Introduction
Multisystem inflammatory syndrome in children following COVID-19 infection was first reported in April 2020 in Europe and North America. Reference Godfred-Cato, Bryant and Leung1–Reference Jiang, Tang and Levin2 Following the first concerning reports, the World Health Organization proposed a case definition of multisystem inflammatory syndrome in chldren. 3 As the pandemic progressed, more reports onmultisystem inflammatory syndorm ein children epidemiology emerged from different parts of the world, and data from these showed that timely diagnosis and treatment resulted in good immediate outcomes and low overall mortality. Reference Jiang, Tang and Levin2,Reference Jiang, Tang, Irfan, Li, Zhang and Bhutta4 However, some reports have reported higher mortality, including one from our group. Reference Mohsin, Abbas and Chowdhary5 The clinical spectrum of multisystem inflammatory syndrome in chldren includes mild to severe categories and three main phenotypes, including Kawasaki disease, viral myocarditis, and toxic shock syndrome. Reference Mohsin, Abbas and Chowdhary5 A large majority (80-85%) of children develop cardiovascular involvement, including myocarditis with shock, myocardial dysfunction, valvar regurgitation, and pericardial effusion. Reference Clark, Sanchez-de-Toledo and Bautista-Rodriguez6 Rapid recovery of both myocardial dysfunction and inflammatory markers has been demonstrated following the acute phase of the illness. There are some reports which show favourable outcomes at short- to medium-term follow-ups (up to a year). Reference Aziz, Sadiq and Qureshi7–Reference Matsubara, Chang and Kauffman10 However, these differ by geography, follow-up protocol, and age.
Our hospital is one of the largest tertiary healthcare centres in Pakistan. During the pandemic, we developed a dedicated interdisciplinary multisystem inflammatory syndrome in children team comprising of paediatric cardiologists, infectious disease physicians, and intensivists, and designed a consensus based treatment algorithm in context to our resources. (Supplement Fig 1 S1) Our follow-up programme was also developed to monitor the general course and cardiac outcomes of these patients. We describe the short- and medium-term outcomes of children admitted with multisystem inflammatory sndrome at our centre during the first two years of the pandemic.
Materials and methods
Study setting, design, and population
In this single centre, prospective longitudinal cohort study, all patients aged 1 month-18 years admitted at our hospital (a private tertiary care hospital in Pakistan located in the largest city) from April 2020 to March 2022, fulfilling the World Health Organisation criteria for multisystem inflammatory syndrome in children were enrolled and followed. Our centre is one of the largest referral centre in the country. Many patients are referred from other secondary and tertiary care centres across Pakistan; hence, the patient population presenting is with a higher acuity and higher severity of illness.
Patients were divided into 3 phenotypes of Kawasaki disease-like illness, viral myocarditis, and toxic shock syndrome-like illness (Table 1) 11,Reference Chuang, Huang and Lin12 as described by our group previously and were categorised into clinical severity levels of mild, moderate, or severe depending on organ involvement and level of inotropic and ventilatory support required Reference Mohsin, Abbas and Chowdhary5 (Table 2). Definitions of all conditions are presented in Table 2.
IQR = interquartile range, SD = standard deviation, PICU= paediatric intensive care unit, PCR = polymerase chain reaction, CRP = c reactive protein, LDH = lactate dehydrogenase, BNP = beta-natriuretic peptide, LVEF = left ventricular ejection fraction, GLS = global linear strain *Normal ranges provided in Supplementary Table 2.s.
The tests required for diagnosis included baseline investigations [complete blood count, inflammatory markers (C-reactive protein, procalcitonin), coagulation (prothrombin time, partial thromboplastin time), d-dimers, fibrinogen], virology studies [SARS-CoV2 polymerase chain reaction, SARS-CoV2 antibodies, bacterial cultures, and imaging (Chest xray and echocardiography)] at baseline and then repeated on discharge if the patient met the case definition of multisystem inflammatory syndrome in children.
