater than 50% as considerable heterogeneity or if the Chi² test was significant (see Data synthesis).

Assessment of reporting biases

As per the original protocol, we planned to assess publication bias and small‐study effects using a funnel plot if there were 10 or more studies included in the meta‐analysis (Morão 2011). However, we did not perform this analysis because fewer than 10 studies were included in each meta‐analysis.

Data synthesis

We generated meta‐analytic estimates for outcomes reported by two or more studies. We performed the analysis using Review Manager 5 software (Review Manager 2014). Because the population is varied, we included all types of procedures. Due to this variation, the intervention effect could have varied across the different studies. We therefore expected that a random‐effects model would be suitable for the meta‐analyses. However, a smaller value of the I² statistic (less than 50%) prompted consideration of the use of a fixed‐effect model. We performed all analyses according to the intention‐to‐treat (ITT) principle.

Subgroup analysis and investigation of heterogeneity

We planned to perform subgroup analyses for age (children, adults (16 years of age or older)), type of procedure (diagnostic, therapeutic) and medical specialty (surgical, non‐surgical) (Morão 2011). However, there was not enough evidence to conduct any subgroup analysis.

Sensitivity analysis

We planned to perform sensitivity analyses by trials with a low risk of bias versus moderate or high risk of bias (Morão 2011). However, we rated most studies to be of either low or very low quality, and this was not appropriate.

'Summary of findings' table

We used the principles of the GRADE system (Guyatt 2008) to assess the quality of the body of evidence associated with the following specific outcomes: level of sedation on a sedation assessment scale; numerical rating of scale of anxiety or number of participants rated as anxious; incomplete procedures; anterograde amnesia (recalled procedures); disinhibition or excitation; discomfort/pain; allergic or anaphylactoid reactions; and we constructed 'Summary of findings' tables using the GRADE software.

The GRADE approach appraises the quality of a body of evidence based on the extent to which one can be confident that an estimate of effect or association reflects the item being assessed. The quality of a body of evidence is based on within‐study risk of bias (methodologic quality), the directness of the evidence, heterogeneity of the data, precision of effect estimates and risk of publication bias. The GRADE approach specifies four levels of quality (high, moderate, low, very low). The highest quality rating is for randomized trial evidence and the lowest is for triple‐downgraded randomized trials, downgraded observational studies or case series and case reports.

Results

Description of studies

Results of the search

Figure 1 summarizes the search results to January 2016. The searches identified 4033 hits. We retrieved 163 papers for consideration and included 30 trials in this review.

Included studies

We include 30 trials with 2319 participants (Characteristics of included studies) that compared midazolam via the intravenous, oral and intranasal routes of administration, either to placebo or to another medication for sedation before a procedure. The included trials were conducted in both adult and paediatric populations. 16 trials enrolled participants having gastrointestinal endoscopy procedures (Bell 1988; Bhalla 2006; Bianchi Porro 1988; Cole 1983; Córdova 1992; Fakheri 2010; Gilvarry 1990; Kuganeswaran 1999; Lavies 1988; Lazaraki 2007; Lee 1989; Sainpy 1984; Takrouri 1988; Tolia 1990; Whitwam 1983; Yuno 1996) and there were three trials involving bronchoscopy (Aktogu 1994; Korttila 1985; Rolo 2012). Diagnostic imaging was performed in five trials (Akil 2005; D'Agostino 2000; Hollenhorst 2001; Stokland 2003; Wheeler 2001), one trial (Coll‐Vinent 2003) was conducted with participants undergoing cardioversion, and another (De Alencar 2010) for participants undergoing minor office‐based plastic surgery. Four trials were conducted with children undergoing minor procedures that required motion control, including lumbar puncture (Derakhshanfar 2013), suturing (Everitt 2002; Younge 2001) and Kirschner wire removal (Templeton 2010).

There was geographic variability across the included trials. Trials were performed in the United Kingdom (Gilvarry 1990; Templeton 2010; Whitwam 1983; Younge 2001), USA (Cole 1983; D'Agostino 2000; Kuganeswaran 1999; Lavies 1988; Tolia 1990; Wheeler 2001), Turkey (Akil 2005; Aktogu 1994), India (Bhalla 2006), Italy (Bianchi Porro 1988), Spain (Coll‐Vinent 2003), Mexico (Córdova 1992), Brazil (De Alencar 2010), Iran (Derakhshanfar 2013; Fakheri 2010), Australia (Everitt 2002), Germany (Hollenhorst 2001), Finland (Korttila 1985), Greece (Lazaraki 2007), Jamaica (Lee 1989), Portugal (Rolo 2012), France (Sainpy 1984), Sweden (Stokland 2003), Jordan (Takrouri 1988) and Japan (Yuno 1996).

