Objective: Guidelines recommend implementing early enteral nutrition (EN) (EEN) in critically ill children. The aim of the study was to determine if EEN for critically ill children is associated with improved clinical outcomes compared with delayed enteral nutrition (DEN), prioritizing associations adjusted for severity of illness. PROSPERO (CRD42021286271). Data Sources: MEDLINE, Embase, CINAHL, and CENTRAL databases to October 2024. Study Selection: The population was critically ill children, the intervention was EEN, the comparator was DEN, the outcome was mortality or clinical outcomes, and the study designs included randomized control trials (RCTs), quasi-experimental, observational cohort, or case-control. Data Extraction: Screening, extraction, and risk of bias assessment using the Newcastle-Ottawa Scale and Cochrane Risk of Bias and Grading of Recommendations Assessment, Development, and Evaluation (GRADE) assessment were conducted in duplicate by two reviewers. Eighteen of 8478 screened studies were included. Data Synthesis: Mortality outcomes were pooled and meta-analyzed using random-effects models. Secondary outcomes were described qualitatively, and directions of associations were reported. Thirteen studies (1 RCT, 12 cohort) reported mortality; however, only three adjusted for illness severity. In the adjusted analysis, receiving EEN was associated with reduced mortality (adjusted odds ratio 0.36 (95% CI, 0.14–0.91), I 2 = 78.6%, n = 5864). The certainty of evidence, as assessed by GRADE, was very low due to indirectness. In the qualitative review of 18 studies (1 RCT, 17 cohort studies, n = 9829), EEN had an association with reduced length of stay, length of invasive respiratory support, improved nutrition adequacy, reduced maximum pediatric logistic organ dysfunction score, and infection. No harmful effects of EEN were found after adjusting for confounding variables. Conclusions: EEN was associated with beneficial outcomes. However, the inclusion of mostly cohort studies with limited confounding adjustment, the small number of studies, the presence of between-study heterogeneity and residual confounding, and heterogeneity in measured outcomes and assessment methods resulted in very low certainty of evidence.
Gilbert et al. (Wed,) studied this question.