Before being labelled as multisystem inflammatory syndrome in children, all cases were reviewed and adjudicated by two clinicians (a cardiologist and an intensivist being present in all cases) independently. Some of the data of eleven patients included in this study was published by our group in a previous paper. Reference Mohsin, Abbas and Chowdhary5
Hospital management
Patients were either kept in COVID-19-designated wards, special care units (paediatric ICU (PICU), depending on the severity of the disease. The multidisciplinary multisystem inflammatory syndrome in children team was taken on board, and patients were treated according to our institute-based protocol and were discharged on tapering doses of steroids and aspirin. (Supplementary Figure 1)
Echocardiography
Two-dimensional echocardiography was performed using the E95 ultra-edition of the General Electronic (GE) system. Standard protocols based on the American Society of Echocardiography guidelines Reference Köroğlu, Atasever and Duman14 were used for measurements of cardiac systolic function. A two-dimensional speckle-tracking analysis was performed offline to assess the systolic global longitudinal strain and global longitudinal strain rate. We defined the global longitudinal strain according to prevalent adult data, as there is a dearth of paediatric normative data. We defined left ventricular systolic dysfunction by global longitudinal strain < −17% Reference Mor-Avi, Lang and Badano15 . Coronary artery z-scores were derived from normative data (Boston z-score system) and classified as shown in Supplementary Table 2. Echo-brightness was described as the appearance of echo-bright densities surrounding the coronary lumen extending for at least 1 cm along the artery. Reference Mor-Avi, Lang and Badano23 Valvulitis was diagnosed based on any valvar regurgitation greater than trivial based on the American Society Echocardiography guidelines.
An echocardiogram was performed for all patients on admission and a follow-up echocardiogram for surviving and discharged patients. An echocardiogram was repeated at the first follow-up and then as needed based on the cardiologist’s discretion.
Outpatient follow-up
Outpatients’ follow-up protocol was designed by paediatric cardiology team with contribution from intensivist and infectious disease team. Patients were guided to follow paediatric cardiology clinic and, if needed, paediatric infectious disease clinics. Patients were advised to follow up in paediatric cardiology clinics at 2 weeks, 3 months, 6 months, 1 year, and 2 years after discharge. Follow-up complete bloos count, ferritin, D dimer, and echo were recommended if patients were presenting with symptoms post-discharge. Follow-ups were discontinued for patients when laboratory or echocardiogram parameters returned to baseline.
Data collection
Data were extracted from the medical records of patients and were collected on a structured proforma, which included basic demographic details, clinical and laboratory variables, imaging data, including echocardiographic reports, and outcomes (survival) during admission and each follow-up.
Data analysis
Data were entered and analysed using STATA version 15.0. Results are presented as median with interquartile range for age, paediatric ICU length of stay, hospital length of stay, vasoactive inotropic score and laboratory parameters and as frequency with percentage for phenotypes, presenting system involvement, severity category, therapies in hospital and laboratory parameters. The weight of all patients was described as mean ± standard deviation. The proportion for patients tested for different laboratory parameters was also mentioned in Table 1. Appropriate statistical tests (Pearson chi-squared for categorical variable, two sample t-test for continuous normally distributed data, and Wilcoxon rank sum test for non-normal data) were applied. A p-value of<0.05 was considered significant.
Results
Demographics, clinical characteristics, and inpatient treatment
A total of 51 patients fulfilling the WHO MIS-C criteria were identified during the study period. (Fig 1) The median (IQR) age of the study population was 72 months (11-144) and 43 (84.3%) were males. Only 2 (3.9%) patients had preexisting comorbid conditions, with both being congenital heart defects. 11/51 (21.5%) were transferred from other hospitals. The most common system wise presentation included the gastrointestinal system in 31 (60.8%) children, followed by the cardiovascular system in 27 (52.9%) children and the central nervous system in 20 (39.2%) children. Severe disease was present in 38 (74.5%) children, whereas 8 (15.7%) children had moderate disease and 5 (9.8%) children had mild disease. The frequency of phenotypes included Kawasaki Disease (atypical Kawasaki Disease), Viral Myocarditis, and Toxic Shock Syndrome with 13 (25.5%), 21 (41.2%), and 17 (33.3%) children respectively.
