Comparison of Pediatric Patients With and Without Multisystem Inflammatory Syndrome Associated With COVID-19: Retrospective Cohort From Ecuador

Abstract

Background: 

Multisystem inflammatory syndrome in children (MIS-C) has been associated with severe acute respiratory syndrome coronavirus 2 infection in the pediatric population cared for in the pediatric intensive care unit. We aimed to compare patients with pediatric acute respiratory distress syndrome (PARDS) with those who also present a MIS-C diagnosis (PARDS vs. PARDS + MIS-C).

Methods: 

One hundred and sixty-seven children (0–15 years) admitted to the pediatric intensive care unit COVID-19 ward of a national reference children’s hospital in Quito, Ecuador, from June 2020 to June 2021 who developed PARDS with or without MIS-C. To diagnose PARDS, the criteria of the Pediatric Acute Lung Injury Consensus Conference Group were used, and to diagnose MIS-C, the criteria of the Centers for Disease Control and Prevention were used. Additionally, the PRISM score was used to calculate the mortality risk of study patients on admission.

Results: 

Of the 167 patients with PARDS, ~59% also developed MIS-C. Patients with PARDS + MIS-C had higher risks than patients without MIS-C in the following: frequency of associated bacterial infections (81.6% vs. 55.1%), mortality risk (36.7% vs. 11.6%), use of respiratory support (invasive mechanical ventilation: 92% vs. 86%), use of vasopressors/inotropes (90.8% vs. 30.4%), renal complications (36.7% vs. 8.7%), septic shock (84.7% vs. 20.3%), multiorgan failure (39.8% vs. 1.4%) and mortality at discharge (39.8% vs. 4.3%). Logistic regression failed to find an association between MIS-C and age, race, sex, ≥3 signs/symptoms and ≥2 comorbidities.

Conclusions: 

Patients with PARDS + MIS-C had a more severe clinical picture than patients without MIS-C. The findings provide useful information to improve the management of PARDS patients with and without MIS-C in Ecuador.

The disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has represented a new challenge for humanity since its emergence in 2019. Studies carried out so far refer to a greater incidence of the disease in the adult population, where there is also a greater record of severe cases. Nevertheless, approximately 2% of cases occur in children, an incidence that should not be underestimated.¹ The literature reports pediatric clinical cases characterized by manifestations of multisystemic inflammatory response syndrome and shock after acute SARS-CoV-2.² Other authors speak in favor of cases similar to Kawasaki disease, a vasculitis more frequent in this age group and which presents with certain skin lesions, in addition to toxic shock syndrome. These entities have a tendency to develop pediatric acute respiratory distress syndrome (PARDS) that requires admission to pediatric intensive care units (PICU) for the administration of mechanical ventilatory support and treatment with vasoactive and ionotropic agents, requiring in some cases therapy with intravenous immunoglobulin, corticosteroids, antibiotics and anticoagulants.^3,4

In 2020, several authoritative institutions in the field of health such as the Centers for Disease Control and Prevention, the World Health Organization and the Royal College of Pediatrics and Child Health participated in the identification and study of multisystem inflammatory syndrome in children (MIS-C), which is described as an entity associated with SARS-CoV-2 infection in the pediatric population.⁵ Although there are some publications that address the clinical characteristics, different therapeutic lines followed and results obtained in patients with MIS-C, most of them come from European countries such as the United Kingdom and Spain. In the case of the American continent, the main exponents are the United States, Chile, Peru and Colombia; there are no reports of the situation related to the health emergency in the field of Ecuadorian pediatrics, especially in cases with the entities described above and whose severity is relevant.^6–8 The Baca Ortiz Pediatric Hospital (HPBO) also lacks MIS-C data despite being a national reference institution for critically ill pediatric patients. Hence the need to carry out this study, so that the clinical and epidemiological characteristics, the therapeutic behaviors adopted and the main results obtained in patients with MIS-C in this tertiary care center can be known and described, to verify whether the behavior of this entity is similar to that reported in the literature. Comparing patients with PARDS to those who also present a diagnosis of MIS-C (PARDS + MIS-C) is the objective of the present investigation.

