The study was conducted at Skåne Pediatric University Hospital, a tertiary care center in southern Sweden serving primarily urban and suburban populations and providing universal healthcare, similar to other Nordic and many European countries. The hospital has two PEDs, one in Malmö and one in Lund. The study period was from January 1, 2014, to December 31, 2018. This study was approved by the Regional Ethics Committee of Lund, Sweden (Dnr 2017/967).

There are no available codes for BRUE or ALTE in the Swedish version of the International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD-10). Apnea has been a common clinical term and diagnosis code to describe apneic episodes characterized by cyanosis or pallor, changes in muscle tone, absent, irregular, or decreased breathing, and alterations in mental status. Thus, to identify infants aged <1 year with episodes which could be compatible with BRUE we retrospectively identified all infants under 1 year of age with: 1) “apnea” registered as chief complaint by the triage nurse in the PEDs' patient electronic registration systems or 2) with ICD-10 code P284 (Other apnea in newborn) registered as diagnosis. In the PEDs' electronic registration system, only one alternative can be selected from the drop-down menu. Except for seizures, there are no other alternatives that can register such episodes in the electronic registration system of PEDs. Infants identified by both methods were registered only once.

The infants' electronic medical records were reviewed, and infants with insufficient information on their initial visits were not considered eligible for the study. We included all infants with apneic episodes characterized by cyanosis or pallor, changes in muscle tone (increased or decreased), absent, irregular, or decreased breathing, and alterations in mental status.

We collected data on the infants' characteristics, previous chronic diseases, symptoms, investigation results, and initial diagnosis at the first visit. We also evaluated the subsequent visits. All diagnoses after the first visit or hospitalization, as well as any new chronic diagnoses after any subsequent visit, were registered regardless of their association with the symptoms at the initial visit. Data collection commenced in September 2019, ensuring that at least nine months had passed since the eligible emergency visit or diagnosis. Thus, the follow-up period was almost six years for the first included child and nine months for the last. The Data were entered into the Research Electronic Data Capture (REDCap) program hosted by Lund University, Sweden (15). Descriptive statistical analyses were conducted using Stata BE version 18.0.

Apneic episodes were defined as any events involving cyanosis or pallor, changes in muscle tone (increased or decreased), absent, irregular, or decreased breathing, and alterations in mental status.

Serious conditions were defined as: (1) any chronic disease (neurological, inborn error of metabolism, respiratory), (2) infections requiring treatment or supportive care, (3) admission to intensive care unit, or (4) trauma or child abuse.

To create a BRUE sub-cohort, we applied the diagnostic criteria described by Tieder et al. and Merritt et al. (5, 6), with the exception of repeated events and abnormal vital signs, for which data were unavailable (Table 1). Infants were excluded from the BRUE sub-cohort if they had an RTI-related ICD diagnosis or signs suggestive of RTI at the physical examination, such as runny nose, sneezing, or cough. These infants were categorized into a separate RTI sub-cohort. Infants with reported symptoms suggestive of RTI, such as runny nose, sneezing, or cough within the past five days but without observed signs or a RTI diagnosis were included in the BRUE sub-cohort and the RTI sub-cohort.

To further stratify the BRUE sub-cohort into lower- and higher-risk groups, we used the risk criteria outlined by Tieder et al. (5). Family history and postconceptual age ≥45 weeks as risk factors were not always available and therefore excluded from the stratification. Infants meeting any of the remaining high-risk criteria were classified as higher risk (Table 1).

We also performed an alternative risk stratification within the BRUE sub-cohort, excluding age ≤60 days as a risk factor. This yielded two higher-risk groups: (1) infants meeting all available high-risk criteria (“all high-risk factors”), and (2) infants meeting all criteria except age ≤60 days (“all high-risk factors minus age”) (Figure 1).

The primary outcome measures included 1) mortality or any diagnosis after the initial visit or admission (“first diagnosis”), 2) mortality or diagnosis of any chronic condition at any subsequent visit from the time of the eligible visit until the time of data collection (“subsequent chronic diagnosis”), and 3) revisits or re-admissions after the eligible visit.

We identified 968 infants who presented with apnea or who had an apnea related ICD code. Of these, 628 were excluded because they were not assessed as having an episode of cyanosis or pallor, changes in muscle tone (increased or decreased), absent, irregular, or decreased breathing, and alterations in mental status, had incomplete medical records, or were older than 365 days of age (Figure 1). The final cohort included 340 infants, of which 171 (50.3%) females. The median age at the first visit was 38 (IQR 17–79.5) days, 228 (67.1%) were ≤60 days old, and 11 (3.2%) were born before 32 weeks of gestation. Twenty infants (5.9%) had preexisting chronic diseases. No deaths occurred during the follow up period. Median follow up time was 3.0 (IQR 1.9–4.4) years.