In treatment, 34 (66.7%) children received intravenous immunoglobulins (IVIG) and 49 (96.1%) children received methylprednisolone. Enoxaparin was administered to 35 (68.6%) children. A total of 42 (82.4%) patients required admission to the PICU, of which 16 (31.4%) required mechanical ventilation and 37 (72.6%) required inotropes (epinephrine, norepinephrine, vasopressin, and/or milrinone). The median (IQR) vasoactive inotropic score was 9.5 (6.0-14.0), which was higher in children who died [14.5 (8.0-18.0)] compared to those who survived [9 (6.0-14.0)] (p = 0.27). The median (IQR) lengths of PICU and hospital stay in patients who survived were 3.0 (2.0-7.0) and 7.0 (5.0-9.0) days respectively. (Table 1)
Laboratory profiles
The hematological, inflammatory, and cardiac laboratory profiles of the patients at admission and discharge are described in Figure 1. Of the 51 patients, 17 (33.3%) tested PCR positive and 36 (70.6%) were found to have reactive SARS-CoV-2 antibodies. (Table 1) At admission, laboratory parameters showed evidence of acute inflammation in most children, with median (IQR) levels of ferritin were 1169 (534.75-1704.5) in 43/48 (89.5%), C- reactive protein 83.0 (24.0-175.0) in 42/47 (89.3%), Lactate Dehydrogenase 468.0 (365.0-1270.0) in 45/47 (95.7%). Similarly, median and IQR values of N-terminal pro-B-type natriuretic peptide were 8656.0 (2538.0-31166) in 40/41 (97.6%) patients and troponins levels were 0.157 (0.025-2.0) in 37/43 (86.0%) patients. Neutrophilia and lymphopenia were evident in all admitted children, with a median (IQR) absolute neutrophil count of 81.6 (66.0-88.2) and lymphocyte count of 11.4 (7.5-22.0). A significant reduction in all inflammatory parameters was seen at discharge, with median (IQR) neutrophil and lymphocyte counts falling to 64 (40.0-76.0) and 24.5 (15.4-42.5) respectively. The median (IQR) serum ferritin 592.0 (355.5-919.3), CRP 13.0 (3.0-35.3), LDH 364.5 (261.0-693.8) also significantly declined. (Fig 1)
Echocardiogram profile: (Supplementary table 2)
Among all 51 patients, decreased left ventricular ejection fraction was present in 30 (58.8%), decreased global longitudinal strain in 17 (33.3%), valvulitis in 38 (74.5%), pericardial effusion in 23 (45.1%), and coronary artery dilation in 3 patients. The median left ventricular ejection fraction and GLS were 55.5 (41.0-62.0) and -14 (-16- 12) respectively. (Fig 2) Out of 30 patients with decreased LVEF, eight had an LVEF of 51-55%, 15/30 had LVEF between 31-50%, and seven had LVEF<30% at admission. Upon discharge, 83% (25 out of 28) showed improved LVEF, but two patients with severely depressed LVEF died.
Outcomes & mortality characteristics
Forty-three (84%) children were discharged alive from the hospital while 6 (11.8%) died during their hospital stay. Two (4%) children left against medical advice. (Supplementary Figure S2) Of the six deaths, two children had the Toxic Shock Syndrome phenotype and four had the Viral Myocarditis phenotype, and all of them had severe disease. The median (IQR) age in months of the non-survival group was lower [9.0 (3.5-22.0)] compared to patients who survived [84.0 (12.0-144.0)], and mortality was significantly higher (p = 0.004) in patients who presented with respiratory symptoms. In the six patients who did not survive, the median (IQR) length of PICU stay was 5.0 (1.0-14.0), and all of them required mechanical ventilation, methylprednisolone, enoxaparin, and inotropic support, whereas only 4 (66.7%) received IVIG. Furthermore, as shown in Table 1, inflammatory markers in patients who did not survive were found to be significantly higher in comparison to the patients who survived, including ferritin (p-value<0.01) and LDH (p-value = 0.03). N-terminal pro-B-type natriuretic peptide was also significantly higher in the non-survival group (p-value = 0.03). There was no significant difference in echocardiogram features between the two groups.
Short and medium-term follow-up
Only 19 (44.2%) children followed up in outpatient clinic at 3 months, 5 (11.6%) following up at 4-6 months and 7-12 months, and 9 (20.9%) following up at 13-24 months. (Fig 1)
A follow-up echocardiogram was carried out only for patients in which cardiac parameters had not returned to baseline on discharge and continued until they normalized on follow-up visits. Global Longitudinal Strain normalized to a median of -18 by the 3-month follow-up and showed an improving trend thereafter with a GLS of -20 by the 1-year follow-up. By the 1-year follow-up, the number of patients with valvulitis and pericardial effusion, and coronary artery dilation decreased from 38 to 5 and from 23 to 0 and 3 to 0 respectively (Fig 3). There was one death during the follow-up period, however, it was unrelated to the study that was due to complications multiple congenital defects from a syndromic condition.
