Multisystem Inflammatory Syndrome in Children and Kawasaki Disease: Two Different Illnesses with Overlapping Clinical Features

In late April 2020, alarming news emerged from Europe that a group of children with evidence of recent severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection had developed a severe illness, manifesting fever, hypotension requiring inotropic support, severe abdominal pain, and myocardial dysfunction with marked elevation in cardiac damage markers. This syndrome has been named pediatric multisystem inflammatory syndrome in Europe, and multisystem inflammatory syndrome in children (MIS-C) by the US Centers for Disease Control and Prevention. As case series began to be reported, the similarities in the clinical features among the cases were striking, as were the frequent additional laboratory features of lymphopenia, thrombocytopenia, and cytokine storm with marked elevation in serum inflammatory markers including IL-6.1, 2, 3, 4, 5, 6, 7, 8 Most of these children recovered, although infrequently patients have required extracorporeal membrane oxygenation, with a fatal outcome from complications of this therapy reported rarely.^{3 , 7} Patients were managed in different ways, often with corticosteroid therapy, intravenous gamma globulin, and, less often, anticytokine therapies, and the vast majority appeared to require intensive care therapy only for a matter of days, regardless of the management strategy.^{1 ,} 5, 6, 7, 8 Some patients had 1 or more clinical features that can be observed in many illnesses of childhood, including Kawasaki disease, such as conjunctival injection, oral erythema, and rash. Classic Kawasaki disease diagnostic criteria were rarely present. Moreover, the median age of the cases was 9-10 years in the largest series reported to date, which is in marked contrast to Kawasaki disease, which occurs predominately in children 5 years of age or younger and with a peak incidence at ∼10 months of age.^{1 ,} 9, 10, 11 Asian children have the highest attack rates of Kawasaki disease, whereas the highest rates of MIS-C have been in children of African descent.^{3 , 12} These marked epidemiologic differences make it clear that the 2 conditions are not the same.

Because of a concern that MIS-C might potentially encompass a wider range of clinical features than those observed in the reported cases, the US Centers for Disease Control and Prevention developed a broad case definition. Unfortunately, the case definition as it currently stands is problematic, because patients with many infectious and inflammatory conditions of childhood that are not MIS-C fulfill the case definition. This includes patients with acute coronavirus disease 2019 (COVID-19) infection (eg, acute infection with fever, rash, diarrhea, and a minimally elevated C-reactive protein level), classic Kawasaki disease (eg, a child with Kawasaki disease who has rash as one of the diagnostic features and mild hepatitis or diarrhea), other viral infections (eg, one of many that could cause fever, rash, cough, and an increase in neutrophils in peripheral blood), systemic-onset juvenile idiopathic arthritis (eg, fever, rash, an increase in neutrophils in peripheral blood, elevated acute phase reactants), and so on. Moreover, on my recent clinical service, I noted an increase in patient transfers from referring hospitals of children with low-grade fever for 1 day, rash, and mild abdominal discomfort, for evaluation for possible MIS-C. These children arrived appearing well, with completely normal blood pressure and heart rate and minimally elevated acute phase reactants. Presumably, all the media publicity about MIS-C is raising concern among practitioners about missing this diagnosis, potentially resulting in an increase in hospital admissions for routine minor childhood illnesses, and a likely inflated number of reported cases. Moreover, a rapid or presumptive diagnosis of MIS-C, which remains uncommon, also has the potential to result in premature diagnostic closure in children who fulfill the broad case definition but in reality have a potentially life-threatening, non-MIS-C illness.

It is problematic that some similar clinical features of MIS-C and incomplete Kawasaki disease can lead to diagnostic uncertainty in individual patients. This scenario reminds me of my clinical experience in the late 1980s and early 1990s in Chicago, when we were experiencing a different viral epidemic, owing to measles. During that time, it was startling to see how closely the clinical features of Kawasaki disease resembled those of measles. Our approach at that time to differentiate the conditions was to perform a measles IgM antibody test, run daily in our hospital during the epidemic, and if the test was negative, to treat the patient for Kawasaki disease. Measles is one of the few infectious diseases that virtually always is symptomatic, so a positive test result confirmed the diagnosis. The present COVID-19 epidemic does not offer such a straightforward solution to the diagnostic dilemma. Not only is there no diagnostic test for Kawasaki disease at present, but asymptomatic or mildly symptomatic SARS-CoV-2 infection is prevalent enough in children in many areas of the US and abroad at the present time that even a positive reverse transcriptase polymerase chain reaction or serologic test for SARS-CoV-2 antibody does not necessarily mean that a child's presenting illness is related to SARS-CoV-2. With routine reverse transcriptase polymerase chain reaction screening of hospitalized patients, we find occasional positive results in children admitted with fractures, for routine surgical procedures, and with a variety of other clinical problems not related to SARS-CoV-2 infection. In addition, antibody positivity is very likely to be increasing over time in the population. Understandably, diagnostic confusion can result, and incorrect diagnoses may be assigned. To add to this confusion, some patients with MIS-C are reported to have developed mild coronary artery dilation or, rarely, aneurysms.1, 2, 3, 4 ^{, 9 , 12} That mild transient coronary artery dilation can develop as a result of cytokine storm with high IL-6 levels has been demonstrated in systemic onset juvenile idiopathic arthritis, and mild transient coronary dilation could be the result of a similar cytokine storm in MIS-C.¹³ Persistent coronary artery aneurysms and their complications, however, have previously only been attributed to Kawasaki disease in pediatric patients. Because of the clinical similarities in the 2 conditions, it is possible that patients reported as having MIS-C with persisting aneurysms actually had Kawasaki disease. Fortunately, patients with Kawasaki disease and MIS-C both improve with intravenous gamma globulin and corticosteroid therapy, although the efficacy of any therapy for MIS-C is evolving and is as yet is unproven. If SARS-CoV-2 can result in persisting coronary artery aneurysms, it would be noteworthy, because it would be the first virus proved to do so.

