A Prospective Study of Brief Resolved Unexplained Events: Risk Factors for Persistent Symptoms Free

Infants admitted to our institution with BRUE were approached during their hospitalization and enrolled after informed consent. Recruitment took place from April 2017 to May 2021. Inclusion criteria included any infant <1 year of age who presented with an episode of cyanosis or pallor, absent decreased or irregular breathing, marked changes in tone, or an altered level of responsiveness and were hospitalized with an admitting diagnosis of BRUE and no other clear explanation for their symptoms at the time of hospital admission. Exclusion criteria included previous BRUE hospitalization or any preexisting medical diagnosis. Only families self-identifying their preferred language of care as English were enrolled because questionnaires were only available in English. Charts were reviewed and medical teams approached to ensure each subject met the inclusion criteria, and the admitting diagnosis of BRUE, as defined by the AAP, was confirmed for each subject by 2 general pediatric physicians.

One parent for each enrolled subject completed baseline symptom and medical history questionnaires at enrollment. Parents also completed the Infant Gastroesophageal Reflux Questionnaire Revised (I-GERQ-R), a validated measure of infant reflux on which scores >16 signify clinically significant GERD.³⁹  Medical charts were reviewed for clinical data, comorbidities, and rates/results of testing. Recognizing that many families are given an explanatory diagnosis for their BRUE at hospital discharge, we reviewed discharge summaries to identify the most common explanatory diagnoses and determine if any diagnosis was associated with persistent symptoms.^10,12 

Swallow evaluations were reviewed to determine if subjects underwent videofluoroscopic swallow study (VFSS) or clinical feeding evaluation (CFE), defined as bedside assessment of feeding that does not include instrumental/radiologic evaluation. Per protocol at our institution, all infants typically undergo CFE before VFSS. VFSS results were considered the gold-standard test to diagnose aspiration because of the high rate of silent aspiration in children.^23,40,41  VFSS were considered abnormal if there was aspiration or laryngeal penetration for any consistency. Laryngeal penetration was considered abnormal because these patients have similar outcomes to those with overt aspiration and respond to thickening.^32,42–44 

To obtain follow-up data after the index hospitalization, 1 survey per household was sent every 2 weeks for 2 months. Study data were collected and managed using REDCap electronic data capture tools hosted at our institution.^45,46  Survey questions asked parents to report on the type and frequency of persistent symptoms. Symptom persistence was defined as any parental report of the symptom since the previous questionnaire. All questions were asked in the following format: “Has your child had episodes of XX symptom in the past 2 weeks?” Questionnaires also assessed medication changes (proton pump inhibitor and H2 receptor antagonist use), feeding changes (transition from breast to bottle, hypoallergenic diets, and thickened feedings), and repeat medical evaluations (hospitalizations and emergency department and clinic visits). Levels of parental anxiety in relation to their children’s symptoms were assessed based on validated Generalized Anxiety Disorder 2-item scores in which values ≥3 suggest clinically significant anxiety.⁴⁷ 

We reviewed charts for repeat hospital and clinic visits in the 6-month period after BRUE. This period was selected because follow-up appointments and diagnostic studies that may provide a definitive diagnosis can take 6 months to occur after hospitalization. Only hospital visits (admissions and emergency room visits) for recurrent BRUE or related symptoms (eg, respiratory distress) were included.

Our primary outcome was the frequency of persistent symptoms after BRUE and the risk factors associated with persistent symptoms. Our secondary, hypothesis-generating outcome was the impact of commonly prescribed therapies on these symptoms. Descriptive statistics were used to summarize patient characteristics and clinical factors, presented as mean (SE) for continuous variables and percentage (frequency) for categorical variables. Proportions were compared with Fisher’s exact test. Generalized estimating equations with the log-link function, binomial distribution, exchangeable correlation matrix, and robust estimators of variance were used to assess the relative risk of persistent symptoms with potential risk predictors, including acid suppression, hypoallergenic formula, and thickened feeds.⁴⁸  Treatments were included as independent variables occurring before each symptom assessment. Covariates adjusted for included age at admission, sex, history of prematurity, and breastfeeding status. Because not all subjects that received thickening had VFSS, analyses were stratified by aspiration status. To take into account potential selection bias due to incomplete follow-up, we conducted a sensitivity analysis in which we used inverse probability weighting to weight study participants in the analysis based on their inverse probabilities of completing follow-up during the study period. On the basis of our previous work on thickening efficacy and rates in BRUE, we estimated that 83 subjects would provide 85% power to detect a 40% difference in proportion to persistent symptoms.^3,33,42  All statistical tests were 2-sided with P <.05 considered statistically significant. Data were analyzed using SPSS version 27 and SAS version 9.4.

