key: cord-1031407-myapg66m
authors: Lapp, Stacey A; Abrams, Joseph; Lu, Austin T; Hussaini, Laila; Kao, Carol M; Hunstad, David A; Rosenberg, Robert B; Zafferani, Marc J; Ede, Kaleo C; Ballan, Wassim; Laham, Federico R; Beltran, Yajira; Hsiao, Hui-Mien; Sherry, Whitney; Jenkins, Elan; Jones, Kaitlin; Horner, Anna; Brooks, Alyssa; Bryant, Bobbi; Meng, Lu; Hammett, Teresa A; Oster, Matthew E; Bamrah-Morris, Sapna; Godfred-Cato, Shana; Belay, Ermias; Chahroudi, Ann; Anderson, Evan J; Jaggi, Preeti; Rostad, Christina A
title: Serologic and Cytokine Signatures in Children With Multisystem Inflammatory Syndrome and Coronavirus Disease 2019
date: 2022-02-24
journal: Open Forum Infect Dis
DOI: 10.1093/ofid/ofac070
sha: 1658c6c5ecc0436d1ecdc0b1f6bc72a59f2a721e
doc_id: 1031407
cord_uid: myapg66m

BACKGROUND: The serologic and cytokine responses of children hospitalized with multisystem inflammatory syndrome (MIS-C) vs coronavirus disease 2019 (COVID-19) are poorly understood. METHODS: We performed a prospective, multicenter, cross-sectional study of hospitalized children who met the Centers for Disease Control and Prevention case definition for MIS-C (n = 118), acute COVID-19 (n = 88), or contemporaneous healthy controls (n = 24). We measured severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike receptor-binding domain (RBD) immunoglobulin G (IgG) titers and cytokine concentrations in patients and performed multivariable analysis to determine cytokine signatures associated with MIS-C. We also measured nucleocapsid IgG and convalescent RBD IgG in subsets of patients. RESULTS: Children with MIS-C had significantly higher SARS-CoV-2 RBD IgG than children with acute COVID-19 (median, 2783 vs 146; P < .001), and titers correlated with nucleocapsid IgG. For patients with MIS-C, RBD IgG titers declined in convalescence (median, 2783 vs 1135; P = .010) in contrast to patients with COVID-19 (median, 146 vs 4795; P < .001). MIS-C was characterized by transient acute proinflammatory hypercytokinemia, including elevated levels of interleukin (IL) 6, IL-10, IL-17A, and interferon gamma (IFN-γ). Elevation of at least 3 of these cytokines was associated with significantly increased prevalence of prolonged hospitalization ≥8 days (prevalence ratio, 3.29 [95% CI, 1.17–9.23]). CONCLUSIONS: MIS-C was associated with high titers of SARS-CoV-2 RBD IgG antibodies and acute hypercytokinemia with IL-6, IL-10, IL-17A, and IFN-γ.

Following the onset of the coronavirus disease 2019 (COVID- 19) pandemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), a novel multisystem inflammatory syndrome in children (MIS-C) was first described in Western Europe in April 2020 [1] . This syndrome temporally follows SARS-CoV-2 infection by 2-6 weeks and is characterized by fever, systemic inflammation, multiorgan involvement, and severe disease requiring hospitalization [2] . Although the majority of patients recover without long-term sequelae [3] , some develop myocardial dysfunction, shock, and respiratory failure requiring intensive care [2, [4] [5] [6] . A variety of treatment approaches have been adopted for MIS-C, which include intravenous immunoglobulin, corticosteroids, immunomodulating agents, aspirin, and anticoagulants [2, 4, 5] , all of which have potential risks and uncertain benefits. Distinguishing MIS-C from alternative etiologies and identifying biomarkers of severity at the time of presentation could better inform patient management.

To date, the pathogenesis of MIS-C is poorly understood. We and others have previously reported the development of high titers of SARS-CoV-2 binding and neutralizing antibodies in patients with MIS-C [7] [8] [9] [10] . Evidence of immune cell activation, mucosal inflammation, auto-antibody formation, and cytokine storm has also been reported [8, 11] . While previous studies have described the generalized hypercytokinemia observed in MIS-C [10, 12] , the clinical correlates, predictive value, and time course of individual and collective cytokines are incompletely understood. In this study, we aimed to describe the distinguishing serologic and cytokine signatures associated with MIS-C diagnosis and clinical outcomes. The detailed clinical data from the cohort are the subject of a future manuscript (in review). Herein, we report the serologic and cytokine signatures of the subset of enrolled participants with MIS-C or COVID-19 who contributed blood samples for analysis.