Discussion
We report the short and medium-term outcomes of children with MIS-C from our interdisciplinary 2-year follow-up program and elaborate on factors that may affect these outcomes. To the best of our knowledge, this is the largest longitudinal study from Pakistan.
The age and genders of our cohort were akin to trends seen in another countries Reference Kahn, Berg and Berntson17–Reference Li, Zhang, Zhang, Chen, Wang and Zhu20 . Over two-thirds [34 children (74.5%)] of our patients were found to have severe disease, and of the 6 children who did not survive, 4 had clear evidence of cardiovascular involvement, categorized as the Viral Myocarditis phenotype. Previously, it has been observed that the Viral Myocarditis phenotype is likely to present with fulminant myocarditis, which may account for the higher mortality observed in this group. Reference Mohsin, Abbas and Chowdhary5,Reference Cheung, Zachariah and Gorelik21
PICU admission was required in 82.4% of the patients, which is significantly higher compared to high income countries, which can be due to delayed presentation, more severity of disease at presentation and preference of the caring team due to new disease.
A variety of anti-inflammatory agents were used, including high-dose steroids (96.1%) and intravenous immunoglobulin (66.7%). Compared to studies from other regions where IVIG was the most commonly used immunomodulator, our IVIG use was less due to the fact that patient pay out of pocket for their treatment and its availability was also limited during pandemic. However, steroids are easily available and were administered to all patients. This is in line with current recommendations which suggest that using combination therapy of intravenous immunoglobulin and steroids early on is associated with a decreased risk of cardiovascular dysfunction later, as well as improved patient outcomes. Reference Mahmoud, El-Kalliny, Kotby, El-Ganzoury, Fouda and Ibrahim27 However it is also important to note that an earlier randomized controlled trial (RCT) conducted showed comparable results between IVIG and steroids for length of stay. 28 A recent RCT from Recovery trial group in England showed that steroids used as first line therapy in children with MIS-C reduce hospital length of stay. 28 All the patients in our cohort who died had received inotropic support, with the median vasoactive inotropic score being higher in these patients than in those who survived, though this difference was not statistically significant.
Current guidelines support the use of anticoagulation with low molecular weight heparins (LMWHs) in children with MIS-C who have mild- to moderate- ventricular dysfunction, coronary dilation with z-score 2.5–10, elevated D-dimers more than 5-10x the upper normal limit, or any new significant rhythm abnormalities, as well as prophylactic-dose anticoagulation for around 30 days post-discharge. Reference Mahmoud, El-Kalliny, Kotby, El-Ganzoury, Fouda and Ibrahim27 Though majority (68.6%) of our patients received enoxaparin during their hospital stay, most did not receive it post-discharge, largely due to a lack of follow-up. Moreover, although the efficacy of immunomodulatory therapy with biologic agents like Interleukin-1, Interleukin-6, and Tumor Necrosis Factor (TNF) inhibitors is still being investigated, there is data to suggest that therapy with these agents in patients refractory to standard combination therapy may be beneficial. Reference Cantarutti, Battista and Stagnaro22 Unfortunately, none of the patients in our cohort were able to benefit from these medications, largely due to unavailability and unaffordability. The findings suggest that nations with elevated per capita income have established protocols and monitoring guidelines for MISC patients, necessitating substantial resources, particularly in the absence of shared financial responsibilities. This challenge is more pronounced in countries with lower per capita income. Additionally, in these regions, there is no indication that the detailed protocols enhance outcomes or predict results in a manner influencing treatment decisions. The outcomes from this cohort also explore the progression of MIS-C patients over time, raising a direct question about the rationale for developing such intricate protocols.
Similar to earlier research conducted in Pakistan, our cohort exhibited a higher mortality rate compared to high-income countries, possibly attributed to factors such as delayed hospital presentation, limited resources (such as extracorporeal membrane oxygenation), and a lack of experience in managing a novel disease, particularly in the early stages of the pandemic. Further investigations into healthcare infrastructure shortcomings contributing to elevated mortality are necessary to pinpoint areas for improvement.
