Multisystem Inflammatory Syndrome in Children: Two Years’ Worth of Learning

Abstract

Multisystem inflammatory syndrome in children (MIS-C) is a life-threatening sequela of SARS-CoV-2 infection. Limited data are available regarding risk-stratification or long-term outcomes in MIS-C. This study sought to determine associations between serologic markers and severity of illness and understand long-term cardiac outcomes. This series includes 46 cases (mean age 8.1 years; 63.0% male) of MIS-C. Pearson’s chi-squared analysis showed an erythrocyte sedimentation rate (ESR) greater than 30 mm/h and 50 mm/h were disproportionately associated with pediatric intensive care unit (PICU) admission (χ² = 4.44, P = .04) and use of vasopressors (χ² = 6.06, P = .01), respectively. Ferritin less than 175.6 ng/mL was associated with use of vasopressors (χ² = 5.28, P = .02). There was a negative correlation between ESR and ejection fraction (EF) (r = -0.39, P = .009). Most patients with abnormal echocardiograms had resolution of abnormalities within 30 days. Therefore, inflammatory markers may be helpful in predicting which patients may require specific interventions or experience cardiac dysfunction, but MIS-C does not appear to be associated with complications at 1 year.

Introduction

The first known case of coronavirus disease 2019 (COVID-19, caused by SARS-CoV-2) in the United States was reported in Washington State on January 31, 2020. ¹ Initial reports regarding the severity of COVID-19 in children and adolescents were reassuring, suggesting that pediatric cases were less severe than those seen in adults. ² Later that spring, multiple case reports were published, beginning in Italy and the United Kingdom, regarding a severe Kawasaki-like inflammatory syndrome associated with a history of or exposure to COVID-19.^3,4 These cases were reported several weeks after the initial surges of COVID-19 cases in those regions, just as similar cases were reported in New York.^4-8 The Centers for Disease Control and Prevention (CDC) issued a Health Alert defining what was termed “multisystem inflammatory syndrome in children” (MIS-C) on May 14, 2020. ⁹

MIS-C is characterized by a cytokine storm ¹⁰ and subsequent multisystem organ dysfunction, including gastrointestinal, cutaneous, and cardiac abnormalities.^5,7,11,12 Patients with MIS-C, on average, have been older and had a higher incidence of multisystem organ involvement and shock, particularly depressed cardiac function and less likely coronary involvement in comparison to Kawasaki’s disease.^13-15 Given the significant inflammatory component and concern for cardiac dysfunction, the American College of Rheumatology recommended using immunomodulating agents to treat MIS-C patients, including intravenous immunoglobulin (IVIG) and corticosteroids. ¹⁶ In MIS-C cases refractory to these therapies, additional immunomodulating drugs such as the interleukin antagonist Anakinra have been used with some success.^17,18 Recent data have indicated that COVID-19 vaccination may reduce the incidence of MIS-C in pediatric patients ¹⁹ ; nevertheless, vaccine hesitancy and breakthrough cases of SARS-CoV-2 in the vaccinated remain causes for concern,^20,21 demonstrating a need for continued lines of research related to MIS-C and other COVID-19 sequelae. Despite advancements in the literature regarding MIS-C over the last 2 years, there remain multiple unanswered questions regarding how to risk-stratify patients to predict which patients are more likely to require intensive care or other interventions. This observational study sought to review MIS-C cases at a large community hospital to determine what inflammatory markers are most associated with severe illness in MIS-C. A description of patients’ clinical outcomes following discharge was considered a secondary outcome.

Methods

The procedure for this review was submitted and approved by the Institutional Review Board at WakeMed Health and Hospitals (WakeMed; IRB Number 1921316-1) including granting a waiver of consent.

Case Identification

Over the course of the pandemic, pediatric hospitalists at WakeMed in Raleigh, North Carolina, have compiled an internal database of possible MIS-C cases based on clinical presentation, clinical judgment, and response to medical interventions. Health information management provided the research team with the medical record numbers (MRNs) from the electronic health record (EHR) of all patients whose hospital stay was coded with the International Classification of Diseases (ICD-10) code for MIS-C or who received IVIG between May 2020 and October 2021. Duplicate MRNs were identified and excluded. Between the hospitalist database and EHR query, 97 MRNs were slated for chart audit.