This was a multicenter cross-sectional study conducted in collaboration with 4 pediatric medical centers and approved by their respective institutional review boards: Children's Healthcare of Atlanta Hospitalized patients meeting the case definition for MIS-C or acute COVID-19 were prospectively enrolled following informed consent and assent as appropriate for age. The MIS-C case definition included any patient <21 years or age with fever, laboratory evidence of inflammation, and evidence of clinically severe illness requiring hospitalization, with multisystem organ involvement (cardiovascular, dermatologic, gastrointestinal, hematologic, neurologic, renal, or respiratory) who tested positive for SARS-CoV-2 or had recent exposure to COVID-19. MIS-C cases were adjudicated by the study sites to ensure all met the CDC case definition and that no more likely alternative diagnoses were identified during the acute hospitalization. The acute COVID-19 case definition included any patient <21 years of age with positive SARS-CoV-2 testing by any method who was hospitalized for >24 hours with no alternative diagnosis, AND had at least 2 of the following symptoms (fever, chills, rigors, myalgia, headache, sore throat, new olfactory or taste disorder) OR at least 1 of the following (cough, shortness of breath and difficulty breathing) OR severe respiratory illness with at least 1 of the following (clinical or radiographic evidence of pneumonia, acute respiratory distress syndrome [ARDS]) [13] .

At the time of enrollment, blood was prospectively collected and/or residual blood was obtained from the clinical laboratories and stored at -80°C. Only samples collected prior to the administration of intravenous immunoglobulin and convalescent plasma were included, so as not to impact serological analyses. Receipt of steroids or other immunomodulatory regimens was not considered exclusionary. Participants were enrolled from 10 June 2020 until 21 April 2021 and specimens were shipped on dry ice to Emory University for analysis. A subset of patients enrolled at the Emory University site provided convalescent samples from 28 to 170 days following symptom onset. From June to October 2020, 24 contemporaneous healthy pediatric controls provided specimens following enrollment into a separate institutional review board-approved outpatient phlebotomy study at Emory.

Clinical data were abstracted from the medical record by trained abstractors using a case report form developed for MIS-C national surveillance with minor revisions. The detailed clinical data describing the cohort from the 4 hospitals will be the subject of a future manuscript (in review). The cohort herein represents the subset of patients from the clinical cohort who had MIS-C or COVID-19 and contributed blood samples for analysis.

The SARS-CoV-2 spike protein receptor-binding domain (RBD) protein was kindly provided by Dr Jens Wrammert [14] . RBD immunoglobulin G (IgG) enzyme-linked immunosorbent assays (ELISAs) were performed as previously described [7] . Plates were developed using o-Phenylenediamine substrate, and absorbance was read at 490 nm. Absorbance curves were generated using nonlinear regression analysis, and end-point titers were interpolated from curves by using a baseline value calculated from the pooled plasma of 8 prepandemic healthy controls. The lower limit of detection (LLOD) was 100, and undetectable titers were assigned a value of 85. ELISAs were analogously performed for SARS-CoV-2 nucleocapsid protein IgG (Sino Biological) for a subset of patients with MIS-C and acute COVID-19.

Serum or plasma samples were analyzed for 10 cytokines using a custom U-PLEX panel (Meso Scale Diagnostics) following the manufacturer's protocol. Cytokines on the panel included interferon gamma (IFN-γ), interleukin (IL) 1β, IL-2, IL-4, IL-6, IL-8, IL-10, IL-13, IL-17A, and tumor necrosis factor alpha (TNF-⍺). These cytokines were chosen based on their associations with systemic inflammation, antiviral response, COVID-19, or MIS-C in prior reports [11] . Cytokine concentrations (pg/mL) were interpolated from a standard curve. For statistical analyses, undetectable cytokine results for a given cytokine were assigned a value of 0.2 times the LLOD.