Upon admission, we observed elevated levels of inflammatory markers, specifically notable increases in Lactate Dehydrogenase and ferritin in the non-survival group. These findings align with prior studies, indicating that heightened levels of these laboratory markers could serve as indicators for identifying children at an increased risk of severe disease and poorer outcomes. Reference Cantarutti, Battista and Stagnaro22–Reference Fink Ericka, Alcamo Alicia and Lovett26,Reference Acevedo, Piñeres-Olave and Niño-Serna30,Reference Valverde, Singh and Sanchez-de-Toledo32–Reference Abrams, Oster and Godfred-Cato34 Additionally, our patients displayed biochemical evidence of myocardial injury, with notable elevations in pro-BNP and troponins, consistent with current literature. Reference Farooqi, Chan and Weller9,Reference Li, Zhang, Zhang, Chen, Wang and Zhu20,Reference Valverde, Singh and Sanchez-de-Toledo32,Reference Abrams, Oster and Godfred-Cato34–Reference Radia, Williams and Agrawal35 However, there was a significant reduction in these parameters upon discharge, with values returning to normal in subsequent follow-ups, indicating no long-term inflammatory dysregulation in our patients.
Farooqi et al. conducted one of the first studies on the long-term follow-up of patients with MIS-C, tracking 45 individuals for up to 9 months following diagnosis. Despite an initial high prevalence of echocardiographic abnormalities (80% mild, 44% moderate-severe), only 18% had residual mild abnormalities at the 1–4-week follow-up. NT-proBNP and Troponin-T levels normalized in most patients during this period. Of the 24 patients returning for the 4–9-month follow-up, only 2 exhibited remaining abnormalities: 1 with mild ventricular dysfunction and 1 with mild atrioventricular valve regurgitation Reference Matsubara, Chang and Kauffman10,Reference Yasuhara, Masuda and Watanabe18 . Similarly, in our cohort, 83% of patients demonstrated the normalization of left ventricular ejection fraction (LVEF) upon discharge. Additionally, improvements in valvar regurgitation and pericardial effusion were observed in 86% and 100% of patients, respectively, over a one-year period. We found that the normalization of many laboratory parameters occurred very rapidly following the acute event. Elevated levels of CRP, D-dimer and neutrophils were found in most of the included children during the acute phase. Reference Sai, Kumar, Arun Babu, Chaitra, Satapathy and Kalidoss24 Even though, follow-up patient population in our study was poor, this is because follow-up was only carried out in symptomatic patients at discharge which was low. This contrasts with the study in United States of America where post discharge, symptoms persisted in 67%. Forty percent had at least one return emergency visit and 24% had a hospital readmission. Reference Fink Ericka, Alcamo Alicia and Lovett26
The literature reports a wide range of incidence for coronary abnormalities in the range of 4-24%. Some studies indicate progressive coronary involvement during the follow-up period. However, in our cohort, only three patients developed coronary artery aneurysms, and all experienced rapid resolution within six months after discharge. This suggests a relative sparing of and quick recovery for the coronary arteries following the acute illness. Reference Matsubara, Chang and Kauffman10,Reference Cantarutti, Battista and Stagnaro22–Reference Bulut, Ekici and Kara23,Reference Jhaveri, Ahluwalia and Kaushik36,
Deformation parameters such as global longitudinal strain, left peak atrial strain, longitudinal early diastolic strain rate, and right ventricular free wall strain have been identified as robust predictors of myocardial injury, unlike left ventricular ejection fraction (LVEF), which may not reliably indicate subclinical myocardial injury and diastolic dysfunction in certain patients throughout the illness. Reference Feldstein, Tenforde and Friedman16,Reference Sanil, Misra and Safa19,Reference Matsubara, Kauffman and Wang37 Notably, our cohort exhibited improvement not only in LVEF but also in GLS. A study by Arifuddin et al Reference Arifuddin and Hazari38 showed that while KD is more prevalent in Asian countries, the incidence of Multisystem Inflammatory Syndrome in Children (MIS-C) was comparatively lower in Asian countries. However, a systematic review comparing high-income countries to low-income countries revealed that MIS-C cases in Low- and Middle-Income Countries (LMICs) had a lower percentage of PICU admission and mechanical ventilation rate but a higher risk of death, attributed to the scarcity of healthcare resources. Reference Jiang, Tang, Irfan, Li, Zhang and Bhutta39 Another study highlights that much before official guidelines were released, clinicians had already gradually adopted to the use of IVIG, corticosteroids and aspirin in their practices. In their study over 70% of MIS-C children in the USA were treated with combination therapy of corticosteroids and IVIG. With IVIG preferred over corticosteroids in younger populations. Reference Jiang, Tang, Irfan, Li, Zhang and Bhutta39 A national consensus study in the United Kingdom favored IVIG as the first-line treatment for MIS-C. Reference Harwood, Allin and Jones40
