All eight Swiss PICUs provided data on organizational characteristics and the situation in their units during the study period. These included general characteristics (total number of beds, total number of admissions, and affiliation to adult hospitals) and characteristics related to the SARS-CoV-2 pandemic (percentage of scheduled admissions, cancellation of interventions, staff recruited to adult wards, and percentage of PIMS-TS patients).

For comparing workloads the Nine Equivalents of nursing Manpower use Score (NEMS) and Riker Sedation-Agitation Scale (SAS) as well as a Swiss derivative of the two, the Swiss Society of Intensive Care Medicine (SSICM) shift categories were used (23–25). These scores are used as standard practice in all PICUs in Switzerland and filled out per patient per shift immediately after every nursing working shift. The routine length of a nursing staff shift is 9 h and planned or unplanned double shifts were not practiced. If unplanned absences occurred, units reduced their bed availability for this shift. NEMS score is well-validated and easier to use than the Therapeutic Intervention Scoring System, therefore the Swiss society of intensive care medicine has established using this scoring system nationwide (26–30). NEMS includes data on interventions as well as of therapies and reflects the workload of nurses and physicians. The score includes the following nine items: basic monitoring, intravenous medication, mechanical ventilator support, supplementary ventilator care, single vasoactive medication, multiple vasoactive medication, dialysis techniques, specific interventions in the intensive care unit (ICU), and specific interventions outside the ICU (23). The higher the NEMS Score, the more nursing manpower was needed during a certain shift or during the whole ICU stay. SAS is used to evaluate a patient's level of sedation and agitation. During each shift each patient is given a score from one to seven ranging from an unarousable patient to a dangerously agitated patient (24). SSICM shift categories integrate NEMS and SAS, category 1A of the SSICM defines the most, and category 3 the least, work-load intensive shift for ICU nurses (25). Details SSICM shift categories are shown in Table 1.

In order to compare the critically ill SARS-CoV-2 pediatric patients to the regular PICU population, data on all patients hospitalized in the eight Swiss PICUs in 2019 was used. The whole year instead of the corresponding time period in 2019 was used to avoid a seasonal bias in the group representing regular PICU patients. This data was drawn from the Minimal Dataset of the Swiss Society of Intensive Care Medicine (MDSi) (31). The Pediatric Index of Mortality 2 (PIM2) was used as an indicator of predicted case severity of the patients with PIMS-TS and the general PICU population (32). Data of the patients with PIMS-TS was retrospectively gathered in all eight Swiss PICUs on patients hospitalized from February 24, 2020, the date on which the first SARS-CoV-2 PCR positive patient was reported in Switzerland, up until the June 15, 2020. Included were all patients with age below 20 years and a positive SARS-CoV-2 polymerase chain reaction (PCR) result or the diagnosis of PIMS-TS according to either the Royal College of Child Health and Pediatrics (33), the World Health Organization (34) or the Centers for Disease Control and Prevention case definitions (35), during the PICU stay. Patient data was collected from chart notes and anonymized in the respective centers with a custom made questionnaire, one center's data was obtained from the Risk Stratification in COVID-19 patients in the ICU (RISC-19-ICU) registry (36). Information was gathered also on the use of anti-inflammatory medications including corticosteroids, intravenous immunoglobulin, anakinra, and tocilizumab, which all have been licensed for use in other conditions but were used off label for PIMS-TS.

General epidemiological data on the SARS-CoV-2 pandemic in Switzerland was drawn from the official website of the Swiss Federal Office of Public Health (FOPH) (37). As no reliable serological data was available at the whole population level, the total number of SARS-CoV-2 PCR positive tests in Switzerland was used in calculations to represent number the SARS-CoV-2 cases despite our PIMS-TS group mostly being PCR negative and seropositive. Data on the demographics of Switzerland was obtained from the website of the Swiss Federal Statistical Office (21).

Acute respiratory distress syndrome (ARDS) was defined according to the Pediatric Acute Lung Injury Consensus Conference definition for pediatric acute respiratory distress syndrome (38). Shock was defined as an arterial systolic blood pressure below the age adapted 5th percentile or a systolic blood pressure below two standard deviations of the age adapted mean, and/or the need for vasoactive support to maintain blood pressure in the range (39, 40). Myocardial injury was defined as ejection fraction reduced below 55% measured using the biplane Simpson method according to the current pediatric echocardiographic recommendations (41, 42). Renal dysfunction was defined according to the pRIFLE classification (43). Hepatic dysfunction was defined as the elevation of liver enzymes above the age adapted reference and elevation of prothrombin time (44). An abnormal prothrombin time or activated partial thromboplastin time (according to age adapted reference values of the local laboratories) was classified as a coagulation disorder (45).