MATERIALS AND METHODS

A retrospective cohort study was conducted with all patients admitted to the PICU COVID-19 ward of HPBO in the period June 2020 to June 2021 for coronavirus 2 (SARS-CoV-2)–related diseases and who developed PARDS with or without MIS-C. The present study was approved by the Institutional Review Board of Universidad San Francisco de Quito (#2021-122TPG). PARDS was assessed as per the Pediatric Acute Lung Injury Consensus Conference criteria,⁹ which include factors such as age (excludes patients with perinatal lung disease), time (onset of symptoms within 1 week of known insult), origin (respiratory failure cannot be explained by fluid overload or cardiac failure) and imaging (unilateral or bilateral new-onset infiltrate, or an infiltrate suggestive of acute parenchymal lung disease). Depending on oxygenation with invasive mechanical ventilation, PARDS can be classified as mild (oxygenation index ≥4 to <8 or oxygen saturation index ≥5–7.5), moderate (oxygenation index ≥8 to <16 or oxygen saturation index ≥7.5–12.3) or severe (oxygenation index ≥16 or oxygen saturation index ≥12.3). Oxygenation with noninvasive mechanical ventilation may consist of full-face mask ventilation or continuous positive airway pressure ≥5 cm H₂O and Partial pressure of oxygen/Fraction of inspired oxygen ≤300 or Oxygen saturation/Fraction of inspired oxygen ≤264.^9,10

To classify a patient with MIS-C, the Centers for Disease Control and Prevention criteria were followed: age (less than 21 years); fever (confirmed, with greater than 38°C for the last 24 hours, or the report of subjective sensation of fever in this period); evidence of inflammation on laboratory results, which includes but is not limited to C-reactive protein elevation, erythrosedimentation, fibrinogen, procalcitonin, D-dimer, ferritin, lactate dehydrogenase, interleukin-6, neutrophilia, lymphocytopenia and hypoalbuminemia; multisystem involvement (2 or more organ systems involved): cardiovascular (shock, elevated troponin, elevated brain natriuretic peptide, abnormal echocardiogram, arrhythmia), respiratory (pneumonia, PARDS, pulmonary embolism), renal (acute renal failure), neurologic (seizures, stroke, aseptic meningitis), hematologic (coagulopathy), gastrointestinal (abdominal pain, vomiting, diarrhea, elevated liver enzymes, ileus, gastrointestinal bleeding) or dermatologic (erythroderma, mucositis, other rashes) and severe illness requiring hospitalization and absence of evidence of other causes for this condition. In addition, recent or current infection or exposure to SARS-CoV-2 and any of the following findings must be present: positive reverse transcription polymerase chain reaction for SARS-CoV-2, positive serology, positive antigen test or exposure to a suspected or confirmed COVID-19 case in the last 4 weeks before the onset of symptoms.¹¹ On the other hand, to predict the risk of mortality, the PRISM score, which was first presented by Pollack et al,¹² was used and is based on physiological variables that are subdivided into several parameters. Depending on the score obtained, the mortality risk can be low (1–20 points, with a probability of death of 35%), moderate (21–30 points, with a probability of death of 40%–80%) or high (>30 points, with a probability of death greater than 80%).¹³ We included 167 patients between the ages of 1 month and 15 years who were admitted to the PICU of the HPBO, in the COVID-19 ward, during the period June 2020 to June 2021 and who developed PARDS or PARDS + MIS-C. Cases with incomplete clinical histories were excluded.

Statistical Analysis

Descriptive statistics were used to summarize the baseline and clinical characteristics of the study participants. Continuous variables were reported as median and interquartile range (IQR) due to skewness, and categorical variables were reported as counts and percentages. Parametric (χ²) and nonparametric (Wilcoxon and Fisher exact test) tests were used to assess differences between MIS-C and non-MIS-C patients. In addition, a multivariable logistic regression analysis was used to model key sociodemographic and clinical-related factors as exposure variables with the outcome of MIS-C/non–MIS-C, measures of association are presented as odds ratios with their 95% confidence intervals. The full regression model was tested for collinearity among independent variables by variance inflation factor. We assessed the goodness-of-fit of our model by several tests including likelihood ratio test, c-statistic, and Hosmer-Lemeshow test. A 2-tailed P value <0.05 was used to define statistical significance. Statistical analyses were conducted using IBM SPSS v25.0 SPSS v25.0 and RStudio for Mac v.1.4.1106 developed by R foundation for statistical computing, Vienna, Austria.