Regarding the characteristics of the interventions, for intravenous midazolam, four trials used weight‐based calculation with a dose of 0.1 mg/kg (Córdova 1992; Korttila 1985; Sainpy 1984; Tolia 1990). Other trials used smaller doses, including 0.07 mg/kg (Bianchi Porro 1988; Lee 1989; Whitwam 1983), 0.06 mg/kg (Aktogu 1994) and 0.05 mg/kg (Rolo 2012). One trial used a higher dose of 0.2 mg/kg (Coll‐Vinent 2003). Other trials did not use participants' weight to calculate doses. These trials used either 2.5 mg (Bell 1988), 5 mg (Bhalla 2006; Cole 1983;), 10 mg (Gilvarry 1990) or 15 mg (De Alencar 2010). Some trials used smaller doses for elderly participants (Bell 1988; Bhalla 2006; Cole 1983). Other trials reported only the mean or range of dose that was administered instead of the planned method of titration (Lavies 1988; Lazaraki 2007; Takrouri 1988). For oral midazolam, only one trial did not use a weight‐based dose calculation (Kuganeswaran 1999). Weight‐based doses of oral midazolam used were 0.5 mg/kg (D'Agostino 2000; Derakhshanfar 2013; Wheeler 2001), 0.6 mg/kg (Akil 2005), 0.7 mg/kg (Younge 2001) and 1 mg/kg (Everitt 2002; Templeton 2010). For intranasal midazolam, Hollenhorst 2001 used a standard dose of 4 mg whereas Stokland 2003 used a dose of 0.2 mg/kg up to 5 mg.

The comparator arms were: intravenous diazepam in 14 trials with 1069 participants (Aktogu 1994; Bhalla 2006; Bell 1988; Bianchi Porro 1988; Cole 1983; Córdova 1992; Gilvarry 1990; Korttila 1985; Lavies 1988; Lee 1989; Sainpy 1984; Takrouri 1988; Tolia 1990; Whitwam 1983); intravenous placebo in five trials with 493 participants (Bhalla 2006; Fakheri 2010; Lavies 1988; Rolo 2012; Yuno 1996); intravenous etomidate in one trial with 17 participants (Coll‐Vinent 2003); intravenous fentanyl in one trial with 126 participants(Lazaraki 2007); intravenous flunitrazepam in one trial with 86 participants (Takrouri 1988); intravenous propofol in one trial with 17 participants (Coll‐Vinent 2003); oral chloral hydrate in four trials with 268 participants (Akil 2005; D'Agostino 2000; Derakhshanfar 2013; Wheeler 2001); oral diazepam in two trials with 122 participants (De Alencar 2010; Everitt 2002); oral diazepam and clonidine in one trial with 34 participants (De Alencar 2010); oral ketamine in one trial with 59 participants (Younge 2001); oral placebo in three trials with 176 participants (Akil 2005; Kuganeswaran 1999; Templeton 2010); and intranasal placebo two trials with 149 participants (Hollenhorst 2001; Stokland 2003).

For three trials, we pooled two different groups, as the trials compared midazolam with: both a placebo and with chloral hydrate (Akil 2005); or both placebo and diazepam (Bhalla 2006; Lavies 1988). Three trials (Coll‐Vinent 2003; De Alencar 2010; Takrouri 1988) compared midazolam with two different medications. One trial compared two different doses of midazolam with another medication (Korttila 1985). For this review, we considered only the outcomes reported from the higher dose of midazolam used, as this dose was comparable with the doses used in the other included trials. Two included articles presented results from the same trial (Bhalla 2006).

Financial support was provided by industry for two trials (Cole 1983; Kuganeswaran 1999). Yuno 1996 and Templeton 2010 reported receiving funding for their trials from non‐industry sources. No other trials reported the source of funding in their publications.

Excluded studies

We needed to review a large number of papers in full text, as it was unclear from the title and abstract whether or not analgesia was administered concurrently with midazolam and whether or not the sedation was administered intraprocedurally or just before the procedure. We excluded 122 articles that we reviewed in full text. A selection of 18 of these excluded articles are included in the Characteristics of excluded studies table, to display the common reasons for exclusion. These include intraprocedural sedation used in addition to preprocedural sedation (Mui 2005; Muttu 2005), the concomitant use of analgesia or other sedative medication with midazolam (Brouillette 1989; Dere 2010; Nascimento 2007; Salmon 1992; Sajedi 2006; Sherry 1989; Tamayo 1993), routine use of flumazenil (Ristikankare 1999; Ristikankare 2000a; Ristikankare 2000b; Uygur‐Bayramiçli 2002), wrong research design (Sandler 1992; Tesoro 2007; Weinstein 2010), midazolam used in control group if initial sedation was ineffective (Bonta 2003), and placebo being administered by a different route to midazolam (Yildirim 2006).

Studies awaiting classification

We await more information in order to classify a further 10 studies (Characteristics of studies awaiting classification). We attempted to contact the authors of four of them (Bardhan 1984; Green 1984; Ogden 1993; Theroux 1993) to clarify details but we either could not locate their current contact details or we did not receive a response from the authors. The remaining six studies awaiting classification still require data extraction because they were not published in English (Frisancho 1996; Mendes 1986; Mignonsin 1994; Münte 2002; Thakur 2003; Wild 1988).

Ongoing studies

We identified three ongoing studies from our search of the clinical trials databases that might be eligible for inclusion in future updates of the review (NCT00563069; Puttapitakpong 2015; NCT01925898) (see Characteristics of ongoing studies).

Risk of bias in included studies

We present summaries of the judgements of the risk of bias of included trials in Figure 2 and Figure 3. Details of the included trials are in the Characteristics of included studies tables. The overall risk of performance bias and detection bias was low for 50% of included trials. For randomization sequence generation and allocation concealment the quality assessment yielded low risk of bias for approximately 25% or less of the included trials. The risk of attrition bias for the primary outcomes was low for more than 75% of trials.

