Association of Intravenous Immunoglobulins Plus Methylprednisolone vs Immunoglobulins Alone With Course of Fever in Multisystem Inflammatory Syndrome in Children

Key Points

Question  Is there an association between treatment with intravenous immunoglobulins (IVIG) plus methylprednisolone vs IVIG alone and course of fever in multisystem inflammatory syndrome in children (MIS-C) associated with severe acute respiratory syndrome coronavirus 2?

Findings  This retrospective cohort study included 111 children with MIS-C. After propensity score matching, the rate of treatment failure (defined by the persistence of fever 2 days after the introduction of first-line therapy or recrudescence of fever within 7 days) for those who received IVIG plus methylprednisolone vs IVIGs alone was 9% vs 51%, a difference that was statistically significant.

Meaning  Combined treatment with methylprednisolone vs IVIG alone was associated with a better course of fever in MIS-C.

Importance  Multisystem inflammatory syndrome in children (MIS-C) is the most severe pediatric disease associated with severe acute respiratory syndrome coronavirus 2 infection, potentially life-threatening, but the optimal therapeutic strategy remains unknown.

Objective  To compare intravenous immunoglobulins (IVIG) plus methylprednisolone vs IVIG alone as initial therapy in MIS-C.

Design, Setting, and Participants  Retrospective cohort study drawn from a national surveillance system with propensity score–matched analysis. All cases with suspected MIS-C were reported to the French National Public Health Agency. Confirmed MIS-C cases fulfilling the World Health Organization definition were included. The study started on April 1, 2020, and follow-up ended on January 6, 2021.

Exposures  IVIG and methylprednisolone vs IVIG alone.

Main Outcomes and Measures  The primary outcome was persistence of fever 2 days after the introduction of initial therapy or recrudescence of fever within 7 days, which defined treatment failure. Secondary outcomes included a second-line therapy, hemodynamic support, acute left ventricular dysfunction after first-line therapy, and length of stay in the pediatric intensive care unit. The primary analysis involved propensity score matching with a minimum caliper of 0.1.

Results  Among 181 children with suspected MIS-C, 111 fulfilled the World Health Organization definition (58 females [52%]; median age, 8.6 years [interquartile range, 4.7 to 12.1]). Five children did not receive either treatment. Overall, 3 of 34 children (9%) in the IVIG and methylprednisolone group and 37 of 72 (51%) in the IVIG alone group did not respond to treatment. Treatment with IVIG and methylprednisolone vs IVIG alone was associated with lower risk of treatment failure (absolute risk difference, −0.28 [95% CI, −0.48 to −0.08]; odds ratio [OR], 0.25 [95% CI, 0.09 to 0.70]; P = .008). IVIG and methylprednisolone therapy vs IVIG alone was also significantly associated with lower risk of use of second-line therapy (absolute risk difference, −0.22 [95% CI, −0.40 to −0.04]; OR, 0.19 [95% CI, 0.06 to 0.61]; P = .004), hemodynamic support (absolute risk difference, −0.17 [95% CI, −0.34 to −0.004]; OR, 0.21 [95% CI, 0.06 to 0.76]), acute left ventricular dysfunction occurring after initial therapy (absolute risk difference, −0.18 [95% CI, −0.35 to −0.01]; OR, 0.20 [95% CI, 0.06 to 0.66]), and duration of stay in the pediatric intensive care unit (median, 4 vs 6 days; difference in days, −2.4 [95% CI, −4.0 to −0.7]).

Conclusions and Relevance  Among children with MIS-C, treatment with IVIG and methylprednisolone vs IVIG alone was associated with a more favorable fever course. Study interpretation is limited by the observational design.