The current study was approved by our institutional review board.

We screened 263 potential subjects and enrolled 124. One hundred one were ineligible because of comorbidities (38), previous BRUE hospitalization (23), or the need for an interpreter (40). Thirty-eight families were not available during hospitalization (21), families declined (10), or teams declined (7). All enrolled subjects completed baseline questionnaires, and 73% (90) completed at least 1 follow-up assessment. Supplemental Fig 2 shows the study population flow diagram.

Table 1 reveals subject characteristics at presentation. Of the subjects, 75% (93) qualified as higher-risk BRUE based on AAP criteria, including 89 subjects with age ≤60 days and 5 subjects with a history of <32 weeks’ gestation with a corrected gestational age of <45 weeks.

Hospitalization characteristics are in Supplemental Table 5. Most common explanatory diagnoses provided to families were gastroesophageal reflux (42%, n = 52), oropharyngeal dysphagia (7%, n = 8), laryngomalacia (6%, n = 7), and feeding difficulties (4%, n = 5). There was a poor correlation between an abnormal I-GERQ-R score and the reporting of reflux as an explanatory diagnosis (r = 0.048, P = .74), highlighting the difficulties with GERD diagnoses.

Parent-reported outcomes are in Table 2, and Figure 1 reveals symptoms and anxiety over time. The mean length of follow-up was 6.3 (0.3) weeks. 86% (77/90) of parents reported persistent symptoms during the follow-up period. Notably, 65% (56/86) reported ongoing choking with feeds and 12% (10/86) reported recurrent BRUE spells. Parent-reported symptom numbers ranged from 0 to 16 concerning episodes per day. Of the 66 families that reported symptom numbers, 46% (30) reported >1 symptom occurrence per day on >1 questionnaire and 11% (7) actually reported 5 or more symptoms per day.

Parents also reported elevated anxiety levels, with 31% (22/70) with an abnormal Generalized Anxiety Disorder 2-item score. Subjects with elevated anxiety had longer length of stay during their BRUE hospitalization (3.7 [0.9] vs 1.95 [0.3] nights, P = .02).

In a comparison of subject characteristics for those that completed or did not complete follow-up, there were racial differences seen, but no other significant differences between the groups, as shown in Supplemental Table 6.

Subjects had 2.7 (0.2) tests during their hospitalization and only 11% were contributory (ie, providing a plausible explanation for BRUE symptoms). No subject had a pH-impedance study performed during the study period. The most common consultation obtained was feeding team consultation, with 49% (61) of subjects undergoing CFE. Subjects reporting abnormal anxiety levels had not had a different number of diagnostic tests performed (2.9 (0.4) tests in those with anxiety vs 2.5 (0.3) in those without, P = .41).

Twenty-four subjects underwent VFSS during admission and 9 underwent VFSS at 112.6 (41) days after admission. Overall, 67% (22/33) of VFSS were abnormal with 27% (9/33) showing aspiration and 39% (13/33) laryngeal penetration; 77% (17/22) had a change in management, including nipple flow rate in 4 and thickening in 14. One hundred percent of those with aspiration had aspiration that was silent (ie, not accompanied by overt coughing or choking) on VFSS.

Families whose infants underwent VFSS did not have different anxiety scores compared with those without VFSS (odds ratio [OR] 0.75, 95% confidence interval [CI] 0.21–2.68, P = .76), but those that underwent CFE did have a higher risk of an abnormal anxiety score (OR 3.73, 95% CI 1.24–11.21, P = .02).