RESULTS

General Characteristics of the Population

The study population consisted of 167 patients under 15 years of age with a diagnosis of PARDS. Of these, 98 (~59%) also developed MIS-C (PARDS + MIS-C) (Table 1). Between the 2 study groups, no significant differences were identified in most variables except in the presence of concomitant bacterial infection and the risk of mortality according to the PRISM score (37% vs. 12%; P < 0.001), which was most prevalent among patients with MIS-C (Table 1).

TABLE 1. - General Characteristics of the Study Population

Characteristic	Total PARDS (n = 167)	MIS-C (n = 98)	Non–MIS-C (n = 69)	P Value*
Age, median (IQR)	3 (0–14)	3 (0–14)	4 (0–14)	0.692
Age group, n (%)				0.718
Infant	65 (38.9)	35 (35.7)	30 (43.5)
Preschool	31 (18.6)	18 (18.4)	13 (18.8)
School	45 (26.9)	28 (28.6)	17 (24.6)
Adolescent	26 (15.6)	17 (17.3)	9 (13.0)
Sex, n (%)				0.09
Masculine	95 (56.9)	61 (62.2)	34 (49.3)
Feminine	72 (43.1)	37 (37.8)	35 (50.7)
Ethnicity, n (%)				0.07
Black	10 (6)	3 (3.1)	7 (10.1)
Indigenous	13 (7.8)	10 (10.2)	3 (4.3)
Mestizo	144 (86.2)	85 (86.7)	59 (85.5)
Nutritional status, n (%)				0.455
Underweight	90 (53.9)	53 (54.1)	37 (53.6)
Normal weight	68 (40.7)	38 (38.8)	30 (43.5)
Overweight	9 (5.4)	7 (7.1)	2 (2.9)
Main symptom, n (%)				0.10
Fever	95 (56.9)	58 (59.2)	37 (53.6)
Cough	29 (17.4)	18 (18.4)	11 (15.9)
Difficulty breathing	37 (22.2)	18 (18.4)	19 (27.5)
Diarrhea	6 (3.6)	4 (4.1)	2 (2.9)
≥3 symptoms, n (%)				0.10
Yes	161 (96.4)	96 (98.0)	65 (94.2)
Comorbidities, n (%)				0.8
Epilepsy	14 (8.4)	9 (9.2)	5 (7.2)
Neurodevelopmental delay	6 (3.6)	3 (3.1)	3 (4.3)
Infantil cerebral palsy	18 (10.8)	10 (10.2)	8 (11.6)
Type 1 diabetes	1 (0.6)	1 (1.0)	0 (0.0)
Asthma	2 (1.2)	0 (0.0)	2 (2.9)
Chronic lung disease/brochopulmonary dysphasia	8 (4.8)	5 (5.1)	3 (4.3)
Cardiac	18 (10.8)	10 (10.2)	8 (11.6)
Hematologic/oncologic	7 (4.2)	5 (5.1)	2 (2.9)
Other	20 (12)	10 (10.2)	10 (14.5)
≥2 comorbidities, n (%)				0.980
Yes	70 (41.9)	41 (41.8)	29 (42)
Viral coinfection, n (%)				0.712
Yes	13 (7.8)	7 (7.1)	6 (8.7)
Bacterial coinfection, n (%)				<0.001
Blood	20 (12)	15 (15.3)	5 (7.2)
Respiratory	118 (70.7)	80 (81.6)	38 (55.1)
No	29 (17.4)	3 (3.1)	26 (37.7)
Pediatric risk of mortality (PRISM Score), n (%)				<0.001
Low	28 (16.8)	8 (8.2)	20 (29)
Moderate	95 (56.9)	54 (55.1)	41 (59.4)
High	44 (26.3)	36 (36.7)	8 (11.6)

^*All continuous variables were analyzed using Wilcoxon t test. All categories were analyzed using the χ² test or Fisher exact test, as appropriate.

