Application of the Berlin definition in children with acute respiratory distress syndrome

ABSTRACT

PURPOSE: The revised Berlin definition (BD) showed better predictive validity for mortality in adults with acute respiratory distress syndrome (ARDS). We examined the validity of BD for pediatric ARDS as compared to the American-European Consensus Conference definition (AECCD). METHODS: This single-center, retrospective study included 127 patients aged 1 month to 19 years who were admitted to the medical intensive care unit due to acute lung injury (ALI, n=31) or ARDS (n=96) using the AECCD. All patient characteristics and mortality rates were compared between the individual severity groups according to the BD and AECCD. RESULTS: Sixty-four patients (50%) died. Mortality rates increased across the severity groups according to both definitions (26% in mild, 42% in moderate, and 75% in severe by the BD [P<0.001]; 26% in ALI non-ARDS and 58% in ARDS by the AECCD [P=0.002]). The mortality risk increased only for 'severe ARDS' (hazard radio for mortality, 2.56; 95% confidence intervals [CI], 1.14-5.78; P=0.023) after adjusting for confounding factors. The BD better predicted mortality, with an integrated area under the receiver operating characteristic curve (iAUC) of 0.651 (95% CI, 0.571-0.725), than the AECCD, with an iAUC of 0.584 (95% CI, 0.523-0.637). The pediatric risk of mortality (PRISM) III and pediatric index of mortality 3 scores were significantly different across BD severity groups, whereas only PRISM III scores were different according to the AECCD. CONCLUSION: The BD applied to children with ARDS. It could be adopted to severity classifications and predict pediatric ARDS mortality better than the AECCD.