ABSTRACT
Aim
During the COVID‐19 pandemic, Sweden chose a more open public health approach. A general concern was that this approach would affect the population negatively. We describe the Swedish national MIS‐C cohort and risk factors for severe disease defined as admission to the intensive care unit and cardiogenic shock.
Methods
Nationwide population‐based cross‐sectional study. All children and adolescents from 0 to 19 years of age with MIS‐C from March 2020 to August 2022 were included. Multivariable logistic regression with ICU admission and cardiogenic shock as outcomes was used to evaluate risk factors for severe disease.
Results
A total of 338 MIS‐C cases were identified; 69 patients (20%) required intensive care and 38 patients (11%) developed cardiogenic shock. Clinical findings associated with ICU admission were cardiac involvement: OR 14.00 (6.28–34.70), airway involvement: OR 6.75 (3.23–14.80), hyperinflammation: OR 3.25 (1.34–7.97), oedema: OR 2.40 (1.07–5.44) and diarrhoea: OR 2.23 (1.07–4.80). Clinical findings associated with cardiogenic shock were hyperinflammation: OR 5.10 (2.16–12.00), airway involvement: OR 4.87 (2.12–12.00), higher age (13–19 vs. 0–5 years): OR 3.85 (1.17–15.60), obesity: OR 3.55 (1.14–10.5) and diarrhoea: OR 2.48 (1.09–5.92).
Conclusion
In Swedish national MIS‐C cohort, the rate of ICU admission was low and outcome excellent. Identifying risk factors for ICU admission and cardiogenic shock may facilitate early detection of risk patients.
Keywords: COVID‐19, MIS‐C, risk factors, severity, Sweden
Abbreviations
- CRP
C‐reactive protein
- ECMO
extracorporeal membrane oxygenation
- ICD10
International Classification of Diseases
- ICU
intensive care unit
- LMWH
low‐molecular‐weight heparin
- MAS
Macrophage Activation Syndrome
- MIS‐C
multisystem inflammatory syndrome in children
- NT‐proBNP
N‐terminal pro‐B‐type natriuretic peptide
- pedSRQ
Swedish National Quality Registry for Paediatric Rheumatology
- SAR‐CoV‐2
severe acute respiratory syndrome coronavirus 2
Summary.
This nationwide Swedish cohort study identified clinical and laboratory predictors of severe MIS‐C, defined as ICU admission and cardiogenic shock.
Despite a non‐restrictive public health approach during the pandemic, Sweden had low ICU admission rates and excellent outcomes among children with MIS‐C, without having investigated whether the strategy affected these outcomes.
Early identification of high‐risk patients may support timely and aggressive treatment strategies to prevent critical illness.
1. Introduction
Multisystem Inflammatory Syndrome in Children (MIS‐C) is a severe hyperinflammatory complication of COVID‐19 [1]. Though relatively rare, MIS‐C can lead to severe morbidity and even mortality [1]. Between 32% and 68% of MIS‐C cases were reported in various studies to require intensive care unit (ICU) admission [2, 3, 4, 5], with cardiogenic shock described to affect between 50% and 80% of cases [6]. The mortality rate for MIS‐C has been estimated to 1%–2% [1, 7, 8], although higher mortality rates have been reported [9, 10, 11]. The risk factors associated with admission to ICU for patients with MIS‐C have been the subject of various studies [5, 9, 11, 12, 13].
The reported incidence of MIS‐C has varied by country and over time [12, 13]. In Sweden, Rhedin et al. reported an incidence of 6.8 per 100,000 children from March 2020 to December 2021 [14]. In a recent publication, Eleftheriou et al. found a decreasing MIS‐C incidence over the three pandemic waves, possibly due to increased immunity, vaccination and milder SARS‐CoV‐2 variants [15].
Sweden provides a unique setting for studies of COVID and especially MIS‐C because of the country's open public strategy during the pandemic, its specific health care system properties, as well as the aggressive treatment strategy for MIS‐C that was generally adopted all over Sweden. Numerous studies have explored MIS‐C, but to our knowledge there is only one prior national prospective cohort study, conducted by Nygaard et al. [16] Here we present our nationwide population‐based Swedish study with a focus on patient characteristics, acute disease presentation, and risk factors associated with ICU admission and cardiogenic shock.
2. Patients and Methods
2.1. Study Design and Setting
This study employed a prospective, population‐based cohort design, including all Swedish children aged 0–19 years diagnosed with MIS‐C who were enrolled in the Swedish National Quality Registry for Paediatric Rheumatology (pedSRQ) (https://barnreumaregistret.se/) from March 1, 2020 to August 31, 2022. Patients who fulfilled World Health Organization (WHO) criteria for MIS‐C were classified according to International Classification of Diseases (ICD10) code U10.9 [17]. The study was approved by the Swedish Ethical Review Authority (DNR 2020–06175).
2.2. Study Population
A total of 366 patients were registered in pedSRQ. Patients were included in the study if they were admitted to a paediatric ward and registered as having MIS‐C in pedSRQ, were under 20 years old at diagnosis, and gave patient or parental consent for sharing pseudonymized patient data for research purposes (n = 356). Patients not meeting World Health Organization (WHO) diagnostic criteria [17] or who were later diagnosed with alternate conditions were excluded (n = 18). Most cases were discussed and re‐evaluated in our bi‐weekly national web‐based meetings to improve our expertise and to minimise risks of bias in the register.