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

Reporting of methods used for randomization sequence generation and allocation concealment was unclear in the majority of trials. As such, it is unclear as to the impact that potential selection bias might have on the estimates of the effects.

Blinding

Overall, there was a low risk of bias from blinding, due to the double‐blinded design used for most trials.

Incomplete outcome data

The trials were generally of short duration in an environment that was conducive to a low attrition rate for intra‐ and post‐procedural data that were collected before the participant was discharged. As such, there is low risk of attrition bias for the primary outcomes set for this review. However, one trial (Everitt 2002) reported high attrition rates for the 'quality of recovery' outcome, which was measured with a post‐discharge survey, meaning there is a high risk of attrition bias, but only for this secondary outcome.

Selective reporting

We found no definite evidence of selective reporting. However, we did not seek trial protocols because most included trials were published prior to the establishment of clinical trial registries. It is therefore unclear whether the outcomes infrequently reported or absent from the included trials, such as allergic or anaphylactoid reactions and sedation reversal, were collected but not reported.

Other potential sources of bias

We did not identify any other definite source of potential bias.

Effects of interventions

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

Summary of findings for the main comparison. Intravenous midazolam compared to diazepam for sedation before procedures.

Intravenous midazolam compared to diazepam for sedation before procedures
Patient or population: adults and children requiring sedation before gastrointestinal endoscopy and bronchoscopy  Settings: hospitals in UK, USA, Mexico, India, Italy, Finland, Jamaica, France, Jordan and Turkey  Intervention: intravenous midazolam  Comparison: intravenous diazepam
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Diazepam	Midazolam
Level of sedation on a sedation assessment scale				75 (1 study)	very low¹	The mean level of sedation in the midazolam group was 3.2 and the mean level of sedation in the diazepam group was 2.7 on a scale that ranged from 0 to 4 (higher scores indicating more sedation). Measured with a scale that ranged from 0 ‐ 4 (higher scores indicating the participant was more sedated).
Numeric rating of anxiety or rated as anxious	167 per 1000	133 per 1000  (65 to 270)	RR 0.80   (0.39 to 1.62)	175  (2 studies)	low²	Effect estimate calculated for number of participants rated as anxious
Incomplete procedure				170 (1 study)		All procedures were completed in both groups
Anterograde amnesia (defined by number of participants who recalled the procedure)	481 per 1000	216 per 1000  (144 to 318)	RR 0.45   (0.3 to 0.66)	587  (9 studies)	low³
Disinhibition or excitation						No studies reported on this outcome
Discomfort/pain	202 per 1000	121 per 1000  (48 to 301)	RR 0.60   (0.24 to 1.49)	415  (5 studies)	low²
Allergic or anaphylactoid reaction						No studies reported on this outcome
The basis for the assumed risk* is the control group risk across studies or the average risk for pooled data and the control group risk for single studies. The corresponding risk (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; RR: Risk ratio.
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Intravenous midazolam compared to placebo for sedation before procedures
Patient or population: adults requiring sedation before gastrointestinal endoscopy and bronchoscopy  Settings: hospitals in India, Iran, UK, Portugal and Japan  Intervention: intravenous midazolam  Comparison: placebo
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Intravenous midazolam
Level of sedation on a sedation assessment scale				100 (1 study)	low¹	Participants who were randomized to midazolam were more sedated (the mean score on the Ramsay scale (1 to 6 with higher scores indicating the participant was more sedated) was 2.77 ± 1.19 in the midazolam group and 1.72 ± 0.50 in the placebo group.
Numeric rating of anxiety or rated as anxious				100 (1 study)	low¹	Authors of this trial reported that fewer participants who received midazolam were anxious (3/50 in midazolam group; 15/50 in placebo group) but results of statistical tests were not reported.
Incomplete procedures						No studies reported on this outcome
Anterograde amnesia (defined by number of participants who recalled the procedure)						No studies reported on this outcome
Disinhibition or excitation						No studies reported on this outcome
Discomfort/pain				167 (1 study)	very low²	There was no difference in the number of participants who had discomfort/pain during upper gastrointestinal endoscopy (3/85 in midazolam group; 4/82 in placebo group; P = 0.876). Measured in the trial as 'uncomfortable'
Allergic or anaphylactoid reaction						No studies reported on this outcome
The basis for the assumed risk* is the control group risk across studies or the average risk for pooled data and the control group risk for single studies. The corresponding risk (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; RR: Risk ratio.
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Oral midazolam compared to chloral hydrate for sedation before procedures
Patient or population: children requiring sedation before procedures that require motion control, including echocardiography, lumbar puncture, micturating cystourethrograms and neuroimaging  Settings: Paediatric ICU in USA, emergency departments in USA and Iran and Medical Imaging department in Turkey  Intervention: oral midazolam  Comparison: oral chloral hydrate
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Chloral hydrate	Oral midazolam
Level of sedation on sedation assessment scale				160 (1)	moderate¹	61 participants (76.25%) in the chloral hydrate group were rated at the highest level of sedation compared with 12 (15%) in the midazolam group (scores ranged from 1 = agitated to 4 = eyes closing spontaneously but response to minor stimuli)