Children account for only 1% to 2% of hospitalized patients with coronavirus disease 2019 (COVID-19).¹ However, in April 2020, severe systemic hyperinflammatory disease was reported in children in Europe and the United States, occurring 2 to 4 weeks after severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection.^2-6 This novel entity, named multisystem inflammatory syndrome in children (MIS-C)⁷ or pediatric multisystem inflammatory syndrome temporally associated with COVID-19,⁸ is associated with a wide range of clinical features including persistent fever, digestive symptoms, rash, bilateral nonpurulent conjunctivitis, mucocutaneous inflammation signs, and frequent cardiovascular involvement.^2,4-6,9 MIS-C is often associated with hemodynamic failure, with acute cardiac dysfunction requiring hemodynamic support in 60% to 75% of cases,^5,6 sometimes associated with death.^2-6

Many children with MIS-C have received empirical treatment based on Kawasaki disease guidelines, with intravenous immunoglobulin (IVIG) alone or combined with corticosteroids.^2-4,6,10 In some studies, children have required second-line treatment, such as tumor necrosis factor inhibitor or interleukin 1 inhibitor, which underscores the importance of defining optimal initial therapy.^11,12

However, evidence for the most effective therapies for MIS-C is still lacking.^10,13 In the absence of evidence, a British Delphi consensus study proposed treating MIS-C with IVIG as initial therapy.¹⁴

The goal of this retrospective cohort study was to compare the outcomes of children with MIS-C associated with SARS-CoV-2 infection treated with IVIG and methylprednisolone vs IVIG alone.

The study was approved by the INSERM ethics committee for evaluation (IRB00003888). A written information form validated by the ethics committee was given to all participants. Oral consent was obtained from study participants; no family members or participants refused to participate.

The reporting of all suspected MIS-C cases in France became mandatory since the first descriptions of this entity in April 2020. This reporting was coordinated by the French National Public Health Agency, with a methodology previously published.^5,15 All French pediatric hospitals were mandated to report any suspected case of MIS-C to the French National Public Health Agency, without waiting for the SARS-COV-2 antibody test result.¹⁶ An electronic case report form for each patient was stored in a secure database based on clinical and biological files and shared by each pediatric hospital. The following data were recorded: demographic characteristics, comorbidities, initial symptoms and clinical signs, biological and microbiological parameters, radiography findings, treatments, and course during hospitalization.¹⁷ Then 2 research members (N. O. and F. A.) classified each case as confirmed MIS-C or not following World Health Organization (WHO) criteria (Box).¹⁸

Box Section Ref ID

All children with confirmed MIS-C associated with SARS-CoV-2 infection fulfilling WHO criteria up to October 22, 2020, were included in the study. The final date of follow-up was January 6, 2021.

The primary outcome was treatment failure, defined as persistence of fever for 2 days (48 hours) after the introduction of initial therapy or recrudescence of fever within 7 days after the initial therapy in children who received IVIG and methylprednisolone vs IVIG alone as first-line therapy. Time was measured from the start of administration of IVIG or methylprednisolone. This outcome was similar to the primary outcome used in therapeutic studies of Kawasaki disease^19-21 and has been associated in Kawasaki disease with increased risk of further cardiovascular complications.^20,21 IVIG and methylprednisolone was considered initial therapy when the beginning of administration of the 2 therapies occurred within 24 hours of one another. Fever was defined as a temperature of 38 °C (ie, ≥100.4 °F) or greater.²²

Secondary outcomes were second-line therapy, defined by a secondary treatment, such as steroids or biological agents, for MIS-C prescribed at least 24 hours after the initial therapy¹⁴; hemodynamic support after first-line therapy; occurrence of acute left ventricular dysfunction, defined by left ventricular ejection fraction less than 55% after first-line therapy; and duration of stay in the pediatric intensive care unit (PICU). For hemodynamic support and acute left ventricular dysfunction, only outcomes occurring at least 1 day after the initiation of first-line therapy were considered. Hemodynamic support was defined as vasoactive or inotropic amine but not the escalation of a previously prescribed drug.

From available data, and prior to any data collection, assuming a risk of treatment failure of 50% in the IVIG alone group^11,12 and that the risk with IVIG and methylprednisolone would be reduced to 20%, with a total of 76 patients, the study was estimated to have 80% power to detect such a difference, assuming 2-sided tests.