2.3. Data Sources and Variables
The pedSRQ is an electronic registry that collects prospective data on children with paediatric rheumatic diseases in Sweden since 2009. In 2020, a custom interface was designed to make it possible to enter MIS‐C cases into the registry. Data extracted from the registry included demographic information, date of MIS‐C diagnosis, patient characteristics, clinical symptoms during the course of MIS‐C, complications/classification of MIS‐C, level and length of care, therapies and outcome.
Weight categories were calculated via the ISO BMI, which considers age and gender and categorises weight status as: underweight (ISO BMI < 18.5), normal weight (ISO BMI 18.5–24.9), overweight (ISO BMI 25–29.9), or obese (ISO BMI > 30) [18, 19]. We did not separate underweight from normal weight, thus defining normal weight as ISO BMI < 25.
Mucocutaneous involvement was identified by at least one sign such as changes in mucous membranes, non‐purulent conjunctivitis, redness of the lips, or a skin rash. Cardiac involvement was assessed by ECG and echocardiogram and characterized by conditions such as arrhythmia, pericarditis, myocarditis, and/or reduced ejection fraction. Manifestations of airway involvement included a sore throat, cough, and/or breathing difficulties. CNS involvement was defined by neurological symptoms including irritability, seizures, and/or encephalopathy. Lastly, we created the co‐variable hyperinflammation by modifying the 2016 Classification Criteria for Macrophage Activation Syndrome (MAS) associated with Systemic Juvenile Idiopathic Arthritis, by Ravelli et al. [20] Our definition included the triad of serum C‐reactive protein (CRP) > 200 mg/L, platelet count < 110 × 109/L and raised liver transaminases. Serum ferritin levels were not included in the definition, as the threshold in our dataset (2000 mg/L) was significantly higher than the 684 mg/L included in the Ravelli criteria.
Outcome variables included ICU admission confirmed by either the documented date of admission to the ICU or the initiation of ECMO therapy. Cardiogenic shock was defined by the requirement for inotropic support. Our study divided the MIS‐C cases according to COVID‐19 variants (Alpha, Delta and Omicron) and also whether patients were admitted to a general paediatric ward or to an ICU.
2.4. Statistical Analysis
Continuous variables were presented using median and range while categorical variables were described using count and percentages. Univariable analyses were conducted using Wilcoxon signed rank test for continuous variables and Chi‐square or Fisher's exact tests for categorical variables. For analysis of ICU admission rates over time, we used Pearson's chi squared test to describe differences in distribution. Multivariable logistic regression models were employed to assess the adjusted impact of baseline variables and patients' characteristics on ICU admission and shock. Variables included in the models were sex, age groups, weight groups, diarrhoea, oedema, hyperinflammation, airway affection, cardiac affection and mucocutaneous involvement. Sex, age, weight and symptom profile were considered possible risk factors based on clinical experience. Results were presented as Odds Ratios (ORs) with 95% confidence intervals (95% CI). The proportion of MIS‐C patient admitted to ICU/PICU during different time‐period was analysed with Chi‐square test. All analyses were performed using R statistical software version 4.2.2 [21].
3. Results
3.1. Most Patients Were Treated on General Paediatric Ward
In total, 338 patients diagnosed with MIS‐C fulfilled the inclusion criteria. Patient characteristics are presented in Table 1. There was a dominance of boys versus girls (63% vs. 37%). The median age at diagnosis was 9.3 years (range 0.3–19). The age distribution was 0–6 years (n = 92; 27%), 7–12 years (n = 138; 41%) and 13–19 years (n = 108; 32%). The median duration from first symptom(s) until diagnosis was 5 days. Fourteen percent of the children were overweight, and 8% were obese.
TABLE 1.
Baseline characteristics of 338 MIS‐C patients, stratified by ICU admission.
| All patients; n (%) | ICU; n (%) | Non‐ICU; n (%) | p | |
|---|---|---|---|---|
| Patients; n (%) | 338 | 69 (20) | 269 (80) | |
| Sex; n (%) | ||||
| Male | 212 (63) | 39 (56) | 173 (64) | 0.233 |
| Female | 126 (37) | 30 (44) | 96 (36) | |
| Age at diagnosis; median years (range) | 9.3 (0.3–19) | 10.9 (0.3–19) | 8.90 (0.6–18) | 0.062 |
| Age groups; n (%) | ||||
| 0–6 years | 92 (27) | 17 (25) | 75 (28) | 0.360 |
| 7–12 years | 138 (41) | 25 (36) | 113 (42) | |
| 13–19 years | 108 (32) | 27 (39) | 81 (30) | |
| Days from symptom onset to diagnosis | ||||
| > 6 days | 79 (23) | 15 (22) | 64 (24) | 0.870 |
| Comorbidities; n (%) | ||||
| Any (except obesity and overweight) a | 57 (17) | 13 (19) | 44 (16) | 0.623 |
| Overweight | 46 (14) | 15 (22) | 31 (11.5) | |
| Obesity | 28/337 (8) | 9/68 (13) | 19/268 (7) | 0.110 |
| Overweight/Obese | 74/337 (22) | 24/68 (35) | 50/268 (19) | 0.003 |
| SARS‐CoV‐2; n (%) | ||||
| PCR | 119 (35) | 22 (32) | 97 (36) | 0.279 |
| Serology | 290 (86) | 62 (90) | 228 (85) | 0.279 |
| PCR and/or serology | 303 (90) | 66 (96) | 237 (88) | 0.066 |
| Outcome | ||||
| Outpatients | 6 (1.8) | 0 | 6 (2.2) | |
| Days at ward (patients) b ; median (IQR) | (332); 9 (1–48) | 10 (2–48) | 263/269, (6 outpatients); 9 (1–38) | |
| Days in ICU; median (range) | NA | 3 (0–18) c | NA | |
| Days in ECMO; median (range) | NA | 6 (5–8) | NA | |
| Case‐fatalities; number (%) | 0 (0) | 0 (0) | 0 (0) | |
Abbreviations: ECMO, extracorporal membrane oxygenation; ICU, intensive care unit.
