Severity predictors for multisystemic inflammatory syndrome in children after SARS-CoV-2 infection in Vietnam

Multisystemic inflammatory syndrome in children (MIS-C) might manifest in a broad spectrum of clinical scenarios, ranging from mild features to multi-organ dysfunction and mortality. However, this novel entity has a heterogenicity of data regarding prognostic factors associated with severe outcomes. The present study aimed to identify independent predictors for severity by using multivariate regression models. A total of 391 patients (255 boys and 136 girls) were admitted to Vietnam National Children’s Hospital from January 2022 to June 2023. The median age was 85 (range: 2–188) months, and only 12 (3.1%) patients had comorbidities. 161 (41.2%) patients required PICU admission, and the median PICU LOS was 4 (2–7) days. We observed independent factors related to PICU admission, including CRP \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\ge $$\end{document}≥ 50 (mg/L) (OR 2.52, 95% CI 1.39–4.56, p = 0.002), albumin \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\le $$\end{document}≤ 30 (g/L) (OR 3.18, 95% CI 1.63–6.02, p = 0.001), absolute lymphocyte count \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\le $$\end{document}≤ 2 (× 109/L) (OR 2.18, 95% CI 1.29–3.71, p = 0.004), ferritin ≥ 300 (ng/mL) (OR 2.35, 95% CI 1.38–4.01), p = 0.002), and LVEF < 60 (%) (OR 2.48, 95% CI 1.28–4.78, p = 0.007). Shock developed in 140 (35.8%) patients, especially for those decreased absolute lymphocyte \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\le $$\end{document}≤ 2 (× 109/L) (OR 2.48, 95% CI 1.10–5.61, p = 0.029), albumin \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\le $$\end{document}≤ 30 (g/L) (OR 2.53, 95% CI 1.22–5.24, p = 0.013), or LVEF < 60 (%) (OR 2.24, 95% CI 1.12–4.51, p = 0.022). In conclusion, our study emphasized that absolute lymphocyte count, serum albumin, CRP, and LVEF were independent predictors for MIS-C severity. Further well-designed investigations are required to validate their efficacy in predicting MIS-C severe cases, especially compared to other parameters. As MIS-C is a new entity and severe courses may progress aggressively, identifying high-risk patients optimizes clinicians' follow-up and management to improve disease outcomes.


Case definition
During the study period, we included 391 children aged from 1 month to 18 years old who met the MIS-C case definition of the United States Centers for Disease Control and Prevention (US.CDC) 23 .Accordingly, any illness in a person aged less than 21 years that meets the clinical and the laboratory criteria, which are as follows: (1) fever > 38.0 °C for ≥ 24 h (subjective or documented fever); (2) laboratory evidence of inflammation, including one or more of the followings: high values of C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), fibrinogen, procalcitonin, d-dimer, ferritin, lactic acid dehydrogenase (LDH), or interleukin 6 (IL-6); elevated neutrophils, reduced lymphocytes, and low albumin; (3) evidence of at least two organs injury (respiratory, cardiovascular, hematologic, renal, gastrointestinal, or neurological involvement); (4) current or recent SARS-CoV-2 infection diagnosed by a positive reverse transcription polymerase chain reaction (RT-PCR) or positive serological tests (IgM, IgG or IgA), or exposure to a suspected or confirmed COVID-19 case within the 60 days prior to the onset of symptoms.;(5) no alternative plausible diagnosis and exclusion of any other microbial infections, including bacterial sepsis, staphylococcal or streptococcal shock syndromes, and viral infections associated with myocarditis, such as enterovirus.Patients suspected of or confirmed with other microbial (bacterial, viral, fungal, and parasite) infections were excluded.

Data on clinical and laboratory characteristics
Data on clinical and demographic characteristics, laboratory and therapeutic variables, and echocardiographic findings within 24 h of admission and daily were routinely collected using standard case report forms (CRFs) (available on the Supplementary 1 for further details).The treatment was supported by strategy guidance from the American College of Rheumatology Clinical Guidance and the Rheumatology Study Group of the Italian Society of Pediatrics 15 .