Table 1 reveals therapies before hospitalization and Supplemental Table 5 reveals those provided at discharge. Management changes during follow-up are in Table 2. A total of 26% (32) received acid suppression at any time before, during, or after hospitalization. An abnormal I-GERQ-R score was not associated with treatment with acid suppression, hypoallergenic feeds, or thickening (OR 1.38, 95% CI 0.28–6.7, P = 1.0 for acid suppression, OR 1.83, 95% CI 0.38–8.78, P = .7 for thickening, and OR 5, 95% CI 0.9–27.82, P = .08 for hypoallergenic feeds). Treatment with any therapy was not associated with decreased anxiety (OR 1.24, 95% CI 0.41–3.8, P = .78 for acid suppression, OR 2.8, 95% CI 0.81–9.73, P = .11 for thickening, and OR 0.73, 95% CI 0.24–2.24, P = .78 for hypoallergenic feeds). Supplemental Tables 7–9 reveal a comparison of characteristics between subjects treated or not treated with each therapy, only notable for increased proportion of those with prematurity who were given hypoallergenic formula and fewer breastfed infants receiving thickening.

Table 3 reveals subsequent health care utilization, defined as repeat hospital visit, primary care clinic visit, or subspecialty clinic visit. Fifteen percent (18) required repeat hospital evaluation for recurrent BRUE or BRUE-type symptoms and many required evaluations in subspecialty clinics. Ninety percent (40/44) of outpatient gastrointestinal (GI) visits, 35% (5/13) of neurology visits, 27% (3/11) of cardiology visits, 22% (2/9) of pulmonary visits, 94% (16/17) of otolaryngology visits, and 28% (7/25) of feeding visits were preceded by inpatient consultation by the respective subspecialty.

Subjects whose parents reported abnormal anxiety were at increased risk for hospital evaluation (OR 4.41, 95% CI 0.95–20.5, P = .04). Those treated with acid suppression or hypoallergenic feeds had no difference in subsequent health care utilization (OR 1.56, 95% CI 0.46–5.3, P = .52 for acid suppression and OR 1.04, 95% CI 0.29–3.73, P = 1.0 for hypoallergenic formula).

No differences were seen in repeat hospital visits or clinic visits between subjects with and without parent-reported follow-up data. Fifteen percent (13/90) of those with follow-up had repeat hospital visits versus 15% (5/34) of those without follow-up (P = 1.0). Forty-seven percent (42/90) of those with follow-up had clinic visits versus 41% (14/34) of those without follow-up (P = .69).

Table 4 reveals risk factors for persistent symptoms. An explanatory diagnosis of reflux was associated with increased symptom risk (OR 1.26, 95% CI 1.12–1.42, P <.001) but not decreased anxiety (OR 1.39, 95% CI 0.49–3.92, P = .54). However, an abnormal I-GERQ-R score at admission was not associated with persistent symptoms (OR 0.93, 95% CI 0.18–4.69, P = 1.0).

To evaluate the impact of an aspiration diagnosis, adjusted results were stratified by aspiration status with 18% (22) known to aspirate, 9% (11) known to not aspirate, and 73% (91) with aspiration status unknown. Thickening was associated with a decreased risk of persistent symptoms, even when aspiration status was unknown. Ninety-two percent (11/12) of those receiving thickening had a decline in symptom numbers compared with only 50% (16/32) not receiving thickening (P = .03). Subjects that underwent clinical feeding evaluation without definitive VFSS, however, were not at a decreased risk of persistent symptoms (OR 0.97, 95% CI 0.37–2.55, P = 1.0) or repeat hospitalization (OR 0.8, 95% CI 0.29–2.19, P = .8). CFEs also did not predict or prevent persistent choking spells, when looking at this symptom in isolation (OR 0.71, 95% CI 0.29–1.74, P = .5).

Similar results were found using inverse probability weighting to account for missing follow-up data, as in Supplemental Table 10.

BRUE management remains a challenging problem in clinical practice because there are no prospective studies to determine if symptoms persist after the index hospitalization.^3,10,11  To our knowledge, this is the first prospective longitudinal cohort study and the first study of parent-reported symptoms after BRUE. Key findings are that the majority of infants in our cohort continued to have concerning symptoms that were distressing to their parents months after their BRUE hospitalization, and, over this time period, many subjects underwent management changes.