Of the 340 infants, a specific diagnosis was not established in 253 (74.4%). Among the 87 infants with specific initial diagnoses, viral respiratory tract infection was the most common diagnosed in 63 (18.5%), followed by gastroesophageal reflux (GER) in 10 (2.9%). There were 9 (2.6%) cases with serious conditions during the first encounter. These included 1 patient with enteroviral meningitis, 1 with pneumonia (positive for metapneumovirus at viral nasopharyngeal test), 1 coagulase-negative staphylococcal sepsis admitted to the neonatal intensive care unit, and 1 with a cytomegalovirus infection. Three patients were diagnosed with a seizure disorder, however 1 of these diagnoses was withdrawn after 1.5 years (Table 2). Additionally, there were 2 infants that were admitted to the pediatric intensive care unit: one diagnosed with a RTI (influenza) and one without a specific diagnosis. Both infants were unwell at the initial presentation.

Thirteen (3.8%) out of the 340 infants were diagnosed with a previously unknown chronic disease: 6 with neurological diagnoses (5 epilepsy diagnoses and 1 patient with delayed motor development, fractures and vitamin deficiency), 2 with GER, 1 with child abuse, 1 with gastritis, 1 with 18q-deletions syndrome, 1 with juvenile oligoarthritis, and 1 with osteogenesis imperfecta. No infants died during the initial study period or follow up. The follow-up time varied from 9 months to nearly six years, with median 3.0 (IQR 1.9–4.4) years from the initial visit to data extraction.

Among the 340 infants younger than 1 year with apneic episodes, 188 met BRUE diagnostic criteria (Table 1), of which 98 (52.1%) were females. The median age at the first visit was 32 (IQR 16–70.5) days, 132 (70.2%) were ≤60 days old, and 1 (0.5%) was born before 32 weeks of gestation. Four infants (2.1%) had preexisting chronic diseases. With the available high-risk BRUE criteria (Table 1), 143/188 (76.1%) infants were stratified as higher risk and 45/188 (23.9%) as lower risk. For 98 infants age was the only higher risk-factor, thus by excluding age as a risk factor, 45/188 (23.9%) were higher risk and 143/188 (76.1%) were lower risk.

All the 9 previously mentioned infants with serious conditions during the first encounter had clinical characteristics and/or received an explanatory diagnosis that excluded them from a BRUE diagnosis and the exclusion of age as a risk factor did not alter their stratification.

Three (1.6%) infants from the BRUE sub-cohort were diagnosed with a previously unknown chronic disease within nine months to nearly six years of their initial visit. All 3 infants were higher risk when all risk factors were considered, and 2 (one with an epilepsy diagnosis three years after the event and one case of child abuse) of them would have been stratified as lower risk if age was not a risk factor.

Of the whole cohort, 203 (59.7%) infants were hospitalized after the first PEDs visit of which 166/203 (81.8%) were admitted for ≤2 days (Table 3). Eleven (3.2%) infants had a planned revisit within 10 days, while 17 (5.0%) had an unplanned revisit. These visits led to 5 readmissions. None of the revisits led to a new chronic diagnosis.

In the BRUE sub-cohort, 95 (50.5%) infants were hospitalized after the first PEDs visit, of which 87 (91.6%) were admitted for ≤2 days (Table 3). Of the 8 infants admitted three or more days, 7 were considered higher risk with all BRUE criteria, while 5 of these 7 would be re-classified as lower risk if age was not considered a risk factor. However, in all these 5 cases the investigations during admission were without any pathological finding.

The RTI sub-cohort consisted of 158 infants who had a history of RTI symptoms, RTI symptoms on physical examination in the PED or received a viral RTI diagnosis after the first visit or hospitalization. Of these, 40 infants had only a history of RTI symptoms and did not exhibit RTI symptoms at the PED, and consequently they were not considered to have an RTI by the treating physician. These 40 children were nevertheless included in the BRUE cohort as higher risk. Overall 111 (70.3%) infants in the RTI sub-cohort were hospitalized after the first PEDs visit, of which 90 (81.1%) were admitted for ≤2 days (Table 3).

Our study has some limitations. First is the identification method. Due to the lack of a specific BRUE diagnose code in the Swedish ICD-10 version, we used apnea as ICD-10 diagnosis or contact reason to identify infants with episodes compatible with BRUE. It is likely that such episodes (i.e., episodes of cyanosis or pallor, changes in muscle tone, absent, irregular, or decreased breathing, and alterations in mental status) could have been coded with another ICD-10 diagnosis, especially in infants with a specific diagnosis. However, using a second identification method (i.e., contact reason), we believe that we have reduced the likelihood of missing a significant number of eligible patients. However, by limiting inclusion to patients with apneic events we introduced a selection bias towards apneic BRUE.

Second, the duration of the episodes of changed breathing was not considered, which resulted in inclusion of infants with a duration of the episodes of less than 20 s.

Third, the lack of all required data for BRUE risk-stratification probably resulted in the misclassification of higher-risk infants as lower risk. However, the low rates of serious underlying diseases and complicated clinical courses strengthen our hypothesis that infants with BRUE are over-classified as having a higher risk.