Numeric rating of anxiety or rated as anxious	The mean rating of anxiety in D'Agostino 2000 was 2.5 The mean rating of anxiety in Akil 2005 was 49.4	The mean rating of anxiety in the D'Agostino 2000 trial (33 participants) was  1.1 lower  (on a scale of 1 ‐ 5 with higher scores indicating less anxiety) The mean rating of anxiety in the Akil 2005 trial (35 participants) was 1.83 lower (on the Spielberger's Trait Anxiety Inventory)		88  (2)	very low²	We did not conduct meta‐analysis because this outcome was measured differently in the trials (could have been answered by parents or children in Akil 2005).
Incomplete procedures	56 per 1000	226 per 1000  (108 to 474)	RR 4.01   (1.92 to 8.4)	268  (4)	moderate¹
Anterograde amnesia (defined by number of participants who recalled the procedure)						No studies reported on this outcome
Disinhibition or excitation						No studies reported on this outcome
Discomfort						No studies reported on this outcome
Allergic or anaphylactoid reaction						No studies reported on this outcome
The basis for the assumed risk* is the control group risk across studies or the average risk for pooled data and the control group risk for single studies. The corresponding risk (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; RR: Risk ratio.
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Oral midazolam compared to placebo for sedation before procedures
Patient or population: children requiring sedation before micturating cystourethrograms and Kirschner wire removal and adults undergoing flexible sigmoidoscopy  Settings: X‐ray department in Turkey, orthopaedic outpatient department in UK and endoscopy suite in USA  Intervention: oral midazolam  Comparison: placebo
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Midazolam
Level of sedation on a sedation assessment scale				99 (1)	moderate¹	Reported that level of sedation became statistically significantly different at 10 minutes after administration of medication but mean and standard deviations for midazolam and placebo group level of sedation were not reported in the article. Scores ranged from 0 = awake to 3 = asleep, responsive only to direct verbal or physical stimulus.
Numeric rating of anxiety or rated as anxious	The mean rating of anxiety in the Kuganeswaran 1999 trial was 4.2 The mean rating of anxiety in Templeton 2010 was 2.6 The mean rating of anxiety in Akil 2005 was 52.8	The mean anxiety score in Kuganeswaran 1999 (99 participants) was 2.52 lower (minimum score 0, maximum score 10; higher score indicates greater anxiety) The mean anxiety score in Templeton 2010 (42 participants) was 0.90 lower (minimum score 0, maximum score 8; higher score indicate greater anxiety) The mean anxiety score in Akil 2005 (35 participants) was 5.20 lower (on the Spielberger's Trait Anxiety Inventory)		176  (3)	very low²	We did not conduct meta‐analysis because of clinical heterogeneity (children and adults undergoing different procedures).
Incomplete procedure				179 (3 studies)	very low³	We did not conduct meta‐analysis because of clinical heterogeneity (children and adults undergoing different procedures). There was one incomplete procedure in the midazolam group in one of the three trials that reported on this outcome.
Anterograde amnesia (defined by number of participants who recalled the procedure)						No studies reported on this outcome
Disinhibition or excitation						No studies reported on this outcome
Discomfort/Pain				99 (1 study)	moderate¹	Statistically significant reduction in discomfort/pain (mean 2.56 (SD 0.49) in midazolam group; mean 4.62 (SD 1.49) in placebo group; P < 0.005; scores ranged from 0 to 10; higher score indicated more pain).
Allergic or anaphylactoid reaction						No studies reported on this outcome
The basis for the assumed risk* is the control group risk across studies or the average risk for pooled data and the control group risk for single studies. The corresponding risk (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; RR: Risk ratio.
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Intranasal midazolam compared to placebo for sedation before procedures
Patient or population: Children requiring sedation before voiding cystourethrograms and adults undergoing MRI  Settings: Medical imaging departments in Germany and Sweden  Intervention: intranasal midazolam  Comparison: placebo
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of Participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Intranasal midazolam
Level of sedation on a sedation assessment scale				54 (1 study)	moderate¹	Deeper level of sedation was observed in the midazolam group (mean 3.15 (SD 0.36) in midazolam group; mean 2.56 (SD 0.64) in placebo group; P < 0.001). Level of sedation measured 15 minutes after medication by one of the authors using a five‐point sedation scale (1 = agitated, non‐co‐operative; 2 = alert, restless; 3 = calm, eyes spontaneously open; 4 = drowsy, responds to minor stimulation; 5 = asleep, rousable but does not respond to minor stimulation).
Numeric rating of anxiety or rated as anxious				54 (1 study)	moderate¹	Reduction in a numerical rating of anxiety among participants who received midazolam prior to magnetic resonance imaging procedure (mean 17.3 (SD 18.58) in midazolam group; mean 49.3 (SD 29.46) in placebo group; P < 0.001). Numerical rating of anxiety measured 15 minutes after medication on a Visual Analogue Scale of Anxiety comprised an undivided 100‐mm line, with 0 meaning “I am not anxious at all,” and 100 meaning “I am extremely anxious.”
Incomplete procedure	81 per 1000	11 per 1000  (2 to 91)	RR 0.14   (0.02 to 1.12)	149  (2 studies)	low²
Anterograde amnesia (defined by number of participants who recalled the procedure)						No studies reported on this outcome
Disinhibition or excitation						No studies reported on this outcome
Discomfort/pain						No studies reported on this outcome
Allergic or anaphylactoid reaction						No studies reported on this outcome
The basis for the assumed risk* is the control group risk across studies or the average risk for pooled data and the control group risk for single studies. The corresponding risk (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; RR: Risk ratio.
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.