The main analysis involved propensity score matching.^23,24 The propensity score was calculated with a multivariable logistic regression model to establish each patient’s probability of receiving combination therapy with corticosteroids according to baseline characteristics. The following baseline characteristics were used to generate the propensity score: age, sex, comorbidities, hospital center, gastrointestinal symptoms, lower respiratory tract symptoms, neurological symptoms, initial acute left ventricular dysfunction (left ventricular ejection fraction <55%), intensity of inflammatory syndrome (C-reactive protein level > or ≤150 mg/L), positive SARS-CoV-2 antibody test result, initial PICU care before first-line MIS-C therapy, and initial hemodynamic support before first-line MIS-C therapy (eTable 1 in Supplement 1). All of these variables, except the SARS-CoV-2 antibody test result, were assessed at the initial presentation (ie, before the start of the first-line MIS-C therapy).

Patients who received IVIG and methylprednisolone were matched to those who received IVIG alone by their propensity score using nearest-neighbor matching, with a minimum caliper of 0.1. The ratio was 1 patient receiving IVIG and methylprednisolone matched with 2 patients receiving IVIG alone. The balance between the 2 treatment groups for each covariate was assessed with a standardized difference less than 0.1, which was considered acceptable.²³ Conditional logistic regression analysis was performed with the matched cohort to test the association between treatment groups and each outcome, with findings expressed as absolute risk differences, odds ratios (ORs), and 95% CIs. The analysis was also adjusted for center by using random-effects modeling, which involved fitting a mixed-effects logistic regression model.²⁵

Six sensitivity analyses were performed to assess the robustness of the study findings. First, the data were analyzed using inverse probability of treatment weighting, an alternative to propensity score matching to account for indication bias in nonrandomized design.^23,24 Unlike propensity score matching, this strategy has the advantage of including all the patients in the final analysis.²⁴ Second, a propensity score–matched analysis was conducted with center as a fixed effect.²⁵ Third, a propensity score–matched analysis was conducted with double adjustment on the most likely confounding variables²⁶ of initial hemodynamic support and initial left ventricular dysfunction. This strategy has been proposed to remove residual confounding after propensity score matching for the main potential confounders and adjusts for these variables both for propensity score calculation and in the final conditional logistic regression analysis with the matched cohort.²⁶ Fourth, a propensity score–matched analysis was conducted including the duration of fever before first-line MIS-C therapy and the delay between hospital admission and start of first-line MIS-C therapy as an additional baseline covariate to account for potential differences in the delay between disease onset and the start of first-line therapy between the 2 treatment groups. Fifth, a propensity score–matched analysis was conducted including mechanical ventilation and the vasoactive inotropic score²⁷ as additional baseline covariates to account for potential remaining differences in the initial severity of illness. Sixth, a logistic multivariable regression analysis adjusted on the variables included in the propensity score was conducted.

Because of the small sample size, subgroup analyses were conducted with the inverse probability of treatment weighting approach. Prespecified subgroups for analyzing the primary outcomes included presence or absence of initial acute left ventricular dysfunction, defined by left ventricular ejection fraction less than 55%, and age 10 years and older or younger than 10 years. This age categorization was based on the receiver operating characteristic curve to define the optimized cut-off value. To test for significant differences in effect size among subgroups, an interaction term was included in the main propensity score model for each subgroup.²⁸

A 2-sided P < .05 was considered statistically significant. Because of the potential for type I error due to multiple comparisons, findings for analyses of secondary end points should be interpreted as exploratory. All analyses were performed with R version 3.6.1 (http://www.R-project.org).

Among the 181 pediatric patients reported to the French National Public Health Agency with suspected MIS-C, 111 fulfilled WHO criteria for MIS-C associated with SARS-CoV-2 infection (Figure 1); 70 did not meet the criteria and were not included (eTable 2 in Supplement 1). Five children who met the criteria were not included because they did not receive either of the 2 treatments (eTable 3 in Supplement 1). The median age of the remaining 106 children was 8.6 years (interquartile range, 4.7 to 12.1), and 58 (52%) were female. Most children (n = 104; 94%) had gastrointestinal manifestations, and 52 (47%) had initial left ventricular dysfunction. In total, 74 children (67%) were initially admitted to a PICU, 46 (41%) received hemodynamic support, and 29 (26%) received ventilatory support. No deaths were recorded. Among the 111 children, 100 had positive results from SARS-CoV-2 antibody testing or nasopharyngeal reverse transcriptase–polymerase chain reaction and 11 had contact with an individual with SARS-CoV-2 infection. Other baseline characteristics not included in the propensity score are reported in eTable 4 in Supplement 1.

