Cardiovascular involvement in multisystem inflammatory syndrome in children and midterm follow-up from a pediatric tertiary center in India

Abstract

BACKGROUND

In multisystem inflammatory syndrome in children (MIS-C) with coronavirus disease 2019, there was paucity of data from low-income and middle-income countries on cardiovascular involvement and its longitudinal outcomes. We planned to estimate the pattern of cardiovascular involvement among children with MIS-C and its mid-term outcomes.

AIM

To determine association between cardiovascular abnormalities and clinical and laboratory parameters. To study the time-line for resolution of various abnormalities.

METHODS

In this prospective study done in a tertiary care hospital, 270 were recruited from June 2020 to January 2022. Baseline demographic data and clinical presentation were recorded. Laboratory parameters and echocardiography were done at admission. Follow-up was done at 2 weeks, 3 months, 6 months and 1 year after diagnosis. Descriptive statistics were used for parametric and non-parametric data. Risk factors were identified by multivariate regression analysis.

RESULTS

The 211 (78.2%) had cardiac involvement and 102 needed intensive care unit (ICU) admission. Cardiovascular abnormalities observed were shock 123 (45.6%), coronary dilatation 28 (10.4%), coronary aneurysm 77 (28.5%), left ventricular (LV) dysfunction 78 (29.3%), mitral regurgitation (MR) 77 (28.5%) and pericardial effusion 98 (36.3%). Coronary artery aneurysm/dilatation during follow-up at 2 weeks and 1 year were 25.7% and 0.9% respectively. Multivariate regression analysis revealed breathlessness [odds ratio (OR) = 3.91, 95%CI: 1.25-12.21, P = 0.019] and hi-flow nasal cannula (HFNC) support (OR = 8.5, 95%CI: 1.06-68.38, P = 0.044) as predictors of cardiovascular involvement. Higher mean age (OR = 1.16, 95%CI: 1.02-1.32, P = 0.026), breathlessness (OR = 4.99, 95%CI: 2.05-12.20, P < 0.001), gallop (OR = 4.45, 95%CI: 0.41-2.52, P = 0.016), MR (OR = 3.61, 95%CI: 1.53-8.53, P = 0.004) and invasive ventilation (OR = 4.01, 95%CI: 1.28-12.58, P = 0.017) were predictive of LV dysfunction. Altered sensorium (OR = 4.96, 95%CI: 2.23-11.02, P < 0.001), headache (OR = 6.61, 95%CI: 1.46-29.92, P = 0.014), HFNC (OR = 7.03, 95%CI: 2.04-24.29, P = 0.002), non-rebreathing mask usage (OR = 21.13, 95%CI: 9.00-49.61, P < 0.001) and invasive ventilation (OR = 5.64, 95%CI: 1.42-22.45, P = 0.014) were risk factors for shock. Anemia was a risk factor for coronary involvement (OR = 3.09, 95%CI: 1.79- 5.34, P < 0.001).

CONCLUSION

Significant number of children with MIS-C had cardiovascular involvement contributing to higher ICU management. Although shock resolved quickly, resolution of ventricular function and coronary abnormalities were slower, and hence warrants a structured long-term follow-up protocol.

Key Words: Multisystem inflammatory syndrome in children; Cardiovascular; Midterm follow-up; Coronary artery aneurysm; Shock; Left ventricular dysfunction

Core Tip: Multisystem inflammatory syndrome in children associated with severe acute respiratory syndrome coronavirus can affect the cardiovascular system leading to morbidity and mortality. This prospective study analysed the spectrum of cardiovascular involvement and aimed at determining the associations with clinical and laboratory parameters. Of the 211 patients with cardiac involvement, shock was the predominant finding followed by pericardial effusion, left ventricular (LV) dysfunction, mitral regurgitation and coronary aneurysm. LV function normalized in majority of children within two weeks. Most of the coronary abnormalities normalized within 6 months. None of the laboratory parameter was predictive of cardiac involvement.

