Intensivists’ Reported Management of Critical Bronchiolitis: More Data and New Guidelines Needed

This anonymized, cross-sectional survey study was approved by the institutional review board at Penn State Milton S. Hershey Medical Center, United States. Potential participants were pediatric critical care providers (attending physicians, fellows, advanced practice providers) participating in the following established PICU research networks groups: from North America, the Pediatric Acute Lung Injury and Sepsis Investigators and its Bronchiolitis Advancing Care and Outcomes Network subgroup; from Latin America, the Red Colaborativa Pediátrica de Latinoamérica and the Brazilian Research Network in Pediatric Intensive Care; from Asia, the Pediatric Acute & Critical Care Medicine Asian Network; and from Oceania, the Australian and New Zealand Pediatric Intensive Care Society.

Ten pediatric intensivists representing an international multilingual group of providers were recruited via the Bronchiolitis Advancing Care and Outcomes Network investigators’ subgroup and convened to conduct the survey. The survey was developed by all the authors through an iterative, consensus-building process that involved group discussion, with the authors all agreeing on the final survey as administered. The initial questions were developed by reviewing the existing bronchiolitis guidelines to ensure that the various aspects of clinical practice (eg, imaging, medication use, etc) were reflected in the survey. The author group then further expanded the list of diagnostic and therapeutic modalities on the basis of clinical experience and expertise, as well as their knowledge of the existing bronchiolitis literature. For example, the American Academy of Pediatrics (AAP) guidelines recommend that “laboratory studies should not be obtained routinely,”⁹  but do not specify which laboratory studies might be obtained. The survey was iteratively tested and revised by the authors to address flow, salience, clarity, acceptability, relevance, completeness, administrative ease, face and content validity, and redundancy. The survey was then further revised on the basis of feedback from members of the scientific committees of the various research networks who endorsed and distributed the survey. The survey was initially developed in English and then translated to Portuguese and Spanish by native-language Portuguese- or Spanish-speaking members of the study team.

The survey was available between November 2020 and March 2021. Although the survey was initially developed as a document, it was administered electronically in English, Portuguese, and Spanish using REDCap electronic data capture tools hosted at Penn State Health Milton S. Hershey Medical Center and Penn State College of Medicine.²²  The electronic versions of the survey in all 3 languages were also tested by the author group to ensure it functioned as designed. In addition to being endorsed and distributed via the Pediatric Acute Lung Injury and Sepsis Investigators, the Red Colaborativa Pediátrica de Latinoamérica Network, the Brazilian Research Network in Pediatric Intensive Care, the Pediatric Acute & Critical Care Medicine Asian Network, and the Australian and New Zealand Pediatric Intensive Care Society e-mail lists, the survey was also endorsed and distributed by the World Federation of Pediatric Intensive and Critical Care Societies. Study investigators and participating networks also posted public survey links using Twitter. To further increase the number of respondents, e-mail recipients were asked to forward the survey links to other PICU providers at their institution. The survey was anonymous and the exact number of potential respondents was unable to be tracked.

The survey asked respondents to consider the clinical management of previously healthy infants aged between 2 and 12 months with a diagnosis of acute viral bronchiolitis. This age range was chosen to align with international definitions of bronchiolitis^9–12  and to minimize overlap with neonatal sepsis in younger infants and with reactive airway disease or asthma in children aged >12 months. Although many factors may influence care decisions, the case presented was purposefully “generic” to provide an understanding of the range and breadth of care provided to infants with bronchiolitis across the world. The final survey is available as a file in Supplemental Information.

Our primary outcome was the frequency with which PICU providers reported it would be beneficial to have treatment guidelines for infants requiring noninvasive ventilatory support (NIV), including high-flow nasal cannula (HFNC), continuous positive airway pressure and bilevel positive airway pressure, and invasive respiratory support via endotracheal tube. Secondary outcomes were the frequency with which PICU providers reported they perform different types of interventions (diagnostic and therapeutic) for infants with bronchiolitis, as measured on a 5-point Likert scale (Rarely [<5% of time], Occasionally [5%–24% of time], Often [25%–75% of time], Very Often [76%–95% of time], and Always [>95% of time]). We opted to define 25% to 75% of the time as often because this definition would include therapies done frequently enough that they are commonplace in the PICU and that a large number of children are therefore exposed to them. Respondents were also asked to provide demographic information about themselves and their clinical practice environments.