A member of the research team (Alexander), who is not a pediatric hospitalist, independently reviewed each MRN to determine if the patient met the Centers for Disease Control and Prevention (CDC) case definition of MIS-C. ⁹ To meet the case definition, patients needed to present with at least 24 hours of fever, evidence of inflammation, multisystem organ involvement, history of SARS-CoV-2 infection (determined by a history of exposure, reverse transcription–polymerase chain reaction [RT-PCR], or antibody assay), and no reasonable alternative diagnosis. For verification, the cases identified through systems data and chart review that the pediatric hospitalist group did not initially refer were confirmed by 2 other members of the research team (Lykes and Nassef) along with a random sample of 10% of the pediatric hospitalist cases. Of the 97 MRNs identified, 46 (47%) cases of MIS-C were identified by chart review. For the 51 (53%) cases that were excluded, 44 (86%) had no history of SARS-CoV-2 exposure or laboratory evidence of SARS-CoV-2 infection, 8 (16%) lacked elevated serologic inflammatory markers, 15 (29%) did not meet the temperature cutoff, 6 (12%) lacked evidence of multisystem involvement, and 18 (35%) had another, more plausible, diagnosis listed in their records. For a summary of exclusions, see Figure 1.

Figure 1.

Possible cases of MIS-C were identified by pediatric hospitalists, hospital visits coded with the ICD10 code for MIS-C, or if patients received IVIG in the audit timeframe. Some charts (*) did not meet multiple aspects of the CDC criteria.

Abbreviations: MIS-C, multisystem inflammatory syndrome in children; IVIG, intravenous immunoglobulin; MRN, medical record number; CDC, Centers for Disease Control and Prevention.

Data Collection and Analysis

During chart audits, various data were extracted based on criteria set by the pediatric hospital medicine group at WakeMed; this decision was based on broader literature regarding the treatment of Kawasaki’s disease and, more recently, MIS-C. These criteria included standard labs drawn on admission for suspicion of MIS-C (complete metabolic panel [CMP], complete blood count [CBC], C-reactive protein [CRP], Westergren’s erythrocyte sedimentation rate [ESR], troponin, D-Dimer, ferritin, brain natriuretic peptide [BNP], SARS-CoV-2 polymerase chain reaction [PCR] and serologies), data suggestive of the severity of a patient’s illness (lactate, admission to the pediatric intensive care unit [PICU], supplemental oxygen therapy, etc.), echocardiography given concerns regarding the development of cardiac complications,^4,11 and data from their outpatient follow-up (cardiology follow-up, steroid course). For those admitted to the PICU, the institution involved will triage those that show clinical signs of shock (e.g., frank hypotension) or signs which raise sufficient clinical concern that they may develop shock (namely, SIRS criteria).

Using the laboratory assays as continuous variables, linear regression analysis was performed to determine if a significant correlation existed between any 2 variables. Laboratory assays were also used to create multiple nominal variables, each representing whether a patient was above or below different serologic cutoffs (multiples of each assay’s upper limit of normal). Pearson’s chi-squared analysis was performed to determine if these cutoffs were disproportionately associated with PICU admission, supplemental oxygen therapy, pressors, or abnormal echocardiography findings. These data were analyzed using STATA BE Version 17.0 (StataCorp LLC, College Station, TX) with statistical significance defined as an alpha level of 0.05.

Results

Presentation

Of the 46 cases of MIS-C identified, 29 (63%) patients were male, and 17 (37%) were female. The average patient age at admission was 8.1 years (SD = 5.1). Patients were admitted for an average of 3.9 days (SD = 1.9). Of the patients, 18 (39%) were treated in the PICU during their hospital course, 8 (17%) were treated with supplemental oxygen therapy, and 14 (30%) were treated with vasopressors. Summary of these data among others can be found in Table 1.

Table 1.

Summary of Categorical Data on Patients Admitted With MIS-C.

Characteristic	No. of patients (N = 46)
Female (%)	29 (63)
Male (%)	17 (37)
Age at presentation
Mean (SD) [range]	8.1 (5.1) [0-19]
Median (IQR)	7 (8)
Length of Stay
Mean (SD) [range]	3.9 (1.9) [1-9]
Median (IQR)	3.5 (1)
PICU admission	18 (39.1)
Treatment with supplemental oxygen	8 (17.4)
Treatment with vasopressors	14 (30.4)
Treatment with diuresis	11 (23.9)
Readmission within 30 days of discharge	5 (10.9)

Abbreviations: MIS-C, multisystem inflammatory syndrome in children; IQR, interquartile range; PICU, pediatric intensive care unit.

On admission, 18 (39%) patients had an acute kidney injury, 11 (24%) had transaminitis, 34 (74%) had lymphopenia, and 15 (33%) had thrombocytopenia. A summary of studies from the patients’ laboratory profiles at admission can be found in Table 2.

Table 2.

Summary Statistics for Labs Drawn on Admission Based on the Pediatric Hospital Medicine Protocol for the Patients Under Investigation for MIS-C (N = 46).