INTRODUCTION

Multisystem inflammatory syndrome in children (MIS-C) associated with severe acute respiratory syndrome coronavirus (SARS-CoV-2) is a multisystem disease which affects a small proportion of children leading to serious illness requiring hospitalization and long-term complications[1]. MIS-C is a hyperinflammatory syndrome that develops approximately 2-6 weeks after SARS-CoV-2 infection. Children can present with features of gastrointestinal, cardiovascular, hematological, mucocutaneous, respiratory or neurological involvement. Two thirds of the children with MIS-C had cardiovascular involvement[1-3]. The brunt of the disease happens in heart causing shock, severe myocarditis and coronary aneurysms[4]. About two-thirds of children with MIS-C required intensive care management in Western countries and in India[3,5,6]. Systematic review of few Indian studies revealed mortality rate of 10.8% and majority had cardiovascular involvement[5]. Most common cardiac findings were left ventricular (LV) dysfunction and coronary dilatation[7,8]. Data from low-income and middle-income countries (LMICs) are comparatively scarce.

We describe the clinical features of children with cardiovascular system involvement in MIS-C and their outcomes after 1 year of diagnosis.

MATERIALS AND METHODS

This study prospectively enrolled children diagnosed as MIS-C as per World Health Organisation (WHO) criteria at a pediatric tertiary care centre in Chennai from June 2020 to Jan 2022[9]. We had previously published the varied profile and presentation of MIS-C cases from our institution and their outcomes. The study was approved by the Institutional Ethics Committee and informed consent obtained from the parents/legal guardians. Children who were treated elsewhere, whose initial details could not be traced and children who had co-infections were excluded. Children with pre-existing congenital and acquired cardiac lesions were also excluded. Baseline demographic characteristics like age, sex, weight and height were obtained. Presenting symptoms and physical examination findings were recorded. Laboratory blood parameters measured were complete blood count, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), alanine aminotransferase (ALT), aspartate aminotransferase (AST), serum albumin, blood urea, serum creatinine, ferritin, D-dimer, interleukin-6 (IL-6), procalcitonin, N-terminal pro-brain natriuretic peptide (NT-proBNP) and troponin I. Duration of intensive care unit (ICU) stay, inotropic support and type of respiratory support-non-rebreathing mask (NRM), hi-flow nasal cannula (HFNC), bilevel positive airway pressure (BIPAP) and invasive ventilation were recorded. Clinical parameters were defined as per Indian Academy of Paediatrics, advanced life support guidelines[10]. Anemia was defined as hemoglobin < 11 g/L up to 5 years, < 11.5 g/L for 5-11 years and < 12 g/L for 12-14 years as per WHO definition[11]. Thrombocytopenia was defined as platelet count < 150 × 10⁹/L while platelet count > 450 × 10⁹/L was considered as thrombocytosis. Lymphocyte count < 3.0 × 10⁹/L was considered lymphopenia. Echocardiogram was performed by pediatric cardiologist using GE VIVID S6 ultrasound machine with phased array cardiac transducers with two range of frequencies (3-8 MHz and 2-4 MHz). Structural cardiac anatomy, coronary artery dimension [left main coronary artery (LMCA), left anterior descending artery (LAD), left circumflex artery (LCX) and right coronary artery (RCA)], LV systolic function–ejection fraction (EF) and fractional shortening by M-mode, valvular regurgitation, pericardial effusion were analysed. Children with shock, pericardial effusion, reduced LV EF, coronary involvement in the form of coronary dilatation or aneurysm, arrhythmia, mitral regurgitation (MR), elevated troponin I and elevated brain natriuretic peptide or NT-proBNP were considered to have cardiac involvement[12]. Shock was defined as hypotension with poor perfusion requiring fluid resuscitation > 10 mL/kg and/or inotropic support. Children were subdivided into two groups based on the above criteria as with and without cardiac involvement. Coronary arteries were normal if the Z score was < 2, dilated if the Z score was 2-2.5 and aneurysmal if the Z score was > 2.5. Coronary artery aneurysms (CAA) were sub-categorised as small, medium and giant based on their Z scores of 2.5-5, 5-10 and > 10 respectively[13]. LV systolic dysfunction was defined as EF less than 55%[14]. SARS-CoV-2 immunoglobulin G antibody titre more than 10 AU was considered as reactive. Antibody tests were done using Indian council of medical research approved assay kits. Children were followed up for a period of one year with repeat echocardiogram at 14 days, 3 months, 6 months and 1 year of diagnosis. Follow up data was evaluated among children who completed one year of diagnosis by January 2023. We planned to analyse the pattern of cardiovascular involvement in MIS-C and its association with clinical and laboratory parameters. Follow-up data was used to determine the time taken for normalization of cardiovascular abnormalities.