Descriptive statistics were used to summarize demographic characteristics and reported practices of the respondents. For analysis, we generated dichotomous outcome measures by combining responses provided on the 5-point Likert scale into practices chosen as Often or more frequent (≥25% of time) and Occasionally or less frequent (≤24% of time), and compared reported frequency of the various interventions between NIV and intubated patients. We also evaluated reported practice by years of experience, comparing providers with 10 or more years of experience (“more experienced,” [ME]) with those with <10 years of experience (“less experienced,” ≤). Ten years was chosen as the cutoff because it represented the median years of experience of respondents. This definition of ME versus LE was used for all experience analyses. To understand the international variations in care, we also compared reported practices of respondents practicing in the United States with those practicing in other parts of the world. χ² and Fisher’s exact tests were used as appropriate for comparisons. P values <.05 were deemed to be statistically significant. Bonferroni corrections were performed as required for analyses that included multiple comparisons.²³  Analyses were performed using Microsoft Excel, 2019.

In total, 657 PICU providers responded to the survey, including 344 in English, 204 in Spanish, and 109 in Portuguese. Respondent attrition was noted as the survey progressed and 527 of 657 (80%) of respondents completed the entire survey. Respondent demographics and clinical practice environments are presented in Table 1.

Overall, most participants (434 of 524; 66%) reported being at least “somewhat familiar” with the AAP guidelines (Table 1), with fewer respondents (118–184 of 524; 18%–28%) reporting familiarity of the other national guidelines. Respondents practicing in countries with national guidelines were more likely to be familiar with them: 73% (168 of 230) of respondents practicing in the United States were at least somewhat familiar with the AAP guidelines⁹ ; 90% (9 of 10) of respondents practicing in Australia and New Zealand were at least somewhat familiar with the Australasian Bronchiolitis Guidelines¹² ; 86% (6 of 7) practicing in Canada were at least somewhat familiar with the Canadian Pediatric Society Statement¹⁰ ; and all (4 of 4) practicing in the United Kingdom were “very familiar” with the National Institute for Health and Care Excellence Guidelines.¹¹  Overall, half of respondents (270 of 527; 51%) believed that currently available guidelines for the treatment of bronchiolitis should be applied to infants requiring NIV (including HFNC), and 91% (480 of 527) agreed it would be beneficial to have specific guidelines for infants requiring NIV (including HFNC). Nearly half (234 of 527; 44%) believed that currently available guidelines should be applied to infants requiring invasive respiratory support, with 89% (467 of 527) agreeing it would be beneficial to have specific guidelines for infants with bronchiolitis requiring invasive support.

Figure 1 details the frequency with which respondents reported performing diagnostic studies on admission for nonintubated and intubated patients. Except for testing for severe acute respiratory syndrome-related coronavirus 2, respondents indicated higher usage of all diagnostics at PICU admission in intubated patients as compared with nonintubated patients (all P < .001). Clinicians with less experience were more likely to obtain a C-reactive protein level (168 of 225 [75%] LE versus 172 of 265 [65%] ME; P = .025) and a sputum culture (168 of 222 [76%] LE versus 166 of 256 [64%] ME; P = .013) on admission for intubated patients; there were no other significant experience-related differences noted (Supplemental Table 3). Respondents practicing outside the United States were significantly more likely to report ordering most laboratory analyses on PICU admission for both intubated and nonintubated infants (Supplemental Table 4). Almost all (505 of 518 (97%)) respondents reported a preference to obtain a chest x-ray on PICU admission at least often for intubated patients and most (427 of 515) 83%) for nonintubated patients, regardless of experience level or practice country. Figure 2 details the frequency with which respondents reported performing daily diagnostic studies and general management for nonintubated and intubated patients. LE providers were significantly more likely to order chest physiotherapy for intubated patients (182 of 223 [82%] LE versus 195 of 265 [74%] ME; P = .046). There were no other experience-related differences noted in reported daily management (Supplemental Table 3). Respondents practicing outside the United States were significantly more likely to report ordering most laboratory analyses for the routine management of both intubated and nonintubated infants, and were also significantly more likely to obtain daily chest x-rays for nonintubated children (Supplemental Table 5).