Clinical data and laboratory results were stored in a central REDCap electronic data capture tool hosted by the CDC [15, 16] . Statistical analyses were performed using R version 4.0.2 software [17] , including the packages CatPredi [18] and glmnet [19] . Log-transformed antibody titers and cytokine concentrations were statistically compared using Fisher exact tests for categorical variables and Mann-Whitney U tests for continuous variables. Pearson correlation coefficients were calculated where applicable, using log-transformed serology titer and cytokine values.

To identify a cytokine signature that distinguishes MIS-C from COVID-19, we performed multivariable analysis of dichotomized cytokine concentrations using least absolute shrinkage and selection operator (LASSO) for variable selection [19] . Cytokine levels were dichotomized as elevated or not elevated using an algorithm that identified cutpoints that maximized discriminatory power [20] . LASSO was utilized to identify the 4 cytokine measurements that had the strongest combined ability to discriminate between MIS-C and COVID-19. Cytokines identified as being strong indicators of MIS-C were additionally assessed for associations with clinical findings and outcome metrics within patients with MIS-C.

Patients were prospectively enrolled and samples collected from 20 June 2020 to 16 April 2021, including 118 patients with MIS-C, 88 with acute COVID-19, and 24 healthy controls (Supplementary Table 1 ). One patient with MIS-C did not have an RBD IgG titer measurement, and another patient did not have cytokine levels tested; RBD IgG titers and cytokine levels were available for all other patients. Patients with MIS-C had a median age of 10 (interquartile range [IQR], [6] [7] [8] [9] [10] [11] [12] [13] [14] years, and 38.1% were female; 43.4% were non-Hispanic Black, 34.5% non-Hispanic White, 19.5% Hispanic ethnicity, and 2.7% Asian. Patients with COVID-19 had a median age of 14 (IQR, 3-17) years, and 62.5% were female; 35.3% were non-Hispanic Black, 29.4% non-Hispanic White, 34.1% Hispanic ethnicity, and 1.2% Asian. Healthy controls had a median age of 8 (IQR, 6-12) years, and 54.2% were female; 79.1% were non-Hispanic Black, 4.2% were non-Hispanic White, 8.3% Hispanic ethnicity, and 8.3% identified as other race. Almost half of MIS-C patients (48.7%) reported a preceding COVID-19-like illness a median of 21 days prior to MIS-C onset. Selected laboratory results and clinical outcome metrics are shown in Supplementary  Table 1 . Compared to patients with acute COVID-19, patients with MIS-C had significantly higher peak C-reactive protein, D-dimer, ferritin, brain natriuretic protein (BNP), proBNP, and troponin levels, and significantly lower nadirs of platelet count and absolute lymphocyte count, which are consistent with our previous data [7] . Among our cohort, patients with MIS-C were also significantly more likely to require vasopressors and to have an adverse cardiac outcome (defined as decreased cardiac function, myocarditis, pericardial effusion, mitral regurgitation, or coronary artery dilatation or aneurysm) compared to patients with COVID-19. The median duration of hospitalization was 5 (IQR, 4-7) days for patients with MIS-C, 63.6% required intensive care, and all survived.

The majority of patients with MIS-C (98.3%) had elevated SARS-CoV-2 RBD IgG titers. Patients with MIS-C had significantly higher RBD IgG titers than patients with acute COVID-19 (median endpoint titer, 2783 vs 146; P < .001) or healthy controls (median, 2783 vs 85; P < .001), although patients with COVID-19 had a wide range of titers ( Figure 1A , Supplementary Table 2 ). None of the contemporaneous healthy pediatric controls had detectable RBD IgG antibodies. Nucleocapsid protein IgG antibodies were measured for a subset of patients with acute MIS-C (n = 13) and acute COVID-19 (n = 14), and these titers correlated strongly with RBD IgG titers (Pearson R = 0.89 for log-transformed titer values in MIS-C patients, P < .001; R = 0.66 in COVID-19 patients, P = .010) ( Figure 1B ). Among patients with acute MIS-C, those aged 0-5 years and 6-12 years had significantly higher SARS-CoV-2 RBD IgG titers than patients 13-20 years of age (P < .001); however, RBD titers did not correlate with patient sex, race, or ethnicity (Supplementary Figure 1) .

