Logo image
Back
Superimposed Pressure Predicts Mortality in Acute Respiratory Failure during Spontaneous Breathing: insights from the CT-COVID19 multicenter study group
Journal article   Open access   Peer reviewed

Superimposed Pressure Predicts Mortality in Acute Respiratory Failure during Spontaneous Breathing: insights from the CT-COVID19 multicenter study group

Emanuele Rezoagli, Davide Signori, Yi Xin, Sarah Gerard, Aurora Magliocca, Francesca Graziano, Giovanni Vitale, Linda Mussoni, Jonathan Montomoli, Matteo Subert, …
CHEST critical care, Vol.4(1), 100231
03/2026
DOI: 10.1016/j.chstcc.2025.100231
url
https://doi.org/10.1016/j.chstcc.2025.100231View
Published (Version of record) Open Access

Abstract

Early lung imaging may improve patient management and prognostication in acute respiratory failure. We aimed to explore whether quantitative assessment of lung injury by computed tomography (CT) predicts outcome in spontaneously breathing patients with COVID-19 acute respiratory failure. This is a large retrospective, multicenter, cohort study including patients presenting to the Emergency Department with a clinical diagnosis of COVID-19 respiratory failure and undergoing early lung CT scan at hospital admission. Lung injury was characterized by the severity of lung involvement as follows: 1) absence, unilateral, or bilateral infiltrates; 2) number of lung quadrants affected by infiltrates (0-4); level of global and regional 3) superimposed pressure (SP) and 4) gas/tissue ratio (G/T). Baseline, laboratory and clinical characteristics were described by the presence or absence of laterality of lung infiltrates. Association of 90-day mortality and lung CT characterization was explored using Cox multivariable models and areas under receiving operating characteristics. Subphenotypes including CT assessment were explored by latent class analyses. Eight-hundred and eight patients were included. Bilateral infiltrates were associated with higher global and regional SP and G/T and a higher 90-day mortality (38%) compared with unilateral infiltrates (18%) or no lung infiltrates (11%). Involvement by laterality, quadrants, degree of global SP and G/T were all associated with the degree of hypoxemia on admission and 90-day mortality. Among other CT-derived variables of lung injury, SP characterized a subphenotype with a robust relationship with 90-day mortality. Characterization of lung injury severity by early lung CT describes the severity of hypoxemia. The adjunct of CT global SP to clinical and laboratory parameters identified a subphenotype with high 90-day mortality prediction. Early lung CT may enhance population enrichment and improve prognostication in non-intubated patients with acute respiratory failure.
 Computed Tomography artificial intelligence COVID-19 mortality respiratory failure spontaneous breathing superimposed pressure

Details

Metrics

10 Record Views
Logo image