FormalPara Take-home message

The GFRUP group presents a guideline of 40 recommendations regarding the management of infants under 12 months of age with severe bronchiolitis. Compared to the different ways to manage patients with severe bronchiolitis described in the literature, our original work proposes an overall less invasive approach in terms of monitoring and treatment.

Introduction

Bronchiolitis is a common lung infection in young children and infants. Every year, a large proportion of infants is affected [1,2,3]. Approximately one-tenth of these children is admitted [3], and between 2 and 6% of them present a severe form and are referred to pediatric intensive care units (PICUs) [1, 2, 4, 5]. In contrast to the mild to moderate forms, the management of this subgroup of patients is poorly codified by national [6] and international [7, 8] recommendations.

The definition of severe acute bronchiolitis is mainly clinical and based on low levels of evidence. In this work, we chose severity criteria from the English, Canadian and Australasian recommendations [7,8,9] remodeled in the French recommendations from the Haute Autorité de Santé (HAS) 2019 [6]. The presence of at least one of the following criteria defines severe bronchiolitis: poor general appearance, heart rate > 180/min or < 80/min, respiratory rate > 70/min or < 30/min or apneas, chest recessions and/or nasal flaring, poor food intake (< 50% in the previous 12 h), and SpO2 < 90% on room air.

Many of the patients who meet the criteria for severe illness are admitted to intensive care units. The purpose of this work is to provide a set of recommendations for the population of infants under 12 months of age who are admitted to these units. This initial management proposal should be adapted according to the infant's condition and evolution, as flexibility is a key feature of the working method used by the experts involved in developing the recommendations. In general, the paucity of high-quality studies in this field suggested the need to proceed with great humility in drafting them. It should also be noted that when a patient meets the diagnostic criteria for acute respiratory distress syndrome, readers should follow those specific recommendations [10, 11].

In this context, the Groupe Francophone de Réanimation et Urgence Pédiatriques (French‐speaking group of paediatric intensive and emergency care; GFRUP) has formalized expert recommendations on the management of acute severe bronchiolitis in infants under 12 months of age admitted to a PICU (Table 1). They are based on the analysis of the literature that targets this population and on the recommendations of expert representatives of the diversity of French-speaking ICUs. As no experts come from low-income countries, the implementation of these recommendations should be discussed and adapted to the particular conditions of these countries.

Table 1 2022: Table of recommendations
Full size table

Methods

These guidelines were drawn up by a group of experts convened by the GFRUP. The work has been registered on PROSPERO (number CRD42021236932). The group’s agenda was defined beforehand. The organizing committee first identified five domains of expertise (1) intensive care admission criteria, (2) environment and monitoring, (3) feeding and hydration, (4) ventilatory support, and (5) adjuvant therapies) and designated five expert-leaders (FB, RP, PD, SE, and GE). Within each expert group, the questions to be addressed were defined and validated by the experts. The questions were formulated according to a Patient-Intervention-Comparison-Outcome (PICO) format. The literature was analyzed, and the guidelines were formulated using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology [12]. SJ served as methodologist. A level of evidence was defined for each bibliographic reference cited according to the type of study, and this level of evidence could be reassessed in light of the methodological quality of the study. An overall level of proof was determined for each criterion, taking into account the level of evidence of each bibliographic reference, the between-study consistency of the results, the direct or indirect nature of the results, and cost analysis.

A high overall level of proof led to a “strong” recommendation (should be done: GRADE 1 + , should not be done: GRADE 1−). A moderate, low, or very low overall level of proof led to a “conditional” recommendation (should probably be done: GRADE 2 + , should probably not be done: GRADE 2− When the literature was inexistent or insufficient, the response to a question was a recommendation in the form of an expert opinion (the experts recommend…). The proposed recommendations were presented and discussed one-by-one by all the experts. Each expert then reviewed and anonymously rated each recommendation using a scale of 1 (complete disagreement) to 9 (complete agreement). The collective rating was established using a GRADE grid. To approve a recommendation, at least 50% of the experts had to be in agreement and less than 20% in disagreement. For an agreement to be strong, at least 70% of the experts had to be in agreement. In the absence of strong agreement, the recommendations were reformulated and rated again, with a view to reaching a consensus. Only expert opinions that elicited strong agreement were kept.

Literature search

A literature search was independently performed from MEDLINE databases via PubMed, Web of Science, Cochrane, and Embase. To be included in the analysis, publications had to be written in English or French and published after 2000. The main search terms used were bronchiolitis and infant/child/pediatric. The detailed search strategy, including terms and the algorithm, is available in Supplementary Material 1. In the absence of data published after 2000, older references were checked and added exceptionally.

Results

These guidelines encompass 40 recommendations (3 strong, 14 conditional, 23 expert opinions). For two questions, no recommendation could be applied. After one round of scoring and amendments, strong agreement was reached for all recommendations, which are summarized in Fig. 1 and Table 1.