Convalescent samples were available for analysis from a subset of patients with MIS-C (n = 13) and COVID-19 (n = 14). For patients with MIS-C, RBD IgG declined from the acute stage to early convalescence (median, 2783 to 1135, P = .010; median follow-up, 50 [IQR, 41-58] days) (Supplementary Table 2 ). In contrast, patients with COVID-19 had significant increases in titer from the acute to the convalescent stage (median, 146-4795, P < .001; median follow-up, 42 [IQR, 37-52] days) ( Figure 1C ). SARS-CoV-2 RBD IgG titers correlated with the duration of time post-symptom onset in patients with acute COVID-19 (R = 0.38 for log titers, P < .001) ( Figure 1D ), in contrast to patients with MIS-C for whom it did not correlate with time. For convalescent MIS-C samples, RBD IgG titers trended toward decreasing with longer duration of time postsymptom onset (R = -0.56, P = .048) (Supplementary Figure 2) .

All cytokines in the 10-plex panel were significantly elevated in patients with MIS-C compared to healthy controls and to patients with acute COVID-19 ( Figure 2) , with the exception of IL-13 (Supplementary Table 2 ). In convalescence, cytokines for both MIS-C and acute COVID-19 normalized to approximately the level of healthy controls ( Figure 2) .

The 4-variable model produced through LASSO identified the following cytokine measurements as best in differentiating MIS-C from acute COVID-19: IL-6 >25 pg/mL, IL-10 >10 pg/ mL, IL-17A >4 pg/mL, and IFN-γ >250 pg/mL (Supplementary Figure 3 ). For MIS-C patients, 54% had elevated levels (ie, values above the aforementioned thresholds) for at least 3 of the 4 cytokines, with 32% having elevated levels for all 4 cytokines. Conversely, only 1% of COVID-19 patients had elevated levels of at least 3 cytokines, and no COVID-19 patients had elevated levels of all 4 cytokines (Figure 3) . Two-thirds of COVID-19 patients (67%) did not have elevated levels of any of these 4 cytokines. Similarly, none of the healthy controls had elevated levels of any of these 4 cytokines.

We performed stratified analysis among patients with MIS-C to determine the association between this cytokine signature with categorical and continuous clinical outcome metrics. Compared to patients with elevated levels (IL-6 >25 pg/mL, IL-10 >10 pg/mL, IL-17A >4 pg/mL, or IFN-γ >250 pg/mL) for 2 or fewer cytokines, patients with elevated levels of at least 3 of the 4 cytokines were significantly more likely to have prolonged hospitalization ≥8 days ( between having at least 3 of 4 cytokines elevated and intensive care unit admission, decreased cardiac function, any severe cardiac outcome, shock, or specific organ involvement. Similarly, no single cytokine was predictive of these outcomes. Patients with elevated levels of at least 3 of the 4 cytokines did have significantly increased peak D-dimer compared to those with ≤2 elevated cytokine levels ( 

In this prospective, multicenter, cross-sectional study, we identified serologic and cytokine signatures of MIS-C, which P <.001 P <.001 P <.001 P <.001 10 000 IFN-γ (pg/mL) were hallmarked by high titers of SARS-CoV-2 RBD IgG antibodies and elevations in IL-6, IL-10, IL-17A, and IFN-γ. Overall, 98.3% of patients with acute MIS-C in our cohort had elevated SARS-CoV-2 RBD IgG antibodies, consistent with previously published results in smaller single-center cohorts [7, 9] . Patients with acute MIS-C had significantly higher SARS-CoV-2 RBD IgG titers than patients with acute COVID-19, who experienced a broad range of titers that Categorical and continuous outcome metrics were compared among pediatric patients with multisystem inflammatory syndrome who had 3-4 cytokines elevated vs those who had ≤2 cytokines elevated.