Our data suggest that the majority of infants actually remain symptomatic for at least 2 months after hospital discharge. Many of the reported symptoms were significant symptoms, with 65% of families reporting choking. These rates are significantly higher than those reported from population studies of infants with feeding difficulties, which have reported rates of 3% to 7%.^49,50  Additionally, 12% of our cohort reported recurrent BRUE spells, and 15% of infants returned to the hospital for care. This high rate of persistent symptoms and health care utilization suggests the need for an improved understanding of how to advance management for this condition.

These persistent symptoms might drive parental anxiety. Previous studies have questioned whether parents felt reassured after discharge from the hospital and our prospective data also reveal that anxiety persists in this group.^8,9  Thirty-one percent of parents reported clinically significant anxiety related to their child’s symptoms, suggesting that, in the face of persistent symptoms, parents might not be reassured by their hospitalization, performance of tests, or the message of reassurance conveyed by hospital teams. Those with elevated anxiety levels were also notably at increased risk for repeat hospitalization. Although there are no studies of parental anxiety levels after infant hospitalizations, studies of parental anxiety after the birth of an infant have found a prevalence of ∼15%.^51,52  Fortunately, in our cohort, there was a decrease in the proportion of those with anxiety over the 2-month study period, but even over 2 months, 10% of parents still reported clinically significant anxiety.

Providing a diagnosis of GERD, the most commonly provided diagnosis in this cohort, also did not result in reduced anxiety. In a recent large retrospective multicenter study, 20% of hospitalized infants with a BRUE episode were given a GERD diagnosis and, in our cohort, 42% were given a reflux diagnosis and 30% were started on acid suppression medication at discharge or in subsequent encounters.^10,12  Despite the common practice of implicating GERD, multiple physiology studies looking at the etiology of apneic events have failed to reveal a refluxate-apnea correlation.^13–18,20  Although none of the infants in our cohort underwent impedance studies, our results also suggest a poor correlation between clinicians’ provision of a reflux diagnosis and an abnormal parent-reported I-GERQ-R score, raising the possibility that BRUE physiology may be more complex than GERD.^39,53 

Making the diagnosis of GERD not only does not reduce parental anxiety but also has significant implications for treatment; patients may be inappropriately treated with acid suppression, which can actually worsen outcomes by increasing gastrointestinal and respiratory infectious risk and delaying the correct diagnosis in symptomatic infants.^{3,27–29,31,54}  In this study, patients treated with acid suppression did not experience symptomatic improvement, their parents did not have reduced anxiety, and these patients had a significantly higher rate of persistent symptoms. Based on these data combined with previous studies, we now have more prospective evidence that acid suppression should likely not be prescribed in infants with BRUE.

Although our data suggest that GERD attribution and acid suppression therapies might not be beneficial, it is notable that VFSS was high yield in this cohort. Although we recognize that patients undergoing VFSS may be an enriched population due to presenting symptoms, similar results have also been reported by authors of retrospective studies in this patient population.^3,23  All subjects with aspiration in our cohort had silent aspiration, which is also consistent with previous studies highlighting the high prevalence of silent aspiration (defined as no coughing or other symptoms during an aspiration event) in infants, ranging from 81% to 89% (among those with aspiration), and particularly in BRUE patients.^23,40,41,55  This finding highlights why clinical feeding evaluations, despite being recommended as first-line in the higher-risk BRUE framework, are not sensitive enough to definitively diagnose and treat aspiration; the sensitivity of clinical feeding evaluations to diagnose aspiration in infants with BRUE is only 50%. Therefore, a VFSS should be strongly considered if there is concern for aspiration.^3,6,23,41  Interestingly, in our cohort, obtaining CFE was associated with increased anxiety, but VFSS was not, perhaps suggesting that undergoing definitive swallow evaluation is not only more sensitive to diagnose aspiration but more reassuring to families. Although VFSS also has limitations, including radiation exposure and the need for patient participation, it remains the gold standard in the diagnosis of oropharyngeal dysphagia in children, particularly in the infant population with a known higher prevalence of silent aspiration among those with aspiration.^40,41,55  Therefore, underrecognized swallowing dysfunction might play a role in some proportion of persistent symptoms after BRUE, and treating oropharyngeal dysphagia might improve outcomes.