Study	Dose of midazolam	Dose of diazepam
Aktogu 1994	0.06 mg/kg	0.15 mg/kg
Bhalla 2006	5 mg (3 mg if older than 65)	5 mg (3 mg if older than 65)
Bell 1988	2.5 mg or 1 mg for elderly (mean 6.0 ± 2.8)	5 mg or 2 mg for elderly (mean 11.5 ± 6.7)
Bianchi Porro 1988	0.07 mg/kg	0.15 mg/kg
Cole 1983	5 mg bolus with 2.5 ‐ 3.75 mg increments at 30‐ ‐ 60‐second intervals as required. Half doses for elderly	5 mg bolus with 2.5 ‐ 3.75 mg increments at 30‐ ‐ 60‐second intervals as required. Half doses for elderly
Córdova 1992	0.10 mg/kg	0.15 mg/kg
Gilvarry 1990	10 mg	20 mg
Korttila 1985	0.1 mg/kg	0.2 mg/kg
Lavies 1988	2.5 ‐ 7.5 mg	2.5 ‐ 10 mg
Lee 1989	0.07 mg/kg	0.15 mg/kg
Sainpy 1984	0.1 mg under 65 yrs and 0.085 mg over 65 yrs infused in 30 seconds	0.2 mg under 65 yrs and 0.15 mg over 65 yrs infused in 30 seconds
Takrouri 1988	mean 5.8 mg	mean 5 mg
Tolia 1990	0.1 ‐ 0.15 mg/kg	0.2 ‐ 0.4 mg/kg
Whitwam 1983	0.07 mg/kg	0.15 mg/kg

Study	Oral midazolam	Oral chloral hydrate
Akil 2005	0.6 mg/kg (max 15 mg)	25 mg/kg (max 0.5 g)
D'Agostino 2000	0.5 mg/kg (max 10 mg)	75 mg/kg (max 2 g)
Derakhshanfar 2013	0.5 mg/kg midazolam. Additional dose up to 8 mg for inadequate sedation	80 mg/kg chloral hydrate followed by further dose 20 mg/kg if required 20 minutes later
Wheeler 2001	0.5 mg/kg midazolam; second dose 30 minutes after the initial dose, either 0.25 mg/kg	75 mg/kg oral chloral hydrate; 30 minutes after the initial dose 25 mg/kg

First author	Journal/Conference Proceedings etc	Year

RCT	Relevant participants [age >16, type of procedures (diagnostic, therapeutic), medical specialty (surgical, non‐surgical)]	Relevant interventions [comparisons between midazolam with placebo, midazolam with other drugs, timings and doses of midazolam administration]	Relevant outcomes (Midazolam's efficacy and level of sedation) (duration of sedation, onset time of sedation, offset time of sedation, retrograde and anterograde amnesia duration, oversedation, disinhibition/excitation, quality of recovery, discomfort, pain, allergic or anaphylactoid reactions, use of flumazenil)
Yes/No/Unclear	Yes/No/Unclear	Yes/No/Unclear	Yes/No/Unclear

Code each paper	Author(s)	Journal/Conference Proceedings etc	Year
A
B
C

Outcomes	Reported in paper (circle)	Subgroups	Information available in paper (circle)
Primary outcome effective sedation	Yes/No	Age (<16)	Yes/No
		Type of procedure ‐ diagnostic	Yes/No
		Type of procedure ‐ therapeutic	Yes/No
		Surgical specialty	Yes/No
Secondary outcomes	Yes/No
Outcome 1 duration of sedation	Yes/No
Outcome 2 onset time of sedation	Yes/No
Outcome 3 offset time of sedation	Yes/No
Outcome 4 retrograde and anterograde amnesia	Yes/No
Outcome 5 oversedation	Yes/No
Outcome 6 discomfort, pain	Yes/No
Outcome 7 disinhibition/excitation	Yes/No
Outcome 8 quality of recovery	Yes/No
Outcome 9 allergic or anaphylactoid reactions	Yes/No
Outcome 10 need for sedation reversal	Yes/No

Allocation concealment Process used to prevent foreknowledge of group assignment in a RCT, which should be seen as distinct from blinding
State here method used to conceal allocation and reasons for grading	Grade (circle)
Comment on allocation concealment by review authors or included study quote concerning allocation:	Low risk of bias
	High risk of bias
	Unclear