Fig. 1
figure 1

Suggestion for severe bronchiolitis management. CPAP continuous positive airway pressure, HFNC high-flow nasal cannula, ICU intensive care unit, IV intravenous, NIV noninvasive ventilation, pCO2 partial pressure in carbon dioxide, RSV respiratory syncytial virus, SpO2 pulse oximetry, US ultrasound, WG weeks of gestation

Full size image

Recommendations

Intensive care admission criteria

Recommendation

R1—The experts suggest that the following severity factors be considered when deciding on PICU admission:

- Presence of apneas

- Hypercapnic acidosis with venous pCO2 > 60 mmHg and/or pH < 7.30 if blood gas testing is available

- Altered alertness and/or hypotonia

- Hypoxemia with SpO2 < 92% under standard oxygen therapy (i.e., nasal cannulas at low-flow up to 2 L/min)

- Significant increase in clinical work of breathing or respiratory fatigue

For patients already hospitalized, dynamic assessment of these parameters is important to consider (expert opinion, strong agreement)


Rationale


The literature mainly comprises retrospective data whose interpretations have been impacted by the significant changes in the management of severe bronchiolitis in recent years.


The clinical elements that seem to indicate the need for intensive care admission are the presence of apneas, hypoxemia and/or hypercapnia [13,14,15,16,17,18]. On the other hand, pCO2 thresholds to indicate PICU admission or ventilatory support initiation remain heterogeneous. According to the main studies conducted in this population [19,20,21,22,23,24,25,26,27,28,29,30,31], the mean venous or capillary pCO2 on admission is 59.6 mmHg for pH values < 7.30. For the other clinical parameters (respiratory rate, heart rate, SpO2), no threshold can be proposed.


Recommendation

R2—When available, additional factors that should probably be considered as risk factors for intensive care admission are:

- History of prematurity especially if < 32 weeks of amenorrhea

- Age below 6 weeks

- Identification of respiratory syncytial virus (RSV) as opposed to other viruses

- Evidence of viral or bacterial co-infection

- Presence of atelectasis or consolidation on radiography (if performed, see R6 recommendation) or lung ultrasound (posterior consolidation > 1 cm and/or high lung ultrasound score) (GRADE 2 + , strong agreement)


Rationale


For the most part, the factors of vulnerability are similar to those in patients admitted to wards [6].

The factors that appear to be more specifically associated with PICU admission are a history of prematurity [16, 32,33,34,35,36,37,38], identification of respiratory syncytial virus (RSV) as opposed to other viruses [38, 39], evidence of viral or bacterial co-infection [39,40,41], and the presence of atelectasis or consolidation on chest radiography (only if available according to recommendation R6) [13, 14, 36]. Lung ultrasound has demonstrated its utility as a bedside tool to stratify risk, and a dedicated score (LUSBRO) based on previous findings in bronchiolitis and the neonate was created [42]. Posterior or paravertebral consolidation > 1 cm or lung derecruitment assessment using different ultrasound scores was associated with PICU admission [13, 14, 36, 43,44,45,46].


The international guidelines emphasize the fragility linked to certain pathologies such as bronchopulmonary dysplasia, congenital heart disease with hemodynamic effects (shunt), neuromuscular pathologies, and immune deficiencies [7,8,9].


Recommendation

R3—The experts suggest the use of clinical scores for the initial patient assessment and follow-up (Expert opinion, strong agreement)


Rationale


The multiplicity of parameters to be considered when deciding on PICU admission justifies the use of scores. Although most of the scores are not specific to severe bronchiolitis and cannot predict the need for or the duration of ventilatory support, they are widely used as outcome or inclusion criteria in studies because they may provide a more objective evaluation (and follow-up) by decreasing inter-individual subjectivity [20, 22, 47,48,49,50]. Among the scores identified, the modified Wood's Clinical Asthma Score (mWCAS) [51] and the Wang score [52] appeared to be the most cited in the bronchiolitis studies, and the recent critical bronchiolitis score (CBS) [53] may be promising because it was developed specifically for severe bronchiolitis in the PICU.

Environment and monitoring

Paraclinical tests at the initial phase and during hospitalization

Recommendation

R4—Virological identification tests should probably not be performed (GRADE 2−, strong agreement)


Rationale


Although this criterion may be considered a risk factor for PICU admission, there is no evidence that virus identification provides an individual benefit regarding the disease course of RSV or other respiratory virus. There is also no evidence of a collective benefit of virological screening that would allow zoning of patients to avoid viral spread [54, 55]. Nevertheless, the identification of a virus can sometimes rule out differential diagnoses. It can also have epidemiological significance.


Recommendation

R5—The experts suggest that serum sodium level should not be systematically checked. However, the test should be performed if there are clinical signs suggestive of hyponatremia (e.g., altered alertness) (Expert opinion, strong agreement)


Rationale


Although hyponatremia is common in patients admitted with acute bronchiolitis [56, 57], there are no studies in this area. This recommendation is consistent with the Australian and American recommendations [8, 55].

