Introduction

Multi-system inflammatory syndrome in children (MIS-C) is a newly described condition characterized by the dysregulated response of the immune system after severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. During the first waves of the coronavirus disease 2019 (COVID-19) pandemic, peaks of MIS-C cases were reported worldwide [1, 2]. In Spain, in August 2020, a national guideline defined the treatment of MIS-C. The first line is IVIG or steroids. In severe cases or without response, it is recommended to combine them [3]. Before this guideline, because of the inflammatory nature of this condition and the overlapping features with Kawasaki disease, attending physicians started treating MIS-C with either steroids, intravenous immunoglobulin (IVIG), or their combination [3,4,5,6,7]. However, there is no high level of evidence that supports the use of one of the three therapeutical options versus the others [8].

The present study aimed to analyze the effect of steroids, IVIG, or their combination on the probability of discharge and of needing second-line treatment over time. As a secondary endpoint, we analyzed the effect of the drugs on persistent fever after 2 days of treatment.

Methods

We assessed children with MIS-C based on the WHO definition [9], and enrolled prospectively in the Epidemiological Study of COVID-19 in Children of the Spanish Society of Pediatrics (EPICO-AEP) [1]. They were diagnosed from 1 March 2020 to 1 June 2021. The study was approved by the Ethics Committee of Hospital 12 de Octubre, Madrid (code 20/101), and other participating hospitals. Participants were enrolled after signed or verbal consent from parents/guardians and by the consent of patients older than 12 years.

Baseline characteristics were described using frequencies for categorical variables and medians (interquartile range [IQR]) for continuous variables in the total population and stratified by treatment received. Groups were compared using chi-square or Fisher’s tests for categorical variables, and Mann–Whitney U tests for continuous variables.

IVIG dose was 2 g/kg, and steroids were 1–2 mg/kg/day of methylprednisolone or equivalent. Patients were discharged after clinical stability and fever resolution. We used a Markovian multi-state model with the clock-forward approach and unidirectional arrows to build a multi-state model. Three transitions were defined: initiation of treatment to hospital discharge (t1), initiation of treatment to second-line therapy (t2), and second-line therapy to hospital discharge (t3). For each transition, we estimated the time-to-event probability (discharge or second-line therapy) according to treatment initiation therapy (IVIG, steroids, or IVIG plus steroids) using a Cox model weighted with propensity score matching to balance baseline characteristics such as age at admission, presence of comorbidities, and period of recruitment. The weights for the propensity score matching were calculated using generalized linear models by estimating the average treatment effect on the treated using.

Time was measured in days from treatment initiation. A treatment was considered as second line if initiated > 2 days after the first therapy. A multivariable weighted logistic regression analysis was used to estimate the probability of persistent fever (temperature ≥ 38 °C) after 2 days of treatment for each treatment initiation strategy. All the analyses were performed using R software.

Results

Overall, 150 children with MIS-C were enrolled. Only patients initially treated with steroids, IVIG, or IVIG plus steroids were analyzed (n = 132/150; 88%). A total of 18/150 (12%) patients were excluded from the study, including 2 patients treated with steroids plus tocilizumab, 1 treated with tocilizumab, 1 treated with IVIG plus steroids plus anakinra, and 14 with no clear information about initial treatment. In total, 82/132 patients (62.1%) were admitted to a pediatric intensive care unit (PICU), 45/132 patients (34.1%) in the first period, and 87/132 (65.9%) in the second period.

The features of this population according to treatment received are described in Table 1 and according to the period in Supplementary Table 1. The unadjusted and adjusted sample after propensity score matching is summarized in Supplementary Fig. S1. Thirty of 132 patients (22.7%) were initially treated with steroids alone, 29/132 (21.9%) with IVIG alone, and 73/132 (55.3%) with IVIG plus steroids (Supplementary Fig. S2). Thirty-four of 132 (25.7%) patients needed second-line treatment over time: 9/30 (30%) children steroids, 12/29 (41.4%) IVIG, and 13/73 (17.8%) IVIG plus steroids, and 70/132 (53.0%) had persistent fever after 2 days of treatment. Forty-three of 70 (61.4%) patients did not receive a second-line treatment despite persistent fever; 29/43 (67.4%) were initially treated with IVIG plus steroids. In all these cases, fever remised without additional treatment after 48 h. Three patients (2.2%) died.

Table 1 Description of the study population according to treatment received. Continuous variables are presented as medians (interquartile range) and categorical variables as frequencies (proportion). IVIG intravenous immunoglobulin, PICU pediatric intensive care unit
Full size table

The probability of discharge and of needing second-line treatment was different according to treatment initiation (Fig. 1). The probability of early discharge was 61% higher for patients who initiated treatment with IVIG than for those initiating with IVIG plus steroids (hazard ratio [HR] 1.61, 95% confidence interval [CI]: 1.06–2.44, p = 0.024), whereas no significant differences were observed between patients initiating with steroids versus IVIG plus steroids (HR 1.55, 95% CI: 0.74–3.24, p = 0.246).

