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

Alon Geva,Janet R Hume,Peter M Mourani,Courtney M Rowan,Charlotte V Hobbs,Keiko M Tarquinio,Sule Doymaz,Gwenn E Mclaughlin,Katharine N Clouser,Katherine Irby,Kenneth D Mandl,Steven M Horwitz,Leora R Feldstein,Heda Dapul,John S Giuliano,Elizabeth H Mack,Philip C Spinella,Natalie Z Cvijanovich,Sabrina M Heidemann,Michele Kong,Jane W Newburger,Becky J Riggs,Adrienne G Randolph,Shira J Gertz,Aalok R Singh,Mia Maamari,Christopher L Carroll,John K Mcguire,Aline B Maddux ,Vijaya L Soma ,Cindy Darnell Bowens ,Laura Loftis ,Natasha Halasa ,Margaret M Newhams ,Mark W Tenforde ,Manish Patel ,Christopher Babbitt ,Mary Allen Staat ,Stephanie P Schwartz ,Julie C Fitzgerald ,Kevin M Havlin ,Tamara T Bradford ,Mark Hall ,Cameron C Young ,Mary Beth Son ,Jennifer E Schuster ,Overcoming Covid-19 Investigators

doi:10.1016/j.eclinm.2021.101112

Abstract

BackgroundMultisystem inflammatory syndrome in children (MIS-C) consensus criteria were designed for maximal sensitivity and therefore capture patients with acute COVID-19 pneumonia.MethodsWe 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.FindingsOf 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.InterpretationUsing 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.FundingThis work was funded by the US Centers for Disease Control and Prevention (75D30120C07725) and National Institutes of Health (K12HD047349 and R21HD095228).

Full Text