Recommendation

R6—The experts suggest that chest radiography should not be routinely performed, but should be reserved for patients with clinical signs of ventilatory complications or for the search for differential diagnosis (Expert opinion, strong agreement)


Rationale


Although many children have radiographic chest abnormalities, these do not correlate with disease severity, although one retrospective study suggested an association between initial atelectasis and severity [58]. Fever was shown to be a good predictor of radiographic abnormalities [59]. On the other hand, it seems that the abnormalities identified on radiography are associated with a greater prescription of antibiotics without improving the course of the disease [60].


Recommendation

R7—The experts suggest that lung ultrasound be performed as an alternative to chest radiography (Expert opinion, strong agreement)

Rationale


Regarding lung ultrasound as a diagnostic tool, several clinical studies—recently quoted in the evidence-based guidelines from the POCUS Working Group of the European Society of Pediatric and Neonatal Critical Care (ESPNIC)—strongly suggested that the lung ultrasound score is helpful for descriptive purposes in viral bronchiolitis, with good concordance between operators [44, 61]. Two prospective observational studies strongly suggested that, unlike the chest X-ray, a high value of the lung ultrasound score recruitment score might well predict PICU admission and/or the need for respiratory support [42, 43].

Recommendation

R8—The experts cannot comment on the place given to blood gas testing in the initial assessment, although it is widely performed in practice. The experts suggest that clinical assessment should be the preferred method for judging the severity of the patient’s condition (Expert opinion, strong agreement)

Rationale


However, if blood gas testing is performed, we suggest capillary sampling, which is easier to perform and less painful than arterial sampling, given that the measurement of pH and pCO2 is reliable and well correlated with arterial sampling [62].

Patient conditioning and installation

Recommendation

R9—The experts suggest not systematically inserting a peripheral venous line (Expert opinion, strong agreement)


Rationale


Practice studies show a rate of peripheral venous line placement in 38–93% of the patients admitted to PICUs [55, 63, 64]. However, no comparative study has examined the management of severe bronchiolitis with and without peripheral venous access. The randomized trial published by Oakley et al. considered all forms of bronchiolitis regardless of severity [65].

The use of a peripheral venous line could be considered in cases of severe bronchiolitis with a probable need for hydration and/or intravenous drug therapy or in cases of emergency.


Recommendation

R10—Patients should probably be placed in the supine position (except for patients supported by invasive ventilation). Studies on prone positioning are needed (Grade 2−, strong agreement)

Rationale


Given the recommendations for the prevention of unexpected infant death and in the absence of data on the potential benefit of prone positioning, HAS 2019 recommends that infants with bronchiolitis be placed in bed flat on their back [6]. There is no evidence to justify a different recommendation for infants with severe bronchiolitis, apart from patients on invasive ventilation who should be positioned on their back with 30°–60° inclination, which is similar to the adult recommendations [66].

A physiological study compared prone versus supine positioning in children with severe bronchiolitis with noninvasive ventilatory support [67]. This study concluded that there is a probable benefit of the prone position, as it significantly improves the work of breathing. An ongoing study should identify a potential benefit of prone position in terms of ventilation time and hospital length of stay (NCT 03976895).

Patient follow-up and monitoring

Recommendation

R11—The experts suggest monitoring the trend of respiratory parameters and/or clinical scores to assess disease evolution (Expert opinion, strong agreement)


Rationale


Management of infants with acute severe bronchiolitis requires monitoring of respiratory (SpO2, respiratory rate) and hemodynamic (heart rate, blood pressure) parameters [53, 55, 68]. Studies performed on patients with severe bronchiolitis use these parameters as criteria of judgment and the evolution of these parameters as a factor of the failure or success of a treatment, suggesting their consideration for the clinical monitoring of patients [69, 70]. Similarly, clinical scores are used to assess the evolution of the disease and the effectiveness of management and/or therapy [69, 71, 72].


Recommendation

R12—The experts suggest that pCO2 (and PtcCO2) should not be systematically monitored (Expert opinion, strong agreement)

Rationale


Increased pCO2 appears to be a risk factor for transfer to the PICU and also helps assess disease progression [69, 73]. Notably, pCO2 has been shown to be a factor associated with high-flow nasal cannula (HFNC) failure [28, 74, 75]. However, no studies have compared routine pCO2 monitoring versus no monitoring on patient outcome improvement.

Transcutaneous CO2 measurement (PtcCO2) is feasible in clinical practice with good correlation with venous pCO2, although it requires at least one blood gas [76]. In a prospective double-blind study, it was shown that PtcCO2 did not have a significant predictive value for respiratory deterioration in patients [72].

The experts, therefore, suggest that pCO2 should not be systematically monitored, but that its measurement should be reserved for cases of clinical deterioration.