Management in the PICU

In relation to management in the PICU, 89% of patients received conventional invasive mechanical ventilation. Patients with MIS-C were more likely than non-MIS-C patients to be treated with invasive mechanical ventilation (92% vs. 86%) and high frequency ventilation (8% vs. 3%; Table 2).

TABLE 2. - Management in the Pediatric Intensive Care Unit

Characteristic	Total PARDS (n = 167)	MIS-C (n = 98)	Non–MIS-C (n = 69)	P Value*
Respiratory support, n (%)				0.004
High-flow oxygen nasal cannula	1 (0.6)	0 (0)	1 (1.4)
Noninvasive ventilation	7 (4.2)	0 (0)	7 (10.1)
Invasive mechanical ventilation	149 (89.2)	90 (91.8)	59 (85.5)
High frequency ventilation	10 (6)	8 (8.2)	2 (2.9)
Support to other organs, n (%)				<0.001
Vasopressor/inotropes support	110 (65.9)	89 (90.8)	21 (30.4)
Neuromuscular blockade	67 (40.2)	51 (52)	16 (23.2)
Dialysis/hemodialysis	4 (2.4)	3 (3.1)	1 (1.4)
Other drugs, n (%)
Antibiotics	167 (100)	98 (100)	69 (100)	-
Antiviral	15 (8.9)	14 (14.3)	1 (1.4)	0.004
Corticosteroids	153 (91.6)	98 (100)	55 (79.7)	<0.001
Immunomodulator (azithromycin)	95 (56.9)	85 (86.7)	10 (14.5)	<0.001
Intravenous immunoglobulin	44 (26.3)	41 (41.8)	3 (4.3)	<0.001
Enoxaparin	89 (53.3)	65 (66.3)	24 (34.8)	0.001
Antifungal	9 (5.4)	8 (8.2)	1 (1.4)	0.612

^*All continuous variables were analyzed using Wilcoxon t test. All categories were analyzed using the χ² test or Fisher exact test, as appropriate.

Support to other organs was given more frequently in the group of patients with MIS-C. Thus, vasopressor/inotropes (66%), neuromuscular blockade (40%) and hemodialysis (2.4%) were the support modalities more commonly used among MIS-C patients. Regarding the use of other medications, statistically significant differences were also identified between both study groups (P < 0.05) in the use of antivirals (14.3% of the MIS-C group vs. 1.4% of the non-MIS-C group), corticosteroids (100% of the MIS-C group vs. 78% of the non-MIS-C group), azithromycin as an immunomodulator (87% of the MIS-C group vs. 15% of the non-MIS-C group), immunoglobulin (IV) (42% of the MIS-C group vs. 4% of the non-MIS-C group) and enoxaparin (66% of the MIS-C group vs. 35% of the non–MIS-C group; Table 2).

Complementary and Laboratory Tests

All patients analyzed had a chest radiograph with radiological signs of unilateral or bilateral new-onset infiltrate upon admission to the PICU, and only 27% in the MIS-C group and 30% in the non–MIS-C group had a positive reverse transcription polymerase chain reaction test for COVID-19 on admission. The rest of the study patients fulfilled the inclusion criteria of exposure to a suspected or confirmed COVID-19 case in the 4 weeks before the onset of symptoms. In addition, all biomarkers that are part of the diagnostic criteria for MIS-C (D-dimer, total CPK, troponin T, procalcitonin, pro-BNP peptide and interleukin-6) reached significantly higher values among patients with MIS-C (P < 0.05; Table 3).

TABLE 3. - Radiography and Laboratory Tests

Characteristic	Total PARDS (n = 167)	MIS-C (n = 98)	Non–MIS-C (n = 69)	P Value*
Chest radiograph positive on admission, n (%)
Yes	167 (100)	98 (100)	69 (100)
RT-PCR for COVID-19, n (%)				0.581
Positive	47 (28.1)	26 (26.5)	21 (30.4)
Negative	120 (71.9)	72 (73.5)	48 (69.6)
D-dimer, ng/mL, median (IQR)	1567 (205–54,221)	2998.5 (318–54,221)	678 (205–25,671)	<0.001
Total CPK, U/L, median (IQR)	212 (3–102,124)	406.5 (16–102,124)	127 (3–9672)	<0.001
Troponin T, pg/mL, median (IQR)	73 (2–10,678)	90.5 (2–10,678)	45 (2–3456)	0.045
C-reactive protein, mg/dL, median (IQR)	16 (0–16)	32 (0.1–16)	7.0 (0.1–204)	<0.001
Procalcitonin, ng/dL, median (IQR)	16 (0–5665)	56.5 (0.1–5665)	5 (0.1–778)	<0.001
Pro-BNP peptide, pg/mL, median (IQR)	525 (3–65,211)	638 (99–65,211)	373 (3–5488)	0.002
Interleukin-6, pg/mL, median (IQR)	43 (1–12,770.0)	67 (5–12,770.0)	12 (1–2001)	<0.001

