COVID-19 is usually mild in pediatric patients (2). Children presenting with severe shock-like syndrome, incomplete Kawasaki disease (KD), or toxic shock syndrome (3) were reported in the United Kingdom and Italy in April 2020. Subsequently, children with a similar clinical presentation were also reported in the rest of Europe, America, and South Africa (4). The entity was identified as multisystem inflammatory syndrome in children (MIS-C) associated with SARS-CoV-2 infection with a spectrum of manifestations, such as KD, toxic shock syndrome, sepsis, and macrophage activation syndrome.

The World Health Organization (WHO) (5), the Center for Disease Prevention and Control (CDC) (6), and the Royal College of Pediatrics and Child Health of the United Kingdom (RCPCH) (7) issued definitions for case identification. Systematic reviews based on case reports, case series, and observational studies have been published; however, these include children from different backgrounds and ethnicities, mainly from Europe and North America. MIS-C appears to vary between regions, with few cases reported in children from Asia (8). Poor accessibility to the health system and delay in diagnosis and treatment in countries with limited resources may lead to a poor prognosis in children with MIS-C in Latin America and the Caribbean region (LAC). Some research has been conducted in LAC, but the topic is largely unexplored. Existing data come from specific collaborative networks of intensive therapists, cardiologists, rheumatologists, and pediatric infectious disease practitioners in the region, but these findings have not been routinely collected or analyzed. To our knowledge, there are no published systematic reviews that include surveillance data and epidemiological records about MIS-C in Latin American children.

The current study aimed to describe the clinical course, laboratory findings, epidemiology, treatment, and use of resources of MIS-C in children from LAC through a systematic review and a meta-analysis incorporating data from regional surveillance systems.

For this systematic review and meta-analysis, we followed the Cochrane methods and the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (9) statement for reporting results. The protocol for the present systematic review was registered (10) in the University of York’s PROSPERO database (CRD42021242505).

Risk of bias was assessed by two reviewers. Disagreements were resolved by consensus of the entire team. If consensus could not be reached, the conflict was resolved by a third reviewer.

For cohort and cross-sectional studies, the NIH instrument was used to score 14 items. Six studies were assessed as fair quality and three as good quality, with a score as having a moderate-to-low risk of bias. The best scores were for research questions and population selection. On the other hand, the included case series had a quality assessment as poor in five, fair in seven and good quality in one, with a high risk of bias. The only included case-control study received a fair quality score with a moderate-to-high risk of bias. In summary, most of the included studies were rated as being of low-to-moderate methodological quality with a moderate-to-high risk of bias. Risk of bias is shown in Supplementary Tables 2–4.

We applied an arc-sine transformation to stabilize the variance of proportions following the Freeman-Tukey variant of the arc-sine square root of transformed proportions method (11).

The search strategy yielded 464 potentially relevant studies in the databases. Figure 1 shows a diagram of the study selection process. Twenty-three full-text studies that met the inclusion criteria for data synthesis were included. All were observational studies with 13 case series, seven cohort studies, two cross-sectional studies, and one case-control study. In addition, surveillance system reports were included for separate analysis. The most frequent reasons for exclusion were wrong study design (N = 15), wrong country (N = 7), wrong outcomes (N = 6), adult population (N = 3), and three studies were left out because they reported fewer than three cases. The included studies collected data from 592 patients from Peru, Brazil, Argentina, Chile, Costa Rica, Colombia, Mexico, Venezuela, and other countries. The main characteristics of clinical studies of MIS-C included are described in Table 1. The mean age reported was 6.6 years (IQR, 6–7.4 years); 355 (60%) were male (95% CI, 55–65). A list of excluded studies during the full text assessment can be found in Supplementary Table 5 (65–67).

In our systematic review of MIS-C in LAC and analysis of data from regional surveillance systems performed we found significant heterogeneity. Most studies were case series. Countries included in our study were Argentina, Brazil, Chile, Colombia, Costa Rica, Mexico, Peru, and Venezuela. Health-Ministry data on MIS-C are very patchy and even the PAHO reports are deficient because limited information is available for Latin America and the Caribbean countries, related probably to the lack of mandatory notification of MIS-C in some countries.

Data from the epidemiological surveillance system show that during 2020, the highest number of MIS-C cases was observed around 3 weeks after the highest number of COVID-19 cases in children and adolescents in Chile. A 2–5-week time lag has been observed between the peak of COVID-19 cases within communities and the increase in MIS-C cases. This suggests that acquired immunity may play a role in its development.