Among the 111 children, 34 received IVIG and methylprednisolone and 72 received IVIG alone as first-line therapy. Distribution of children by participating centers is reported in eTable 5 in Supplement 1. Among the 106 children who received IVIG and methylprednisolone or IVIG alone, none received any other immunomodulatory treatment before the initial therapy. The dosage of IVIG was 2 g/kg for all patients. A total of 30 of 34 patients in the IVIG and methylprednisolone group received methylprednisolone at 0.8 to 1 mg/kg every 12 hours (maximum of 30 mg for 12 hours) for 5 days; the 4 remaining children received a bolus of 15 to 30 mg/kg/d of methylprednisolone for 3 days. As compared with children who received IVIG alone, those who received IVIG and methylprednisolone had a more severe initial presentation with more frequent initial acute left ventricular dysfunction (22/34 [65%] vs 28/72 [39%]), initial PICU care (31/34 [91%] vs 42/72 [58%]), and initial hemodynamic support requirement (21/34 [62%] vs 23/72 [32%]).

Among the 34 children who received IVIG and methylprednisolone, 3 (9%) did not respond to treatment; among the 72 children who received IVIG alone, 37 (51%) did not respond to treatment.

Among the 34 children who received IVIG and methylprednisolone, 3 (9%) received a second-line therapy (a second IVIG course in 2 children and an interleukin 1 inhibitor in 1 child) (eTables 6 and 7 in Supplement 1). Two children had acute left ventricular dysfunction after initial therapy, and 2 required hemodynamic support.

Among the 72 children who received IVIG alone, 33 (46%, all of whom had treatment failure) received second-line therapies: a second IVIG course alone in 14, IVIG and methylprednisolone in 11, methylprednisolone alone in 2, and a biological agent in 6 (an interleukin 1 inhibitor in 4 and an interleukin 6 inhibitor in 2). Sixteen children showed acute left ventricular dysfunction after initial therapy, and hemodynamic support was introduced or added for 17 (24%) (eTables 6 and 7 in Supplement 1).

Among the 106 treated children, 32 in the IVIG and methylprednisolone group and 64 in the IVIG group were matched based on the propensity score (Figure 1). The treatment groups differed in several baseline characteristics, but after matching, the balance was satisfactory (Table 1 and the eFigure in Supplement 1). There were no missing data for all baseline covariates included in the propensity score.

IVIG and methylprednisolone compared with IVIG alone was associated with a lower rate of treatment failure (3/32 [9%] vs 24/64 [38%]; absolute risk difference, −0.28 [95% CI, −0.48 to −0.08]; OR, 0.25 [95% CI, 0.09 to 0.70]; P = .008; Table 2).

Treatment with IVIG and methylprednisolone vs IVIG alone was associated with a lower rate of second-line treatment (3/32 [9%] vs 20/64 [31%]; absolute risk difference, −0.22 [95% CI, −0.40 to −0.04]; OR, 0.19 [95% CI, 0.06 to 0.61]; P = .004), secondary acute left ventricular dysfunction (absolute risk difference, −0.18 [95% CI, −0.35 to −0.01]; OR, 0.20 [95% CI, 0.06 to 0.66]), and hemodynamic support (absolute risk difference, −0.17 [95% CI, −0.34 to −0.004]; OR, 0.21 [95% CI, 0.06 to 0.76]) (Table 2). The duration of PICU stay was also significantly shorter (median, 4 vs 6 days; difference in days, −2.4 [95% CI, −4.0 to −0.7]; P = .005) (Table 2).

All sensitivity analyses gave similar results (eTable 8 in Supplement 1), including the inverse probability of treatment weighting that included all 106 treated children. The association between IVIG and methylprednisolone treatment and lower rate of treatment failure compared with IVIG alone remained similar for children older and younger than 10 years and with or without initial acute left ventricular dysfunction (Figure 2, interaction test: P = .78 for age and P = .74 for initial acute left ventricular dysfunction).

No long-term cardiovascular complication or persistent inflammatory syndrome was reported in patients up to January 6, 2021.

Among children with MIS-C, treatment with IVIG and methylprednisolone vs IVIG alone was associated with a more favorable fever course. The combination therapy was also associated with less severe acute complications, including acute left ventricular dysfunction and hemodynamic support requirement.

