We describe 186 patients younger than 21 years of age who met the criteria for MIS-C associated with SARS-CoV-2 infection from across the United States. The majority of patients (70%) had laboratory-confirmed antecedent or concurrent SARS-CoV-2 infection, and most had no documented underlying conditions. Cardiovascular involvement was common, with almost half receiving vasopressor or vasoactive support and 1 in 12 having coronary-artery aneurysms. Most patients were cared for in an intensive care unit, and 20% received invasive mechanical ventilator support. Although most discharged patients survived, 28% were still hospitalized as of May 20, 2020, and 4 patients (2%) died, 2 of whom had previously been healthy.

Evidence supporting a causal link with SARS-CoV-2 includes a strong temporal association with Covid-19 activity, confirmation of SARS-CoV-2 infection through nucleic acid or antibody testing in the majority of patients, and hyperinflammatory manifestations similar to those in adults with Covid-19.24-26 Almost one third of the patients tested negative for SARS-CoV-2 by RT-PCR but had detectable antibodies. In a small subgroup of the patients in our series, a median interval of 25 days was reported between the onset of Covid-19 symptoms and hospitalization for MIS-C. Although not sufficient to establish causality, these findings suggest that a substantial proportion of the patients in this series were infected with SARS-CoV-2 at least 1 to 2 weeks before the onset of MIS-C.27

These patients met a case definition developed by the CDC that was designed to be sensitive.8 Although more than one third of the patients had Kawasaki’s disease–like clinical features, 60% of the patients would not have met complete or incomplete criteria for Kawasaki’s disease. Patients with Kawasaki’s disease–like features were more likely to be younger than 5 years old, similar to what is seen among patients with Kawasaki’s disease reported in the literature.9,10 Some of the Kawasaki’s disease–like features, including fever, erythroderma, and delayed desquamation, are also seen in toxic shock syndrome, which can have manifestations of multiorgan involvement and has been associated with other viruses.14,28

Although both Kawasaki’s disease and MIS-C can have cardiovascular involvement, the nature of this involvement appears to differ between the two syndromes. We observed cardiovascular involvement leading to vasopressor or vasoactive support as a common and severe subphenotype, most frequently in older children and adolescents. Approximately 5% of children with Kawasaki’s disease in the United States present with cardiovascular shock leading to vasopressor or inotropic support,10 as compared with 50% of the patients in our series. Myocardial dysfunction is a prominent extrapulmonary manifestation of Covid-19 that has been associated with increased mortality in adults.29,30 Coronary-artery aneurysms are a common feature of Kawasaki’s disease, affecting approximately one quarter of patients within 21 days after disease onset.31 In our series, a maximum z score of 2.5 or higher in the left anterior descending or right coronary artery was reported in 8% of the patients overall and in 9% of patients with echocardiograms. However, images may have been technically limited in some studies, and we were unable to evaluate the evolution of coronary dimensions after discharge. Cases of fatal and nonfatal myocardial infarction in otherwise healthy young adults have been attributed to undiagnosed Kawasaki’s disease during childhood.32 However, some patients had no Kawasaki’s disease findings, which underscores the importance of performing echocardiography in all patients presenting with MIS-C. Until more is known about long-term cardiac sequelae of MIS-C, providers could consider following Kawasaki’s disease guidelines for follow-up, which recommend repeat echocardiographic imaging at 1 to 2 weeks and 4 to 6 weeks after treatment for patients whose disease course is uncomplicated and more frequent echocardiography for patients with coronary-artery z scores of 2.5 or higher.10 Long-term monitoring for other potential sequelae of MIS-C will also be critical.