Statistical analysis

Categorical variables were presented as number and percentage. Continuous variables were presented as mean with standard deviation or median and interquartile range (IQR) appropriately. For study parameters, descriptive statistics like proportions, mean and standard distribution were used for parametric data and median (IQR) for non-parametric data. Categorical data were analyzed using odds ratio with 95%CI and student t test with P < 0.05 taken as significant. Multivariate regression analysis was undertaken to identify the individual risk factors. Friedman test followed by Wilcoxon Signed Rank test was used to compare the Z score of coronary measurements at different time points. Data was analyzed using Epi InFo version 7 and Statistical Package for the Social Sciences version 20.0 for Windows (IBM Corporation ARMONK, NY, United States).

RESULTS

Among the 300 children admitted with features of multisystem inflammation during the study period, 270 children were included for the study as per the WHO definition. Among the study group 211 children (78.2%) had cardiovascular involvement. The median (IQR) age was 6 years (2-9 years). Male:female ratio was 1.25. At admission, coronary dilatation was seen in 28 (10.4%), hyperechogenic coronaries in 42 (15.6%), coronary aneurysm in 77 (28.5%), pericardial effusion in 98 (36.3%), LV dysfunction in 78 (29.3%), MR in 77 (28.5 %) and shock in 123 (45.6%). Comparison of study parameters were done among the children with and without cardiac involvement (Table 1). Median (IQR) of NT-proBNP was 4875 (559-16006) and troponin I was 18.6 (0.01-141.8) among the children with cardiac involvement. Majority of children with cardiac involvement required some form of respiratory support like oxygen through NRM (n = 96), HFNC (n = 52), invasive ventilation (n = 53), BIPAP (n = 17). Regression analysis revealed breathlessness and HFNC as significant factors associated with cardiac involvement in MIS-C children (Table 2). Median (IQR) inotrope duration was 64.5 hours (43-120 hours). Inotropes were used for initial stabilisation in 38.9% (95%CI: 33.04-45.99). ICU care was required in 102 children and median (IQR) duration of ICU care was 5 days (2-9 days). Figure 1 depicts the prevalence of coronary artery dilatation and various categories of CAA as per Z score. Mean coronary artery Z score at various time periods are shown in Figure 2. Mean Z score for LMCA at admission and at 12 months was 1.59 ± 1.89 and 0.42 ± 0.72 respectively (P < 0.001). Mean Z score for LAD decreased from 1.25 ± 1.48 at admission to 0.44 ± 0.77 at 12 months (P < 0.001). RCA mean Z score declined from 0.99 ± 1.64 at admission to 0.34 ± 1.17 at 1 year (P < 0.001). Frequency of coronary involvement at admission among LMCA, LAD and RCA was 28.3%, 32.9% and 20.9 % respectively (Table 3). Normalization of various cardiovascular abnormalities during follow-up is given in Figure 3. Number of children who were assessed at admission, 2 weeks, 3 months, 6 months and 1 year were 270, 237, 229, 221 and 219 respectively.

Figure 1 Pattern of coronary artery involvement as per Z score classification at admission and during follow-up. CAA: Coronary artery aneurysms.

Figure 2 Mean Z Score of left main coronary artery, left anterior descending artery and right coronary artery at various time periods. A: left main coronary artery Z score; B: Left anterior descending artery Z score; C: Right coronary artery Z score. LMCA: Left main coronary artery; LAD: Left anterior descending artery; RCA: Right coronary artery.

Figure 3 Time to normalization of various cardiovascular parameters and the number of children affected. A and B: Time to normalization of various cardiovascular parameters; C: Number of children affected. LV: Left ventricular.

Table 1 Univariate analysis of clinical and laboratory parameters in those with and without cardiac involvement, n (%).

