Paediatric esketamine evidence points to context-specific doses, not universal safety threshold

A targeted review of paediatric perioperative esketamine finds that intravenous doses of 0.2–0.5 mg·kg−1 are the most extensively studied for preventing emergence delirium or agitation in selected settings, particularly adenotonsillectomy under sevoflurane anaesthesia. The analysis does not establish a universal therapeutic window: benefits varied by endpoint, route, timing, procedure, anaesthetic technique, co-medication and outcome definition, while the upper boundary for either efficacy or safety remains undefined. Instead, the authors identify a context-dependent studied range and argue that direct dose comparisons, paediatric exposure–response modelling and longer safety follow-up are needed before indication-specific limits can be defined.

The review used a targeted search of PubMed/MEDLINE and reference lists to identify evidence published from January 2006 to 10 August 2026. Of 166 records retrieved, the authors included 14 peer-reviewed paediatric studies and five directly relevant systematic reviews or meta-analyses. Two reviewers screened records independently and resolved disagreements by discussion, but the team did not perform a formal risk-of-bias assessment, certainty grading, publication-bias analysis or quantitative synthesis. It also did not search trial registries, Embase, Web of Science, Scopus, the Cochrane Library or non-English databases, so the conclusions apply to the evidence captured by this strategy rather than to every available study.

Most randomised trials and meta-analyses indicate that esketamine can reduce emergence delirium or emergence agitation, with the clearest evidence emerging from children undergoing adenotonsillectomy under sevoflurane anaesthesia. One four-arm trial assigned 164 children to placebo or intravenous esketamine doses of 0.1, 0.3 or 0.5 mg·kg−1 before induction. Emergence agitation occurred in 62.5%, 56.8%, 25.6% and 18.9% of the respective groups, indicating progressively lower observed incidence as the assigned dose increased within that procedure. The 0.5 mg·kg−1 group had a longer time to tracheal tube removal, but the reduction in agitation did not reverse. Because the trial did not test higher doses or establish a concentration–effect relationship, it supports a dose-related signal up to 0.5 mg·kg−1 without locating an upper efficacy or toxicity boundary.

Other randomised studies reported similar directional findings with a 0.2 mg·kg−1 intravenous bolus given at the end of anaesthesia or during surgery, while pooled analyses generally found fewer emergence outcomes after esketamine. Those pooled results, however, combine studies with different procedures, doses, administration schedules, comparators and assessment tools, making them useful for estimating an average preventive effect rather than defining a precise dose threshold. A prospective observational study of 230 preschool children reached the opposite conclusion, associating esketamine exposure at a mean dose of 0.46 mg·kg−1 with more emergence delirium and a longer stay in the post-anaesthesia care unit. Because treatment was not randomised, clinicians may have selected esketamine for children with different underlying risks, and the study compared no prespecified dose levels, it cannot show that a particular dose caused the excess events or identify a dose–response inflection point.

The apparent contradictions are better understood as differences between endpoints than as evidence for a single reversal in drug response. Esketamine is also evaluated for procedural sedation, attenuation of propofol injection pain and postoperative analgesia, each of which may require a different depth and duration of drug action. A dose that suppresses a brief response during gastroscope insertion, for example, cannot define the lower threshold for sustained postoperative analgesia. Likewise, a regimen producing procedural sedation cannot establish the upper limit for emergence prophylaxis. The review therefore treats the emergence results, injection-pain dose estimates and analgesic studies as separate bodies of evidence rather than evidence for one common effect.

Measurement also contributes to the apparent variability. Emergence delirium and emergence agitation overlap but are not interchangeable: the Pediatric Anesthesia Emergence Delirium scale was designed to assess delirium, whereas the Watcha and Cravero instruments place more emphasis on agitation. Pain, fear, residual sedation and dissociative behaviour can influence scores, and studies differ in their thresholds and assessment times. These differences limit direct comparison of incidence and treatment effects. Anaesthetic context may matter as well. Sevoflurane exposure, preschool age, preoperative anxiety and postoperative pain are established correlates of emergence delirium, while procedures such as endoscopy, bronchoscopy, strabismus surgery, hypospadias repair and neurosurgery have different baseline risks and anaesthetic conditions. Results from one setting therefore cannot be assumed to estimate the same effect in another.