Feeding and hydration

Nutrition

Recommendation

R13—The experts suggest that gastric enteral nutrition (or oral nutrition when possible) should be preferred to intravenous hydration, regardless of the type of ventilatory support (Expert opinion, strong agreement)


Rationale


Several studies have shown that enteral feeding compared with intravenous hydration during a fasting period does not increase the risk of gastric distension, vomiting, or aspiration, regardless of the type of ventilation [32, 65, 77,78,79,80,81,82,83]. A cohort study in children with noninvasive ventilatory support (with and without bronchiolitis) showed a very low incidence of inhalation pneumonitis (1.5%) [84]. Oral or enteral feeding compared with fasting or interrupted feeding is also associated with a reduction in the hospital length of stay [65, 78, 79, 85], which is in line with recent European recommendations (ESPNIC) [86]. Despite the lack of data, oral nutrition may be associated with a higher risk of inhalation than gastric nutrition, especially for patients with an increased work of breathing.

However, it should be noted that most studies did not distinguish between hydration and oral/enteric nutrition versus intravenous hydration, which may make it difficult to interpret their results.


Recommendation

R14—There is no evidence to support continuous enteral nutrition over discontinuous nutrition (No recommendation, strong agreement)

Rationale


No studies have evaluated the superiority of continuous gastric administration over the discontinuous mode. The comparison is only valid for the gastric route because the choice of the transpyloric route requires a continuous flow [87,88,89]. Practices remain heterogeneous in this area [83, 87, 90].


Recommendation

R15—The experts suggest that oral or enteral nutrition should not be thickened (Expert opinion, strong agreement)

Rationale


No clinical studies have evaluated the effect of routine thickening of oral nutrition in infants with bronchiolitis. In patients supported by noninvasive or invasive ventilation, no studies have evaluated the thickening of enteral nutrition. The increased viscosity and osmolarity of thickened nutrition could potentially increase the risk of tube obstruction and affect gastric emptying time, which is already impaired in critically ill children.

Recommendation

R16—The experts suggest that the recommendations for energy and protein intake in the critically ill infant be implemented (Expert opinion, strong agreement)

Rationale


Fortified nutrition has been shown to achieve significantly higher energy and protein targets, but at the same time it increases the risk of overnutrition [88, 91]. A retrospective study showed a positive association between protein intake and hospital length of stay and ventilation [92]. Nevertheless, these results need to be kept in perspective because of numerous biases, and they are also at odds with the results of larger studies conducted in general intensive care populations [93, 94].

Thus, to date, in the absence of strong scientific evidence, the nutritional strategy proposed to patients with severe bronchiolitis may be identical to that recommended for critically ill children [86]. Pragmatically, intubated and sedated infants could be fed through a gastric tube with their usual milk (breast milk or formula adapted to the age and possible history), aiming at about 65% of the nutritional references for the population, i.e., about 65 kcal/kg/day, as the Schofield equations are unreliable for this weight range. If water restriction is necessary, fortification of breast milk or the use of infant formula pre-enriched in energy and protein would help limit enteral water intake without compromising nutritional objectives. No studies have assessed the energy requirements of infants but, based on expert opinion, continuous enteral nutrition on a gastric tube could be proposed for infants supported by noninvasive ventilation, with a target of 85–100 kcal/kg/day (or 120–130 mL/kg/day of formula). The total energy expenditure is probably less reduced in these infants compared to those on invasive ventilation support. Intakes could be increased later during the recovery phase according to the observed weight loss.

Hydration

Recommendation

R17—When enteral feeding is not possible, the experts suggest using an isotonic rather than a hypotonic solution. There is no evidence for choosing a balanced isotonic solution over an unbalanced solution (Expert opinion, strong agreement)


Rationale


Infants with bronchiolitis are at risk of hyponatremia [57], which may be increased by administration of hypotonic solution [56, 95].

In patients admitted for bronchiolitis, hyponatremia is associated with a poorer outcome, including longer hospital stay and increased risk of mechanical ventilation [96].

For this reason, and although no randomized controlled trials have compared the two types of intravenous fluids in infants admitted to the PICU for bronchiolitis, it seems reasonable to use isotonic intravenous fluids as a maintenance infusion in children for whom enteral nutrition is not possible or has been delayed, in accordance with American and British recommendations [7, 97].

Furthermore, the superiority of a balanced crystalloid solution over an unbalanced solution has not been evaluated. The potentially deleterious effects of isotonic saline put forward as a reason for preferring balanced solutions relate primarily to their use as a vascular filling solution in the perioperative period or in states of shock, and not as a maintenance solution.


Recommendation

R18—The experts suggest monitoring the fluid balance (input/output balance and/or weight) on a daily basis to avoid fluid overload (Expert opinion, strong agreement)


Rationale


In patients with bronchiolitis, fluid overload at 24 h of admission is associated with increased duration of mechanical ventilation, as well as longer ICU and hospital length of stay, regardless of the route of fluid administration (enteral or intravenous) [31].

Children admitted to intensive care are also at risk of energy and protein deficits, particularly because of a deficit in nutritional intake that is exacerbated by the associated fluid restriction [98]. However, in a pilot study including 23 children ventilated for respiratory infection, a strategy of restricted fluid intake was feasible without limiting protein and caloric intake compared to a more liberal strategy [99].

To our knowledge, no randomized controlled trial has evaluated the value of monitoring fluid balance in children admitted to a PICU for bronchiolitis.