Clinical parameters	With cardiac involvement (n = 211)	Without cardiac involvement (n = 59)	Odds ratio (95%CI)	P value
Gender (female)	94 (78)	26 (21.6)	0.98 (0.52-1.82)	0.534
Pain abdomen	103 (48.80)	29 (49.97)	0.986 (0.55-1.766)	0.54
Altered sensorium	78 (36.97)	12 (20.34)	2.29 (1.164-4.745)	0.010
Breathlessness	64 (30.3)	4 (6.78)	5.95 (2.231-20.06)	< 0.001
Cracked lips	94 (44.55)	21 (35.59)	1.45 (0.800-2.676)	0.139
Diarrhea	195 (77.21)	31 (52.54)	0.895 (0.499-1.600)	0.409
Edema	100 (47.39)	17 (28.81)	2.22 (1.197-4.232)	0.007
Fever (> 38 °C)	209 (99.0)	56 (94.92)	5.55 (0.622-67.95)	0.071
Oliguria	43 (20.38)	5 (2.47)	2.75 (1.019-9.36)	0.022
Rash	97 (45.97)	33 (55.93)	0.671 (0.358-1.247)	0.113
Red eye	129 (61.14)	36 (61.02)	1.005 (0.55-1.815)	0.55
Red lips	80 (37.91)	22 (37.29)	1.02 (0.56-1.88)	0.528
Vomiting	161 (76.3)	45 (76.27)	1.008 (0.49-1.95)	0.559
Non-rebreathing mask	96 (45.71)	0	Undefined	< 0.001
Hi-flow nasal cannula	51 (24.17)	1 (1.92)	18.3 (3.4-382.5)	< 0.001
Bilevel positive airway pressure	16 (7.88)	1 (1.69)	4.94 (0.86-106.8)	0.07
Invasive ventilation	53 (25.24)	0	Undefined	0.001
Thrombocytopenia	108 (51.18)	14 (24.14)	3.28 (1.711-6.52)	< 0.001
Thrombocytosis	62 (29.52)	17 (28.81)	1.03 (0.55-1.994)	0.527
Anemia	128 (60.66)	11 (28.81)	3.79 (2.04-7.248)	< 0.001
Hyponatremia	170 (80.59)	36 (61.01)	2.638 (1.401-4.938)	0.002
Continuous variables
Age (years)	6 (2-9)	5 (2.5-9)		0.912
Heart rate/minute	130 (112-150)	110 (100-130)		< 0.001
C-reactive protein (mg/L)	48 (12-112)_	48 (12-137)		0.785
Erythrocyte sedimentation rate (mm/hour)	45 (28-80)	49.5 (31-88)		0.49
Highest platelet count (103/L)	294 (178-486)	343 (215-502)		0.132
Lowest platelet count (103/L)	127 (65-225)	250 (120-320)		0.001
Neutrophil-to-lymphocyte ratio	4 (2.2.-8)	3 (2-5.3)		0.150
White blood cell count (103/L)	15.65 (11-21.6)	12.4 (8.9- 17.2)		0.004
Aspartate aminotransferase (U/L)	41 (26-90)	31 (24-53)		0.047
Alanine aminotransferase (U/L)	32 (18-66)	25 (14-47)		0.042
Urea (mg/dL)	31.5 (21-54_	21 (16.5-27.5)		0.001
Creatinine (mg/dL)	0.4 (0.3-0.6)	0.4 (0.3-0.5)		0.087
Ejection fraction	58 (48-62)	63 (60-66)		< 0.001
Fractional shortening	31 (26-33)	33.5 (33-35)		0.017
interleukin-6 (pg/mL) (n = 95)1	87.6 (19.2-320	29.9 (9.36-142)		0.08
D-dimer (mg/L) (n = 125)1	3.1 (1.57-5.1)	1.83 (0,88-3.29)		0.028
Procalcitonin (ng/mL) (n = 70)1	15.8 (2.2-44.5)	0.76 (0.25-5.62)		0.014
Troponin I (ng/mL) (n = 49)1	18.6 (0.01-141.8)	0.01		-
N-terminal pro-brain natriuretic peptide (pg/mL) (n = 49)1	4875 (559-16006)	26		-
Ferritin (ng/mL) (n = 166)1	575 (293-1335)	281 (254-576)		0.001
Coronavirus disease 2019 antibody titre	63.8 (24.8-95)	65.8 (23.5-102.4)		0.58
Hospital stay (days)	10 (7-13)	8 (7-11)		0.070
Prehospital illness (days)	5 (3-7)	5 (4-7)		0.333

¹Number of children who had undergone the test.