Understanding the pathogenesis of MIS-C will be necessary to inform clinical management and prevention efforts. In our case series, a majority of patients were treated with immunomodulatory drugs, most commonly intravenous immune globulin (77%) and systemic glucocorticoids (49%). Our study was not designed to evaluate or compare the long-term effectiveness of any therapies; however, most discharged patients survived. The selection of candidate treatment methods must be informed by whether organ damage is mediated by ongoing viral replication in the affected tissues, an exuberant host inflammatory response, or both.6,33 Antiviral agents may be beneficial in the former situation, whereas immunomodulatory agents are generally preferred for immune dysregulation.33 In adult patients, the onset of severe Covid-19 coincides with a decline in viral load in the respiratory tract and an increase in markers of hyperinflammation,3-5 although virus replication in nonrespiratory tissues cannot be ruled out. The clinical and the laboratory features of hyperinflammation, the timing of onset in relation to SARS-CoV-2 infection, and the similarities with the disease pattern in adults with Covid-195 support the hypothesis that MIS-C is a consequence of immune-mediated injury triggered by SARS-CoV-2 infection.

Although available data suggest that MIS-C is an uncommon complication of SARS-CoV-2 infection in children and adolescents, an essential question is why MIS-C develops in some patients in this age group and not in others. Potential age-specific differences among patients with MIS-C could result from differences in SARS-CoV-2 infection related to the likelihood of exposure or to differences in nasal expression of angiotensin-converting enzyme 2 (ACE2), the receptor used by SARS-CoV-2 for cell entry.34 In the sentinel MIS-C case series in London, five of the eight patients were of black or Afro-Caribbean descent.7 The percentage of patients in our series who were black or Hispanic was also higher than in the U.S. population overall35 and was similar to the reported percentage of children who are black or Hispanic among those who are 0 to 17 years of age and have Covid-19 in the U.S. population.36 Susceptibility to Kawasaki’s disease and its response to treatment may be influenced by the gut biome,37 as well as by signaling pathways and genetic variants.10,38 Exploration of potential host factors is also needed to identify potential determinants of developing MIS-C.10

We observed more cases of MIS-C in our sites in April than in March. The recent emergence of reported cases in several countries during the descent of the Covid-19 epidemic in those locations7,12,22,39 suggests that the increased detection of MIS-C in the later part of our surveillance period reflects a delayed onset after infection rather than an increase in community transmission. Surprisingly, published reports of illness similar to MIS-C occurring in China are lacking, with most reports of hospitalized children with Covid-19 in China indicating that their illness is nonsevere.1,40 Reasons for this observation are unclear and may involve differences in rates of infection in children, host factors, early treatment with immunomodulators, or incomplete reporting.

Our case series study has certain limitations. The working MIS-C case definition was intended to be sensitive, and as a result, some cases may have had a different underlying cause. Cases of MIS-C were identified only from the reporting hospitals in each state, and the results are not generalizable beyond the surveillance population. In the absence of a comparison group, caution is warranted in interpreting our case-series data to infer risk factors for MIS-C, particularly with respect to factors such as age, sex, and race or ethnic group. Although we used a standardized case report form, clinical management differs among centers, and therefore we may not have captured certain variables completely, including detailed echocardiographic data on coronary-artery outcomes and quality of imaging. Because of the limitations of a retrospective chart review and clinical testing, we were unable to accurately assess the time of onset of SARS-CoV-2 infection or the influence of false negative results on respiratory testing, nor were we able to test nonrespiratory specimens, including stool. Our pooled statewide results for SARS-CoV-2 test positivity may not accurately reflect test positivity or the incidence of Covid-19 in the catchment populations of our participating hospitals. Finally, to avoid duplicate reporting, we excluded 27 patients who had been included in the New York Department of Health report22; this resulted in the exclusion of 2 additional previously healthy patients who died, which may have influenced our estimates. We therefore include results based on the entire sample of 213 patients in the Supplementary Appendix. Taken as a whole, the evidence from our investigations, from early reports from New York State and Europe, and from the literature on adult patients suggests that the cases of MIS-C are part of a spectrum of Covid-19–related disease with severe immune-mediated pathology.

We report the emergence of a life-threatening hyperinflammatory syndrome across the United States that involves damage to multiple organ systems in predominantly previously healthy children and adolescents during the Covid-19 pandemic.