Fortunately, thickening of feeds, a mainstay of aspiration therapy (and an effective therapy for regurgitation), has been shown to decrease symptoms and hospitalization risk in infants with swallowing dysfunction and, particularly, in infants with silent aspiration.^{30,32,33,36,56,57}  Our stratified analysis in the current study also suggests that thickening is beneficial after BRUE even if aspiration status is unknown. Thickening is typically well-tolerated and safe and can be used to treat both regurgitation and oropharyngeal dysphagia; however, it is important to consider thickening under the guidance of a speech-language pathologist to determine the appropriate thickening level and nipple flow rate.^34,35,37,58  Providers might therefore consider obtaining VFSS or recommending empirical thickening in infants with persistent symptoms after BRUE because this might be a high-impact, low-risk therapy.

Key strengths of our study include prospective longitudinal design and parent-reported outcomes. Previous retrospective studies have been unable to address the question of persistent symptoms, but the current study highlights the need for continued clinical monitoring after discharge. The high proportion of subjects in our cohort who received management changes during the follow-up period suggests that care providers are making attempts to address these persistent symptoms, but an evidence-based strategy is needed. Our results suggest potential strategies to improve outcomes after BRUE, including considering aspiration in infants with persistent symptoms, avoiding acid-suppression medication, and treatment with thickened feeds when clinically appropriate. Prospective randomized controlled trials will be necessary to better evaluate these therapies and are greatly needed.

Potential study limitations include those related to study design. Classification bias is possible if any subjects were mislabeled as BRUE, but our institution has a clinical pathway to ensure this diagnosis was made appropriately. We also confirmed the admitting diagnosis with 2 general pediatric physicians. Recognizing that many infants with BRUE are ultimately given an explanatory diagnosis but that this diagnosis might not be provided until later in their hospital course, we based inclusion on the admitting diagnosis of BRUE. We acknowledge this might have broadened the group included in our cohort; however, this approach is consistent with other studies in BRUE.^10,12  Additionally, because we did not assign therapies, confounding by indication is possible because patients may have differed at baseline; patients with more severe symptoms may have been more likely to be prescribed therapies or to have VFSS, and those that underwent VFSS might have been more likely to have persistent symptoms. It is possible that thickening might have had other effects beyond treating aspiration, such as decreasing reflux, but validated reflux scores were not higher in those receiving thickening. Generalizability might be a concern given recruitment from our tertiary care center, and it is likely our findings are more applicable to the higher-risk group, but our subject demographic patterns mimic general pediatric patient populations from other institutions, and clinical characteristics were consistent with those in other cohorts.¹⁰  In the future, it would also be important to include parents who are not proficient in English because these families may report symptoms or anxiety levels differently or have a different experience of clinical care; this limitation might also affect generalizability. The clinical significance of the persistent symptoms that were concerning to parents is also not known; we included multiple symptoms that could be seen as part of the BRUE spectrum, including choking, as opposed to just recurrent BRUE; because nothing is known about the presence of symptoms posthospitalization in these infants, we felt that it was important to capture all symptoms that might serve as a warning for future BRUE. Although some symptoms, such as a single choking episode, might be common in otherwise healthy young infants, the persistence of symptoms appears to be relatively uncommon and, therefore, would represent a symptom of concern.^49,50  Although studies suggest these symptoms are not typical for infants, future BRUE studies might benefit from a healthy control comparison group. It is possible that families that agreed to participate or completed questionnaires were more or less symptomatic, but we had a high rate of families agreeing to participate with 73% completing follow-up assessments, and our analysis with inverse probability weighting to account for missing follow-up data revealed similar results.

Results of this prospective study suggest that persistent symptoms are common after BRUE. Although gastroesophageal reflux was frequently provided as an explanatory diagnosis, aspiration might be underrecognized in infants with persistent symptoms after BRUE. Acid suppression did not impact outcomes, but thickening was associated with symptom improvement in this observational study. Future randomized controlled trials are needed to better evaluate the potential efficacy of these therapies.

The authors gratefully acknowledge permission for use of I-GERQ-R from the Innovation Institute at the University of Pittsburgh.