Table 2 Multivariate regression analysis.

Condition	Parameter	Odds ratio	95%CI	Coefficient	SE	P value
Cardiovascular involvement	Breathlessness	3.91	1.25-12.21	1.36	0.58	0.019
	HFNC	8.50	1.06-68.38	2.14	1.06	0.044
Left ventricular dysfunction	Higher mean age	1.16	1.02-1.32	0.15	0.07	0.026
	Breathlessness	4.99	2.05-12.20	1.61	0.46	< 0.001
	Gallop	4.45	0.41-2.52	1.49	0.46	0.016
	Mitral regurgitation	3.61	1.53-8.53	1.28	0.44	0.004
	Invasive ventilation	4.01	1.28-12.58	1.39	0.58	0.017
Shock	Altered sensorium	4.96	2.23-11.02	1.60	0.41	< 0.001
	Headache	6.61	1.46-29.92	1.89	0.77	0.014
	Female gender	0.30	0.14-0.67	-1.19	0.40	0.003
	HFNC	7.03	2.04- 24.29	1.95	0.63	0.002
	Non-rebreathing mask	21.13	9.00-49.61	3.05	0.44	< 0.001
	Invasive ventilation	5.64	1.42-22.45	1.73	0.71	0.014
Coronary involvement	Anemia	3.09	1.79-5.34	1.13	4.07	< 0.001

HFNC: Hi-flow nasal cannula.

Table 3 Coronary arteries involved as per Z score in echocardiography at admission and follow-up, n (%).

Coronary dimension	Site	At admission	2 weeks	3 months	6 months	12 months
Normal	LMCA	190 (71.7)	201 (84.8)	215 (93.5)	216 (98.2)	218 (99.1)
	LAD	175 (67.1)	186 (78.5)	207 (91.6)	217 (98.2)	-
	RCA	207 (78.7)	209 (88.2)	218 (95.2)	217 (98.6)	218 (99.1)
Only dilatation	LMCA	25 (9.4)	2 (0.8)	3 (1.3)	1 (0.45)	-
	LAD	14 (5.4)	14 (5.9)	3 (1.33)	1 (0.45)	-
	RCA	10 (3.8)	10 (4.2)	3 (1.3)	1 (0.45)	-
Mild aneurysm	LMCA	46 (17.4)	31 (13.1)	10 (4.4)	2 (0.9)	1 (0.45)
	LAD	60 (22.9)	30 (12.7)	14 (6.2)	2 (0.9)	-
	RCA	41 (15.6)	14 (5.9)	6 (2.6)	1 (0.45)	1 (0.45)
Moderate aneurysm	LMCA	4 (1.51)	3 (1.3)	1 (0.43)	1 (0.45)	1 (0.45)
	LAD	11 (4.2)	5 (2.1)	2 (0.88)	1 (0.45)	-
	RCA	4 (1.5)	2 (0.84)	-	-	-
Giant aneurysm	LMCA	-	-	1 (0.43)	-	-
	LAD	1 (0.4)	2 (0.84)	-	-	-
	RCA	1 (0.4)	2 (0.84)	2 (0.87)	1 (0.45)	1 (0.45)

LMCA: Left main coronary artery; LAD: Left anterior descending artery; RCA: Right coronary artery.

Comparison of the study parameters with shock, LV dysfunction and initial coronary involvement was done. Altered sensorium (P < 0.001), breathlessness (P = 0.004), echogenic coronaries (P = 0.006), female gender (P = 0.004), headache (P = 0.010), vomiting (P = 0.013), need for HFNC (P < 0.001), NRM (P < 0.001), BIPAP (P = 0.005), ventilatory support (P < 0.001), inotrope support (P < 0.001), thrombocytopenia (P < 0.001), LV dysfunction (P < 0.001), shorter prehospital illness (P = 0.001), elevated IL-6 (P = 0.014), neutrophil-to-lymphocyte (NLR) ratio (P < 0.001), NT-proBNP (P = 0.001), and death (P < 0.001) were significantly associated with shock. Children with thrombocytosis had lesser shock and this was statistically significant (P < 0.001) in univariate analysis.