Figure 3 details the frequency with which respondents reported prescribing medications and therapies for nonintubated and intubated patients. Short-acting β-2 agonists, systemic corticosteroids, antibiotics, diuretics, and sedatives were significantly more likely to be used in intubated than nonintubated patients. Providers with fewer years of experience were more likely to prescribe inhaled short-term β-2 agonists (126 of 223 [57%] LE versus 106 of 261 [41%] ME; P < .001), hypertonic saline nebulized treatments (103 of 223 [46%] LE versus 86 of 261 [33%] ME; P = .004) and antibiotics (129 of 221 [58%] LE versus 127 of 263 [48%] ME; P = .034) often or more to intubated patients. LE providers were also more likely to prescribe systemic corticosteroids to nonintubated patients (54 of 222 [24%] LE versus 41 of 262 [16%] ME; P = .023) but less likely to use as needed bolus doses of sedation for nonintubated patients (36 of 224 (16%) LE versus 71 of 258 [28%] ME; P = .004) (Supplemental Table 4). Additional details about provider reported practice by experience level are reported in Supplemental Table 4. Significant variation occurred between United States and non-US providers; notably, providers practicing outside the United States were significantly more likely to report prescribing systemic corticosteroids and antibiotics for both intubated and nonintubated patients (Supplemental Table 6).

Isotonic fluids were reported as the most commonly prescribed type of intravenous fluids for both nonintubated (528 of 596 [89%] isotonic, 54 of 596 [9%] hypotonic, 14 of 596 [2%] other) and intubated (488 of 550 [89%] isotonic, 51 of 550 [9%] hypotonic, 11 of 550 [2%] other) patients. Providers reported starting intravenous fluids at a median rate of 100% of maintenance (interquartile range [IQR] 0.75–1) for nonintubated patients and 80% of maintenance (IQR 0.75–1) for intubated patients. Table 2 presents information about respondents’ reported practice regarding enteral feeds. Although work of breathing was the most common variable providers reported as affecting their decision to initiate enteral feeds for nonintubated infants, hemodynamic status was the most common variable for intubated infants. Respondents’ transfusion thresholds were the same (median 7 g/dL; IQR 7–7) for both intubated and nonintubated patients. Respondents reported the same oxygen saturation thresholds (median 90%; IQR 88–92), below which they would intensify respiratory support for both intubated and nonintubated patients.

This cross-sectional survey of an international cohort of PICU providers indicates that specific guidelines for infants with critical bronchiolitis requiring either noninvasive or invasive levels of respiratory support are needed. We also observed significant variation in providers’ reported practices when caring for infants with critical bronchiolitis. Several practices had substantial rates of both high (>75%) and low (<25%) usage, including measuring C-reactive protein at admission, placing a central venous catheter in intubated patients, and short-acting β-2 agonist use. As would be expected, providers indicated they would be significantly more likely to perform most of the diagnostic and therapeutic interventions for infants who are intubated because of bronchiolitis.

A recent similar survey-based study of Canadian PICU providers’ reported management of critical bronchiolitis indicated higher use of nebulized epinephrine (75% vs 12%–16% in our study) but lower reported use of short-acting β-2 agonists (36% vs 43%–50% in our study). The increased use of racemic epinephrine may reflect previous evidence that it can decrease hospital admission.²⁴  Interestingly, in that previous study, Canadian PICU providers did not report differences in prescribed medications for intubated compared with nonintubated patients, except for antibiotics which, similar to our findings, were reported prescribed more frequently for intubated patients.²⁵  However, caution must be taken when comparing the results of the Canadian PICU providers’ survey; not only because of a difference in the era, but also because the Bradshaw et al survey asked for providers’ practice on a more detailed hypothetical case as opposed to asking how often they use them overall for a less specified infant with bronchiolitis. Nonetheless, when compared with data about actual management of patients with critical bronchiolitis, many medications, including albuterol and antibiotics, appear to be used more frequently in actual practice than PICU providers reported in our survey.^18,26,27 

