Data-driven clustering identifies features distinguishing multisystem inflammatory syndrome from acute COVID-19 in children and adolescents

• Type of Material: Article
• Authors: Alon Geva, Children's Hospital Boston; Manish M Patel, Centers for Disease Control and Prevention; Margaret M Newhams, Children's Hospital Boston; Cameron C Young, Children's Hospital Boston; Mary Beth F Son, Children's Hospital Boston; Michele Kong, The University of Alabama at Birmingham; Aline B Maddux, University of Colorado School of Medicine; Mark W Hall, Nationwide Children’s Hospital; Becky J Riggs, Johns Hopkins School of Medicine; Aalok R Singh, New York Medical College; John S Giuliano, Yale School of Medicine; Charlotte V Hobbs, University of Mississippi Medical Center; Laura L Loftis, Texas Children's Hospital Houston; Gwenn E McLaughlin, University of Miami Leonard M. Miller School of Medicine; Stephanie P Schwartz, University of North Carolina at Chapel Hill; Jennifer E Schuster, Children's Mercy Kansas City; Christopher J Babbitt, Miller Children's and Women's Hospital of Long Beach; Natasha B Halasa, Vanderbilt University Medical Center; Shira J Gertz, Saint Barnabas Medical Center; Sule Doymaz, SUNY Downstate Health Sciences University; Janet R Hume, University of Minnesota Twin Cities; Tamara T Bradford, LSUHSC School of Medicine; Katherine Irby, Arkansas Children's Hospital; Christopher L Carroll, Division of Critical Care; John K McGuire, Seattle Children's Hospital; Keiko Tarquinio, Emory University; Courtney M Rowan, Indiana University School of Medicine; Elizabeth H Mack, Medical University of South Carolina; Natalie Z Cvijanovich, UCSF Benioff Children's Hospital Oakland; Julie C Fitzgerald, University of Pennsylvania Perelman School of Medicine; Philip C Spinella, Washington University School of Medicine in St. Louis; Mary A Staat, University of Cincinnati College of Medicine; Katharine N Clouser, Hackensack Meridian School of Medicine; Vijaya L Soma, NYU Grossman School of Medicine; Heda Dapul, NYU Grossman School of Medicine; Mia Maamari, UT Southwestern Medical School; Cindy Bowens, University of Louisville Health Sciences Center; Kevin M Havlin, Central Michigan University; Peter M Mourani, University of Colorado School of Medicine; Sabrina M Heidemann, Central Michigan University; Steven M Horwitz, Rutgers Robert Wood Johnson Medical School; Leora R Feldstein, Centers for Disease Control and Prevention; Mark W Tenforde, Centers for Disease Control and Prevention; Jane W Newburger, Children's Hospital Boston; Kenneth D Mandl, Children's Hospital Boston; Adrienne G Randolph, Children's Hospital Boston
• Language: English
• Date: 2021-10-01
• Publisher: Elsevier Inc
• Publication Version: Final Published Version
• Copyright Statement: © 2021 The Authors
• License: Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International
• Final Published Version (URL): http://dx.doi.org/10.1016/j.eclinm.2021.101112
• Title of Journal or Parent Work: iScience
• Volume: 40
• Keywords: inflammatory syndrome; COVID-19
• Research Categories: Health Sciences, Medicine and Surgery

Abstract

Background: Multisystem inflammatory syndrome in children (MIS-C) consensus criteria were designed for maximal sensitivity and therefore capture patients with acute COVID-19 pneumonia. Methods: We performed unsupervised clustering on data from 1,526 patients (684 labeled MIS-C by clinicians) <21 years old hospitalized with COVID-19-related illness admitted between 15 March 2020 and 31 December 2020. We compared prevalence of assigned MIS-C labels and clinical features among clusters, followed by recursive feature elimination to identify characteristics of potentially misclassified MIS-C-labeled patients. Findings: Of 94 clinical features tested, 46 were retained for clustering. Cluster 1 patients (N = 498; 92% labeled MIS-C) were mostly previously healthy (71%), with mean age 7·2 ± 0·4 years, predominant cardiovascular (77%) and/or mucocutaneous (82%) involvement, high inflammatory biomarkers, and mostly SARS-CoV-2 PCR negative (60%). Cluster 2 patients (N = 445; 27% labeled MIS-C) frequently had pre-existing conditions (79%, with 39% respiratory), were similarly 7·4 ± 2·1 years old, and commonly had chest radiograph infiltrates (79%) and positive PCR testing (90%). Cluster 3 patients (N = 583; 19% labeled MIS-C) were younger (2·8 ± 2·0 y), PCR positive (86%), with less inflammation. Radiographic findings of pulmonary infiltrates and positive SARS-CoV-2 PCR accurately distinguished cluster 2 MIS-C labeled patients from cluster 1 patients. Interpretation: Using a data driven, unsupervised approach, we identified features that cluster patients into a group with high likelihood of having MIS-C. Other features identified a cluster of patients more likely to have acute severe COVID-19 pulmonary disease, and patients in this cluster labeled by clinicians as MIS-C may be misclassified. These data driven phenotypes may help refine the diagnosis of MIS-C.