Quantitative or qualitative differences in immunity may drive clinical severity in COVID-19. Although longitudinal studies to record the course of immunological changes are ample, they do not necessarily predict clinical progression at the time of hospital admission. Here we show, by a machine learning approach using serum pro-inflammatory, anti-inflammatory and anti-viral cytokine and anti-SARS-CoV-2 antibody measurements as input data, that COVID-19 patients cluster into three distinct immune phenotype groups. These immune-types, determined by unsupervised hierarchical clustering that is agnostic to severity, predict clinical course. The identified immune-types do not associate with disease duration at hospital admittance, but rather reflect variations in the nature and kinetics of individual patient’s immune response. Thus, our work provides an immune-type based scheme to stratify COVID-19 patients at hospital admittance into high and low risk clinical categories with distinct cytokine and antibody profiles that may guide personalized therapy.

Quantitative or qualitative differences in immunity may drive and predict clinical severity in COVID-19. We therefore measured modules of serum pro-inflammatory, anti-inflammatory and anti-viral cytokines in combination with the anti-SARS-CoV-2 antibody response in COVID-19 patients admitted to tertiary care. Using machine learning and employing unsupervised hierarchical clustering, agnostic to severity, we identified three distinct immunotypes that were shown post-clustering to predict very different clinical courses such as clinical improvement or clinical deterioration. Immunotypes did not associate chronologically with disease duration but rather reflect variations in the nature and kinetics of individual patient's immune response. Here we demonstrate that immunophenotyping can stratify patients to high and low risk clinical subtypes, with distinct cytokine and antibody profiles, that can predict severity progression and guide personalized therapy.