In general, existing bronchiolitis guidelines recommend against diagnostic imaging, laboratory studies, and using medications when treating infants with acute viral bronchiolitis. A recent systematic review identified 32 bronchiolitis clinical practice guidelines, with widespread agreement amount avoiding unnecessary diagnostic tests but multiple areas of disagreement about various aspects of care, including variability in pharmacologic management.²⁸  With this study, we have demonstrated that many PICU providers generally do not adhere to the less is more approach. This likely reflects PICU providers’ concern that infants with critical bronchiolitis may have additional organ dysfunction because of the level of their respiratory disease, as well as the need to safely monitor and adjust therapies for infants requiring mechanical ventilation and fluid management. PICU providers likely also use these additional studies to more fully evaluate for and rule out the multitude of other diseases that may “masquerade” as bronchiolitis, including bacterial pneumonia, cardiac disease, and metabolic derangements, as well as the need to identify when an infant initially intubated for viral bronchiolitis may have progressed to having pediatric acute respiratory distress syndrome, which may lead to differences in clinical management. The differences noted in PICU providers’ reported practices, including for example how those outside of the United States were overall more likely to order bloodwork for their patients, may reflect differences in local vaccination rates, pathogens, patient comorbidities, or even access to certain resources. These needs will need to be heavily considered, with appropriate input from an international group of PICU providers, should this work progress toward international critical bronchiolitis guidelines. In review of the existing guidelines, intensive care and/or intensive care treatments are only rarely referenced or discussed. Only 14 of 32 (44%) of the guidelines identified in a systematic review make any mention of PICU-level respiratory support by describing indications for initiating continuous positive airway pressure, but no other types of positive pressure are mentioned.²⁸  The AAP guidelines include comments about HFNC decreasing rate of intubation, antibiotics being justified in children who require intubation and mechanical ventilation, and that 3% saline-nebulized treatments are not studied in PICU settings; otherwise, there is no mention of caring for infants with critical bronchiolitis.⁹ 

The variability in PICU providers’ practice should serve as a wake-up call to the PICU community about the lack of consistency in our management, which may be because of the lack of adequate data. Additional critical bronchiolitis studies are needed. Only about half of PICU providers that responded to the survey reported that they use a treatment pathway, standardized respiratory scores, or a weaning protocol. In our study cohort, providers with less experience were also more likely to obtain certain diagnostic studies and prescribe β-2 agonists, hypertonic saline treatments, corticosteroids, and antibiotics to certain patients than providers with more experience, also indicating additional variability in care on the basis of comfort or experience. Given that bronchiolitis is an increasingly common²  reason for admission to the PICU, further studies to develop and implement evidence-based treatment guidelines and protocols for infants with critical bronchiolitis could then lead to PICU-specific measures of quality of care for these patients, which could lead to more efficient and cost-effective care. The notable differences in management styles of PICU providers working within and outside the United States are of unclear etiology, and could be because of management “style” largely influenced by training, actual differences in disease presentations, or variation in local support mechanisms and/or resources. Regardless, these differences need to be taken into consideration when evaluating the generalizability of critical bronchiolitis research and designing future studies.

This study had multiple limitations. Firstly, because of the lack of already validated surveys for critical bronchiolitis management, we had to use a de novo survey and rely on face validity. This limitation was mitigated by having the survey reviewed and revised by multiple critical bronchiolitis experts, as well as by the scientific committees of various pediatric critical care research networks. Secondly, although this survey measured PICU providers’ reported practices, reported practice may not correlate with actual provider practice. Although the survey asked for respondents’ country of practice, it did not ask for institution-level information, which may have allowed for additional practice analyses. Additionally, to better understand the variation in management for “uncomplicated” bronchiolitis, we used case presentations that were purposefully simplified. Given that children with existing comorbidities have different risks for needing PICU-level care, it would be difficult to generalize the results of this study to all children admitted to the PICU with respiratory viral illnesses. Finally, given the anonymous nature of the survey and the fact that it was shared internationally via multiple research networks’ e-mail lists, as well as via social media, to maximize the number of respondents, we were unable to provide an actual response rate. Although our respondents’ demographics indicated a wide range in providers’ practice environments and clinical experience, we were unable to control for potential sampling bias that may have influenced which providers opted to complete the survey.

Although practices vary, a multinational group of PICU providers reports performing diagnostic and therapeutic interventions for infants with critical bronchiolitis more frequently than recommended by current bronchiolitis clinical guidelines, with interventions occurring more frequently for infants requiring invasive support. PICU providers agree on the need for additional research studies to help inform clinical guidelines specifically for infants with critical bronchiolitis.