Abstract
OBJECTIVE: To determine the prevalence, clinical characteristics, outcomes, and mortality risk factors of pediatric acute respiratory distress syndrome (PARDS) among children with sepsis, and to compare pulmonary and extrapulmonary PARDS phenotypes. METHODS: This retrospective cohort study analyzed children aged 0-14 years with Phoenix-defined sepsis admitted to a tertiary pediatric intensive care unit between 2015 and 2023. PARDS was defined according to PALICC-2 criteria. Demographics, illness severity, microbiology, organ support requirements, and clinical outcomes were compared between children with and without PARDS and between pulmonary and extrapulmonary phenotypes. Multivariable logistic regression models were used to identify independent predictors of mortality. RESULTS: Among 279 children with Phoenix-defined sepsis, 161 (57.7%) developed PARDS. Children with PARDS were younger, had higher PELOD-2 and Phoenix severity scores, and required significantly more mechanical ventilation, vasoactive support, and renal replacement therapy compared with those without PARDS. Mortality was substantially higher in the PARDS cohort (36.6% vs. 7.6%). Model-estimated mortality probability increased stepwise with worsening PARDS severity and was highest among children with both septic shock and severe PARDS. Pulmonary PARDS accounted for two-thirds of cases, whereas extrapulmonary PARDS demonstrated a higher inflammatory burden and more bacterial infections. In adjusted analyses, the presence of PARDS, higher PELOD-2 score, and greater cumulative fluid balance were independently associated with mortality. CONCLUSION: PARDS is a common and common complication associated with high risk of pediatric sepsis, associated with severe organ dysfunction, increased support requirements, and markedly elevated mortality. These findings underscore the need for multicenter validation to confirm the epidemiology and risk factors of sepsis-associated PARDS and to guide earlier recognition and management approaches for this high-risk population.