Any comorbidity (Attention‐Deficit/Hyperactivity Disorder (ADHD), asthma bronchiale, eczema, type 1 diabetes, congenital heart disease).
Days at ward exclude outpatients.
Zero days refers to patients who were in ICU < 24 h.
For the majority of patients (n = 263, 78%), the paediatric ward was the highest level of care, whereas 6 patients fulfilled the inclusion criteria but did not require any inpatient care. The median hospital stay was 9 days (range 1–48 days). Of the 338 patients, 20% were admitted to ICU; the majority of these were admitted to the ICU within the first or following day of hospital admission. Three patients required ECMO care. No mortality was observed. The median stay in ICU was 3 days in addition to the number of days at the general paediatric ward (range 0–18 days in ICU; 0 days refers to patients who were in ICU < 24 h), and 6 days for patients on ECMO (range 5–8 days). All patients could be discharged home (Table 1).
3.2. Differences in Symptoms Between ICU and Non‐ICU Group
Changes in mucous membranes represented 12% in non‐ICU versus 3% in the ICU group (p = 0.023) (Table 2). The following findings were statistically more common among patients treated in the ICU versus those not in the ICU: arrhythmia, pericarditis, myocarditis, reduced ejection fraction, sore throat, cough, shortness of breath, irritability, lethargy, diarrhoea, oedema (Table 2).
TABLE 2.
Clinical characteristics and laboratory findings of 338 MIS‐C patients, by ICU admission.
| Clinical characteristics, n (%) | All patients | ICU | Non‐ICU | p |
|---|---|---|---|---|
| n = 338 | n = 69 (20) | n = 269 (80) | ||
| Clinical symptoms | ||||
| Mucocutaneous involvement | 257 (76.0) | 49 (69.6) | 209 (77.7) | |
| Changes in mucous membranes | 35 (10) | 2 (3) | 33 (12) | 0.023 |
| Non‐purulent conjunctivitis | 182 (54) | 38 (55) | 144 (54) | 0.819 |
| Redness of the lips | 62 (18) | 13 (19) | 49 (18) | 0.905 |
| Skin rash | 202 (60) | 36 (51) | 166 (62) | 0.150 |
| Peeling skin | 22 (6) | 5 (7) | 17 (6) | 0.781 |
| Cardiac involvement | 132 (39.1) | 58 (84.1) | 74 (27.5) | |
| (UCG perf) | 332/335 (99) | 69 (99) | 264/266 (99) | 0.584 |
| Arrythmia | 25/331 (8) | 13/68 (19) | 12/263 (5) | < 0.001 |
| Pericarditis | 38/327 (12) | 19/65 (29) | 19/262 (7) | < 0.001 |
| Myocarditis | 71/322 (22) | 38/64 (59) | 33/258 (13) | < 0.001 |
| Reduced EF | 83/330 (25) | 51/67 (76) | 32/263 (12) | < 0.001 |
| Coronary involvement | 48/328 (15) | 11/65 (17) | 37/262 (14) | 0.568 |
| Airway involvement | 126 (37.3) | 51 (73.9) | 75 (27.9) | |
| Sore throat | 82 (24) | 30 (43) | 52 (19) | < 0.001 |
| Cough | 48 (14) | 18 (26) | 29 (11) | 0.001 |
| Shortness of breath | 39 (12) | 23 (34) | 16 (6) | 0.001 |
| CNS involvement | 18 (5.3) | 5 (7.2) | 13 (4.8) | |
| Irritability | 57 (17) | 21 (30) | 36 (13) | < 0.001 |
| Seizures | 3 (1) | 1 (1) | 1 (1) | 0.497 |
| Encephalopathy | 18 (5) | 4 (6) | 13 (5) | 0.744 |
| Headache | 128 (38) | 28 (41) | 100 (37) | 0.603 |
| Lethargy | 102 (30) | 36 (52) | 66 (24) | < 0.001 |
| Other | ||||
| Vomiting | 153 (45) | 34 (49) | 119 (44) | 0.453 |
| Abdominal pain | 219 (65) | 49 (71) | 170 (63) | 0.225 |
| Diarrhoea | 171 (51) | 49 (71) | 122 (45) | < 0.001 |
| Splenomegaly and/or hepatomegaly | 81/263 (31) | 24/59 (41) | 57/204 (28) | 0.062 |
| Lymphadenopathy | 80 (24) | 16 (23) | 64 (24) | 0.916 |
| Edema | 73 (22) | 27 (39) | 46 (17) | < 0.001 |
| Arthralgia | 26 (8) | 2 (3) | 24 (9) | 0.094 |
| Laboratory; median (range) | ||||
| CRP (mg/L) | (28–600) n = 338 | (55–600) n = 69 | (28–520) n = 269 | < 0.001 |
| ESR (mm) | 52 (2–129) n = 299 | 61 (6–115) n = 57 | 51 (2–129) n = 242 | 0.170 |
| Albumin (g/L) | 23 (13–42) n = 334 | 20 (13–38) n = 68 | 24 (14–42) n = 266 | < 0.001 |
| NT‐pro‐BNP (ng/L) | 4169 (45–8 4027) n = 333 | 13 676 (713–8 4027) n = 68 | 3385 (45–3 5500) n = 265 | < 0.001 |