Panel	N	Laboratory assay	Reference range (units)	Case mean (±SD)
Complete metabolic panel		Sodium	136-145 (mmol/L)	134.3 (±3.6)
	46	Creatinine	0.51-1.00 (mg/dL)	0.66 (±0.39)
		Alanine Transaminase	7-52 (IU/L)	39.3 (±32.0)
Complete blood count		White blood cell count	4.0-10.5 (K/µL)	10.9 (±6.5)
		Absolute lymphocyte count	0.4-5.6 (K/µL)	1.4 (±1.8)
	46	Hemoglobin	Female: 12.0-15.0 (g/dL)	11.6 (±1.7)
			Male: 12.1-16.1 (g/dL)
		Platelet count	150-450 (K/µL)	207.8 (±111.2)
	46	C-reactive protein	≤1.0 (mg/dL)	16.2 (±9.9)
	43	Erythrocyte sedimentation rate	0-10 (mm/h)	60.6 (± 27.9)

Abbreviation: MIS-C, multisystem inflammatory syndrome in children.

Admission Through Discharge

During their hospital course, the average maximum temperature (T_max) was 39.4°C (SD = 0.7). Several labs were trended for admitted patients per the pediatric hospital medicine protocol. A summary of the highest recorded values for these labs can be seen in Table 3. While admitted, 44 patients (96%) were treated with IVIG and 10 mg/kg IV methylprednisolone per hospital protocol; the remaining 2 patients (4%) were not treated with this therapy based on the provider’s clinical judgment. Similarly, 41 patients (89%) were treated with aspirin, and 1 patient (2%) was treated with enoxaparin for anticoagulation.

Table 3.

Summary Data of Lab Values Monitored During Inpatient Hospitalization for MIS-C Cases.

Laboratory study	N	Reference range (units)	Case mean (±SD)
Troponin a	44	<0.05 (ng/mL)	0.41 (±1.2)
Ferritin	43	1.7-87.8 (ng/mL)	534.2 (±499.1)
C-reactive protein	46	≤1.0 (mg/L)	18.9 (±10.8)
D-dimer b	39	<231 (ng/mL)	1675.2 (±1144.3)
Brain natriuretic peptide	41	0-47 (pg/mL)	791.9 (±834.7)
Lactate	12	0-2.2 (mmol/L)	3.2 (±2.6)

The values reported are the average of the highest value recorded for each patient.

^aTroponin values of < 0.03 ng/mL are not resulted with an exact value. They are reported in the data set as 0.00 ng/mL.

^bThe D-dimer assay used has a maximum value of 5250 ng/mL which was exceeded in 1 patient in the data set.

Pearson’s chi-squared analysis showed that an ESR above 3 times the upper limit of normal (30 mm/h) at admission was significant for a disproportionate need for admission to the PICU (χ2 = 4.44, P = .04); above 5 times the upper limit of normal (50 mm/h), ESR was also significant for a higher-than-expected need for vasopressors (χ2 = 6.06, P = .01). D-dimer levels greater than 4 times the upper limit of normal (924 ng/mL, χ2 = 4.28, P = .04) were also significant for a higher need for admission to the PICU than expected by random chance. Chi-squared analysis also showed a statistically disproportionate use of vasopressors (χ2 = 5.28, P = .02) at ferritin cutoffs below 2 times the upper limit of normal (175.6 ng/mL). Patients below this cutoff required vasopressors more often than expected by random chance, and patient’s about this cutoff required vasopressors less often.

Of the patients with MIS-C, 14 (30%) had abnormal echocardiogram findings. Four patients (9%) had coronary artery ectasia (z score > 2.0 on echocardiography), 11 (23%) had a below-normal ejection fraction (EF < 55%) (m = 47%, SD = 4.6), and 9 (20%) had a pericardial effusion. Of those with abnormal coronary artery findings, 4 (4, 100%) had dilation of the left anterior descending artery, 3 (3, 75%) had dilation of the left main coronary artery, and 3 (3, 75%) had dilation of the right coronary artery with the largest relative dilation seen in 1 patient’s left anterior descending measuring 4.4 mm (z = 4.09). Pearson’s chi-squared analysis showed that a serologic cutoff for BNP of 3 times the upper limit of normal (141 pg/mL) was associated with a greater number of abnormal echocardiography results than expected (χ2 = 4.90, P = .03). When comparing each patient’s lowest inpatient ejection fraction (EF) and various measured inflammatory markers, there was a statistically significant correlation between ESR and EF (r = -0.39, P = .009, Figure 2). Of the 14 patients with abnormal echocardiography findings, 12 (86%) had resolution of their abnormalities within 30 days from the date of presentation (Median = 13.5 days, interquartile range [IQR]: 2-17 days).

Figure 2.

Scatterplot demonstrating a moderate negative correlation (r = −0.39) between each patient’s lowest recorded ejection fraction while admitted and Westergren’s erythrocyte sedimentation rate (P = 0.009).

The average length of stay was approximately 4 days (mean = 3.9, SD = 1.9). Of the available biomarkers, none were significantly correlated with the length of stay.