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of participants rated as 'anxious'	2	175	Risk Ratio (M‐H, Fixed, 95% CI)	0.80 [0.39, 1.62]
2 Participant co‐operation (not co‐operative)	4	316	Risk Ratio (M‐H, Random, 95% CI)	0.94 [0.48, 1.84]
3 Participant satisfaction	2	91	Mean Difference (IV, Fixed, 95% CI)	2.17 [‐0.51, 4.85]
4 Proceduralist satisfaction	2	91	Mean Difference (IV, Fixed, 95% CI)	1.09 [‐10.43, 12.60]
5 Difficulty performing procedures (rated as difficult to perform procedure)	3	277	Risk Ratio (M‐H, Fixed, 95% CI)	0.66 [0.41, 1.07]
6 Onset of sedation	2	140	Mean Difference (IV, Random, 95% CI)	‐1.80 [‐3.76, 0.16]
7 Anterograde amnesia (defined as the number of participants who recalled procedure)	9	587	Risk Ratio (M‐H, Random, 95% CI)	0.45 [0.30, 0.66]
8 Quality of recovery (rated as delayed recovery)	3	166	Risk Ratio (M‐H, Random, 95% CI)	0.72 [0.08, 6.63]
9 Discomfort/Pain	5	415	Risk Ratio (M‐H, Random, 95% CI)	0.60 [0.24, 1.49]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Vital signs (mean lowest oxygen saturation)	2	207	Mean Difference (IV, Fixed, 95% CI)	‐1.50 [‐1.77, ‐1.23]
2 Vital signs (oxygen desaturation)	3	447	Risk Ratio (M‐H, Fixed, 95% CI)	1.12 [0.79, 1.58]
3 Vital signs (hypotension)	2	140	Risk Ratio (M‐H, Fixed, 95% CI)	0.43 [0.07, 2.78]
4 Vital signs (tachycardia)	2	207	Risk Ratio (M‐H, Fixed, 95% CI)	1.46 [0.74, 2.87]
5 Vital signs (hypertension)	2	207	Risk Ratio (M‐H, Fixed, 95% CI)	1.15 [0.43, 3.13]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incomplete procedure	4	268	Risk Ratio (M‐H, Fixed, 95% CI)	4.01 [1.92, 8.40]
2 Onset time of sedation	2	200	Mean Difference (IV, Random, 95% CI)	8.37 [‐3.49, 20.23]
3 Offset time of sedation	3	233	Mean Difference (IV, Random, 95% CI)	‐12.87 [‐63.24, 37.50]

For continuous data (with a separate copy for each relevant subgroup)
Code of paper	Outcomes	Unit of measurement	Intervention group		Control group		Details if outcome only described in text
Code of paper	Outcomes	Unit of measurement	n	Mean (SD)	n	Mean (SD)
	Primary Outcome effective sedation
	Outcome 1 duration of sedation
	Outcome 2 onset time of sedation
	Outcome 3 offset time of sedation
	Outcome 4 retrograde and anterograde amnesia
	Outcome 5 oversedation
	Outcome 6 discomfort, pain
	Outcome 7 disinhibition/excitation
	Outcome 8 quality of recovery
	Outcome 9 allergic or anaphylactoid reactions
	Outcome 10 need for sedation reversal

For dichotomous data (with a separate copy for each relevant subgroup)
Code of paper	Outcomes	Intervention group (n) n = number of participants, not number of events	Control group (n) n = number of participants, not number of events
	Primary outcome effective sedation
	Outcome 1 duration of sedation
	Outcome 2 onset time of sedation
	Outcome 3 offset time of sedation
	Outcome 4 retrograde and anterograde amnesia
	Outcome 5 oversedation
	Outcome 6 discomfort, pain
	Outcome 7 disinhibition/excitation
	Outcome 8 quality of recovery
	Outcome 9 allergic or anaphylactoid reactions
	Outcome 10 need for sedation reversal

Did this report include any references to published reports of potentially eligible trials not already identified for this review?
First author	Journal/Conference	Year of publication

Did this report include any references to unpublished data from potentially eligible trials not already identified for this review? If yes, give list contact name and details

Trial characteristics
	Further details
Single centre/multicenter
Country/Countries
How was participant eligibility defined?
How many people were randomized?
Number of participants in each intervention group
Number of participants who received intended treatment
Number of participants who were analysed
Drug treatment(s) used
Dose/frequency of administration
Duration of treatment (state weeks/months, etc, if cross‐over trial give length of time in each arm)
Median (range) length of follow up reported in this paper (state weeks, months or years or if not
Time‐points when measurements were taken during the study
Time‐points reported in the study
Time‐points you are using in RevMan 5.0
Trial design (e.g. parallel/cross‐over*)
Other

Random sequence generation
State here method used to generate allocation and reasons for grading	Grade (circle)
Comment on allocation by review authors or included study quote concerning allocation:	Low risk of bias (Random)
	High risk of bias (e.g. alternate)
	Unclear

Blinding
Participant	High/ Low/ Unclear Risk
Outcome assessor	High/ Low/ Unclear Risk
Other (please specify)	High/ Low/ Unclear Risk
Comment on blinding by review authors or included study quote concerning allocation:
Intention‐to‐treat An intention‐to‐treat analysis is one in which all the participants in a trial are analysed according to the intervention to which they were allocated, whether they received it or not.
All participants entering trial
15% or fewer excluded
More than 15% excluded
Not analysed as ‘intention‐to‐treat’
Unclear

Methods	Parallel‐group single‐centre randomized controlled trial conducted in Turkey
Participants	53 children requiring micturating cystourethrogram with sedation (39 girls, 14 boys; mean age of 5.8 ± 3.5 years)
Interventions	Oral midazolam 0.6 mg/kg (max 15 mg) versus chloral hydrate 25 mg/kg (max 0.5 g) and placebo (saline) 15 ‐ 30 minutes before procedure
Outcomes	Measured during procedures: Incomplete procedures Anxiety (measured during the procedure using Spielberger's Trait Anxiety Inventory) Participant co‐operation (measured during the procedure using Houpt behavioural scale; range 1 ‐ 6 with higher scores indicating better co‐operation) Tolerance of procedure (measured during the procedure using Frankl behaviour rating score; range 1 ‐ 4 with higher scores indicating better tolerance) Duration of sedation Onset of sedation Oversedation
Notes	Conflicts of interest or funding sources were not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Not blinded
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Outcomes included in this review were not blinded ‐ only assessment of image quality was performed by a blinded outcome assessor
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No withdrawals
Selective reporting (reporting bias)	Unclear risk	No clear evidence that measured outcomes were not reported (trial protocols were not sought for confirmation)
Other bias	Low risk	None expected