| Troponin T (ng/L) | 21 (2–2 5500) n = 308 | 81 (4–2 5500) n = 66 | 17 (2–2500) n = 242 | < 0.001 |
| Laboratory; n abnormal (%) | ||||
| Hyperinflammation a | 230 (68.0) | 60 (87.0) | 170 (63.2) | |
| Hb (< 92 g/L) | 105/336 (31) | 28/68 (41) | 77/268 (29) | 0.048 |
| Leukopenia (WBC < 5 × 109/L) | 166/334 (50) | 31 (45) | 135/265 (51) | 0.373 |
| Lymphocytes < 1.5 × 109/L | 269/335 (80) | 60 (87) | 209/266 (79) | 0.119 |
| Platelets (< 110 × 109/L) | 143/333 (43) | 41 (59) | 102/264 (39) | 0.002 |
| Bi‐cytopenia b | 145/333 (43) | 38 (55) | 107/264 (41) | 0.030 |
| Ferritin (> 2000 μg/L) | 98/329 (30) | 24 (35) | 74/260 (29) | 0.307 |
| IL‐6 increased c | 75/180 (42) | 27/45 (60) | 48/135 (36) | 0.004 |
| sIL2R increased d | 69/78 (89) | 24/26 (92) | 45/52 (86) | 0.710 |
| Liver transferases (AST, ALT) increased e | 161/333 (48) | 43 (62) | 118/264 (45) | 0.009 |
Note: Underlined symptoms are described in definition of Mucocutaneous, Cardiac, Airway and CNS involvement. / Number of patients with symptom versus (/) total number of patients registered with symptom.
Hyperinflammation defined as a triad of serum C‐reactive protein (CRP) > 200 mg/L, platelet count < 110 × 109/L and raised liver transaminases.
Bi‐cytopenia defined as leukopenia (WBC < 5 × 109/L) and thrombocytopenia (< 110 × 109/L).
IL‐6 increased (interleukin‐6 increased over normal value).
sIL2R increased (soluble interleukin‐2 receptor increased over normal value).
AST (age 6–12 months girls/boys, 0.55–1.1 μkat/L; 1–4 years girls/boys, 0.41–0.93; 5–8 years girls/boys 0.40–0.80; 9–18 years girls/boys, 0.28–0.72), ALAT (6 months −8 years, girls 0.13–0.39 μkat/L; 6 months–8 years, boys 0.12–0.29; 9–18 years girls/boys 0.17–0.51).
3.3. Increased Inflammatory Response in Patients Admitted to ICU
Patients admitted to ICU had higher CRP, lower albumin, and higher levels of NT‐Pro‐BNP and troponin, as well as more frequent anaemia, thrombocytopenia, bicytopenia, increased IL‐6, and increased liver enzymes than patients treated at general paediatric wards. NT‐pro‐BNP and troponin were often elevated and sometimes to very high levels (NT‐pro‐BNP median 4169, range 45–84 027 ng/L and troponin median 21, range 2–25 000 ng/L). Thrombocytopenia or bicytopenia was seen in almost half of the patients at presentation. The majority of patients were lymphopenic (80%). Approximately one third of cases had high serum ferritin (> 2000 μg/L). In those analyzed, sIL2R levels were increased in almost all patients, and IL‐6 was increased in almost half of the patients. Increased liver enzymes were present in around half of the cases (Table 2).
3.4. Majority of Patients Were Treated With Combination Therapy
The combination of intravenous immunoglobulin and glucocorticosteroids was given to the vast majority (88%). Intravenous administration of short‐acting IL‐1 blockade (anakinra) was given to 41% and anti‐IL‐6 (tocilizumab) was given to 1% of patients. The anakinra dosage was 5–10 mg/kg intravenously divided into four daily doses. The ICU group had a higher usage of anakinra and IVIG + corticosteroids + anakinra in comparison to the non‐ICU group. Both patients given anti‐IL‐6 therapy were admitted to the ICU and had failed therapy with IVIG + corticosteroids + anakinra. The six out‐patients did not receive any treatment and resolved their symptoms spontaneously.
Inotropic support for treatment of cardiogenic shock was used in 11% of patients. Five patients needed dialysis. Three patients were treated in ECMO.
Supportive therapy with intravenous albumin was given to a third of patients. Albumin supplementation and low‐molecular‐weight heparin (LMWH) in treatment doses were more common among ICU patients than in patients treated at general paediatric wards (Table 3).
TABLE 3.