A recent research letter²⁹ reported a single-center experience of cardiac evolution in 40 children with MIS-C and suggested that receiving IVIG plus corticosteroids was associated with a shorter cardiac recovery than receiving IVIG alone. Another single-center study reported that following the implementation of a new local protocol including corticosteroids for MIS-C treatment, hospital length of stay decreased.³⁰ Recent UK guidelines for MIS-C management,¹⁴ developed using a Delphi method and in the absence of comparative studies, suggested IVIG alone as first-line therapy, or no therapy in some cases. The findings in the current study may warrant reconsidering these recommendations.

A recent prospective surveillance of COVID-19 in children suggested that older children may be at increased risk of developing severe disease.¹⁷ This may reflect an overlap in young children between the classification of MIS-C and Kawasaki disease,³¹ which seems less severe than MIS-C.^2,11 In the current study, there were no significant interactions in subgroups based on age 10 years as a cut point. While the subgroup analyses are limited by a small sample size and should be interpreted with caution, they suggest that the association of IVIG and methylprednisolone with better outcomes may be similar in older and younger children. Further studies are required to confirm these findings.

Some recent preliminary mechanistic studies also suggested similarities between MIS-C and acute respiratory distress syndrome related to SARS-CoV-2 infection in adults.³² Corticosteroid use is currently one of the few validated therapeutics in severe respiratory adult forms of COVID-19.^33,34 The findings in the current study suggest that corticosteroids may also be beneficial in MIS-C, possibly acting systematically as a potent inhibitor of SARS-CoV-2–induced inflammation. Combined with findings of studies reporting MIS-C cases in young adults, there may be common pathways between severe respiratory adult forms of COVID-19 and MIS-C.^32,35 Additional studies are warranted to understand the mechanisms underlying a possible corticosteroid effect in MIS-C and in severe forms of COVID-19.

The main strength of this study was the use of data from a national surveillance system, propensity score–matched analysis to limit selection bias, and consistency of findings using other statistical approaches to control for potential bias.

This study has several limitations. First, it was not a randomized trial. While propensity score matching and inverse probability of treatment weighting were used to address limitations in the observation design, confounding by indication due to unmeasured covariates may remain. However, given the rarity and severity of MIS-C, conducting randomized trials may be highly challenging, and observational methods such as these may provide the best level of evidence.

Second, it is not certain that all patients had MIS-C. Patients with Kawasaki disease could have been infected with SARS-CoV-2 given the high prevalence of the pandemic in the general population. To limit this risk, study inclusion was based on WHO criteria for MIS-C.¹⁸ Furthermore, population baseline characteristics, including median age, rate of gastrointestinal symptoms, rash, bilateral nonpurulent conjunctivitis, or mucocutaneous inflammation signs, were similar to those in other reports of MIS-C.^2,4,6 The clinical features of some patients included in this study have been previously described^3,5,9,11,36 and are consistent with the literature.^2,4

Third, there was variation in the dosage and routes of steroid treatment used among the 34 children who received IVIG and methylprednisolone as first-line therapy, and the study design did not allow for comparing regimens. In the same way, because no patient received methylprednisolone alone or biological therapy as first-line therapy, other potential therapeutic approaches were not assessed. Further studies are needed to answer these questions.

Fourth, patients in each treatment group may have tended to present to the hospital at a different time in the natural course of this disease. However, the duration of fever before first-line therapy was similar between the 2 groups, and a sensitivity analysis including the duration of fever before first-line therapy as an additional baseline covariate had consistent results.

Fifth, patients with initial MIS-C may show symptoms of septic shock, and initial treatment with IVIG and methylprednisolone has a theoretical risk of worsening an unrecognized bacterial infection. Empirical antibiotic therapy might be initiated until the diagnosis is established to avoid the risk of an untreated infection with corticosteroid exposure.

Sixth, although no deaths occurred in the study population and despite mandatory reporting by the French National Public Health Agency, underascertainment is a possibility.

Among children with MIS-C, treatment with IVIG and methylprednisolone vs IVIG alone was associated with a more favorable fever course. Study interpretation is limited by the observational design.

