Reconciling conflicting evidence on stress ulcer prophylaxis in critically ill children: a systematic review and meta-analysis.

Stress ulcer prophylaxis (SUP) is widely used in critically ill children; however, evidence for benefit is limited and previous reviews reached conflicting conclusions, particularly regarding mortality. Here we aimed to reassess the benefits and harms of SUP in critically ill children and reconcile the conflicting conclusions of previous reviews. We searched the PubMed/MEDLINE, CENTRAL, Scopus, Web of Science, and Igaku Chuo Zasshi databases from inception to May 20, 2026, for randomized and nonrandomized studies comparing SUP with placebo or no treatment in critically ill children. Risk of bias was assessed with outcome-specific risk of bias 2 and Risk Of Bias In Non-randomized Studies of Interventions, version 2 (ROBINS-I V2) and certainty with Grading of Recommendations Assessment, Development, and Evaluation; random-effects models pooled outcomes reported in ≥2 studies. The protocol was registered (PROSPERO CRD420251074125), and reporting followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) 2020 guidelines. Eighteen studies were included (7 randomized, 11 nonrandomized); all nonrandomized studies had a critical risk of bias (confounding by indication) and were excluded from synthesis. The evidence was very uncertain for all outcomes. For the primary outcome, upper gastrointestinal bleeding, no reduction was found (5 trials, n=713; risk ratio [RR], 1.06; 95% confidence interval [CI], 0.76-1.47; I²=0%), with similarly inconclusive estimates for clinically significant gastrointestinal bleeding (RR, 0.81; 95% CI, 0.20-3.26), mortality (RR, 1.12; 95% CI, 0.73-1.71), and ventilator-associated pneumonia (RR, 1.13; 95% CI, 0.78-1.63); the previously reported mortality signal was not reproduced. An exploratory analysis suggested an increased incidence of central line-associated bloodstream infection (CLABSI; 2 trials, n=212; RR, 2.47; 95% CI, 1.24-4.90). By formally appraising every non-randomized study with ROBINS-I V2 rather than excluding them by design, this review reconciles earlier conflicting syntheses: across the largest randomized evidence base to date, prophylaxis conferred no demonstrable benefit, and the earlier mortality signal was not reproduced. With very low overall certainty, the evidence neither showed a benefit nor excluded an exploratory signal of an increased incidence of CLABSI. Accordingly, the current evidence is insufficient to support the routine use of SUP.