Severe outcomes
There are previous reports in the literature of sudden fatality in MIS-C due to cardiovascular compromise during the hyperinflammatory phase of the illness, mostly related to cardiovascular failure, with hypotension and decreased left ventricle contractility 24 .Additionally, severe involvement of the respiratory, cardiac, neurologic, or renal systems in MIS-C cases was significantly associated with PICU admission [24][25][26] .Therefore, our variables

Statistical analysis
We used a convenient method for determining the sample size of our study.All statistical analyses were performed using STATA software version 17.0 (StataCorp LP, College Station, TX, USA).Categorical variables were described as frequencies and percentages, while continuous variables were described as the median and interquartile range.Results were described as odds ratio (OR) and 95% confidence intervals (CIs).The area under the receiver operating characteristic curve was calculated to compare the predictive performance of the difference biomarkers for MIS-C severity.For each variable, the area under the curve (AUC), 95% CI, and the optimal cutoff point were determined.Variables significantly associated with the outcomes in univariate analyses were selected for multivariate logistic progression to identify independent predictors.A two-sided p-value of less than 0.05 was considered statistically significant.

Baseline characteristics of the study population
We analyzed data from 159,410 patients admitted to hospital between January 2022 and June 2023.Of these, 525 patients were reported with MIS-C diagnosis.Among them, 88 patients who unmeet the case definition for MIS-C were excluded.We further excluded 37 patients with insufficient data, 9 patients with microbiologically confirmed infections (5 cases with Staphylococcus aureus infections, 2 cases with Burkholderia pseudomallei infections, and 2 cases with Adenovirus infections).A total of 391 patients (255 boys and 136 girls) were ultimately recruited in this study (See the flow diagram for study participants in the Supplement 2).The median age was 85 (range 2-188) months and only 14 (3.6%) patients had comorbidities.The frequency of children with a COVID-19 infection history was 310 (79.3%) patients, and the median time since the previous infection prior to symptom onset was 36 (29-47) days.We summarized the epidemographic and clinical parameters of the total study population in Table 1.See Supplement 3 for additional details.
Among the study population, 161 (41.2%) patients needed PICU admission, and the median PICU LOS was 4 (2-7) days.Shock developed in 140 (35.8%)patients.In our study, we found that 161 patients were admitted to the PICU, with 140 patients experiencing shock that coincided with one or more of the above indications.The remaining 21 patients, without shock, were admitted for respiratory distress (7 cases), seizure (3 cases), evidence of liver or kidney damage (5 cases), arrhythmia (4 cases), and both arrhythmia and respiratory distress (2 cases) (Data not shown).Overall, two deaths (0.5%) occured in our study population.

Prognostic factors of disease severity
Shock Shock developed in 140 (35.8%)patients within the first 72 h after admission.The predictors of shock were identified based on indicators with significant differences, followed by an analysis of the univariate and multivariate logistic regression as presented in Table 3.

PICU admission
One hundreds and sixty-one (41.2%) patients required PICU admission, and the median PICU LOS was 4 (2-7) days.Similarly, the predictors of PICU admission were identified based on indicators with significant differences, followed by an analysis of the univariate and multivariate logistic regression as presented in Table 4.
Cut-offs of these variables were chosen based on reference values in previous cohort [28][29][30] .