Ventilatory support

Recommendation

R19—The experts suggest the use of a noninvasive ventilatory support protocol (Expert opinion, strong agreement)


Rationale


Four “before-and-after” studies specifically evaluated the value of implementing a protocol for HFNC use in the PICU [26, 28, 100, 101]. The results are homogeneous and these studies, although retrospective, confirm the positive impact of using the protocol on patient outcome [102, 103].

Noninvasive ventilation

Recommendation

R20—Noninvasive ventilatory support is effective to reduce the work of breathing and improve clinical respiratory parameters (GRADE 1 + , strong agreement)


Rationale


Data from eight studies (four on continuous positive airway pressure, CPAP [19, 21, 25, 104] and four on high-flow nasal cannula, HFNC [105,106,107,108]) are convergent and show a significant and lasting reduction in the markers of respiratory failure (respiratory frequency, pCO2) and work of breathing (esophageal pressure–time product) with noninvasive ventilatory support.

A physiological study [25] evaluated different levels of positive airway pressure and showed greater effectiveness in terms of reduction in the work of breathing with a CPAP setting of 7 cmH2O. With regard to HFNC, the work by Milési et al. and Weiler et al. showed that the physiological impact was greater when the flow was close to 2 L/kg/min [105, 108].


Recommendation

R21—The experts suggest noninvasive ventilatory support as a first-line treatment rather than invasive ventilation (Expert opinion, strong agreement)


Rationale


The widespread use of noninvasive ventilatory support as the first-line therapy for managing bronchiolitis reinforces this recommendation despite the lack of strong data [30]. Nowadays, randomized studies comparing noninvasive support versus intubation at the outset are no longer ethically conceivable in view of the good tolerance of CPAP, noninvasive ventilation (NIV) and HFNC described in the literature. It is probably for this reason that there are no randomized prospective studies comparing noninvasive support versus intubation.

One retrospective study showed that management was less invasive and ICU length of stay was shorter in children supported by CPAP compared with intubated patients [109], which has been supported by the results from other studies [110,111,112]. Two other studies, retrospective and with a limited number of patients, did not show a more rapid clinical improvement in patients who received CPAP or NIV compared with patients who were initially intubated [29, 113].

There are no studies comparing the use of HFNC with invasive ventilation.


Recommendation

R22—For the most severe form, continuous positive airway pressure should probably be used as the first-line treatment rather than high-flow nasal cannula (GRADE 2 + , strong agreement)

Rationale


Four randomized controlled trials in the literature compared the efficacy and safety of HFNC versus CPAP. In the largest one, Milési et al. found a higher failure rate of HFNC (51 versus 31%) in a multicenter study [20]. In this work, it is important to note that a large majority of the CPAP failures was explained by poor tolerance of the device, whereas in the HFNC group, support failure was mainly due to respiratory deterioration.

The other three studies, of small size and with variable criteria of judgment, did not confirm these results [71, 114, 115], which explains the conclusions of a recent meta-analysis on this question [116].


Recommendation

R23—Continuous positive airway pressure should probably be initiated at a positive pressure level of 7 cmH2O (GRADE 2 + , strong agreement)


Rationale


Only one physiological study evaluating different levels of positive pressure demonstrated greater efficacy for a CPAP setting of 7 cmH2O [25]. But most data on bronchiolitis in the PICU were based on levels of pressure support close to 7 cmH2O, with significant efficiency and a low adverse effect rate.


Recommendation

R24—The experts suggest the use of noninvasive ventilation with two pressure levels in cases of failure of continuous positive pressure and in the absence of intubation criteria (Expert opinion, strong agreement)


Rationale


There are no studies in the literature that randomly compare two levels of pressure in NIV to another noninvasive ventilation modality. The data are limited to three retrospective studies evaluating the risk factors for failure using this mode [27, 117, 118]. These studies highlight that this technique can be used as initial support or as rescue, with greater efficiency in terms of the reduction in oxygen requirements compared to CPAP [117]. However, we cannot specifically recommend two pressure levels of ventilation. Delacroix et al. suggested that the use of NIV at two pressure levels could be associated with unfavorable outcomes, but the retrospective design of the study makes the interpretation of these results difficult [27].


Recommendation

R25—The choice of interface should take into account the ventilator being used and the expertise of the medical team. The experts are not in a position to suggest a specific type of interface for patients ventilated with continuous positive airway pressure.

In cases of failure, the experts suggest the use of a face mask to improve patient-ventilator synchronization (Expert opinion, strong agreement)


Rationale


A European survey of practice showed that the face mask was widely used [119]. However, only two small single-center randomized controlled trials [120, 121] compared two types of CPAP interfaces. The first compared the helmet versus the face mask, the second the helmet versus nasal cannulas. In both trials, there was no difference in terms of respiratory physiological parameters or failure, but there seemed to be more discomfort with a face mask.

