Latent Class Analysis Reveals COVID-19–related Acute Respiratory Distress Syndrome Subgroups with Differential Responses to Corticosteroids

• DOI: 10.1164/rccm.202105-1302oc
• Publication Date: 2021-09-20T21:49:14Z
• Authors: Pratik Sinha; David Furfaro; Matthew J. Cummings; Darryl Abrams; Kevin Delucchi; Manoj V. Maddali; June He; Alison Thompson; Michael Murn; John Fountain; Amanda Rosen; Shelief Y. Robbin...; Matthew A. Adan; Tejus Satish; Mahesh Madhavan; Aakriti Gupta; Alexander K. Lyas...; Cara Agerstrand; Natalie H. Yip; Kristin M. Burkart; Jeremy R. Beitler; Matthew R. Baldwin; Carolyn S. Calfee; Daniel Brodie; Max R. O’Donnell
• Keywords: Male; phenotyping; Clinical Sciences; Respiratory System; 610; Clinical sciences; Medical and Health Sciences; Cardiovascular medicine and haematology; Cohort Studies; 03 medical and health sciences; Rare Diseases; 0302 clinical medicine; Adrenal Cortex Hormones; latent class analysis; Humans; Lung; Acute Respiratory Distress Syndrome; Aged; Retrospective Studies; Respiratory Distress Syndrome; Biomedical and Clinical Sciences; COVID-19; Original Articles; Middle Aged; COVID-19 Drug Treatment; 3. Good health; Infectious Diseases; Emerging Infectious Diseases; Good Health and Well Being; Latent Class Analysis; ARDS; Female

Rationale: Two distinct subphenotypes have been identified in acute respiratory distress syndrome (ARDS), but the presence of subgroups ARDS associated with coronavirus disease (COVID-19) is unknown. Objectives: To identify clinically relevant, novel COVID-19-related and compare them previously described subphenotypes. Methods: Eligible participants were adults COVID-19 at Columbia University Irving Medical Center. Latent class analysis was used to baseline clinical, respiratory, laboratory data serving as partitioning variables. A developed machine learning model classify patients hypoinflammatory hyperinflammatory Baseline characteristics clinical outcomes compared between subgroups. Heterogeneity treatment effect for corticosteroid use tested. Measurements Main Results: From March 2, 2020, April 30, 483 met study criteria. two-class latent best fit population (P = 0.0075). Class 2 (23%) had higher proinflammatory markers, troponin, creatinine, lactate, lower bicarbonate, blood pressure than 1 (77%). Ninety-day mortality versus (75% vs. 48%; P < 0.0001). Considerable overlap observed these Severe (SARS-CoV-2) RT-PCR cycle threshold not phenotype. corticosteroids 0.0295), improved phenotype worse phenotype, caveat that randomized. Conclusions: We two differential outcomes, similar SARS-CoV-2 PCR value predicting The responses corticosteroids.