Among those with LV dysfunction, univariate analysis revealed significant association with breathlessness (P < 0.001), vomiting (P = 0.042), gallop (P < 0.001), shock (P < 0.001), skin peeling (P = 0.029), rash (P = 0.029), coronary involvement as aneurysm or dilatation (P = 0.006), echogenic coronaries (P = 0.004), MR (P < 0.001), need for HFNC (P < 0.001), BIPAP (P < 0.001), inotrope support (P < 0.001), thrombocytopenia (P < 0.001), bleeds (P = 0.015) and ventilatory support (P < 0.001). Children with LV dysfunction had higher age (P < 0.001), required longer duration of inotrope support (P = 0.001), higher NLR (P = 0.007), higher ferritin (P < 0.001), higher IL-6 level (P = 0.018), elevated D-dimer (P < 0.001), elevated troponin I (P = 0.006), low platelet levels (P < 0.001), and raised procalcitonin (P = 0.024) when compared to those without LV dysfunction. Children with thrombocytosis had lesser occurrence of LV dysfunction and this was statistically significant (P < 0.001).

Children with coronary involvement at admission were compared with those without coronary involvement. Univariate analysis revealed anemia (P < 0.001), breathlessness (P = 0.020), puffiness of face (P = 0.047), oliguria (P = 0.041), LV dysfunction (P = 0.006), MR (P = 0.014), need for BIPAP (P = 0.028), inotrope support (P = 0.031), lower platelets (P = 0.038), elevated ESR (P = 0.046) and longer hospital stay (P = 0.020) to be significantly different between the groups. Logistic regression analysis of various parameters for LV dysfunction, shock and coronary involvement is given in Table 2.

DISCUSSION

In this prospective study involving children with MIS-C with specific focus towards those with cardiovascular system involvement we described the clinical presentation and longitudinal cardiac outcomes. MIS-C has many clinical manifestations similar to Kawasaki disease (KD), albeit there are some dissimilarities. The difference in the hyperinflammatory states of MIS-C and KD may be the reason for the variability in presentation. Majority of children affected with KD are under 5 years of age[15]. The median age of children affected with MIS-C is 6 years to 10 years and is established in this study[16]. Male preponderance was similar to other studies[17].

Pathogenesis of cardiovascular involvement in MIS-C is probably multifactorial[12]. Hyperimmune response to the SARS-CoV-2 in genetically susceptible child causes cytokine storm leading to fever and markedly elevated inflammatory markers like procalcitonin, ferritin and proinflammatory cytokines like IL-6 which was apparent in our case series. Pathophysiology of cardiovascular involvement are likely due to viral invasion of cardiomyocyte, dysregulated inflammatory response, microvascular injury, endothelial injury and hypercoagulability[12]. MIS-C children may suddenly present with shock and severe hemodynamic compromise which was evident in the current study and such presentation are rare in KD[13,18]. Cardiac complications in MIS-C necessitated fluid boluses, inotropic support and mechanical ventilation for hemodynamic stabilization which has been the scenario in many countries[18,19]. LV systolic dysfunction was significant in our group with cardiac involvement and was a common finding in most studies[18,20,21]. Due to logistical constraints in coronavirus disease 2019 pandemic, cardiac biomarkers like NT-proBNP and troponin I was done only in few children and it was elevated. Elevated cardiac enzymes indicate cardiomyocyte injury[21]. Some patients (n = 4) had decreased LV function and normal cardiac enzymes level probably due to an alternate pathogenesis like generalized inflammation or changes in loading conditions[22]. Elevated levels of troponin (n = 30), NT-pro BNP (n = 42) and their association with LV dysfunction and shock were evident in our study and also from meta-analysis of MIS-C patients[23]. Clinical correlate of these findings suggests that as the inflammation increased, more severe was the myocarditis, LV dysfunction and shock. These findings were corroborated by univariate analysis data on LV dysfunction and shock. LV dysfunction causes elevated LV end diastolic pressure and decreased cardiac output leading to pulmonary congestion, breathlessness and hypotension. Hemodynamic stabilization of these children with acute heart failure requires ventilatory support. Higher percentage of children requiring ICU admission, oxygen support and ionotropes indicate that we had more severe forms of MIS-C with severe cardiovascular involvement. Breathlessness and HFNC were predictors of cardiovascular involvement and this validates that significant number of children had severe cardiovascular involvement (LV dysfunction, shock). Children with LV dysfunction often have MR due to abnormal coaptation of leaflets and annular dilatation. Ptak et al[24] had reported older age as an independent predictor of LV dysfunction and hypotension which was also seen in our study. Breathlessness, gallop, MR and invasive ventilation were independent predictors for LV dysfunction in this cohort (Table 2). LV dysfunction would have precipitated occurrence of acute heart failure (breathlessness and gallop) and management of these findings necessitated invasive ventilation in severe cases of our cohort. LV function normalized in majority of children within 2 weeks just like resolution of immunological changes noted in the acute phase of illness which is similar to other longitudinal studies[20,24-26].

