Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Midazolam use in children undergoing ketamine sedation to reduce emergence reaction

Three Part Question

In [children undergoing ketamine sedation] does the [administration of a benzodiazepine] reduce the [incidence of emergence phenomenon]?

Clinical Scenario

A 13-year-old girl with a shoulder dislocation presents to your ED. You plan to reduce it using ketamine sedation but know that ketamine can cause emergence phenomenon in this age group. You wonder if you should use prophylactic midazolam to reduce the risk of this side effect.

Search Strategy

PUBMED, Elsevier, and Cochrane library database searches.

PUBMED search:
(0-18 years old OR child* OR paed* OR Pediat* OR adolescent OR infant) AND (sedat*) AND (ketamine OR ketalar OR ketanest OR ketaset) AND (midaz* OR benz* OR loraz*) AND (emergence OR psychosis OR side effect)

Elsevier Search:
(child OR paediatric OR children) AND (sedation) AND (ketamine) AND (emergence)

Cochrane library Search:
(paediatric OR child):ti,ab,kw AND (sedation):ti,ab,kw AND (emergence):ti,ab,kw AND (ketamine OR ketalar OR ketaset):ti,ab,kw AND (midazolam):ti,ab,kw

Search Outcome

103 results of which 95 were irrelevant and 2 were discounted as they were not in English. 6 papers were relevant and are shown in the table below.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Sherwin et al
March 2000
104 children aged 12 months to 15 years undergoing painful procedures in the ED were randomised to receive either 1.5mg/kg IV ketamine and atropine or 1.5mg/kg IV ketamine and atropine 0.01mg/kg plus 0.05mg/kg midazolam after 2 minutesProspective randomised double-blind placebo-controlled clinical trial (Level 1 evidence)Recovery agitation No difference1. Mixed definition of emergence including both negative phenomena (for example nightmares, reported by one child although child appeared calm) and positive phenomena (hallucinations which were not unpleasant). These definitions are difficult to apply to clinical practice. 2. Low power for low incidence complication.
Wathen et al
Dec 2000
266 children aged 4.5 months to 16 years undergoing fracture reduction or laceration repair in the ED were randomised to receive either 1mg/kg IV ketamine, 5µg/kg glycopyrrolate and 0.1mg/kg midazolam or 1mg/kg ketamine, 5µg/kg glycopyrrolate and normal salineProspective, double-blind, randomised controlled study (level 1 evidence)Distress (Observation Score of Behavioral Distress-Revised score) No difference1. Wide age range. 2. Low power for low incidence complication. 3. Emergence included pleasant dreams which are not of clinical significance 4. Subgroup analysis for age range does not specify number of patients within these groups 5. Wide confidence intervals in subgroup analysis.
Immediate emergenceNo difference in immediate emergence, immediate emergence agitation severity or immediate ‘significant’ emergence (defined as nightmares, hallucinations or severe agitation).
Delayed emergence (at home)No difference in delayed emergence or delayed significant emergence (defined as nightmares or hallucinations).
Subgroup analysis to determine age at which emergence is greatestMore agitation in over 10 year olds when midazolam given with ketamine than ketamine (ketamine 5.7% vs ketamine with midazolam 35.7%; difference –30.0, 95% CI –10.7 to –49.3)
Dilli et al
March 2008
99 children aged 2-14 years undergoing lumbar puncture in the ED for suspected meningitis were randomised to receive either 1mg/kg IV ketamine or 1mg/kg IV ketamine plus 0.1mg/kg IV midazolam. 0.5mg/kg IV ketamine was added if conscious sedation was not achieved within 5 minutes. All children also received 0.01mg/kg atropine.Prospective single- blinded randomised control trial (Level 2 evidence)EmergenceNo difference1. Convenience sampling so possible selection bias. 2. Small sample size. 3. Low power for low incidence complication. 4. Large spread of ages. 5. Sedationists not blinded. 6. Not truly measuring emergence phenomena: defined ‘nightmares’ as the children describing unpleasant visual imagery after recovery. 7. Sample were children with meningitis which may alter threshold for emergence
Brecelj et al
Dec 2013
201 children aged 1-19 years undergoing procedural sedation for elective endoscopy were randomised to either 0.75mg/kg IV ketamine with 0.25mg/kg increments as required (max 1.5mg/kg) or the same dose of ketamine plus 0.1mg/kg midazolamProspective randomised unblinded control trial (Level 2 evidence)Early emergence Fewer episodes when midazolam was given (10 vs 2 patients, p=0.02)1. Low power for low incidence complication 2. Investigators not blinded. 3. No description of severity of emergence reaction, described merely as irritability and/or unpleasant dreams. 4. Parents telephoned more than 1 month after the episode introducing potential recall bias.
Delayed emergence (at home)No difference
Michalczyk et al
March 2013
25 children aged 11 months to 16 years undergoing elective lumbar puncture +/- bone marrow biopsy underwent 84 sedation episodes randomised to receive either 1mg/kg IV ketamine plus 0.5mg/kg increments as needed, or, if previously sedated, the historic IV ketamine dose; or IV ketamine (dosing regimen as above) plus 0.1mg/kg IV midazolam or non-ketamine sedative of either fentanyl/propofol or fentanyl/midazolam (doses not described)Prospective randomised, unblinded trial. (Level 2 evidence)Emergence4 patients had overt emergence/delirium, all of whom were in the ketamine only group (n=39). 1. Small study size. 2. Low power for low incidence complication. 3. Varying dosing regimens for ketamine. 4. Study was designed to investigate CSF opening pressures: emergence was neither a primary nor secondary outcome of the study but was noted as an adverse event. 5. Sedationist not blinded. 6. Doses of ketamine not standardised
Erk et al
Oct 2007
100 children aged 3-10 years old undergoing adenotonsillectomy under 65% nitrous oxide general anaesthesia received either intramuscular ketamine 7mg/kg and atropine 0.015mg/kg or intramuscular ketamine 7mg/kg plus atropine 0.015 7mg/kg and midazolam 0.1mg/kg as a pre-anaesthetic.Prospective double-blind randomised control trial (Level 1 evidence)Early emergence (within the first 4 hours) ‘Moderate’ early emergence reactions (anxiety, agitation or delirium) were significantly lower when midazolam was given (3.6% vs 25%, p<0.05)1. Young age group. 2. Rates of emergence phenomena higher than in other studies. 3. Children also received general anaesthetic which may confound results. 4. Documented that IV ketamine may have been given for maintenance but no note of which patients received IV ketamine and no correlation between this and outcomes.
Incidence of early severe emergence reactions (screaming with shrieks or cries or uncooperative)None seen in either group
Delayed emergence (up to 60 days later)No difference