The study proposal (KEK: 2020-00720), as well as the RISC-19-ICU (KEK: 2020-00322, ClinicalTrials.gov Identifier: NCT04357275) registry have been evaluated by the Cantonal Ethics Committee of Zurich, a member of the Swiss Association of Research Ethics Committees—Swissethics and in line with the Swiss Federal Human Research Act deemed exempt from the need for additional ethics approval. The study complies with the Declaration of Helsinki, the Guidelines on Good Clinical Practice (GCP-Directive) issued by the European Medicines Agency as well as the Swiss law and Swiss regulatory authority requirements and has been designed in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational studies.

Statistical analysis was conducted employing the R environment for statistical computing version 4.0.2. Comparisons of population characteristics were performed using Wilcoxon rank-sum test for continuous not normally distributed and the chi-squared test for categorical variables. Continuous not normally distributed data is presented as median with interquartile ranges (IQR) and nominal data as counts and percentages.

The Swiss PICUs varied in general characteristics as well as in characteristics related to the pandemic. The different medical and logistical aspects affecting the PICUs are listed in Table 2. In addition to some PICUs having to care for patients with this new syndrome, all PICUs were confronted with organizational issues arising from temporary regulations put in place by the federal council due to the pandemic. These general regulations had different consequences for the eight PICUs due to their different characteristics. As shown in Table 2, there were units (PICU 8) not affiliated to adult hospitals and without staff recruitment to adult wards, where nevertheless scheduled interventions were canceled due to regulations set by the federal council. As a consequence, this PICU was not fully occupied and had a relatively low workload. Other PICUs (PICU 5), however, were affiliated to an adult ward and staff recruitment to the adult wards was required. Therefore, cancellation of scheduled interventions was essential to have enough manpower and staff for the emergency admissions to these PICUs (PICU 5). Other PICUs (PICU 1) were also influenced by local circumstances with staff being recruited to adult wards due to staff shortage in the area despite the PICU not being affiliated with an adult hospital.

Four out of eight PICUs had patients admitted with PIMS-TS between February 24, 2020 and June 15, 2020. Patients with PIMS-TS constituted 0.6% of all Swiss PICU admissions (9/1,528) during this time period and 0.3–1.8% of the PICU admissions of PICUs treating patients with PIMS-TS. A comparison of patients with PIMS-TS and general pediatric intensive care unit patients hospitalized in 2019 is shown in Table 3, based on quantitative and qualitative measures of case complexity. Patients with PIMS-TS had higher total NEMS scores (p = 0.0008) and were classified to higher SSICM workload categories more frequently than general PICU patients hospitalized in 2019 (p < 0.0001).

The total of 1,113 children and young adults under the age of 20 were tested positive for SARS-CoV-2 infection between February 24, 2020 and June 15, 2020 in Switzerland, of which 6% (63/1,113) had to be hospitalized [Swiss FOPH (37)]. Ten pediatric patients with a concurrent positive SARS-CoV-2 PCR result or with PIMS-TS were hospitalized in PICUs. The only patient not diagnosed with PIMS-TS was admitted to the PICU due to an acute necrotizing encephalopathy. As the role of SARS-CoV-2 in the pathogenesis and disease progression remained unclear in this case, this patient's data was excluded from the present analysis. The nine remaining patients with PIMS-TS hospitalized in Swiss PICUs constituted 14% (9/63) of all SARS-CoV-2 positive hospitalized and 1% (9/1,113) of all SARS-CoV-2 positive children and young adults under the age of 20 in Switzerland during that period [Swiss FOPH (37)]. The incidence of severe PIMS-TS requiring hospitalization at a PICU was, therefore, 0.5 per 100,000 people under the age of 20 during the study period. No information is available on seropositivity in the general population during the study period. The general characteristics of all the PIMS-TS patients, details on organ dysfunction and applied therapies are presented in Table 4.