Characterizing COVID-19 Clinical Phenotypes and Associated Comorbidities and Complication Profiles

Background There is limited understanding of heterogeneity in outcomes across hospitalized patients with coronavirus disease 2019 (COVID-19). Identification of distinct clinical phenotypes may facilitate tailored therapy and improve outcomes.

Objective Identify specific clinical phenotypes across COVID-19 patients and compare admission characteristics and outcomes.

Design, Settings, and Participants Retrospective analysis of 1,022 COVID-19 patient admissions from 14 Midwest U.S. hospitals between March 7, 2020 and August 25, 2020.

Methods Ensemble clustering was performed on a set of 33 vitals and labs variables collected within 72 hours of admission. K-means based consensus clustering was used to identify three clinical phenotypes. Principal component analysis was performed on the average covariance matrix of all imputed datasets to visualize clustering and variable relationships. Multinomial regression models were fit to further compare patient comorbidities across phenotype classification. Multivariable models were fit to estimate the association between phenotype and in-hospital complications and clinical outcomes.

Main outcomes and measures Phenotype classification (I, II, III), patient characteristics associated with phenotype assignment, in-hospital complications, and clinical outcomes including ICU admission, need for mechanical ventilation, hospital length of stay, and mortality.

Results The database included 1,022 patients requiring hospital admission with COVID-19 (median age, 62.1 [IQR: 45.9-75.8] years; 481 [48.6%] male, 412 [40.3%] required ICU admission, 437 [46.7%] were white). Three clinical phenotypes were identified (I, II, III); 236 [23.1%] patients had phenotype I, 613 [60%] patients had phenotype II, and 173 [16.9%] patients had phenotype III.

When grouping comorbidities by organ system, patients with respiratory comorbidities were most commonly characterized by phenotype III (p=0.002), while patients with hematologic (p<0.001), renal (p<0.001), and cardiac (p<0.001) comorbidities were most commonly characterized by phenotype I. The adjusted odds of respiratory (p<0.001), renal (p<0.001), and metabolic (p<0.001) complications were highest for patients with phenotype I, followed by phenotype II. Patients with phenotype I had a far greater odds of hepatic (p<0.001) and hematological (p=0.02) complications than the other two phenotypes. Phenotypes I and II were associated with 7.30-fold (HR: 7.30, 95% CI: (3.11-17.17), p<0.001) and 2.57-fold (HR: 2.57, 95% CI: (1.10-6.00), p=0.03) increases in the hazard of death, respectively, when compared to phenotype III.

Conclusion In this retrospective analysis of patients with COVID-19, three clinical phenotypes were identified. Future research is urgently needed to determine the utility of these phenotypes in clinical practice and trial design.