D-dimer: normal value: 0–500 ng/mL. Troponin T: normal value: 0–14 pg/mL. C-reactive protein: normal value: 0.00–0.28 mg/dL. Procalcitonin: normal value: <0.5 ng/dL—low risk of sepsis; >2.0 ng/dL—high risk of sepsis. Pro-BNP peptide: normal value: boys (0–15 y) <62 pg/mL; girls (0–15 y) <83 pg/mL. Interleukin-6: normal value: 0.00–7.00 pg/mL. RT-PCR, reverse transcription polymerase chain reaction.

^*All continuous variables were analyzed using Wilcoxon t test. All categories were analyzed using the χ² test or Fisher exact test, as appropriate.

Clinical Outcomes

Table 4 shows the complications and outcomes of the patients studied. MIS-C patients were more likely than non–MIS-C patients to develop acute kidney injury, septic shock, coronary artery dilatation and multiple organ failure. Other complications at PICU discharge such as the need for tracheostomy (4%) or gastrostomy (3%) occurred only among MIS-C patients. PICU length of stay was similar in both groups: 8.5 days (IQR: 1–88) versus 7 days (IQR: 3–44), P = 0.141. Overall, 75% of patients survived to PICU discharge although patients in the non–MIS-C group were more likely to survive as compared with the MIS-C group (96% vs. 60%, P < 0.001; Table 4). On the logistic regression analysis, we did not find any association with the selected explanatory variables and the presence of MIS-C. The regression model fitted the observed data appropriately according to several assessment tests (likelihood ratio test: 0.02; c statistic: 0.61; and Hosmer-Lemeshow test: 0.99), and we were not able to find collinearity among the used variables (Table 5).

TABLE 4. - Clinical Outcomes

Characteristic	Total PARDS (n = 167)	MIS-C (n = 98)	Non–MIS-C (n = 69)	P Value*
Duration of ventilatory support, median (IQR)
High-flow nasal cannula oxygenation†	4 (4–4)	-	4 (4–4)	-
Noninvasive ventilation	3 (2–12)	-	3 (2–12)	-
Conventional invasive ventilation	6 (1–88)	7 (1–88)	6 (2–38)	0.101
High-frequency ventilation	10 (6–28)	9.5 (6–28)	10.5 (10–11)	0.317
Complications, n (%)
Acute kidney injury
Yes	42 (25.1)	36 (36.7)	6 (8.7)	<0.001
Septic shock
Yes	97 (58.1)	83 (84.7)	14 (20.3)	<0.001
Coronary artery dilatation
Yes	32 (19.2)	26 (26.5)	6 (8.7)	<0.01
Multiple organ failure
Yes	40 (24)	39 (39.8)	1 (1.4)	<0.001
Complications at PICU discharge, n (%)
Tracheostomy	4 (2.4)	4 (4.1)	0 (0)	0.09
Gastrostomy	3 (1.8)	3 (3.1)	0 (0)	0.142
Length of stay in the PICU, median (IQR)	7 (1–88)	8.5 (1–88)	7 (3–44)	0.141
Discharge status from PICU, n (%)				<0.001
Deceased	42 (25.1)	39 (39.8)	3 (4.3)
Alive	125 (74.9)	59 (60.2)	66 (95.7)

^*All continuous variables were analyzed using Wilcoxon t test. All categories were analyzed using the χ² test or Fisher exact test, as appropriate.

^†This is a single patient.