Last, two retrospective studies evaluated the feasibility and predictors of failure of the cannula interface and nasopharyngeal tubes in different noninvasive ventilatory modes [122, 123], but these two studies did not compare these interfaces to others. Patient comfort and tolerance should be important factors in the choice of the interface.

High-flow nasal cannula

Recommendation

R26—The high-flow nasal cannula should not be used prophylactically to reduce the risk of admission to the PICU (GRADE 1−, strong agreement)


Rationale


Two large randomized control studies failed to demonstrate the benefit of prophylactic HNFC to avoid admission in PICU [124, 125]. This was confirmed by a recent meta-analysis that evaluated the impact of using HFNC to manage bronchiolitis outside of the PICU on short-term patient outcomes [116]. On the admission to PICU outcome, they reported 1127 patients in the HFNC group and 1096 patients in the standard oxygen therapy group and no significant difference in PICU admissions (OR = 1.1 [0.81–1.42]). Despite some superiority of HFNC in terms of treatment failure on the wards, most data suggest that prophylactic use of HNFC does not modify the underlying disease process in moderately severe bronchiolitis [116, 126].


Recommendation

R27—A flow rate of 1.5–2 L/kg/min should probably be initiated with high-flow nasal cannula and should not exceed 2 L/kg/min (GRADE 2 + , strong agreement)


Rationale


Physiological studies have demonstrated that the efficiency of HNFC to improve the work of breathing increases with a flow rate close to 2 L/kg/min [105,106,107,108]. A higher flow rate (3 L/kg/min versus 2L/kg/min) failed to demonstrate better efficiency and was associated with greater discomfort [22].

Invasive ventilation

Recommendation

R28—The experts are not able to make a recommendation regarding the choice of invasive ventilation mode (No recommendation, strong agreement)


Rationale


No randomized trials have evaluated the short-term outcome of infants intubated for bronchiolitis based on ventilation mode, particularly for neurally adjusted ventilatory assist (NAVA) mode.

Transport and ventilator support

Recommendation

R29—Noninvasive ventilatory support (continuous positive airway pressure or nasal high flow) should probably be used during patient transport.

If a high-flow nasal cannula is used, a humidification heating system should be used.

The experts do not recommend routine intubation for transport (GRADE 2 + , strong agreement).


Rationale


Several teams have reported their experience using noninvasive ventilatory support for transport of patients with bronchiolitis [127,128,129,130,131,132]. All the studies agreed on the safety of the technique during transport and a decreased use of invasive ventilation after PICU admission. Schlapbach et al. showed a significant reduction in the use of intubation for transport using HFNC without affecting intensive care management [130]. A review of the literature on the adverse effects of noninvasive ventilatory support for transport reported the need for intubation during transport in only 0.4% of cases and the use of invasive ventilation within 24 h of transport in 10% of cases (n = 63/650). Other adverse events were anecdotal in this study [133].

Adjuvant therapies

Sedation

Recommendation

R30—The experts suggest that non-medicinal measures (parental presence, installation, optimization of the interface and ventilation, feeding) should be preferred to improve the infant's comfort. If these measures fail, drug sedation may be used (Expert opinion, strong agreement)


Rationale


The discomfort associated with noninvasive ventilatory support in infants with bronchiolitis is common [20, 22, 120, 121] and is the leading cause of ventilation failure [20, 120]. The presence and empowerment of the parents is encouraged in the PICU [134]. No randomized study has investigated the value of sedation in improving the efficacy and tolerance of noninvasive support in this population. The use of hydroxyzine [20, 22] and midazolam [120, 121] has been reported with good tolerance. In two pediatric cohorts (including 30% of patients with bronchiolitis), dexmedetomidine improved the tolerance of noninvasive ventilatory support at the cost of mild side effects such as bradycardia and hypo/hypertension, which resolved when the dosage was reduced [135, 136]. Mild sedation that does not impair airway patency, respiratory drive or muscle strength could, therefore, be of interest to optimize the efficacy of noninvasive ventilatory support in cases of failure of non-drug measures.

Place of corticosteroids

Recommendation

R31—Corticosteroids, whether inhaled or systemic, should probably not be used (GRADE 2, strong agreement)


Rationale


In 2013, a Cochrane review showed that there was no benefit to the use of corticosteroids regarding hospital length of stay [137]. Importantly, the data did not include patients admitted to the PICU.

In our specific population, two studies were performed. The first, including 22 patients and comparing dexamethasone with placebo, showed no difference in cytokine concentration between the two groups, which would potentially explain the lack of a clinical benefit of corticosteroids [138]. The second study, evaluating the same molecule with a higher dose, was stopped prematurely because of a recruitment problem. It did not find any beneficial effect of corticosteroids in terms of duration of invasive ventilation or duration of hospitalization in intensive care [139].

Place of bronchodilators

Recommendation

R32—Inhaled beta2-agonists and/or inhaled epinephrine should probably not be used routinely. In some cases, a therapeutic test could be considered to avoid intubation (GRADE 2−, strong agreement)


Rationale


Only one study looked at patients admitted to the PICU [140]. In this small randomized controlled trial, the authors compared adrenaline, salbutamol and norepinephrine aerosols. The primary endpoint was the rate of hospitalization at day 7. No difference was found in any group compared with placebo.