Emergence reactions, such as agitation, delirium and unpleasant hallucinations, are a recognised side effect of ketamine paediatric procedural sedation. Although many physicians advocate the prophylactic use of midazolam to minimise emergence phenomena, a recent prospective, multicentre, observational cohort study of over 6000 children showed that the risk of serious adverse events and significant interventions is increased when midazolam is given with ketamine (Bhatt et al 2017. Only six studies have been published investigating the use of prophylactic midazolam in paediatric ketamine sedation. Two well designed randomised double-blinded controlled studies by Sherwin et al (2000) and Wathen et al (2000) conclude that there is no significant difference in the rate of emergence phenomenon between ketamine paediatric sedation and ketamine plus midazolam procedural sedation. In subanalysis, Wathen et al describe a higher incidence of emergence reaction in children over ten years old in the midazolam and ketamine arm compared to ketamine alone, though confidence intervals are wide. A 2008 single-blinded study by Dilli et al also found no significant difference in emergence between ketamine and ketamine plus midazolam paediatric procedural sedation. Three studies (Brecelj et al, 2013; Michaelczyk et al, 2013; and Erk et al, 2007) describe lower rates of emergence when midazolam is given with ketamine. All three, however, have significant limitations. The sedationist in the first two studies was not blinded to the study arm, potentially introducing bias; the second study was not designed to assess emergence as an outcome and merely described it as a noted side-effect; and children in the third study, randomised to receive ketamine or ketamine-midazolam combination as a pre-anaesthetic prior to nitrous oxide general anaesthetic, had higher rates of emergence than otherwise described in the literature, probably related to the fact that emergence was assessed during recovery from the anaesthetic. All six studies were poorly powered for a low incidence complication with further difficulty in their interpretation due to nonconformity in the definition of emergence. There are no validated emergence scales; some studies merely describe emergence as a binary outcome, whilst others assess levels of emergence (including mild hallucinations that are not distressing to the patient or family or cause difficulty recovering the child from sedation) using unvalidated scales.

Clinical Bottom Line

The increased risk of adverse events and the lack of robust evidence to the contrary means prophylactic midazolam is not recommended to prevent emergence phenomena in children undergoing ketamine sedation.


  1. Sherwin TS, Green SM, Khan A et al Does adjunctive midazolam reduce recovery agitation after ketamine sedation for pediatric procedures? A randomized double blind placebo controlled trial Ann Emerge Med 2000;35:229-238.
  2. Wathen JE, Roback MG, Mackenzie T, Bothner JP Does midazolam alter the clinical effects of intravenous ketamine sedation in children? A double-blind, randomized, controlled, emergency department trial Ann Emerg Med 2000 Dec;36(6):579-88
  3. Dilli D, Dallar Y, Sorgui NH Intravenous ketamine plus midazolam vs. intravenous ketamine for sedation in lumbar puncture: a randomized controlled trial Indian Pediatr 2008 Nov;45(11):899-904
  4. Brecelj J, Trop TK, Orel R Ketamine with and without midazolam for gastrointestinal endoscopies in children J Pediatr Gastroenterol Nutr 2012Jun;54(6):748-52
  5. Michalczyk K, Sullivan JE, Berkenbosch JW Pretreatment with midazolam blunts the rise in intracranial pressure associated with ketamine sedation for lumbarpuncture in children Pediatr Crit Care Med 2013 Mar;14(3):e149-55
  6. Erk G, Ornek D, Dönmez NF, Ta?pinar V The use of ketamine or ketamine-midazolam for adenotonsillectomy Int J Pediatr Otorhinolaryngol 2007 Jun;71(6):937-41