For patients in whom MIS-C and Kawasaki disease are being considered, some clinical and laboratory features of the 2 conditions may help in discerning the correct diagnosis. Abdominal pain significant enough to prompt advanced imaging and surgical consultation occurs rarely in Kawasaki disease, but is characteristic of MIS-C.¹⁴ Lymphopenia is a typical finding in MIS-C and is reported rarely in Kawasaki disease; the more severe the lymphopenia, the more likely the diagnosis is MIS-C. Although N-terminal pro B-type natriuretic peptide is a biomarker of potential value in Kawasaki disease and some patients have been reported to have levels as high as 7000 pg/mL, the extent of the elevation in Kawasaki disease generally is lower than in MIS-C, for which values in excess of 10 000 pg/mL often are reported; the higher the N-terminal pro B-type natriuretic peptide level, the more likely the diagnosis is MIS-C.^{15 , 16} Myocardial dysfunction as assessed by echocardiography is common in MIS-C and is rare in Kawasaki disease.

Some investigators have proposed that MIS-C, although having very different epidemiologic and some different laboratory features compared with Kawasaki disease, could provide etiologic clues that Kawasaki disease may result from infection with a coronavirus.^{4 , 12} The Table outlines some differing features of Kawasaki disease and coronavirus infections that make an etiologic relationship unlikely. Moreover, if SARS-CoV-2 or a closely related virus was a Kawasaki disease etiologic agent, one would expect the epidemiology of MIS-C and Kawasaki disease to be similar. Of note, MIS-C cases have not been observed in China or Japan, countries with the highest prevalence of Kawasaki disease in the world, which surely should have been expected if there was an etiologic relationship between Kawasaki disease and MIS-C.²⁹

Table.

Distinct features of Kawasaki disease and coronavirus infections

Kawasaki diseases	Coronavirus infections
No virus can be isolated from cultures of clinical specimens17,18	Virus can be isolated from cultures of clinical specimens
No coronavirus is identified by high throughput RNA sequencing of tissues from patients with Kawasaki disease19, 20, 21	Virus can be identified by high-throughput RNA sequencing of tissues from infected patients
No signal exists for serologic cross-reactivity with coronaviruses, even using new highly sensitive VirScan method17,18,20	Serologic cross reactivity occurs with other coronaviruses, particularly those in the same subfamily22
Recurrence is rare; disease is rare in adolescents and adults	Immunity wanes and infections with most coronaviruses generally recur lifelong23
Numerous reverse transcriptase polymerase chain reaction studies investigating coronavirus as potential cause have been negative	Viral RNA is consistently detected in patient samples by reverse transcriptase polymerase chain reaction
Inclusion bodies have been identified in ciliated bronchial epithelium that are targeted by antibodies from patients with Kawasaki disease19; virus-like particles found adjacent to inclusion bodies are about 50 nm in diameter24	No inclusion bodies are identifiable in bronchial epithelium; virus particles are ∼120 nm in diameter
Patients have an antigen-driven immune response that is not directed at coronavirus19	Immune response is directed at coronavirus
Coronary artery aneurysms occur; thrombosis is limited to within aneurysms	Hypercoagulability with vascular thrombosis at multiple sites is characteristic of SARS-CoV-2 infection; coronary artery aneurysms are not reported in acute SARS-CoV-2 infection; autopsy in the only pediatric patient reported to date with cardiac death from SARS-CoV-2 showed eosinophilic myocarditis with no evidence of vascular inflammation25
Epidemiologic and histologic evidence supports the hypothesis of persistent infection26,27	There is no persistent infection
The median age of patients with Kawasaki disease-associated shock is 2.8 years28	The median age of patients with SARS-CoV-2 associated pediatric shock is 9-10 years1,9

Many favor the theory that diverse infectious agents (including coronavirus) can trigger a “final common pathway” of immune dysregulation causing Kawasaki disease in a genetically susceptible individual. However, the idea that a genetically predisposed individual who is immunologically susceptible to multiple infections could be triggered only once (because Kawasaki disease rarely recurs) does not fit with our present understanding of human immunology. This hypothesis would also predict that Kawasaki disease could occur at any age, whereas the age distribution of Kawasaki disease is quite classic for a disease caused by a single ubiquitous infectious agent. One also cannot explain the occurrence of well-documented epidemics and outbreaks of Kawasaki disease with geographic wave-like spread of illness based on a theory of diverse infectious triggers.³⁰

MIS-C cases have appeared during periods of high prevalence of COVID-19 in an individual country, and also seem to decrease after a decrease in COVID-19 cases.³¹ It is hoped that when the SARS-CoV-2 pandemic is finally overcome, potentially by vaccination, MIS-C will no longer be a clinical issue. Although the etiologic agent of Kawasaki disease presently remains unclear, research studies support a ubiquitous RNA virus, different from presently known human viruses, that forms intracytoplasmic inclusion bodies in tissues from patients with Kawasaki disease.^{19 , 24} We have identified a protein epitope of the putative agent recognized by the antibody response in patients with Kawasaki disease, with substantial progress in developing the first-ever serologic assay for Kawasaki disease.¹⁹ The development of a diagnostic test for Kawasaki disease would enable distinguishing Kawasaki disease from the many infectious and inflammatory conditions in the differential diagnosis, which presently should include MIS-C.³² In the meantime, research is urgently needed to identify the pathogenesis of MIS-C, which is currently unknown but bears similarities to the cytokine storm observed in adult patients with COVID-19 during their second week of illness.³³