Abbreviations: BNP, brain natriuretic peptide; CI, confidence interval; CRP, C-reactive protein; ICU, intensive care unit; IQR, interquartile range; proBNP, pro-brain natriuretic peptide; PR, prevalence ratio. a Cytokine cutoffs for this analysis: interleukin (IL) 6, >25 pg/mL; IL-10, >10 pg/mL; IL-17A, >4 pg/mL; and interferon-γ, >250 pg/mL. b Defined as 1 or more of the following: decreased cardiac function, myocarditis, pericardial effusion, mitral regurgitation, and coronary artery dilatation or aneurysm. correlated with time from symptom onset. Interestingly, children with acute MIS-C aged 0-5 and 6-12 years had significantly higher SARS-CoV-2 RBD IgG titers than adolescent patients 13-20 years of age. One possible explanation of this finding is that older children with MIS-C may be more likely to present concurrently with acute COVID-19, prior to their development of high-titer SARS-CoV-2 antibodies [2] . In contrast to patients with acute COVID-19, patients with MIS-C had significant declines in SARS-CoV-2 RBD IgG titers during early convalescence. This may be primarily attributed to the timing of acute MIS-C following SARS-CoV-2 infection, which typically follows peak COVID-19 transmission in the community by 2-6 weeks [4, 21, 22] . Nevertheless, this natural waning of RBD immune response may inform the optimal timing of vaccination post-MIS-C, as susceptibility to reinfection likely increases with time following the initial SARS-CoV-2 infection. In addition to having high titers of SARS-CoV-2 IgG antibodies, patients with MIS-C had significant elevations of multiple proinflammatory and Th1-type cytokines, consistent with a cytokine storm. Although the term cytokine storm lacks a strict definition, it has been described as the dysregulated release of interleukins, interferons, tumor necrosis factors, and other small-molecule mediators that results in broad immune cell activation and an end-organ damage [23] . Hypercytokinemia, and specifically elevations in IL-6, IL-8, and TNF-α, has been associated with severe outcomes and death among adult patients with COVID-19 [24, 25] . However, less is known about the cytokine signatures of pediatric COVID-19 and MIS-C. In this study, we found that IL-6, IL-10, IL-17A, and IFN-γ were strongly associated with the diagnosis of MIS-C. These results overlap with the findings of previously published reports of cytokine analyses in smaller cohorts of patients with MIS-C. For example, Consiglio et al found that IL-6 and CXCL10 contributed to the cytokine storm observed in patients with MIS-C [11] . Diorio et al found that the most discriminatory cytokines among a small cohort of patients with MIS-C and severe COVID-19 were IL-10 and TNF-α [26] . Moreover, Gruber et al found IL-6 and IL-17A to be elevated in MIS-C, in addition to other cytokines involved in inflammation, lymphocytic and myeloid cell chemotaxis and activation, and mucosal immune dysregulation [8] .

The cytokine responses we observed in patients with MIS-C were distinct from acute and convalescent COVID-19. The unique serologic and cytokine signatures we identified could add diagnostic and prognostic value for patients presenting with signs and symptoms compatible with MIS-C. The diagnosis of MIS-C in the United States is currently based upon meeting the CDC case definition, which requires the presence of fever, multiorgan involvement, systemic inflammation, lack of an alternative explanatory diagnosis, and epidemiologic link or positive test for SARS-CoV-2. However, clinical features overlap with other inflammatory syndromes and there is concern that the increasing SARS-CoV-2 seroprevalence, both due to natural infection and vaccination, may confound the interpretation of spike or RBD SARS-CoV-2 antibody titers in cases of suspected MIS-C. In our cohort, we found that almost all patients with MIS-C had high titers of RBD IgG antibodies, and that RBD titers correlated with nucleocapsid IgG titers with the log-titer levels having a strong linear association. Thus, quantitative nucleocapsid serology is likely to retain applicability as a biomarker of MIS-C in the setting of widespread pediatric immunization with spike protein-based vaccines. Nevertheless, few prognostic indicators have been identified for either short-or long-term MIS-C outcomes. In our study, we found that IL-6, IL-10, IL-17A, and IFN-γ were significantly associated with prolonged duration of hospitalization, but not other categorical outcomes. Elevations in at least 3 of the 4 cytokines was also associated with elevated peak D-dimer and decreased platelet nadir, both of which are thought to contribute to MIS-C pathophysiology. All cytokine levels returned to normal within approximately 2 months, suggesting that MIS-C represents a transient state of immune activation and hyperinflammation following SARS-CoV-2 infection.