Immunomodulatory and supportive treatments in 338 MIS‐C patients, by ICU admission.
| Characteristic; n (%) | All patients | ICU | Non‐ICU | p |
|---|---|---|---|---|
| n = 338 | n = 69 (21) | n = 269 (89) | ||
| Immunomodulatory treatment, n (%) | ||||
| None | 16 (5) | 2 (3) | 14 (5) | 0.541 |
| IVIG | 310 (92) | 65 (94) | 245 (91) | 0.401 |
| Steroids | 308 (91) | 66 (96) | 242 (90) | 0.138 |
| IL1 Inhibition (anakinra) | 137 (41) | 50 (73) | 87 (32) | < 0.001 |
| IL6 inhibition (tocilizumab) | 2 (1) | 2 (3) | 0 (0) | 0.005 |
| IVIG + steroids | 296 (88) | 64 (93) | 232 (86) | 0.144 |
| IVIG + steroids + IL1 inhibition | 135 (40) | 49 (71) | 86 (32) | < 0.001 |
| IVIG + steroids + IL1‐IL 6 inhibitions | 2 (1) | 2 (3) | 0 (0) | 0.005 |
| ICU procedures, n (%) | ||||
| Inotropic agents | 38 (11) | 38 (55) | 0 (0) | < 0.001 |
| Dialysis | 5 (1) | 1 (1) | 4 (1) | 1.000 |
| ECMO | 3 (1) | 3 (4) | 0 (0) | NA |
| Supportive treatments, n (%) | ||||
| Albumin infusion | 99 (29) | 29 (42) | 70 (26) | 0.009 |
| Acetylsalicylic acid, low dose | 266 (79) | 50 (73) | 216 (80) | 0.156 |
| LMWH, prophylactic dose | 158 (47) | 36 (52) | 122 (45) | 0.311 |
| LMWH, therapeutic dose | 95 (28) | 29 (42) | 66 (25) | 0.004 |
Note: Combination therapy: IVIG + steroids; IVIG + steroids + IL 1 inhibition; IVIG + steroids + IL1‐IL 6 inhibitions.
Abbreviations: IL, interleukin; IVIG, intravenous immunoglobulin; LMWH, low molecular weight heparin.
3.5. Risk Factors for Admission to the ICU
There was no statistically significant variation in the rate of ICU admissions for patients with MIS‐C over time during different COVID‐19 variants (p = 0.90) (Figure S1).
Risk factors for admission to ICU and/or ECMO were evaluated in logistic regression models. In crude analyses, overweight as well as obesity compared to normal weight was associated with a more than doubled risk of admission to ICU/ECMO. Diarrhoea or oedema as presenting symptoms were associated with an almost threefold increased risk for ICU admission. Hyperinflammation and airway symptoms were associated with a sevenfold increase, and cardiac affection with a 16‐fold increased risk for ICU admission as compared to patients without these symptoms. CNS and mucocutaneous symptoms were not statistically associated with admission to ICU/ECMO and were omitted from the multivariate analysis. In the adjusted model, the association of overweight and obesity with admission to ICU/ECMO was no longer statistically significant whereas the associations of the other risk factors were not significantly changed (Table 4).
TABLE 4.
Crude and adjusted odds ratios for ICU admission or ECMO among 326 MIS‐C patients.
| Risk factor | N (%) | Crude | Adjusted model a | ||
|---|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | ||
| Sex | |||||
| Male | 122 (37.4) | 1.00 | Ref | 1.00 | Ref |
| Female | 204 (62.6) | 0.79 | 0.46–1.38 | 0.91 | 0.43–1.96 |
| Age group | |||||
| 0–6 | 90 (27.6) | 1.00 | Ref | 1.00 | Ref |
| 7–12 | 136 (41.7) | 1.04 | 0.52–2.12 | 0.84 | 0.33–2.19 |
| 13–19 | 100 (30.7) | 1.63 | 0.81–3.33 | 0.93 | 0.35–2.48 |
| Weight group | |||||
| Normal | 254 (77.9) | 1.00 | Ref | 1.00 | Ref |
| Overweight | 45 (13.8) | 2.45 | 1.19–4.90 | 1.05 | 0.39–2.73 |
| Obese | 27 (8.3) | 2.45 | 0.99–5.71 | 1.64 | 0.51–5.18 |
| Symptom b | |||||
| Diarrhoea | 163 (50.0) | 2.90 | 1.65–5.26 | 2.23 | 1.07–4.80 |
| Edema | 72 (22.1) | 2.94 | 1.63–5.27 | 2.40 | 1.07–5.44 |
| Hyperinflammation | 223 (68.4) | 7.00 | 3.51–14.2 | 3.25 | 1.34–7.97 |
| Airway affection | 121 (37.1) | 7.07 | 3.93–13.2 | 6.75 | 3.23–14.8 |
| Cardiac affection | 131 (40.2) | 16.4 | 8.10–37.1 | 14.0 | 6.28–34.7 |
| CNS | 16 (4.9) | 1.82 | 0.56–5.20 | Omitted | |
| Mucocutaneous | 247 (75.8) | 0.63 | 0.35–1.16 | Omitted | |
Model adjusted for sex, age group, weight group, diarrhoea, edema, hyperinflammation, airway affection and cardiac affection.
In the symptom variables, the reference is absence of specific symptom.