Likewise, it has been observed that the number of MIS-C cases reported by PAHO is lower than that reported by the epidemiological reports of the countries within similar temporal parameters. Nevertheless, this may be related to under-registration of cases due to differences in epidemiological records. Also, the reporting of data from the central level may not reflect the behavior of the surveillance system at the subnational level since delays in reports could occur due to the registration of a large number of suspected cases of COVID-19. MIS-C case fatality rate for the same period was lower in Chile than in Brazil according to country surveillance reports. Chilean MIS-C cases and mortality reported by PAHO were slightly lower than those recorded by the country’s surveillance system for the same period. The pooled case fatality rate from 23 reports of 4% was slightly higher than the 1–2% overall reported rate from other regions (44).

Similar to other regions, in LAC countries more than half the patients were male (45). The median age of the included patients was 6.6 years, slightly lower than 8–9 years described in other populations. Apparently, the affected population in LAC is younger than in other regions, such as Europe and North America. It would be interesting to investigate possible causes, such as sociodemographic, ethnic, or other variables that could determine this difference (44, 46–49).

In general terms, the distribution of cases by sex was similar with a frequency of male sex between 51 and 58%. Gender was not related to severity of the clinical course; however, we observed that 73% of patients requiring PICU admission were male. In adults, male sex is a risk factor for severe COVID-19. Specifically, in a Latin-American study including data from patients with COVID-19 and MIS-C, girls had a lower frequency of hospitalization without differences in mortality among patients who developed MIS-C (44, 46–50).

In 59% of patients with MIS-C, close contact with a person with COVID-19 was described. This finding was more frequent than in other reviews reporting only 15–20% (44). The high percentage of household infections in Latin-America may be related to overcrowded housing conditions and other socio-demographic factors. In most epidemiological surveillance registries from Chile and other Latin American countries, most cases were confirmed by clinical or epidemiological criteria with a low frequency of concomitant positive SARS-CoV-2 tests.

Around one-third of the patients had one or more comorbidities. This has also been described in other reviews, with a frequency between 20 and 25%. The most observed comorbidities were obesity and/or overweight, chronic respiratory diseases, onco-hematological diseases, and neurological disorders (64). This finding confirms previous reports showing that MIS-C predominantly affects healthy individuals (44, 49, 51, 52). In most reviews, MIS-C diagnosis was made by serology in 60% to almost 90% of cases, while direct diagnostic methods such as PCR for SARS-CoV-2, were positive in swabs in less than of 40% of cases, similar to our findings. This would support the fact that MIS-C occurs as a post-infection phenomena (47, 48, 51, 53, 54).

Regarding the presenting signs and symptoms, fever was the most frequent followed by gastrointestinal manifestations, rash, and conjunctival injection. Neurological, respiratory, and cardiological manifestations were also reported. In most reports, fever and abdominal or gastrointestinal symptoms (abdominal pain, diarrhea, and vomiting) are described as the most frequent (46, 48, 53). On the other hand, the presenting symptoms of MIS-C may mimic an acute abdomen, including acute appendicitis, as shown in a multicenter Latin-American study, in which 34/1,010 (3.3%) patients with COVID-19 and MIS-C had an intraoperative diagnosis of appendicitis (55). Furthermore, a systematic review of abdominal pain in MIS-C found an incidence of acute abdomen of 19%, which was non-surgical in most cases. Patients who required surgery due to appendicitis or obstruction were 17/72 (23.6%) patients with acute abdomen (55, 56).

Mucocutaneous manifestations, such as rash were reported between 40 and 60%; conjunctivitis in 50%, similar to our study (52). Neurological involvement was less frequent, with headache, neck stiffness, and visual disturbances reported in 30% of cases (50). MIS-C patients generally have less respiratory symptoms than COVID-19 as reported in large series and reviews. In a systematic review, around 50% of the patients showed respiratory symptoms, including cough or dyspnea with radiological findings (51–54). A variable proportion of patients with MIS-C develop hypotension and shock because of acute myocardial dysfunction or hyperinflammatory state. In our study, hemodynamic involvement was less frequent compared to other reviews (48, 54, 57).

