Objectives: Evaluate the relationship between admission Pa o2 and mortality in a large multicenter dataset and among diagnostic subgroups.
Design: Retrospective cohort study.
Setting: North American PICUs participating in Virtual Pediatric Systems, LLC (VPS), 2015-2019.
Patients: Noncardiac patients 18 years or younger admitted to a VPS PICU with admission Pa o2 .
Measurements and main results: Thirteen thousand seventy-one patient encounters were included with an overall mortality of 13.52%. Age categories were equally distributed among survivors and nonsurvivors with the exception of small differences among neonates and adolescents. Importantly, there was a tightly fitting quadratic relationship between admission Pa o2 and mortality, with the highest mortality rates seen among hypoxemic and hyperoxemic patients (likelihood-ratio test p < 0.001). This relationship persisted after adjustment for illness severity using modified Pediatric Index of Mortality 3 scores. A similar U-shaped relationship was demonstrated among patients with diagnoses of trauma, head trauma, sepsis, renal failure, hemorrhagic shock, and drowning. However, among the 1,500 patients admitted following cardiac arrest, there was no clear relationship between admission Pa o2 and mortality.
Conclusions: In a large multicenter pediatric cohort, admission Pa o2 demonstrates a tightly fitting quadratic relationship with mortality. The persistence of this relationship among some but not all diagnostic subgroups suggests the pathophysiology of certain disease states may modify the hyperoxemia association.
Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.