3.6. Risk Factors for Cardiogenic Shock
Risk factors for cardiogenic shock were evaluated in logistic regression models. In crude analyses, being 13–19 years old as compared to 0–6 years old was associated with an almost fivefold increased risk of cardiogenic shock. Being obese as compared to normal weight was associated with a threefold increased risk of cardiogenic shock. Among the symptoms, diarrhoea was associated with an almost threefold increased risk, hyperinflammation with an almost eightfold increased risk, and airway affection with an almost sixfold increased risk of cardiogenic shock as compared to the absence of these respective symptoms. Cardiac affection was associated with a 23‐fold increased risk for cardiogenic shock. Cardiac affection was considered to be a precursor to cardiogenic shock instead of a confounder and was therefore omitted from the multivariate model. Mucocutaneous symptoms were associated with a reduced risk of cardiogenic shock. Oedema and CNS affection were not statistically associated with cardiogenic shock and were omitted from the multivariate model. In the adjusted model, the association of the highest age group was reduced to almost a fourfold risk (p = 0.038), and the association of obesity to cardiogenic shock was similar (p = 0.023), but overweight was no longer statistically significant. The associations of the other risk factors were not significantly changed in the adjusted model except for mucocutaneous symptoms, which were no longer statistically significant (Table 5).
TABLE 5.
Crude and adjusted odds ratios for cardiogenic shock among 335 MIS‐C patients.
| Risk factor | N (%) | Crude | Adjusted model a | ||
|---|---|---|---|---|---|
| OR | 95% CI | OR | 95% CI | ||
| Sex | |||||
| Male | 125 (37.3) | 1.00 | Ref | 1.00 | Ref |
| Female | 210 (62.7) | 1.16 | 0.58–2.43 | 1.35 | 0.59–3.26 |
| Age group | |||||
| 0–6 | 92 (27.5) | 1.00 | Ref | 1.00 | Ref |
| 7–12 | 137 (40.9) | 2.70 | 0.94–9.74 | 2.64 | 0.82–10.4 |
| 13–19 | 106 (31.6) | 4.80 | 1.72–17.1 | 3.85 | 1.17–15.6 |
| Weight group | |||||
| Normal | 261 (77.9) | 1.00 | Ref | 1.00 | Ref |
| Overweight | 46 (13.7) | 1.42 | 0.50–3.47 | 0.88 | 0.27–2.49 |
| Obese | 28 (8.4) | 3.15 | 1.15–7.86 | 3.55 | 1.14–10.5 |
| Symptom b | |||||
| Diarrhoea | 168 (50.1) | 2.72 | 1.33–5.90 | 2.48 | 1.09–5.92 |
| Edema | 73 (21.8) | 1.79 | 0.83–3.68 | Omitted | |
| Hyperinflammation | 228 (68.1) | 7.58 | 3.53–16.2 | 5.10 | 2.16–12.0 |
| Airway affection | 123 (36.7) | 5.95 | 2.86–13.4 | 4.87 | 2.12–12.0 |
| Cardiac affection | 131 (39.1) | 23.3 | 8.14–98.5 | Omitted | |
| CNS | 18 (5.4) | 1.61 | 0.36–5.19 | Omitted | |
| Mucocutaneous | 254 (75.8) | 0.44 | 0.22–0.90 | 0.53 | 0.23–1.26 |
Model adjusted for sex, age group, weight group, diarrhoea, hyperinflammation, airway affection and mucocutaneous symptoms. Cardiac affection variable was excluded due to collinearity with cardiogenic shock.
In the symptom variables, the reference is absence of specific symptom.
4. Discussion
This nationwide prospective study describes the clinical characteristics, treatment and short‐term outcome of 338 Swedish children diagnosed with MIS‐C between March 2020 and August 2022. Despite Sweden's relatively open public health strategy during the COVID‐19 pandemic and limited ICU capacity, we observed low ICU admission rates (20%) and no mortality, without exploring if the Swedish health‐care strategy possibly influenced our results. The findings suggest that consistent national guidelines, early treatment strategies and centralised specialist collaboration may have contributed to favourable outcomes.
Several symptoms were more prevalent among children requiring intensive care, including cardiac dysfunction, respiratory symptoms, diarrhoea and oedema. These findings are consistent with previously published data from other cohorts [1, 2, 22]. Many ICU patients showed signs of capillary leak, including hypoalbuminemia and oedema, which may contribute to clinical deterioration [23]. Neurological symptoms were reported in 30% with irritability and headache being more common in 41% ICU patients, which is in line with reports of CNS involvement in severe MIS‐C [24].
The inflammatory response was pronounced in patients admitted to ICU, with elevated CRP, NT‐pro‐BNP and troponin levels, as well as lymphopenia and cytopenias. Elevated NT‐pro‐BNP levels, even in patients with preserved cardiac function, may reflect systemic inflammation rather than isolated myocardial injury [25]. These hematologic and biochemical findings are consistent with those reported in other cohorts describing severe MIS‐C [7, 26].
The Swedish MIS‐C treatment strategy was characterised by early and widespread use of IVIG and corticosteroids. Anakinra was used in 41% of cases, predominantly in ICU patients, in contrast to stepwise approaches used in many other countries. The high use of biologics likely reflects early national consensus and frequent multidisciplinary discussions. Although our study was not designed to evaluate treatment efficacy, our outcomes compare favourably to international reports and align with data from the RECOVERY trial supporting early corticosteroids and biologics in selected cases [27].