In the current review, 59% of MIS-C cases had Kawasaki-like syndrome. Kaushik et al. reported cardiac involvement in MIS-C patients with Kawasaki-like syndrome, of whom 20% had coronary artery dilatation/aneurysms and hypotension and 28% shock, like our findings. Despite some similar phenotypic characteristics between MIS-C and KD, there are other differential parameters, such as age less than 5 years in children with KD and the fact that approximately 7% of KD patients present with cardiovascular collapse (KDSS). In our study, 49% of the patients evolved to shock (58).

Inflammatory markers (CRP, D-dimer, and ferritin) were elevated and so was the incidence of lymphopenia. Feldstein et al. reported similar findings in patients with MIS-C compared to those with severe acute COVID-19 (8). In addition, in patients with MIS-C who required PICU admission we observed elevated inflammatory markers and lower lymphocyte and platelet counts compared to patients with MIS-C in general. These findings are consistent with other systematic reviews (53).

When comparing the overall analysis of patients included in our study with that of the subgroup of those with severe MIS-C, an increased male predominance (71% vs. 59%), a higher frequency of comorbidities, higher CRP levels, and lower lymphopenia counts are found. Different series of patients with severe MIS-C reported similar features (59, 60). Tripathi et al. reported a comorbidity rate (33%) similar to our study (61). Mortality was higher when compared to that of developed countries (7% vs. 3–4.7%), but lower than that observed in the Colombian MISCO study (9%) (61). It is likely that the delay in access to health care, diagnosis, and patient care explains, at least in part, this higher mortality and heterogeneity between countries in the region. In a report by Farias et al. a high frequency of pre-existing diseases and immunosuppression was found, which may have contributed to the high mortality rate.

IVIG and/or steroids are proposed as first-line treatment in patients with MIS-C, although to date there are no controlled studies comparing IVIG and corticosteroids alone or in combination. Some observational studies have found combination therapy to be beneficial. A French study including 111 patients with MIS-C found a lower rate of treatment failure in those who received combination therapy vs. IVIG as monotherapy (62). A US study in 518 MIS-C patients observed that combination therapy was associated with a lower risk of new or persistent cardiovascular dysfunction compared to IVIG alone. Both studies found a lower requirement for second-line therapy and hemodynamic support within 24–48 h after initial treatment (63). On the other hand, a large study carried out in multiple countries comparing three treatment regimens (combination therapy, IVIG, or corticosteroids alone) found no differences in morbidity and mortality. Nevertheless, the selection and outcome criteria were different in the three studies (45). In our study, most patients received IVIG, in combination with corticosteroids in more than half of the patients.

In agreement with a study by Pereira et al. (35) we observed an increased use of antibiotics, related to the main differential diagnoses of MIS-C, septic shock, and toxic shock syndrome. Antibiotic prescriptions among MIS-C patients may not have been appropriate in most cases and antimicrobial stewardship should be explored in MIS-C patients. Tocilizumab or siltuximab were used only in a small number of patients.

Other frequently used medications as part of treatment were aspirin and anticoagulants. The wide variety of treatments used may have been due to the lack of knowledge of MIS-C, especially in the first months after described in Europe. In addition, many treatments are unavailable in the region due to their high costs. Most importantly, there is still no clear consensus on the treatments to be administered beyond the recommended use of IVIG and corticosteroids.

PICU admissions were less frequent than in other reports (44, 48). One possible explanation is that at the moment that MIS-C started to be reported, no guidelines or alerts to detect and treat children with MIS-C were available. Use of resources in terms of length of hospital and PICU stay were like other reports (48).

Information available from LAC was found to be limited, probably related to the lack of mandatory notification of MIS-C locally. One of the key limitations of the study is that most of the data reported were from case series and were highly heterogeneous. To our knowledge this is the first systematic review with meta-analysis incorporating data from regional surveillance systems from LAC.

The results of the present study provide important information to help understanding MIS -C in children and adolescents from LAC. Epidemiological information in the region was very limited. Optimization of case registration and surveillance is needed. Knowledge about this syndrome and the impact of the recent introduction of COVID-19 vaccination schedules in several Latin American countries is still insufficient.

SRu and AB conceived the study. CV, MR, AF, NV, MB, SRo, and RU-G collected and analyzed the studies included. SRu, CV, MR, AF, NV, MB, SRo, RU-G, and AB drafted the manuscript. All authors contributed and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.