Five randomized controlled studies compared nebulized epinephrine or beta2-agonists versus placebo [141,142,143,144]. No effect of these treatments was found on hospital length of stay, hospitalization rate, or clinical signs of respiratory distress.

Similarly, in a study comparing inhaled epinephrine with dexamethasone versus placebo, no significant effect on admission rate was found [145]. This was confirmed by the work by Kua et al. [146].

In two studies by the same team comparing adrenaline aerosol combined with 3% saline versus placebo combined with 3% saline, the results were contradictory for the main outcome, which was hospital length of stay [147, 148].

Recommendation

R33—The experts suggest that intravenous beta2-agonists should not be used (Expert opinion, strong agreement)


Rationale


We did not identify any studies evaluating the place of this therapy in our population.


Recommendation

R34—Intravenous theophylline should not be used (GRADE 1−, strong agreement)


Rationale


Only one randomized study on this issue was identified, by Turner et al. [149]. The authors compared continuous infusion of theophylline versus placebo in a population with bronchiolitis requiring ventilatory support. The trial was stopped early because of insufficient enrollment. However, the results obtained did not show a positive effect of theophylline in terms of the duration of ventilation required. It should also be noted that the adverse effects of this molecule are not negligible and that the therapeutic index is narrow.

Place of antibiotics

Recommendation

R35—The experts suggest that antibiotic therapy should not be systematically used but should be reserved for suspected pulmonary superinfection or bacterial co-infection (Expert opinion, strong agreement)


Rationale


Apart from severe forms of bronchiolitis, the use of antibiotics has not been shown to be of interest [150,151,152,153]. Bacterial infections are less frequent in infants with bronchiolitis than in infants with fever without bronchiolitis, but their incidence can reach 10–25% depending on the severity status [154,155,156]. Risk factors of bacterial infection in this population include the severity of the respiratory distress, higher general severity scores, and the need for invasive ventilation and vasoactive agents [155, 157, 158], which explains the wide use of this class of drugs in this population [159]. However, the use of antibiotics is not without consequences [160] and should, therefore, be reserved for suspected pulmonary superinfection or bacterial co-infection.

If antibiotic therapy is warranted for suspected pulmonary superinfection, it is recommended that amoxicillin (100 mg/kg/day every 8 h) be used as the first-line therapy in the absence of microbiological documentation, taking into account the child's clinical data.

Place of caffeine in bronchiolitis with apnea

Recommendation

R36—Caffeine should probably not be used routinely in patients presenting with apnea (GRADE 2−, strong agreement)


Rationale


Apneas are frequent in this population [161], and caffeine has sometimes been used, by analogy with apneas in preterm infants, although the pathophysiological mechanism of apneas in viral bronchiolitis is different. The biological plausibility of this intervention is not clear, and the data in the literature are limited. Of the four studies reviewed, the study of Alansari et al., with about 90 patients under four months of age, was the only one that was prospective and randomized [162]. This study did not find a beneficial effect of a single dose of 25 mg/kg of caffeine on the occurrence of apneas, the need for ventilatory support, or the duration of hospitalization. However, this study was criticized for the randomization technique and the subjective method of apnea detection.

The other three observational studies [163,164,165] involved small numbers of patients under three to 12 months of age, with doses ranging from 10 to 25 mg/kg of caffeine. These studies also concluded that there was no beneficial effect on the prevention of mechanical ventilation or the hospital length of stay.

In the absence of a prospective randomized multicenter study targeting the population under 2 months of age with a protocol including maintenance doses, it is nevertheless difficult to affirm the non-effectiveness of caffeine for the prevention or treatment of apneas in this population.

Place of helium

Recommendation

R37—Helium should probably not be used. Despite the beneficial effects on the work of breathing, feasibility makes it difficult to implement (GRADE 2−, strong agreement)


Rationale


In 2015, a Cochrane systematic review with meta-analysis evaluated the value of a helium–oxygen mixture (heliox) in severe bronchiolitis [166]. Seven randomized controlled trials were included in the systematic review, with a total of 447 infants with acute bronchiolitis. No beneficial effect was found in terms of use of ventilation (invasive or not), intubation rate, or PICU length of stay. On the other hand, a significant effect on clinical signs of respiratory work was observed. Subsequently, a single additional randomized trial was published [167], without modifying the conclusions of the meta-analysis. The clinical benefit observed at two hours was no longer observed at 24 h.

It should also be noted that the use of this therapy is complicated by its low availability in routine practice and its cost.

Place of inhaled hypertonic saline

Recommendation

R38—Nebulization of hypertonic saline should probably not be used routinely (GRADE 2−, strong agreement)


Rationale


Hypertonic serum nebulization has been studied in bronchiolitis because it can decrease epithelial edema and improve mucociliary transport. A Cochrane review [168] and a meta-analysis [169] compared hypertonic saline nebulization versus isotonic saline, with a beneficial effect of the hypertonic saline in a population of children with non-severe forms. Nevertheless, the level of evidence remains low due to a high level of statistical heterogeneity and a risk of selection bias for some of the included studies. The very heterogeneous results can be explained in part by differences in the frequency of administration and the concentration of hypertonic saline (3, 6 or 7%) that was tested.