Methods	Single‐centre cross‐over randomized controlled trial conducted in Italy
Participants	23 adults undergoing upper gastrointestinal endoscopy (14 men and 9 women; mean weight 60.7 kilograms; age range 20 to 48 years; mean age 32.5)
Interventions	1) Intravenous midazolam 0.07 mg/kg 2) Intravenous Diazepam 0.15 mg/kg
Outcomes	Measured during the procedure: Vital signs (blood pressure 5 minutes after sedation) Vital signs (heart rate 5 minutes after sedation Level of sedation using a sedation assessment scale 4 Participant co‐operation Measured 2 hours after the procedure: Quality of recovery Measured 24 hours after the procedure: Anterograde amnesia (defined by number of participants who recalled the procedure)
Notes	At least 30 days between procedures We were unable to locate contact details of the authors to access pre‐cross‐over data Conflicts of interest or funding sources were not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Drugs prepared and administered by physician not performing endoscopy or assessments
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Drugs prepared and administered by physician not performing endoscopy or assessments
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No withdrawals
Selective reporting (reporting bias)	Unclear risk	No clear evidence that measured outcomes were not reported (trial protocols were not sought for confirmation)
Other bias	Low risk	None expected

Methods	Parallel‐group randomized controlled trial conducted in 2 sites in the USA
Participants	40 adult participants (American Society of Anesthesiology physical classification status I ‐ II) undergoing upper gastrointestinal endoscopy for the first time (aged 18 ‐ 70 years; authors reported that the groups were comparable in all parameters evaluated: age, sex, weight, race (white/non‐white), psychoactive drug history, duration of procedure, injection speed, data collection site)
Interventions	1) Intravenous midazolam 5 mg bolus with 2.5 ‐ 3.75 mg increments at 30‐ ‐ 60‐second intervals as required. Half doses for elderly 2) Same dose for diazepam
Outcomes	Measured during the procedure: Onset of sedation Vital signs (change in heart rate) Vital signs (change in respiration rate) Vital signs (change in diastolic blood pressure) Discomfort/pain (measured in the study as absolute number with pain) Measured after the procedure: Proceduralist satisfaction Measured the day after the procedure: Participant satisfaction Quality of recovery (number of participants reporting unusual sensations the day after the procedure) Anterograde amnesia (numerical rating from 0 to 100 with a lower number indicating greater amnesia)
Notes	Authors declared that Hoffman‐La Riche Inc. provided financial assistance for the study, but did not disclose the role of the funder in the study
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Both the participants and the endoscopist were masked to treatment allocation
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Both the participant and the endoscopist were masked to treatment allocation
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No withdrawals
Selective reporting (reporting bias)	Unclear risk	No clear evidence that measured outcomes were not reported (trial protocols were not sought for confirmation)
Other bias	Low risk	None expected

Methods	Single‐centre parallel‐group randomized controlled trial conducted in Spain
Participants	32 consecutive adults undergoing cardioversion for supraventricular arrhythmia in an emergency department
Interventions	Randomized to 1 of 4 treatment arms (all intravenous administration): 1) 0.2 mg/kg midazolam 2) 0.2 mg/kg etomidate 3) 1.5 mg/kg propofol 4) 0.2 mg/kg midazolam and flumazenil 0.5 mg bolus followed by 0.5 mg infusion for 1 hour If induction was not achieved within 3 ‐ 5 minutes, supplementary doses of etomidate (0.05 mg/kg), midazolam (0.05 mg/kg) or propofol (0.5 mg/kg) were injected at 1‐minute intervals until the desired effect was obtained
Outcomes	Measured during the procedure: Level of sedation using a sedation assessment scale (Ramsay scale) Vital signs (systolic blood pressure after sedation) Vital signs (desaturation < 90%) Duration of sedation Onset of sedation Measured after the procedure: Offset time of sedation
Notes	Only median values reported for continuous outcomes (skewed distributions because of small sample size). We did not include the midazolam and flumazenil group in our review because combinations of medications with midazolam were excluded.  The authors declared that no outside funding or support was received for the study. No other conflicts of interest were reported.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random‐number table
Allocation concealment (selection bias)	Low risk	Sealed envelopes
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Not blinded
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Not blinded
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No withdrawals
Selective reporting (reporting bias)	Unclear risk	No clear evidence that measured outcomes were not reported (trial protocols were not sought for confirmation)
Other bias	Low risk	None expected

PERMALINK

Midazolam for sedation before procedures

Aaron Conway

John Rolley

Joanna R Sutherland

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

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

'Summary of findings' table

Results

Description of studies

Results of the search

1.

Included studies

Excluded studies

Studies awaiting classification

Ongoing 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. Intravenous midazolam compared to diazepam for sedation before procedures.

Summary of findings 2. Intravenous midazolam compared to placebo for sedation before procedures.

Summary of findings 3. Oral midazolam compared to chloral hydrate for sedation before procedures.

Summary of findings 4. Oral midazolam compared to placebo for sedation before procedures.

Summary of findings 5. Intranasal midazolam compared to placebo for sedation before procedures.