Corresponding Author: François Angoulvant, MD, PhD, Paediatric Emergency Department, Necker-Enfants Malades University Hospital, 149 Rue de Sèvre, 75015 Paris, France (francois.angoulvant@aphp.fr).

Accepted for Publication: January 19, 2021.

Published Online: February 1, 2021. doi:10.1001/jama.2021.0694

Correction: This article was corrected on July 6, 2021, to add the list of nonauthor collaborators as a supplement.

Author Contributions: Drs Ouldali and Angoulvant had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Ouldali, Lachaume, M. Levy, C. Levy, Angoulvant.

Acquisition, analysis, or interpretation of data: Ouldali, Toubiana, Antona, Javouhey, Madhi, Lorrot, Léger, Galeotti, Claude, Wiedemann, Ovaert, Dumortier, Kahn, Mandelcwajg, Percheron, Biot, Bordet, Girardin, Yang, Grimaud, Oualha, Allali, Bajolle, Beyler, Meinzer, M. Levy, Paulet, C. Levy, Cohen, Belot, Angoulvant.

Drafting of the manuscript: Ouldali, Mandelcwajg, Yang, Angoulvant.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Ouldali, Bordet, Angoulvant.

Obtained funding: C. Levy, Belot.

Administrative, technical, or material support: Toubiana, Antona, Javouhey, Lachaume, Dumortier, Mandelcwajg, Girardin, Allali, Meinzer, Paulet, C. Levy, Belot, Angoulvant.

Supervision: Ouldali, Madhi, Mandelcwajg, Girardin, Bajolle, Beyler, C. Levy, Cohen, Belot, Angoulvant.

Conflict of Interest Disclosures: Dr Javouhey reported receiving grants from CSL Behring. Dr C. Levy reported receiving grants from GlaxoSmithKline, Merck Sharp & Dohme, and Sanofi and personal fees from Pfizer and Merck. Dr Cohen reported receiving personal fees from GlaxoSmithKline, Pfizer, Sanofi, and Merck Sharp & Dohme. No other disclosures were reported.

Funding/Support: This study received an unrestricted grant from Pfizer; the French Covid-19 Paediatric Inflammation Consortium received an unrestricted grant from the Square Foundation (Grandir–Fonds de Solidarité Pour L’enfance).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: A list of members in the French Covid-19 Pediatric Inflammation Consortium is in Supplement 2.

Additional Contributions: We are grateful to Santé Publique France, Société Française de Pédiatrie, Groupe de Paediatrie Generale, Groupe de Pathologie Infectieuse Pédiatrique, Groupe Francophone de Réanimation et d’Urgences Pédiatriques, Société Française de Cardiologie, Filiale de Cardiologie Pédiatrique et Congénitale, Société Francophone Dédiée à L’étude des Maladies Inflammatoires Pédiatriques, and Filière de Santé des Maladies Auto-immunes et Auto-inflammatoires Rares for their participation in the French Covid-19 Paediatric Inflammation Consortium study. We thank Isabelle Ramay, BSc, Claire Prieur, BSc, Marine Borg, Aurore Prieur, BSc, Laura Meyet, LLM, Jéremy Levy, BSc, Stéphane Bechet, MSc, and Sofia Abbou, LLM, from ACTIV (Association Clinique et Thérapeutique Infantile du Val-de-Marne), Créteil, France; Cecile Hoffart, MSc, and Maxime Brussieux, BSc, from Clinical Research Centre, Centre Hospitalier Intercommunal de Créteil, Créteil, France; Daniel Levy-Bruhl, MD, Mireille Allemand, Scarlett Georges, BSc, Valerie Olie, PhD, Nolween Regnault, PhD, and Jerome Naud, PharmD, from Santé Publique France, Agence Nationale de Santé Publique, Saint-Maurice; Murielle Herasse, PhD, from Filière de Santé Des Maladies Auto-immunes et Auto-inflammatoires Rares (FAI2R), Lyon, France; and David Skurnik, PhD, from INSERM U1151-Equipe 11. We are grateful to every microbiological laboratory staff member who performed severe acute respiratory syndrome coronavirus 2 reverse transcriptase–polymerase chain reaction and antibody testing in each center. None of the persons listed here received compensation for their role in the study.