These findings are in contrast to those of two randomized trials, which also excluded severe forms and found no benefit [170, 171].

Only one retrospective study focused on patients with a severe form [172]. The authors showed a trend toward a reduction in the duration of ventilatory support for patients ventilated on CPAP, but comparability of the groups was limited and the analyses were conducted in very small subgroups.

Place of respiratory physiotherapy

Recommendation

R39—Respiratory physiotherapy should probably not be systematically performed (GRADE 2−, strong agreement)


Rationale


Despite a potential benefit of respiratory physical therapy on tidal volume and oxygen requirements [173], a meta-analysis published in 2016 and including 12 prospective randomized studies concluded that none of the respiratory physical therapy techniques were efficacious regarding vital parameters or hospital length of stay [174].

The prospective randomized study by Gajdos et al. confirmed this finding in a population of 496 infants [175]. As some patients may respond to this therapy, a therapeutic trial may be acceptable (e.g., in cases with severe atelectasis or when nasopharyngeal de-obstruction is difficult). A prospective study, underway since 2020, is investigating the impact of individualized respiratory physiotherapy that may involve external vibratory devices on infants hospitalized for bronchiolitis (NCT04553822).

Place of ribavirin, DNAse, antileukotrienes, magnesium sulfate, and inhaled nitric oxide

Recommendation

R40—Ribavirin, deoxyribonuclease (DNAse), antileukotrienes, magnesium sulfate, and inhaled nitric oxide should probably not be used (GRADE 2−, strong agreement)


Rationale


Three randomized trials evaluating ribavirin were identified in the literature. The first suggested a reduction in the duration of ventilation, but with a particularly long ventilation time in the control group (10 days), which received potentially deleterious water aerosols as placebo [176]. The other two low-powered studies did not show any significant difference in ventilation and length of stay [177, 178]. It should be noted that ribavirin aerosols may cause complications, in particular obstruction of intubation tubes.

A Cochrane systematic review of deoxyribonuclease (DNAse) use in moderate forms of bronchiolitis was published in 2012 [179]. No benefit was found, and the control group even had a shorter hospital length of stay and a more favorable course of respiratory symptoms. Since no studies have been performed specifically in patients with severe bronchiolitis, the effect of DNAse remains to be determined in this population.

Antileukotriene therapy has not been shown to be effective in mild to moderate acute bronchiolitis [4]. No studies have been performed in patients with severe bronchiolitis.

Regarding magnesium sulfate, we identified four randomized controlled studies evaluating its efficacy in which no positive effect was found in terms of length of hospitalization [180,181,182], although Kose et al. observed an improvement in respiratory symptoms [183].

We identified two randomized controlled studies from the same team comparing the effect of nitric oxide “pulses” versus placebo [184, 185]. No beneficial effect was identified for any of the endpoints, although a post hoc analysis showed a shorter duration of hospitalization in patients hospitalized for more than 24 h.

Future directions

These guidelines highlight a lack of high-level evidence regarding certain crucial points in the management of severe bronchiolitis. Some topics that may seem obvious in daily practice are nevertheless not so obvious in the current literature.

We, therefore, felt it necessary to consider the following research topics to allow the scientific community to take ownership of them and provide new evidence for future revision of these guidelines:

  1. 1.

    Intensive care admission criteria: validation of a clinical score to facilitate and standardize the severity rating and the initial orientation of the patient and to follow the response to the ventilatory support;

  2. 2.

    Environment and monitoring: validation of lung ultrasound in relation to the prognosis and evolution of the disease; pCO2 measurement in relation to the severity rating, ventilation strategy and follow-up; assessment of patient positioning, in particular the prone position;

  3. 3.

    Feeding and hydration: interest of continuous versus discontinuous enteral nutrition, natremia monitoring and management of dysnatremia;

  4. 4.

    Ventilatory support: benefit of indication for two pressure levels in NIV (including BiPAP or NAVA) compared to CPAP in patients requiring NIV, rapid identification of patients who are failing on NIV, and the best strategy to wean the patient from noninvasive ventilatory support;

  5. 5.

    Adjuvant therapies: comparison of sedative approaches.

Conclusion

These guidelines cover the different aspects in the management of severe bronchiolitis in infants admitted to pediatric critical care units. Given the low overall level of evidence, most recommendations are either conditional or expert opinion. Compared to the different ways to manage patients with severe bronchiolitis described in the literature, our original work proposes an overall less invasive approach in terms of monitoring and treatment. Importantly, these guidelines are not in contradiction with the ESPNIC guidelines on respiratory support [11] or with the other guidelines dealing with all forms of bronchiolitis regardless of severity. Finally, this work is the first step before working on the implementation of these recommendations in daily practice.