Multicentric studies had reported that large proportion of children admitted in ICU had hemodynamic abnormality of shock and hypotension which was observed in this study[27]. Elevated levels of NT-pro BNP may also be because of myocardial edema or stunning[27]. Loss of vasomotor tone in resistance vessels may be because of raised IL-6 and probably the cause of shock in majority of children along with LV dysfunction[27]. Many children required fluid boluses, inotropes and respiratory support during the initial phase of management. Shock was corrected in all patients in few days (Figure 3). Larger proportion of children requiring ICU care is similar to a meta-analysis study and other western studies which is due to the presence of LV dysfunction and shock during the initial phase of illness[19-21,23]. Median duration of inotropic usage was similar to a study by Toubiana et al[28]. Length of ICU stay in our MIS-C children was comparable to a multicentric ICU prospective study[29]. Respiratory support by NRM, HFNC and invasive ventilation were independent predictors of shock substantiating significant hemodynamic instability in our study. Altered mental status and headache being a sign of clinical hypoperfusion was a predictor of shock in these children validating the requirement of inotropes and aggressive ICU management in our study. Higher proportion of females presented with shock which was statistically significant and we could not explain the reason.

Though many studies had observed higher prevalence of myocardial injury causing LV dysfunction compared to coronary dilatation, in this study the frequency of coronary involvement (dilatation/CAA) was much higher[2,27]. Prevalence of CAA is higher among MIS-C children (6%-24%) compared to KD[12,21,30]. Studies have observed that CAA in MIS-C generally resolves with time[12]. Pathophysiology for coronary involvement has not been elucidated. Transient coronary dilatation may be because of vasodilatation due to high proinflammatory milieu and destruction of arterial wall may lead to coronary aneurysm[31,32]. We had 1.3 % of children with giant CAA at 2 weeks and only 0.5% at 1 year follow up which was comparable to other reports[20,21,33,34]. The trend of regression of CAA is similar to other studies and in our cohort most of the coronary abnormalities normalized at 6 months follow-up[20,34]. LCX was not involved in any of our cases. RCA was less frequently involved compared to LMCA and LAD and similar findings had been reported in small case series[35]. Coronary artery dimension was maximum at admission and there was a progressive decline during follow up (Figure 2). Children with giant CAA were managed with oral warfarin and antiplatelet drugs as they are at risk of progressive coronary artery occlusion/stenosis and worsening ischemia. Most of the other children with coronary involvement were on oral aspirin. Anemia was the only factor predicting coronary involvement in regression analysis. Lower hematocrit was a predictor for CAA in KD and had been used to predict CAA in few scoring systems in KD[36,37]. Anemia may impair cell-mediated immunity and that may trigger release of cytokines and enzymes by inflammatory cells resulting in coronary arteriopathy[38]. This hypothesis is probably the reason for anemia being predictive of coronary involvement in our cohort. None of the cardiac biomarkers, proinflammatory markers nor any clinical features were predictive of coronary involvement suggesting systemic inflammation alone was probably not the cause for changes in coronary artery. These findings also suggest that LV dysfunction and coronary artery abnormalities are due to two different pathophysiological processes in MIS-C. Longitudinal study by Chakraborty et al[34] had proposed this possibility. Most cohorts have described higher frequency of cardiac involvement compared to coronary involvement[2,27,33].

