Phenotypic Classification of Multisystem Inflammatory Syndrome in Children: A Latent Class Analysis

Abstract

AbstractImportanceMultisystem inflammatory syndrome in children (MIS-C) is an uncommon but severe hyperinflammatory illness occurring 2–6 weeks after SARS-CoV-2 infection. Presentation overlaps with other conditions, and risk factors for severe clinical outcomes differ by patient. Characterizing patterns of MIS-C presentation can guide efforts to reduce misclassification, categorize phenotypes, and identify patients at risk for severe outcomes.ObjectiveTo characterize phenotypic clusters of MIS-C and identify clusters with increased clinical severity.DesignWe describe MIS-C phenotypic clusters inferred using latent class analysis (LCA) applied to the largest cohort to date of cases from U.S. national surveillance. Illness onset ranged from February 2020 through December 2022.SettingNational surveillance comprising data from 55 U.S. public health jurisdictions.ParticipantsWe analyzed 9,333 MIS-C cases. Twenty-nine clinical signs and symptoms were selected for clustering after excluding variables with ≥20% missingness and ≤10% or ≥90% prevalence. We excluded 389 cases missing ≥10 variables and conducted multiple imputation on the remaining 8,944 (96%) cases.Main Outcomes and MeasuresDifferences by cluster in prevalence of each clinical sign and symptom, percentage of cases admitted to the intensive care unit (ICU), length of hospital and ICU stay, mortality, and relative frequency over time.ResultsLCA identified three clusters characterized by 1) frequent respiratory findings primarily affecting older children (n = 713; 8.0% of cases; median age: 12.7 years); 2) frequent cardiac complications and shock (n = 3,359; 37.6%; 10.8 years); and 3) remaining cases (n = 4,872; 54.5%; 6.8 years). Mean duration of hospitalization and proportion of cases resulting in ICU admission or death were higher in the respiratory (7.9 days; 49.5%; 4.6%; respectively) and shock/cardiac clusters (8.7 days; 82.3%; 1.0%; respectively) compared with other cases (5.3 days; 33.0%; 0.06%; respectively). The proportion of cases in the respiratory and shock/cardiac clusters decreased after emergence of the Omicron variant in the United States.Conclusions and RelevanceMIS-C cases clustered into three subgroups with distinct clinical phenotypes, illness severity, and distribution over time. Use of clusters in future studies may support efforts to evaluate surveillance case definitions and help identify groups at highest risk for severe outcomes.Key pointsQuestionCan phenotypic clusters of multisystem inflammatory syndrome in children (MIS-C) be identified, and are some clusters associated with increased severity?FindingsWe describe clusters inferred using latent class analysis (LCA) on 9,333 MIS-C cases from U.S. national surveillance 2020–2022. LCA identified three clusters characterized by frequent respiratory symptoms, frequent cardiac complications and shock, and remaining clinically milder cases. Mortality and ICU admission were highest in the respiratory and shock/cardiac clusters; prevalence of these two clusters decreased over time.MeaningMIS-C clusters had distinct presentation, illness severity, and distribution over time, highlighting the importance of recognizing the varied presentation of MIS-C.