Another study conducted in Russia demonstrated complete short-term recovery (within 2 weeks) among the majority of patients following MIS-C diagnosis. Trends in cardiac involvement, including features of myocardial dysfunction and coronary artery dilation, were similar to those observed in our study. Reference Glazyrina, Zholobova and Iakovleva25 Likewise, a study from Turkey reported similar improvements in cardiovascular and echocardiographic parameters among children assessed two weeks after. Reference Bulut, Ekici and Kara23 In their cohort, intravenous immunoglobulin (IVIG) was the most commonly used immunomodulator, followed by steroids, with some cases requiring anakinra due to incomplete recovery. Similarly, a study conducted in South India indicated that IVIG was utilized in 94% of their cohort, followed by steroids in 70%. Their treatment approach was based on national consensus guidelines developed by the government. Reference Sai, Kumar, Arun Babu, Chaitra, Satapathy and Kalidoss24
Although global guidelines for the management of MIS-C are still lacking, the approach to management globally emphasizes prompt intervention. A study conducted in Italy described that the majority of patients received immunomodulatory therapy, primarily intravenous immunoglobulin or corticosteroids. In some cases, cytokine blockers, particularly IL-1 inhibitors like Anakinra, were used alongside conventional therapy, especially in MIS-C patients exhibiting severe left ventricular global systolic dysfunction and/or arrhythmias, or in those who did not show improvement within 24–48 hours after the initial therapeutic intervention. Antiplatelet treatment with aspirin was commonly administered, particularly in patients with evidence of coronary involvement or a presentation resembling Kawasaki disease. Therapeutic or prophylactic anticoagulation with heparin was also employed. Inotropic support was utilized for patients with moderate-to-severe ventricular dysfunction. Reference Cantarutti, Battista and Stagnaro22
Limitation and conclusion
Our study has notable limitations. Follow-up data were unavailable for some patients who left against medical advice or were lost to follow-up, reflecting the challenge of out-of-pocket healthcare costs in Pakistan. Additionally, the small sample size from a single tertiary care center limits the generalizability of our findings. The early stages of the pandemic may have led to missed cases due to limited SARS-CoV-2 testing availability. Cost and availability constraints prevented the assessment of certain cardiac parameters, including cardiac MRIs.
In conclusion, our cohort showed a high in-hospital mortality. Although most children with MIS-C experienced severe disease, with evidence of hyperinflammation and cardiac injury the majority recovered within a year of discharge without long term effects. The improvement we saw in our surviving cohort could also be attributed to the fact that the most severely affected children didn’t survive. However, those who survived showed good recovery on follow up.
Our study highlights the importance of understanding population-specific long-term outcomes in MIS-C, serving as a foundational step for future research and the development of specific follow-up guidelines in our context.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1047951124026283.
Acknowledgement
Authors would sincerely like to thank the whole study team and all the patients enrolled in this study as well as their families.
Author contribution
QA and SM orchestrated the conceptualization and methodology. QA and NA managed and curated the data, while SM supervised and undertook analysis. Both QA and SM took part in drafting the initial manuscript and contributed to the final review and editing. FS conducted the literature search and curated the data. FS initiated the first manuscript draft and participated in the subsequent review and editing stages, along with a thorough review. FA was responsible for data collection and analysis. FA also played a role in editing the first draft and contributed to the final review and editing. FK conducted formal analyses and provided input in the final draft’s editing and review. NA managed data curation, as well as drafting and editing the manuscript.
Financial support
None.
Competing interests
None.
Ethical standards
The study was conducted in compliance with the Good Clinical Practices protocol and Declaration of Helsinki principles. This was an observational study, and the collected data was de-identified by eliminating all individual patient identifiers. A waiver of consent was granted by the ethical review board of the Aga Khan University (ERC#2020-5715-14993).
Data sharing statement
De-identified data may be shared upon reasonable requests to the corresponding author.