Route of administration creates another unresolved layer. Adult studies estimate that intranasal esketamine has an absolute bioavailability of approximately 50%–54%, but that figure alone cannot establish that 1.0 mg·kg−1 intranasally is equivalent to 0.5 mg·kg−1 intravenally. Intranasal delivery produces variable absorption, a swallowed fraction, first-pass metabolism and different timing of peak concentration, while an intravenous bolus creates a more immediate systemic peak. The cited adult nasal-spray models were developed in healthy volunteers and adults with depression, not perioperative children, and the targeted search found no validated paediatric perioperative population pharmacokinetic model. Intravenous and intranasal regimens consequently remain distinct evidence categories rather than being converted through a simple bioavailability calculation.

Developmental pharmacology adds uncertainty but does not supply a solution. Maturation of body composition, hepatic metabolic capacity and the formation of active metabolites such as noresketamine may alter distribution and clearance, while developmental changes in N-methyl-D-aspartate receptor expression provide biological plausibility for age-dependent responses. None of these factors has yet been quantified in a paediatric perioperative exposure model, however, and apparent age effects may instead reflect baseline emergence risk, communication ability, vascular access or procedure mix. Adult data, racemic ketamine models and animal studies can therefore generate hypotheses but cannot justify empirical age-based dose escalation or reduction. Likewise, high-dose neonatal mouse studies and prolonged exposure in developing non-human primates do not define a neurotoxicity threshold for a single low-dose perioperative bolus in children, although they support the need for longer clinical surveillance.

Clinical studies have generally concentrated on short-term outcomes such as heart-rate and blood-pressure changes, respiratory events, oxygen desaturation, nausea and vomiting, abnormal behaviour, salivation, recovery time, tracheal tube removal and post-anaesthesia care unit stay. In the four-dose adenotonsillectomy trial, heart rate and mean arterial pressure differed among groups, and 0.5 mg·kg−1 prolonged tube removal, although post-anaesthesia care unit duration and postoperative complications did not differ significantly. Most trials were designed to test efficacy rather than detect uncommon harms, so the absence of a statistically significant safety difference does not demonstrate an upper boundary. Reviews combining dexmedetomidine and esketamine are also encouraging in some settings, but current comparisons do not include factorial designs or prespecified interaction tests; they therefore do not establish pharmacological synergy, a uniform half-dose strategy or a higher safety ceiling.

The review’s central conclusion is that evidence supports context-specific studied ranges rather than a validated universal therapeutic window. The 0.2–0.5 mg·kg−1 intravenous interval is best supported for selected emergence-prevention settings and should not be extrapolated automatically to analgesia, procedural sedation, intranasal administration or other operations. Future trials should randomise prespecified doses and timing within the same procedure and anaesthetic protocol, report emergence delirium, emergence agitation, pain, extubation time, recovery duration and adverse effects separately, and preserve the thresholds and terminology of validated assessment instruments. Population models should link weight, maturation, route, timing, co-medication and metabolite exposure to both efficacy and safety, while follow-up should extend beyond the behavioural assessments commonly limited to 30 days. Until that work is available, the evidence neither validates a universal dose nor shows that doses above 0.5 mg·kg−1 are safe.

Subject of Research: Endpoint- and dose-dependent evidence for perioperative esketamine in children

Article Title: Dose- and endpoint-dependent effects of perioperative esketamine in children: evidence from a targeted review

Article References: Yang, Y., He, T., Mao, W., & Wang, X. (2026). Dose- and endpoint-dependent effects of perioperative esketamine in children: evidence from a targeted review. Frontiers in Pediatrics, 14, Article 1963881. https://doi.org/10.3389/fped.2026.1963881

Image Credits: AI Generated

DOI: 10.3389/fped.2026.1963881

Keywords: paediatric anaesthesia, esketamine, emergence delirium, emergence agitation, dose response, analgesia, procedural sedation, dexmedetomidine, pharmacokinetics, drug safety, Dose-, endpoint-dependent