Intravenous midazolam versus diazepam (comparison 1, outcomes 1.1 to 1.6 and 2.1 to 2.10)

1. Intravenous midazolam versus diazepam doses.

Primary outcomes

Outcome 1.1. Level of sedation on a sedation assessment scale

Outcome 1.2. Numeric rating scale of anxiety or number of participants rated as anxious

1.1. Analysis.

Outcome 1.3. Vital signs

Outcome 1.4. Tolerance of procedure or participant co‐operation

1.2. Analysis.

Outcome 1.5. Participant or proceduralist satisfaction

1.3. Analysis.

1.4. Analysis.

Outcome 1.6. Incomplete procedures/difficulty performing procedures

Methods	Single‐centre parallel group randomized controlled trial conducted in USA
Participants	120 adults undergoing upper gastrointestinal endoscopy (mean age: 58 years in placebo group, 55 in diazepam group, 50 in midazolam group; 55% men in the placebo group, 70% men in the diazepam group, 48% men in the midazolam group)
Interventions	Intravenous administration of: 1) Midazolam 2.5 ‐ 7.5 mg 2) Diazepam 2.5 ‐ 10 mg 3) Placebo
Outcomes	Measured during the procedure: Vital signs Tolerance of procedure
Notes	Conflicts of interest or funding sources were not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Random‐number sequence used but "The study was continued until 40 patients were included in each group"
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Participant was unaware of medication
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No withdrawals
Selective reporting (reporting bias)	Unclear risk	No clear evidence that measured outcomes were not reported (trial protocols were not sought for confirmation)
Other bias	Low risk	None expected

Methods	Multicentre (2 sites) parallel‐group randomized controlled trial conducted in Portugal
Participants	100 adults undergoing fibreoptic bronchoscopy (Mean age was 56 ± 14 years (range 18 ‐ 79 years); 66% were men
Interventions	0.05 mg/kg intravenous midazolam Placebo
Outcomes	Measured during the procedure: Level of sedation on a sedation assessment scale Vitals signs Disinhibition/excitation (measured in the trial as 'agitation') Sedation reversal Measured 1 hour after the procedure: Participant satisfaction (willingness to undergo another procedure with the same medication)
Notes	Conflicts of interest or funding sources were not reported
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Did state it was double‐blind but no specific information provided about how this was achieved
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No withdrawals
Selective reporting (reporting bias)	Unclear risk	No clear evidence that measured outcomes were not reported (trial protocols were not sought for confirmation)
Other bias	Low risk	None expected

Study	Reason for exclusion
Bonta 2003	Both groups could receive midazolam if initial sedation not effective
Brouillette 1989	Meperidine administered in both groups
Dere 2010	Fentanyl used in both groups
Mui 2005	Additional IV midazolam used in placebo group
Muttu 2005	Intra‐procedural sedation used
Nascimento 2007	Meperidine used in all groups if sedation inadequate
Ristikankare 1999	Flumazenil for all participants
Ristikankare 2000a	Flumazenil for all participants
Ristikankare 2000b	Flumazenil for all participants
Sajedi 2006	Fentanyl used in both groups
Salmon 1992	Temazepam for all groups
Sandler 1992	Quasi‐RCT
Sherry 1989	Nitrous oxide used on both groups
Tamayo 1993	Fentanyl used in both groups
Tesoro 2007	Quasi‐RCT
Uygur‐Bayramiçli 2002	Flumazenil at end of procedure
Weinstein 2010	Post hoc analysis
Yildirim 2006	Placebo administered in different route of administration

Methods	RCT
Participants	Endoscopy
Interventions	Intravenous midazolam versus diazepam; Unclear whether intra‐procedural sedation was used
Outcomes	Full details not known
Notes	Could not locate current contact details for authors

Methods	RCT
Participants	GI endoscopy
Interventions	Intravenous midazolam versus diazepam
Outcomes	Full details not known
Notes	Article published in Spanish. Awaiting data extraction from a second person

Methods	RCT
Participants	Bronchoscopy
Interventions	Intramuscular midazolam versus hydrocodonum
Outcomes	Full details not known
Notes	Article published in German. Awaiting data extraction from a second person

Methods	RCT
Participants	MRI scan
Interventions	Intranasal midazolam versus placebo
Outcomes	Full details not known. 4/27 randomized to placebo did not complete procedure, whereas all participants randomized to midazolam completed the procedure.
Notes	Article published in German. Awaiting data extraction from a second person

Methods	RCT
Participants	Children undergoing bone marrow aspiration/biopsy
Interventions	Oral midazolam 1 mg/kg Oral ketamine 10 mg/kg
Outcomes	Full details not known
Notes	The design of this trial is unclear because the abstract reports that only 29 participants were enrolled but 89 procedures are included. It also states it is a non‐cross‐over study. We could not locate current contact details for authors

Methods	RCT
Participants	Children undergoing echocardiography
Interventions	Oral midazolam versus triclofos sodium
Outcomes	Full details not known. Children who received midazolam had a quicker onset of sedation and recovery from sedation
Notes	Article published in French. Awaiting data extraction by 2 people

Methods	RCT
Participants	Children requiring laceration repair
Interventions	Intranasal midazolam versus placebo versus no sedation
Outcomes	Placebo group results were reported with control
Notes	Contacted authors by email on 12th December 2014 for clarification about just the results of the placebo group, but did not receive a response