Baseline predictors for 28-day COVID-19 severity and mortality among hospitalized patients: results from the IMPACC study

Jintong Hou^1,2Ben Samuel Haslund-Gourley²Joann Diray-Arce³Annmarie Hoch³Nadine Rouphael⁴Patrice M Becker⁵Alison D Augustine⁵Al Ozonoff³Leying Guan⁶Steven H. Kleinstein⁶Bjoern Peters⁷Elaine F Reed⁸Matthew C Altman^10,9Charles Langelier^11,12Holden Maecker¹³Seunghee Kim-schulze¹⁴Ruth Montgomery⁶Florian Krammer¹⁴Michael Wilson¹²Walter Eckalbar¹²Steven E Bosinger⁴Ofer Levy³Hanno Steen³Lindsey B Rosen⁵Lindsey R Baden¹⁵Esther Melamed¹⁶Lauren Ehrlich¹⁶Grace McComsey¹⁷Rafick-Pierre Sekaly¹⁸Joanna M. Schaenman⁸Albert C Shaw⁶David A. Hafler⁶David Corry¹⁹Farrah Kheradmand¹⁹Mark A Atkinson²⁰Scott Brakenridge²⁰Nelson I Agudelo Higuita²¹Jordan Metcalf²¹Catherine L Hough²²William Messer²²BALI PULENDRAN¹³Kari Christine Nadeau¹³Mark Morris Davis¹³Ana Fernandez-Sesma¹⁴Viviana Simon¹⁴Monica Kraft²³Chris Bime²³Carolyn S Calfee¹²David Erle¹²Lucy F Robinson¹Charles Cairns²Elias K Haddad²Mary Ann Comunale^2*

• ¹Dornsife School of Public Health, Drexel University, Philadelphia, Pennsylvania, United States
• ²College of Medicine, Drexel University, Philadelphia, Pennsylvania, United States
• ³Boston Children's Hospital, Harvard Medical School, Boston, Massachusetts, United States
• ⁴Emory University, Atlanta, Georgia, United States
• ⁵National Institutes of Health (NIH), Bethesda, Maryland, United States
• ⁶Yale University, New Haven, Connecticut, United States
• ⁷La Jolla Institute for Immunology (LJI), La Jolla, California, United States
• ⁸University of California, Los Angeles, Los Angeles, California, United States
• ⁹University of Washington, Seattle, Washington, United States
• ¹⁰Benaroya Research Institute, Seattle, Washington, United States
• ¹¹Chan Zuckerberg Biohub, San Francisco, California, United States
• ¹²University of California, San Francisco, San Francisco, California, United States
• ¹³School of Medicine, Stanford University, Stanford, California, United States
• ¹⁴Icahn School of Medicine at Mount Sinai, New York, New York, United States
• ¹⁵Harvard Medical School, Boston, Massachusetts, United States
• ¹⁶The University of Texas at Austin, Austin, Texas, United States
• ¹⁷University Hospitals of Cleveland, Cleveland, Ohio, United States
• ¹⁸Case Western Reserve University, Cleveland, Ohio, United States
• ¹⁹Baylor College of Medicine, Houston, Texas, United States
• ²⁰University of Florida, Gainesville, Florida, United States
• ²¹University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma, United States
• ²²Oregon Health and Science University, Portland, Oregon, United States
• ²³University of Arizona, Tucson, Arizona, United States

The coronavirus disease 2019 (COVID-19) pandemic threatened public health and placed a significant burden on medical resources. The Immunophenotyping Assessment in a COVID-19 Cohort (IMPACC) study collected clinical, demographic, blood cytometry, serum receptor-binding domain (RBD) antibody titers, metabolomics, targeted proteomics, nasal metagenomics, Olink, nasal viral load, autoantibody, SARS-CoV2 antibody titers, nasal and peripheral blood mononuclear cell (PBMC) transcriptomics data from patients hospitalized with COVID-19. The aim of this work is to analyze the IMPACC dataset using machine learning to select baseline biomarkers to predict COVID-19 severity and mortality. Severity was best predicted by the baseline SpO2/FiO2 ratio obtained from COVID-19 patients (test AUC: 0.874). Adding patient age, BMI, FGF23, IL-6, and LTA to the disease severity prediction model improves the test AUC by an additional 3%. Next, a clinical mortality prediction model using SpO2/FiO2 ratio, age, and BMI results in a test AUC of 0.83. The performance of the mortality prediction model further increases at least 3.5% by adding laboratory results such as TNFRSF11B and plasma ribitol count. The severity and mortality prediction models developed outperform the Sequential Organ Failure Assessment (SOFA) score among inpatients and perform similarly to the SOFA score among ICU patients. Taken together, this work identifies clinical data and laboratory biomarkers of COVID-19 severity and mortality using machine learning models.