Cardiac involvement was the strongest independent risk factor for ICU admission (OR 14.0), followed by airway symptoms (OR 6.8), hyperinflammation, diarrhoea and oedema. These results confirm observations from other cohorts [4, 5]. Risk factors for cardiogenic shock included older age, obesity, airway involvement and hyperinflammation. The association between obesity and shock has been described previously and may be related to pre‐existing subclinical inflammation [26, 28]. Mucocutaneous involvement appeared to be associated with a reduced risk of shock in univariable analyses but not in the adjusted model.
Despite a relatively low ICU bed density (5.8 per 100 000) compared to countries such as the United States (29.4 per 100 000) [29], outcomes were excellent, with no deaths and all patients discharged home. The centralised case management model, with frequent national case discussions and uniform treatment protocols, likely contributed to this.
The main strength of this study is its population‐based design and use of a national quality register, which enabled inclusion of nearly all diagnosed MIS‐C cases during the study period. Centralised review and consensus‐based management likely improved data consistency and care quality. This study provides one of the few national descriptions of MIS‐C in a setting with a non‐restrictive pandemic response.
Clinical features and laboratory parameters were not uniformly available for all patients, which may have introduced selection bias. Additionally, although data collection was prospective, clinical data were entered by different clinicians across multiple centres, which could introduce variability in reporting. At the very beginning of the inclusion period, a few cases may have been missed, potentially introducing bias due to the novelty of the condition. Also, we did not specifically ask for the individual reason for admission to ICU, and due to the multiple centres involved, some clinical variable definitions, such as cardiogenic shock, had to be pragmatically determined. Furthermore, the capacity to manage severely ill patients outside the ICU may have varied between hospitals, which could have influenced ICU admission thresholds and introduced site‐dependent bias.
5. Conclusion
This Swedish population‐based study of MIS‐C highlights low ICU admission and excellent outcomes in a healthcare setting with limited ICU capacity but strong coordination by the Swedish paediatric MIS‐C consortium bi‐weekly meetings and early treatment strategies with consensus‐based management. We identified clear clinical and laboratory predictors of severe disease. These findings emphasise the importance of early recognition, aggressive immunomodulatory therapy and national coordination. The opportunity to identify risk factors for ICU admission and cardiogenic shock might become a valuable tool for early detection and treatment of high‐risk patients.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Appendix S1: apa70329‐sup‐0001‐AppendixS1.docx.
Acknowledgements
We are grateful to the families who took part in the study. We thank all physicians and nurses who took care and treated patients with MIS‐C in Sweden, the Swedish Paediatric Rheumatologic Society and its Swedish MIS‐C consortium holding regular meetings and discussions, and statistician Julia Eriksson for all discussions and statistical analyses.
Funding: This work was supported by the Swedish Rheumatism Association and by grants provided by Region Stockholm (ALF project) (AC.H.) and by the Swedish governmental funding of clinical research (ALF), the Greta and Johan Kocks Foundation and Swedish Rheumatism Association in Region Skåne (P.K.). None of the funders had any role in the concept, design, or conduct of the study or paper.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. Whittaker E., Bamford A., Kenny J., et al., “Clinical Characteristics of 58 Children With a Pediatric Inflammatory Multisystem Syndrome Temporally Associated With SARS‐CoV‐2,” JAMA 324 (2020): 259–269. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Radia T., Williams N., Agrawal P., et al., “Multi‐System Inflammatory Syndrome in Children & Adolescents (MIS‐C): A Systematic Review of Clinical Features and Presentation,” Paediatric Respiratory Reviews 38 (2021): 51–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Yilmaz Ciftdogan D., Ekemen Keles Y., Karbuz A., et al., “Multisystem Inflammatory Syndrome in Children Associated With COVID‐19 in 101 Cases From Turkey (Turk‐MISC Study),” Journal of Paediatrics and Child Health 58 (2022): 1069–1078. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Merckx J., Cooke S., el Tal T., et al., “Predictors of Severe Illness in Children With Multisystem Inflammatory Syndrome After SARS‐CoV‐2 Infection: A Multicentre Cohort Study,” CMAJ 194 (2022): E513–E523. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Cruz Vidal D., Lee S., Ardoin S. P., et al., “Multisystem Inflammatory Syndrome in Children and Cardiac Involvement: A Quaternary Center Experience,” Pediatric Infectious Disease Journal 43 (2024): e160–e163. [DOI] [PubMed] [Google Scholar]
- 6. Alsaied T., Tremoulet A. H., Burns J. C., et al., “Review of Cardiac Involvement in Multisystem Inflammatory Syndrome in Children,” Circulation 143 (2021): 78–88. [DOI] [PubMed] [Google Scholar]
