Design: Systematic review of 20 studies with ≥250 children each
Population: 58,752 children receiving IV ketamine for ED procedural sedation
Primary outcome: Frequency of sentinel and serious adverse events
Effect: No deaths; sentinel events 0.0041% (1 in 24,390); serious events 0.33%

A systematic review in Annals of Emergency Medicine pooled 20 studies, each with at least 250 children, covering 58,752 intravenous ketamine sedations for procedures in the emergency department.

There were no deaths or permanent harms. Sentinel events — intubation, neuromuscular blockade, chest compressions, aspiration or vasoactive drugs — occurred in about 1 in 24,390 sedations. Broader serious adverse events, including bag–mask ventilation, an oral airway or unplanned admission, occurred in 0.33%. Three factors modestly raised that risk: age 10 or over, an upper respiratory infection, and co-administered opioids. ASA physical status and co-administered benzodiazepines did not.

This is a large body of evidence on paediatric ED ketamine sedation, and it supports what many departments already do. It also gives clinicians concrete numbers to use when consenting families.

Ketamine remains a sedation that needs airway-competent staff and monitoring, because the rare event still happens. Be more cautious in older children, those with a cold, and when an opioid has already been given.

Use intravenous ketamine for paediatric procedural sedation with airway-competent staff and monitoring
Tell families serious airway events are rare, about 1 in 300, and sentinel events far rarer
Take extra care in children aged 10 or over and those with an upper respiratory infection
Avoid adding opioids where possible; they modestly raised the risk of serious events
Do not exclude a child on ASA class alone; it did not predict adverse events

Why it matters

It gives families and clinicians real numbers for a sedation decision made many times a day.

The statistics, in plain English

A rate of 0.0041% means about 4 sentinel events per 100,000 sedations; the interval (0.0010% to 0.016%) is wide because so few occurred. The predictors added roughly one extra serious event per 100 to 160 sedations.