Understanding the pathogenesis of MIS-C could better inform targeted approaches to patient management. Current treatment strategies commonly include intravenous immunoglobulin, systemic corticosteroids, or immunomodulatory monoclonal antibodies, such as the IL-1β inhibitor anakinra. All of these interventions target broad or specific components of the inflammatory cascade, to reduce the risk that the transient state of immune activation leads to end-organ damage. However, the underlying trigger for the systemic hyperinflammatory response remains elusive. A recent study demonstrated that persistence of SARS-CoV-2 in the gastrointestinal tracts of patients with MIS-C was associated with a breakdown of mucosal integrity and subsequent spike protein antigenemia and hyperinflammation [27] . While this explanation could correspond with the prominent gastrointestinal symptoms observed in MIS-C, the finding of SARS-CoV-2 antigenemia has not been universal [28] . The reasons why antigenemia may trigger delayed but not acute systemic hyperinflammation are similarly unclear. Interestingly, Kumar et al found significantly elevated markers of microbial translocation in children with MIS-C, which could contribute to systemic inflammatory responses [29] . Future studies are needed to elucidate MIS-C pathogenesis and to prospectively evaluate the efficacy of various treatment modalities upon MIS-C outcomes.

Strengths of our study include the prospective, multicenter design and large sample size of well-characterized patients with MIS-C. Limitations include the small numbers of convalescent samples available and the lack of long-term follow-up. The inclusion criteria of our study may have limited enrollment of clinically ambiguous cases that did not meet the prespecified case definitions. We did not assess serologic and cytokine profiles of other pediatric hyperinflammatory conditions, such as Kawasaki disease or toxic shock syndrome. From a diagnostic standpoint, the comparison of immune profiles with acute COVID-19 may have greatest applicability to adolescents and/ or adults with MIS-A, who have a more heterogeneous clinical presentation that often overlaps with acute COVID-19 [30, 31] . We were unable to evaluate treatment effects on serologic and cytokine profiles and clinical outcomes, although there were no pediatric deaths due to MIS-C and available convalescent samples showed cytokine normalization. We did not measure neutralizing antibody responses, although we and others have previously shown that these correlate with SARS-CoV-2 RBD IgG binding antibodies [7, 32] . Similarly, we were unable to evaluate associations between serologic and cytokine profiles with nasopharyngeal viral load by reverse-transcription polymerase chain reaction. Although the vaccination status of the children was not prospectively collected, all enrollments preceded the US Food and Drug Administration emergency use authorization of COVID-19 vaccination for children <16 years of age. Enrollments also preceded the surges of SARS-CoV-2 Delta and Omicron variants in the United States. The effects of emerging SARS-CoV-2 variants of concern on MIS-C disease pathogenesis and clinical responses remain areas of active research.

In conclusion, MIS-C was associated with high titers of SARS-CoV-2 RBD IgG antibodies and elevated IL-6, IL-10, IL-17A, and IFN-γ. In convalescence, antibody titers waned and hypercytokinemia resolved, suggesting transient immune activation and hyperinflammation in acute MIS-C.

Hyperinflammatory shock in children during COVID-19 pandemic

COVID-19-associated multisystem inflammatory syndrome in children-United States

6-month multidisciplinary follow-up and outcomes of patients with paediatric inflammatory multisystem syndrome (PIMS-TS) at a UK tertiary paediatric hospital: a retrospective cohort study

Multisystem inflammatory syndrome in U.S. children and adolescents

Multisystem inflammatory syndrome in children in New York State

Acute heart failure in multisystem inflammatory syndrome in children in the context of global SARS-CoV-2 pandemic

Quantitative SARS-CoV-2 serology in children with multisystem inflammatory syndrome (MIS-C)

Mapping systemic inflammation and antibody responses in multisystem inflammatory syndrome in children (MIS-C)

Severe acute respiratory syndromecoronavirus-2 (SARS-CoV-2) antibody responses in children with multisystem inflammatory syndrome in children (MIS-C) and mild and severe coronavirus disease 2019 (COVID-19)