Fig. 1
figure 1

Kaplan–Meier curves for each transition, including IVIG, steroids, and IVIG plus steroids as initiation therapies, to compare each one of the individual treatments versus the combination

Full size image

Despite the apparent benefit of IVIG on early discharge, the probability of needing second-line treatment was 3 times higher in patients who started on IVIG than on those starting on IVIG plus steroids (HR 3.11, 95% CI: 1.13–8.55, p = 0.028). This was not observed in children who started on steroids (HR 1.70, 95% CI: 0.58–5.00, p = 0.332).

Patients treated with steroids or IVIG plus steroids had a similar probability of persistent fever after treatment initiation (odds ratio [OR] 0.60, 95% CI: 0.31–1.14, p = 0.126). Patients treated with IVIG alone had a higher probability of persistent fever after treatment initiation (OR 4.4, 95% CI: 2.05–9.82, p < 0.001) than patients treated with IVIG plus steroids. When we compared patients initially treated with IVIG or steroids, the former had a 4 times higher probability of persistent fever after treatment initiation (OR 4.23, 95% CI: 1.43–13.5, p = 0.011).

Discussion

This study suggests that the benefits of the examined treatment approaches for MIS-C are different for each outcome assessed. For persistent fever 2 days after treatment, steroids, either alone or in combination with IVIG, seem superior to IVIG alone. Regarding days of admission, IVIG alone is superior to the combination, and the combination is superior to IVIG to the need for treatment escalation.

The results from recent retrospective studies seemed to favor the combination of treatments over steroids or IVIG alone, but the evidence is low, as acknowledged by WHO [8]. Studies in the USA found that the combination of IVIG plus steroids was associated with a lower risk of cardiovascular dysfunction, initiation of vasopressors, and adjunctive therapy versus treatment with IVIG alone [6]. In two studies from France, combination therapy seemed better than IVIG alone to accelerate recovery of cardiac function in patients with cardiogenic shock, and also reduction of fever 2 days after initiation of treatment [5, 10]. The BATS study found that treatment escalation was less common among patients treated with the combination than with either of the other two treatments, but no differences were observed for improvement in the rate of mechanical ventilation, inotropic support, death, or the score on an ordinal clinical-severity scale for any of the three treatments [4].

Since published papers have generally favored the combination of treatments, patients with MIS-C in recent waves have been treated mostly with combined treatments. This limits the observational evaluation of the different approaches. Therefore, we focused on the first waves and adjusted by time.

Our study shows that each treatment seems to have pros and cons. The anti-inflammatory actions of steroids may help to decrease persistent fever after 2 days of treatment, but the regimen typically lasts a minimum of 3 days, and a tapering regimen is needed for prolonged treatments, compared with IVIG alone, which is usually administered in less than 48 h and may accelerate the discharge. If the inflammation of some patients needs strong immunomodulation, and benefits from a combined approach, it is reasonable that IVIG plus steroids are superior to steroids or IVIG alone for the endpoint of escalation of treatment.

The reasons for the reduced benefit of the combination therapy in fever defervescence or early discharge are unclear. It is possible that the combination therapy is beneficial for hemodynamic endpoints as suggested in previous research [6, 10], but less for clinical endpoints such as fever or discharge. Fever is a frequent adverse reaction to IVIG, which may have reduced the benefit of the combined treatment concerning the persistence of fever compared to steroids. However, IVIG-associated fever is a rate-dependent phenomenon that usually occurs during or immediately after the infusion, and rarely 2 days later, as measured here.

Different responses to specific types of immunomodulation might be attributable to genetic differences, SARS-CoV-2 variants, re-exposure to differing or mismatched variants, or prolonged and repetitive exposure to virus circulating within a community [7].

There are limitations to the study. The population was small; the study was not specifically powered for some clinical outcomes, including long-term outcomes such as the presence of aneurysms, which have not been analyzed. However, the population was more homogeneous than the population described in other studies. A major limitation was our inability to analyze relevant endpoints such as shock, PICU admission, or inotropic support after treatment initiation, as children were admitted to the PICU in shock, and patients subsequently received IVIG or steroids, or both, as per guidelines for patients with severe disease, only after the endpoints were reached. Furthermore, we did not have information about the timing of inotrope treatment, aneurysm development, or change of ejection fraction before/after the treatment was prescribed, which impacted the analysis of the effectiveness of these outcomes. Other factors, not evaluated in our study, such as the costs or adverse events associated with each treatment, should be considered, balanced with their clinical benefits.

In conclusion, each treatment seems to have pros and cons according to the different outcomes, also considering the availability of the product and economic factors. Clinical trials are necessary to thoroughly evaluate the effectiveness and safety of the available treatments in short- and long-term outcomes.