TABLE 5. - Multivariable Logistic Regression of the Association of Independent Explanatory Factors With MIS-C Patients

Variables	Unadjusted	Adjusted
	OR (95% CI)	OR (95% CI)
Age
<12 y old	1.0*	1.0*
≥12 y old	1.06 (0.36–3.31)	1.14 (0.38–3.72)
Race
Mestizo	1.0*	1.0*
Indigenous	2.31 (0.67–10.64)	2.87 (0.76–15.26)
Afro-Ecuadorians	0.30 (0.06–1.12)	0.29 (0.06–1.10)
Sex
Female	1.0*	1.0*
Male	1.70 (0.91–3.18)	1.51 (0.78–2.91)
≥3 signs/symptoms
No	1.0*	1.0*
Yes	2.95 (0.56–21.75)	3.58 (0.59–31.53)
≥2 comorbidities
No	1.0*	1.0*
Yes	0.99 (0.53–1.85)	0.94 (0.50–1.82)

^*Reference group.

CI indicates confidence interval; OR, odds ratio.

DISCUSSION

The purpose of this research was to describe and compare the clinical and outcome characteristics of pediatric patients with PARDS with those who also had a diagnosis of MIS-C (PARDS + MIS-C) in the PICU of a national reference pediatric hospital in Ecuador.

Patients with PARDS + MIS-C were found to have a higher risk of mortality on admission, concomitant bacterial infections, need for ventilatory and other organ support, higher inflammatory biomarker values, higher number of complications and higher mortality compared with patients with PARDS alone. This is explained by the fact that MIS-C appears to be the consequence of an exacerbated or poorly regulated host immune system response.^14,15 Once the virus enters human cells, the first line of defense against infection should be a rapid and well-coordinated immune response; however, when this mechanism is dysregulated and excessive, hyperinflammation can occur.^16–19 Mortality was higher in the PARDS + MIS-C group (39.8% vs. 4.3%). This is probably because it is an extremely critical condition with the need for respiratory and other organ support due to multiple organ dysfunction. However, mortality has been reported internationally in the range of 2%–13%.^15,20,21 This higher mortality found in our study could be due to delays in the diagnosis and care of patients who went to a health facility before their arrival at HPBO; HPBO is a public hospital of national reference in pediatric care, which handles a high flow of patients in PICU. Regarding the use of ventilatory support and vasoactive/ionotropic support, in this investigation, it was found to be superior in the group of patients with PARDS + MIS-C. This is due to the fact that in these patients there is a more intense inflammatory response, and they presented an important affectation in several organ systems, supporting the assumption that PARDS + MIS-C is a very serious condition, with an important increase in the morbimortality of the patients. On this, Feldstein et al¹⁴ report similar results, arguing that this is due to the hyperinflammatory state in patients with PARDS + MIS-C, which favors a state of multiple organ dysfunction, requiring support to several organ systems.

Clinical Implications

The present investigation provides information on the clinical presentation, evolution and management of a series of cases with MIS-C in the context of an Ecuadorian public institution, which will allow pediatric intensive care medicine professionals to be better prepared to face this complication associated with SARS-CoV-2 infection. The information obtained in this study will contribute to developing management guidelines in HPBO, justifying a greater investment in certain drugs in the PICU of HPBO, and training pediatric physicians in the management of this complication.

Strengths and Weaknesses

Among its strengths, this study was performed in a national reference center in pediatric care, providing the first results on the behavior of PARDS and MIS-C in the Ecuadorian pediatric population. As a weakness, it was identified in this study that biomarker results were not available in all cases, so there is a possibility of information bias. Additionally, since only 1 PICU was included and although it is a national reference center, its results cannot be generalized to the rest of the PICUs in Ecuador. Future research on the subject should consider a multicenter design.

CONCLUSIONS

Patients with PARDS + MIS-C were characterized by higher risk of mortality on admission, need for ventilatory support, use of medications (eg, vasopressor), elevated inflammatory biomarkers (such as D-dimer, total CPK, troponin T, C-reactive protein, procalcitonin, pro-BNP peptide and interleukin-6), more complications and higher mortality in the PICU at discharge compared with patients with PARDS without MIS-C. The findings of the present investigation provide useful information to improve the management of PARDS patients with and without MIS-C in HPBO.