Respiratory distress requiring oxygen support or mechanical ventilation is common in MIS-C and is rarely seen in KD[2,28,29]. Invasive ventilation was required in about 1/4^th of our children which was comparatively lower than in western studies[25,27]. Lower incidence of invasive ventilation had been reported in few centers[29]. Breathlessness is secondary to shock, fluid overload from vascular leak and pulmonary congestion due to LV dysfunction leading to increased frequency of respiratory support. Primary respiratory involvement is not the cause for breathlessness in MIS-C[25]. In our cohort, breathlessness requiring oxygen support reflects that there was considerable number of children with LV dysfunction, shock and vascular leak. Breathlessness and respiratory support in the form of HFNC were independent predictors for cardiac involvement.

The occurrence of MR nearly paralleled LV dysfunction in our case series. The resolution of MR lags slightly from the recovery of LV function. We had four cases with mild MR at 1 year follow up though few studies recorded complete resolution of MR by 6 months[34]. These findings in our study are supportive to the hypothesis of subtle long term cardiac injury in MIS-C which prolongs recovery time of cardiac findings[20,30]. Most children had trace to mild pericardial effusion due to pericarditis with spontaneous resolution during follow up which was similar to contemporary studies[21,26,34]. We did not have any child with arrhythmia or rhythm abnormalities similar to case series in India[8,35]. Studies form other countries have reported rhythm problems[26,27,33].

IL-6 levels were higher in our group with cardiovascular involvement but not statistically significant. Elevated IL-6 has been postulated to be due to macrophage activation and stretched cardiomyocytes and cardiac fibroblasts[27].

A large multicentric case series of MIS-C cases from United States with large number of cardiac involvement reported elevated NLR, elevated CRP and thrombocytopenia[20]. MIS-C children with cardiac involvement in our group had statistically significant thrombocytopenia but not associated with higher NLR or CRP. In our cohort of MIS-C children with cardiac involvement had higher white blood cell count, anemia and low sodium which had been reported earlier[33]. Procalcitonin, ferritin and D-dimer were significantly elevated in cardiac MIS-C group of patients as reported in many studies and systematic review[19,21,23]. Elevated D-dimer is probably due to microvasculitis and our center had reported significant association of D-dimer and other coagulation parameters with mortality in MIS-C[39,40]. MIS-C children with cardiac involvement had significantly lower sodium levels which was probably due to gastrointestinal loss. Metanalysis study by Jiang et al[1] had mentioned higher prevalence of hyponatremia in MIS-C and it had been reported to be associated with mortality in MIS-C in a multicentric study from eastern India[41]. Elevated levels of AST and ALT were seen in cases with cardiovascular involvement which is similar to other MIS-C studies[2,39,41]. None of the laboratory parameters were predictors of cardiac involvement though a multicenter cohort study had troponin and D-dimer as predictors of myocardial involvement[39].

Strengths and limitations

This was a single center study from the largest government tertiary referral center for children in our region and hence we would have received more children with MIS-C complications which would have affected the results. The overall spectrum of MIS-C with cardiovascular manifestations might not have been reflected and hence results cannot be generalized. Laboratory investigations were not standardized at admission due to logistical constraints but were individualized to patients and we did not analyse the laboratory parameters during follow up. As cardiac magnetic resonance imaging was not part of our evaluation, we could have missed subclinical myocardial damage and any residual myocardial damage during follow up. Comprehensive assessment of LV diastolic function was not done. We did not elucidate the temporal progression from viral exposure to disease onset.

We analysed a larger cohort compared to many studies from other LMICs and followed them up for one year which is the strength of our study. Loss to follow up and missing data for echocardiogram was less in our study. Elucidation of immunopathogenesis of this disease will provide further insights for management of severe forms of MIS-C.

CONCLUSION

Cardiovascular involvement was common and heterogenous in MIS-C. Though there were similarities between MIS-C and KD, the prominent clinical signs were largely different. Shock and coronary aneurysms were the most common cardiovascular manifestations. Early diagnosis, timely and appropriate multidisciplinary management is essential for good recovery and better long-term outcomes. Although most cardiac findings improved by one year, few cases had residual cardiac issues which warrants the need for regular pediatric cardiology follow-up.

ACKNOWLEDGEMENTS

We wish to acknowledge Gnanasambandam S and all other faculty of Department of Cardiology, Pediatric Medicine and Allied Departments who supported in patient care.