- 7. Feldstein L. R., Tenforde M. W., Friedman K. G., et al., “Characteristics and Outcomes of US Children and Adolescents With Multisystem Inflammatory Syndrome in Children (MIS‐C) Compared With Severe Acute COVID‐19,” JAMA 325 (2021): 1074–1087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Miller A. D., Zambrano L. D., Yousaf A. R., et al., “Multisystem Inflammatory Syndrome in Children‐United States, February 2020–July 2021,” Clinical Infectious Diseases 75 (2022): e1165–e1175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Sık G., Inamlık A., Akçay N., et al., “Mortality Risk Factors Among Critically Ill Children With MIS‐C in PICUs: A Multicenter Study,” Pediatric Research 94 (2023): 730–737. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Acevedo L., Piñeres‐Olave B. E., Niño‐Serna L. F., et al., “Mortality and Clinical Characteristics of Multisystem Inflammatory Syndrome in Children (MIS‐C) Associated With Covid‐19 in Critically Ill Patients: An Observational Multicenter Study (MISCO Study),” BMC Pediatrics 21 (2021): 516. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. de Farias E. C. F., Pavão Junior M. J. C., de Sales S. C. D., et al., “Factors Associated to Mortality in Children With Critical COVID‐19 and Multisystem Inflammatory Syndrome in a Resource‐Poor Setting,” Scientific Reports 14 (2024): 5539. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Belay E. D., Abrams J., Oster M. E., et al., “Trends in Geographic and Temporal Distribution of US Children With Multisystem Inflammatory Syndrome During the COVID‐19 Pandemic,” JAMA Pediatrics 175 (2021): 837–845. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Fernández‐Sarmiento J., Acevedo L., Niño‐Serna L. F., et al., “Risk Factors Associated With Intensive Care Admission in Children With Severe Acute Respiratory Syndrome Coronavirus 2‐Related Multisystem Inflammatory Syndrome (MIS‐C) in Latin America: A Multicenter Observational Study of the REKAMLATINA Network,” Journal of Intensive Care Medicine 39 (2024): 785–793. [DOI] [PubMed] [Google Scholar]
- 14. Rhedin S., Lundholm C., Horne A., et al., “Risk Factors for Multisystem Inflammatory Syndrome in Children—A Population‐Based Cohort Study of Over 2 Million Children,” Lancet Regional Health Europe 19 (2022): 100443. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Eleftheriou I., Maritsi D., Lampidi S., et al., “Decreasing Incidence of the Multisystem Inflammatory Syndrome in Children Over 3 Pandemic Waves,” Pediatric Infectious Disease Journal 42 (2023): 122–124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Nygaard U., Holm M., Hartling U. B., et al., “Incidence and Clinical Phenotype of Multisystem Inflammatory Syndrome in Children After Infection With the SARS‐CoV‐2 Delta Variant by Vaccination Status: A Danish Nationwide Prospective Cohort Study,” Lancet Child Adolesc Health 6 (2022): 459–465. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. “Multisystem Inflammatory Syndrome in Children and Adolescents With COVID‐19,” https://www.who.int/publications/i/item/multisystem‐inflammatory‐syndrome‐in‐children‐and‐adolescents‐with‐covid‐19.
- 18. “Obesity and Overweight,” https://www.who.int/news‐room/fact‐sheets/detail/obesity‐and‐overweight.
- 19. “BMI‐Kurvan—Svensk Förening För Pediatrisk Endokrinologi och Diabetes,” https://endodiab.barnlakarforeningen.se/tillvaxt/hur‐ar‐tillvaxtkurvorna‐gjorda/bmi‐kurvan/.
- 20. Ravelli A., Minoia F., Davì S., et al., “2016 Classification Criteria for Macrophage Activation Syndrome Complicating Systemic Juvenile Idiopathic Arthritis: A European League Against Rheumatism/American College of Rheumatology/Paediatric Rheumatology International Trials Organisation Collaborative Initiative,” Arthritis & Rhematology 68 (2016): 566–576. [DOI] [PubMed] [Google Scholar]
- 21. “R: The R Project for Statistical Computing,” https://www.r‐project.org/.
- 22. Jiang L., Tang K., Levin M., et al., “COVID‐19 and Multisystem Inflammatory Syndrome in Children and Adolescents,” Lancet Infectious Diseases 20 (2020): e276–e288. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Kahn R., Mossberg M., Berthold E., et al., “Capillary Leak Syndrome Was Associated With More Severe Multisystem Inflammatory Syndrome in Children During the COVID‐19 Pandemic,” Acta Paediatrica 113 (2024): 1404–1411. [DOI] [PubMed] [Google Scholar]
- 24. Sezer M., Çelikel E., Tekin Z. E., et al., “Multisystem Inflammatory Syndrome in Children: Clinical Presentation, Management, and Short‐ and Long‐Term Outcomes,” Clinical Rheumatology 41 (2022): 3807–3816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Fish‐Trotter H., Ferguson J. F., Patel N., et al., “Inflammation and Circulating Natriuretic Peptide Levels,” Circulation. Heart Failure 13 (2020): e006570. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Abrams J. Y., Oster M. E., Godfred‐Cato S. E., et al., “Factors Linked to Severe Outcomes in Multisystem Inflammatory Syndrome in Children (MIS‐C) in the USA: A Retrospective Surveillance Study,” Lancet Child Adolesc Health 5 (2021): 323–331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. RECOVERY Collaborative Group , “Immunomodulatory Therapy in Children With Paediatric Inflammatory Multisystem Syndrome Temporally Associated With SARS‐CoV‐2 (PIMS‐TS, MIS‐C; RECOVERY): A Randomised, Controlled, Open‐Label, Platform Trial,” Lancet Child Adolesc Health 8 (2024): 190–200. [DOI] [PubMed] [Google Scholar]
- 28. Khoury M., Harahsheh A. S., Raghuveer G., et al., “Obesity and Outcomes of Kawasaki Disease and COVID‐19‐Related Multisystem Inflammatory Syndrome in Children,” JAMA Network Open 6 (2023): e2346829. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. “ICU Beds by Country | Europe,” https://tradingeconomics.com/country‐list/icu‐beds?continent=europe.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix S1: apa70329‐sup‐0001‐AppendixS1.docx.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.