Discussion
The present study reported independent predictors of severity in 391 pediatric patients with MIS-C in the tertiary hospital from Vietnam.The results found that lymphocytopenia, hypoalbuminemia, high CRP, and decreased LVEF were independently associated with the risk of severity among children with MIS-C.According to previous systemic reviews, MIS-C might manifest a broad spectrum of scenarios, ranging from mild features to severe complications, including shock, multiorgan failures, and mortality [5][6][7] .Regarding clinical presentations, the most frequent symptoms were fever (100%), followed by mucocutaneous symptoms (91.8%),GI involvement (62.9%), lymphadenopathy (41.4%), and cardiovascular findings (23.8%) (Table 1).Ahmed et al. found out that the most common symptoms of MIS-C are fever, GI symptoms (abdominal pain, vomiting, and diarrhea), rash, and conjunctivitis 31 .Similarly, GI symptoms (especially abdominal pain, vomiting, and diarrhea) were also previously considered prevalent symptoms of this novel entity 28,32,33 .The predominance of GI manifestations in children with MIS-C was striking 33 .A plausible pathogenesis might be the overwhelming systemic immune response after COVID-19 infection, including the effect on the digestive system 34 .Besides, by detecting auto-antibodies targeting gastrointestinal antigens in the plasma of MIS-C patients, it was hypothesized that the GI tract may serve as a viral reservoir for continuous exposure to the SARS-CoV-2 SAg-like motif 35 .In the cohort study, 35.8% developed shock, 11.7% had respiratory distress, 41.2% required PICU admission, and 0.5% were death (Table 1).However, the ICU admission rate was less than in cohorts from Hungary (52.6%) 36 and Turkey (54.5%) 37 .Shock and mortality were documented with a higher proportion in most cohorts from other parts of the world 38,39 .We hypothesized that the variation in shock definition and PICU admission criteria depending on centers, SARS-CoV-2 variants circulating, study sample sizes, and healthcare settings may account for this difference.
Despite being considered as a "new entity" of KD, MIS-C is common in children 9-10 years, whereas classic KD occurs predominantly in children under five years of age 9,18,40 .In line with previous studies, our study identified that the median age group was 7 (3-12) years (Table 1) [41][42][43] .This point can be explained by the findings that younger children are less likely to contract COVID-19 than older children and adolescents due to several hypotheses related to specific properties of young children, including the lower expression in receptors of angiotensin-converting enzyme 2 (ACE2), cross-protection against SARS-CoV-2 infection, and their immature immune system 4,44 .We found that age ≥ three years was associated with an increased risk of shock (OR 3.62, 95% CI 1.91-7.25,p < 0.001) and PICU admission (OR 1.78, 95% CI 1.07-2.98,p = 0.013) among children with MIS-C, though insignificance of the predictor after adjusting in the multivariate analysis (Tables 3 and 4).Other studies also identified older age as a risk factor for severity in children with MIS-C, despite differences in cut-off point  47 .Likewise, the lesser susceptibility to develop severe MIS-C at an early age might not be distinctive in the pediatric population.Recent reviews also highlighted that patients with multisystem inflammatory syndrome in adults (MIS-A) have a higher mortality than MIS-C cases 48 .The overall mortality of MIS-A is between 5 and 7%, whereas that of MIS-C is approximately 1.7% 49 .However, the currently available data is limited, and there is a need for additional studies to enlighten this imparity.www.nature.com/scientificreports/MIS-C pathogenesis is characterized by overwhelming immune system responses, leading to hyper-inflammation and cytokine storm 8,48,50 .This over-inflammatory mechanism is crucial for explanation of with multi-organ damage in MIS-C, in which biochemical changes could be partly demonstrated by the laboratory inflammatory markers, such as lymphocyte count, CRP, PCT, LDH, ferritin, d-dimer, TNF-α, and IL-6… 20,41,50 .Common lab findings indicative of MIS-C were elevated levels of CRP, ESR, d-dimer, ferritin, PCT, LDH, and a decrease in serum albumin and lymphocyte count as a panel of biomarkers in the existing guidelines 14,15 .The inflammatory markers changes might reflect the MIS-C entity's host response patterns and provide clues to identify patients at risk of severity for strict follow-up and timely treatment escalation.In the present study, absolute LYM count ≤ 2 (× 10 9 /L) were recognized to be independently associated with shock (OR 2.48, 95% CI 1.10-5.61)and PICU admission (OR 2.18, 95% CI 1. 29-3.71) on the multivariate model, respectively (Tables 3 and 4).Similarly, an observational study showed lymphopenia was common in classic MIS-C with multiple organ involvement and shock 51 .Bar-Meir et al. reported lymphopenia (lymphocyte count < 1500 µL) was an independent predictor of MIS-C, with an adjusted odds ratio of 24 (95% CI 1.3-326, p = 0.02) 52 .This raised a question about the mechanisms and the role of lymphopenia observed in MIS-C patients.One proposed explanation is the profound lymphopenia caused by marked T cell exhaustion after SARS-CoV-2 infection, which persists for weeks, leading to uncontrolled inflammation and immune dysregulation 53 .We emphasized that further studies are needed to validate whether lymphopenia can predict MIS-C development and severity and whether lymphocyte-targeted therapeutics could improve the disease outcome.We also found that other biomarkers, including d-dimer ≥ 2500 (ng/mL), ferritin ≥ 300 (ng/mL), CRP ≥ 50 (mg/L), and albumin ≤ 30 (g/L), significantly related to the severity scenario in