Prevalence of pediatric delirium among critically ill children: a meta-analysis of observational studies.

BACKGROUND

Pediatric delirium (PD) is a serious neurological complication in critically ill children, associated with prolonged hospitalization, extended mechanical ventilation, and increased mortality. However, the pooled prevalence of PD in this population remains unclear. This meta-analysis of observational studies aimed to determine the overall prevalence of PD among critically ill children and to identify its potential moderating factors.

METHODS

A comprehensive literature search was conducted in PubMed, Web of Science, Scopus, Embase, Cochrane Library, and CINAHL from inception to May 26, 2026. Methodological quality was appraised with the Joanna Briggs Institute (JBI) critical appraisal checklist for prevalence studies. Pooled prevalence of PD was calculated utilizing a DerSimonian-Laird random-effects model. To stabilize variances and improve the normal approximation, a logit transformation was applied for the meta-analysis of prevalence data. Heterogeneity was assessed using Cochran's Q test and quantified with the I² statistic. Subgroup analyses and meta-regression were conducted to explore potential sources of heterogeneity.

RESULTS

This meta-analysis included 40 observational studies, involving 18,875 critically ill children. The pooled prevalence of PD among critically ill children was 38.1% (95% CI: 32.5-43.8%, I2 = 98.7%, 95% PI: 3.6-72.4%). According to the JBI critical appraisal tool, 11 studies (27.5%) were rated as having a low risk of bias, 26 (65.0%) a moderate risk, and 3 (7.5%) a high risk. Subgroup analysis revealed significant variation in pooled prevalence across different assessment tools. Meta-regression suggested that prevalence was not significantly associated with sample size, income status, study design, or assessment tool.

CONCLUSIONS

Our findings indicate that PD is a prevalent neurological complication among critically ill children. However, due to extremely high heterogeneity and a very wide prediction interval, this pooled estimate should not be interpreted as a stable universal figure, but merely as a summary of highly heterogeneous studies. Therefore, caution is warranted when generalizing the pooled estimate to local settings.

SYSTEMATIC REVIEW REGISTRATION: identifier CRD420261339560.