Follow-Up

No deaths were reported in this case series. Following discharge, 7 patients (16%) were lost to follow-up. Of the patients with documented follow-up, 1 (2%) later developed coronary artery ectasia that subsequently resolved with continued oral steroids. For the 33 patients with a specific steroid taper documented at discharge or on follow-up, the average number of days on steroids following discharge was 20.4 days (SD = 5.8). Five (11%) patients were readmitted within 30 days of discharge; the most common reason for readmission was observation due to recurrent fevers without significant biomarker associations between their initial or readmission presentations. There were no clear predictive laboratory abnormalities in the dataset for readmission.

Discussion

This case series describes MIS-C in children and adolescents at a large community hospital in North Carolina between May 2020 and October 2021. Cases were confirmed through chart audits using a strict interpretation of the CDC’s published criteria ⁹ for MIS-C. While the patients described in this case series do not drastically differ from patients in other published case series and public data,^5,7,12,18 this case series underscores the fact that MIS-C is a heterogeneous syndrome with a wide spectrum of presentation.

The primary objective of this observational study was to determine if any serologic markers are associated with the more severe phenotype of MIS-C and may be used to risk-stratify patients. The associations between ESR and D-dimer assays with PICU admission and, for ESR, vasopressor therapy suggest that these specific inflammatory labs may prove useful for risk stratification. The negative correlation of ESR and ejection fraction further supports the use of this assay in risk-stratifying patients, although significant variability still exists in the sample population that is not accounted for using simple linear regression as above. While the associations with these inflammatory markers is not entirely surprising, the authors suspected to find a similar association with elevations in CRP based on clinical experience and its specificity (such as the association seen in Kawasaki’s disease) ²² ; however, a statistically significant association was not observed in this population. Erythrocyte sedimentation rate has been noted to serve as an independent predictor of cardiac disease in previous studies^23,24; although the exact mechanism is unclear for the link between cardiac dysfunction and ESR value, this may suggest the presence of additional serum proteins which would cause additional Rouleaux formation of the blood.^25,26 This could explain the elevation in ESR but not other, more specific protein-based assays such as CRP and serum ferritin.

In the literature, higher serum ferritin has previously been associated with more severe cases of MIS-C ²⁷ ; in this series, however, serum ferritin lower than 2 times the upper limit of normal was more likely to be associated with the need for vasopressors than predicted by random chance. This could suggest that imbalanced inflammatory responses could contribute to the severity of MIS-C in similar ways to those seen in primary COVID-19 infection. ²⁸ This finding should be validated in larger sets of data before guiding clinical decision-making.

The follow-up data in this case series provides an optimistic long-term outlook for patients presenting with MIS-C. In this series, almost all patients were treated with standard therapies (IVIG, methylprednisolone, and aspirin) without significant long-term complications. Even for patients presenting with cardiac manifestations of disease, the majority saw the resolution of their findings within 30 days from the date of presentation. While more potent immunomodulatory drugs are available for refractory cases, ¹⁸ these data support that standard therapies as the first line intervention are sufficient to prevent adverse outcomes.

These findings represent a timely analysis in the setting of recent spikes in SARS-CoV-2 infections following the emergence of new, more infectious variants.^29,30 If trends continue, MIS-C cases will follow each spike in SARS-CoV-2 infections by 4 to 6 weeks. Given the growing concerns that more infectious variants will continue to circulate, becoming endemic in the population, ongoing data will be necessary to assess and update recommendations to MIS-C therapies for the foreseeable future.

As with any observational study, there is no treatment group available for comparison, limiting the conclusions drawn from the data. Our data are also limited to the observational data available, preventing us from understanding exact resolution periods of laboratory and imaging abnormalities. In this series, the use of surrogate measures of clinical status and the low sample size also limits the power of the statistical analysis, making it susceptible to extreme values. A strict interpretation of the CDC case definition was used to meet the case definition. Given the previously mentioned spectrum of disease, however, the CDC definition may be limiting, excluding cases of MIS-C on the less severe end of the spectrum that fail to meet specific aspects of the criteria that may otherwise satisfy the case criteria. ³¹ Future research should seek to remedy these limitations via aggregate, multicenter studies and, where possible, randomized controlled trials to assess different treatment interventions in comparison to one another and which further describe correlation between serum biomarkers and clinical markers of illness (eg, hypotension).

Based on this information, the use of inflammatory markers, particularly ESR and D-dimer, above certain thresholds may guide which patients are more likely to require interventions such as admission to the PICU and vasopressors. Multisystem inflammatory syndrome in children presents with such a broad spectrum of disease but, when treated with IVIG and steroids, does not appear to be associated with significant cardiac complications after discharge.

Author Contributions

All authors were involved in the conception of and analysis involved in this study. Alexander drafted the manuscript which was critically reviewed and revised by all authors.