the univariate model.However, only albumin ≤ 30 (g/L) was a significant predictor of both shock (OR 2.53, 95% CI 1.22-5.24,p = 0.013) and PICU admission (OR 3.18, 95% CI 1.63-6.02,p = 0.001) (Tables 3 and  4) after adjusted in the multivariate model.The pivotal role of albumin in severity prediction can be explained by an increased capillary permeability and escape of albumin to the interstitial space parallel with the context of the overwhelming inflammation in MIS-C 54 .In a study in 76 MIS-C patients, an albumin level < 33.6 (g/L) was independently related to the risk of PICU admission 26 .Sharon et al. concluded that compared to CRP values with a wide range and high variability throughout the illness, albuminemia is a more reliable and accurate inflammatory marker that needs to be monitored regularly, particularly in critically ill children 55 .Similarly, Abrams J.Y. et al. noted that high ferritin, C-reactive protein, and D-dimer were associated with life-threatening manifestations, including shock and heart depression 56 .A d-dimer > 2568 ng/mL is an early predictor of severe MIS-C requiring ICU admission 57 .The surge of these biomarkers may reflect the presence of the known hyperinflammatory condition and cytokine storm theory seen in MIS-C; therefore, immunoregulatory agents, such as IVIG, corticosteroids, interleukin blockers, are the mainstays in the treatment of MIS-C [13][14][15] .To date, IVIG (2 g/kg) plus high dose methylprednisolone (10-30 mg/kg/day) for three consecutive days is the first choice in MIS-C patients with shock, cardiac injury, or neurological damage 15 .The initial guidelines for the MIS-C were extrapolated from KD treatment.However, a more violent inflammatory activation was found in MIS-C, and it raised controversies around whether only IVIG or corticosteroid therapy is sufficient in patients with MIS-C.According to a study comparing efficacy among corticosteroids alone, IVIG alone, and IVIG plus corticosteroids group, there was an insignificant difference in initial responses between patients receiving corticosteroids alone and IVIG plus corticosteroids 58 .However, this outcome might be due to heterogeneous characteristics in disease severity between the groups; further studies on this issue are needed.In addition, the unavailability and high cost of IVIG and biological agents might strain the approach to this therapy, particularly in resource-limited settings.This point partly explains why our study reported that 19.2% received intravenous corticosteroids plus IVIG and only 1.5% received biological therapy (Table 1).
Cardiovascular complications are common manifestations in MIS-C, with an incidence of 40% to 80% 17,46 .Cardiovascular complications in MIS-C might present from mild manifestations to severe ones, such as arrhythmias, pericardial effusion, coronary artery aneurysms, myocarditis, shock, and life-threatening 10,16,17 .Therefore, cardiac markers, including proBNP, NT-proBNP, and troponin, are potential parameters to predict the progression of MIS-C deterioration.Several studies reported elevated BNP or NTpro-BNP, troponin levels, and LVEF related to required ventilation, PICU admission, or mortality 17,59,60 .Similarly, our study showed that proBNP ≥ 200 (pg/mL) and LVEF < 60 (%) were related to an increased risk of shock and PICU admission.However, we noted that LVEF might be more valuable for the severity prediction, with higher odds risk, than the proBNP level (see Tables 3 and 4).We supposed that this could be explained because proBNP peptides are primarily synthesized and upregulated by myocardial stress; therefore, it only reflects indirect cardiac suppression.Additionally, another report noted that there is a delayed improvement of NT-proBNP levels after myocardial function normalization 60 .In the present study, we then assessed the LVEF < 60% as an independent predictor for shock (OR 2.24, 95% CI 1.12-4.51,p = 0.022) and PICU admission (OR 2.48, 95% CI 1.28-4.78,p = 0.007) in the multivariate model (Tables 3 and 4).
Our study had some limitations.First, the data were collected from a single center in a tertiary hospital.Nevertheless, to the best of our knowledge, this paper highlighted the largest-sample study of this novel syndrome published in Vietnam.Second, this study performed the laboratory parameters at a single time and did not evaluate their variations during the subsequent days.However, this may be beneficial as we can eliminate the effects of therapy during treatment.Finally, we only evaluate the risk for short-term outcomes; further, well-designed research is required to define the prognostic factors identifying the long-term effects of MIS-C.

Conclusions
In conclusion, our study emphasized that absolute lymphocyte count, serum albumin, CRP, and LVEF were independent predictors for MIS-C severity.Further well-designed investigations are required to validate their efficacy in predicting MIS-C severe cases, especially compared to other parameters.As MIS-C is a new entity

Table 3 .
Univariate and multivariate logistic regression for predictors of shock (Linktest p < 0.001 for multivariate logistic regression).Bold font indicates statistical significance.WBC white blood cell, LYM lymphocyte, PLT platelet, CRP C-reactive protein, LDH lactic acid dehydrogenase, LVEF left ventricular ejection fraction, proBNP pro B-type natriuretic peptide.

Table 4 .
Univariate and multivariate logistic regression for predictors of pediatric intensive care unit admission (Linktest p < 0.001 for multivariate logistic regression).Bold font indicates statistical significance.WBC white blood cell, LYM lymphocyte, PLT platelet, CRP C-reactive protein, LDH lactic acid dehydrogenase, LVEF left ventricular ejection fraction, proBNP pro B-type natriuretic peptide.