1130: MORTALITY AND CRITICAL DETERIORATION IN HOSPITALIZED CHILDREN

Abstract

Introduction: Critical deterioration in hospitalized children has been defined in the literature as ward to intensive care unit (ICU) transfer followed by events such as mechanical ventilation (MV) or vasopressor administration within 12 hours. This study aims to better understand the association between in-hospital mortality and events occurring within and beyond the 12 hours of ICU transfer. Methods: We conducted a retrospective analysis of all pediatric patients transferred to the ICU from the ED or ward at the University of Chicago Comer Children’s Hospital during 2009-2019. We determined the association between mortality and the occurrence of MV or vasopressor events throughout ICU stay after adjusting for age, gender, race, admitting location, and prior comorbidities calculated using the Children’s Comorbidity Index. Results: Out of 6,741 patients that were transferred to the ICU, 1,020 (15.1%) patients experienced at least one MV or vasopressor event throughout their ICU stay. Patients who experienced these events were similar to patients who did not experience these events in terms of age, gender, race, and ethnicity but had a higher likelihood of at least one prior comorbidity (42% vs. 26%, P < 0.001). In the fully adjusted model, experiencing an event within the first 12 hours was associated with the highest risk of mortality (OR for < 12 hours 10.6 [95% CI 6.9-16.6]). However, experiencing an event beyond the first 12 hours was also associated with elevated mortality risk (OR for 12-24 hours: 1.7 [0.5-4.2], 24-36 hours: 3.7 [1.1-9.6], 36-48 hours: 3.3 [0.8-9.7], >48 hours: 4.3 [2.2-7.7]). Similar results were observed when considering MV or vasopressor events only. Conclusions: In this retrospective study, we demonstrate that critical deterioration events in the ICU are associated with increased mortality in a time-dependent manner. Children who experience MV or vasopressor events within 12 hours of ICU transfer are at the highest risk for mortality and constitute a target cohort for risk prediction models in pre-ICU settings. These events also pose increased mortality risk throughout an ICU stay and thus could offer a valuable target outcome for risk prediction models within the ICU