Baseline predictors for 28-day COVID-19 severity and mortality among hospitalized patients: results from the IMPACC study - PubMed
7 days ago
- #biomarkers
- #machine-learning
- #COVID-19
- The IMPACC study collected extensive clinical and laboratory data from 1102 hospitalized COVID-19 patients to identify baseline predictors for 28-day severity and mortality.
- Key predictors for disease severity included the SpO2/FiO2 ratio, age, BMI, FGF23, IL-6, and LTA, achieving a test AUC of 0.874.
- For mortality prediction, SpO2/FiO2 ratio, age, BMI, TNFRSF11B, and plasma ribitol were significant, with a test AUC of 0.83.
- The study's predictive models outperformed the SOFA score for inpatients and matched its performance in ICU patients.
- SpO2/FiO2 ratio was identified as the most critical predictor for both severity and mortality, with additional biomarkers providing modest improvements.
- The findings offer a baseline for understanding SARS-CoV-2 infection dynamics and host responses to emerging variants.