SARS-CoV-2 immune repertoire in MIS-C and pediatric COVID-19

The immunology of multisystem inflammatory syndrome in children with COVID-19

Plasma biomarker profiling of PIMS-TS, COVID-19 and SARS-CoV-2 seropositive children-a cross-sectional observational study from southern India

COVID-19) 2020 interim case definition

Rapid generation of neutralizing antibody responses in COVID-19 patients

Research electronic data capture (REDCap)-a metadata-driven methodology and workflow process for providing translational research informatics support

The REDCap consortium: building an international community of software platform partners

R: A language and environment for statistical computing

CatPredi: optimal categorisation of continuous variables in prediction models

Regularization paths for generalized linear models via coordinate descent

A new approach to categorising continuous variables in prediction models: proposal and validation

Multisystem inflammatory syndrome in children (MIS-C): report of the clinical and epidemiological characteristics of cases in Santiago de Chile during the SARS-CoV-2 pandemic

Trends in geographic and temporal distribution of US children with multisystem inflammatory syndrome during the COVID-19 pandemic

Is a "cytokine storm" relevant to COVID-19?

Profiling serum cytokines in COVID-19 patients reveals IL-6 and IL-10 are disease severity predictors

An inflammatory cytokine signature predicts COVID-19 severity and survival

Multisystem inflammatory syndrome in children and COVID-19 are distinct presentations of SARS-CoV-2

Multisystem inflammatory syndrome in children is driven by zonulin-dependent loss of gut mucosal barrier

Measurement of SARS-CoV-2 antigens in plasma of pediatric patients with acute COVID-19 or MIS-C using an ultrasensitive and quantitative immunoassay. medRxiv

Systemic inflammation and microbial translocation are characteristic features of SARS-CoV-2-related multisystem inflammatory syndrome in children

Characteristics associated with multisystem inflammatory syndrome among adults with SARS-CoV-2 infection

Clinical characteristics of multisystem inflammatory syndrome in adults: a systematic review

Rapid generation of neutralizing antibody responses in COVID-19 patients

We thank Dr Jens Wrammert for kindly sharing purified receptor-binding domain protein antigen. We thank clinical research coordinators Beena Desai, Kerry Dibernardo, Felicia Glover, Vikash Patel, Maureen Richardson, Amber Samuel, Cindy Terrill, and Jasmine Prater and clinical research nurses Lisa Macoy, Kathy Stephens, and Lori Barganier for their assistance enrolling patients and collecting specimens. We thank Nadine Rouphael and the Hope Clinic laboratory, Theda Gibson, Wensheng Li, and the Emory Children's Center Vaccine Research Clinic laboratory for their assistance processing specimens. We thank the Children's Healthcare of Atlanta research laboratory for their assistance in collecting residual specimens. And last, we thank the study participants and their families for generously donating their time and blood to further our understanding of coronavirus disease 2019 and multisystem inflammatory syndrome in children.Patient consent. All human subjects were prospectively enrolled following written or verbal informed consent and assent as appropriate for age and as approved by each site's local institutional review board. Verbal consent and assent were allowed at some sites for the purpose of limiting exposure to severe acute respiratory syndrome coronavirus 2. All procedures followed were in accordance with the ethical standards of the Helsinki Declaration (1964, amended most recently in 2008) of the World Medical Association.

Financial support. This study was funded by the U.S. Centers for Disease Control and Prevention.Potential conflicts of interest. E. J. A. has received personal fees from AbbVie, MedScape, Pfizer, and Sanofi Pasteur for consulting, and his institution receives funds to conduct clinical research unrelated to this manuscript from MedImmune, Regeneron, PaxVax, Pfizer, GSK, Merck, Novavax, Sanofi-Pasteur, Janssen, and Micron. He also serves on a safety monitoring board for Sanofi-Pasteur and Kentucky BioProcessing. C.A.R. 's institution has received funds to conduct clinical research unrelated to this manuscript from the National Institutes of Health, BioFire, GSK, MedImmune, Micron, Janssen, Merck, Moderna, Novavax, PaxVax, Pfizer, Regeneron, and Sanofi-Pasteur. She is coinventor of patented respiratory syncytial virus vaccine technology unrelated to this manuscript, which has been licensed to Meissa Vaccines. All other authors report no potential conflicts of interest.All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.