key: cord-0035164-h3cjylcl
authors: Javouhey, Etienne; Pouyau, Robin; Massenavette, Bruno
title: Pathophysiology of Acute Respiratory Failure in Children with Bronchiolitis and Effect of CPAP
date: 2013-05-29
journal: Noninvasive Ventilation in High-Risk Infections and Mass Casualty Events
DOI: 10.1007/978-3-7091-1496-4_27
sha: 88097c1e961f5a0154fd54b2d56c0112e30712d3
doc_id: 35164
cord_uid: h3cjylcl

Acute bronchiolitis is the most common lower respiratory tract infection (LRTI) during the first year of life. Respiratory syncytial virus (RSV) infection is the most prevalent virus found in these children, accounting for 60–80 % of cases. The rate of hospitalization is less than 2 %. Up to 8 % of those hospitalized require ventilatory support [1, 2].

Prematurity, young age, and preexisting chronic respiratory and cardiac diseases are the main risk factors for severe bronchiolitis [ 1 , 6 ] . When the WOB increases or persists for a long period, ventilatory support is required to prevent severe hypoxemia or hypercapnic coma.

Noninvasive ventilation (NIV) is most often delivered by continuous positive airway pressure (CPAP) via nasal prongs or mask. Some studies have reported the use of assisted spontaneous breathing (ASB) and biphasic positive airway pressure (BIPAP) as other NIV modalities [ 7 -12 ] . High-fl ow cannulas (HFCs) deliver positive end-expiratory pressure (PEEP) that can attain 3 or 5 cmH 2 O, and some studies suggested that it could obviate the need for endotracheal intubation [ 7 , 8 , 13 -15 ]. There is no strong level of evidence that NIV avoids intubation and is benefi cial for patients compared to intubation. During the last decade, however, increasing numbers of clinical and physiological studies have reported a good experience of NIV as the primary ventilatory support mode. Currently, CPAP is widely used as the fi rst ventilatory support in many centers, with a decreasing rate of intubation.

The objective of this chapter is to summarize the impact of NIV techniques in the management of children with severe bronchiolitis requiring ventilatory support. Physiological knowledge is fi rst discussed followed by clinical studies assessing the impact of NIV on the intubation rate and outcome. We then address the technical and practical aspects of NIV application in children according to age and clinical condition.

In a physiological study of 37 infants, Hammer and colleagues showed that RSV infection could lead to two pulmonary function abnormalities [ 5 ] . The most common is bronchiolitis, an obstructive airway disease characterized by increased airway resistance (respiratory system resistance, R rs ), air trapping [high functional residual capacity/total lung capacity (FRC/TLC)], reduced TLC, and low respiratory system compliance ( C rs ) compared with normal values. Typically, chest radiography of these children shows bilateral perihilar infi ltrates and hyperinfl ation. Of the 37 infants, 10 had a resistive profi le, corresponding to the criteria of acute respiratory distress syndrome (ARDS), with very low C rs and R rs . Radiography revealed bilateral alveolar consolidations. This form corresponded to RSV pneumonia. The mechanism of apnea associated with RSV infection is not completely understood. Immaturity of central ventilatory centers is likely to be one of the explanations, which can explain the high prevalence of apnea in infants born prematurely and infants <2 months of age. The real incidence of apnea varies among studies, from 2.5 to 28.0 %, depending on the case mix. Among children admitted to the pediatric intensive care unit (PICU) the incidence is much higher [ 3 ] . It is probably important to distinguish primary apnea from apnea occurring after several hours of respiratory distress and a high level of WOB. The latter is likely to be due to muscle fatigue, which occurs more rapidly in young infants and those born prematurely.

Obstruction of small airways is the main physiological phenomena of RSV infection in infants. It is the consequence of bronchial and peribronchial infl ammation, plugging of airways by mucus and cellular debris, and bronchial smooth muscle constriction ( Fig. 27.1 ) . Consequently, the airway resistance and respiratory load increase. To preserve their pulmonary function, infants use their accessory respiratory muscles, increasing their WOB. They also increase their respiratory rate, but because of airway obstruction the duration of the expiratory period is too short to expire completely. Air is then trapped in the alveoli, generating dynamic hyperinfl ation and auto-PEEP. The inspiratory time/total respiratory time ( T i / T tot ) ratio increased as T i decrease ( Fig. 27.1 [ 7 , 16 , 17 ] . These measures were obtained using an esophageal and gastric probe with balloons. The PTP es per breath was obtained by measuring the area under the diaphragmatic pressure ( P di ) and the esophageal pressure ( P es ) signal between the onset of inspiration and the end of inspiration. Essouri et al. showed in ten infants with severe bronchiolitis that the median level of auto-PEEP generated was 6.05 cmH 2 O (range 3.9-9.2 cmH 2 O) [ 16 ] . They showed that the decrease in WOB was greater with a CPAP level of 7 cmH 2 O compared to 4 or 10 cmH 2 O. This suggested that application of extrinsic PEEP decreased the pressure gradient between the mouth and alveoli at end-expiration. It allowed air to pass through the airways, reducing the work required for the next inspiration. The Montpellier team in France confi rmed these results and the role of CPAP in a randomized controlled trial (RCT). They compared ten children treated with nasal CPAP at 6 cmH 2 O to nine children managed with oxygen alone [ 17 ] . The T i / T tot and transcutaneous carbon dioxid partial pressure (TcPCO 2 ) were decreased in the CPAP group and the WOB was signifi cantly reduced compared to that in control patients. This improvement of the WOB was correlated with the clinical improvement assessed by the modifi ed Wilson Clinical Asthma Score (mWCAS).

Beasley and Jones [ 18 ] and Soong et al. [ 19 ] were the fi rst to report NIV use, especially CPAP, in infants with severe bronchiolitis. These preliminary studies showed that CPAP was able to decrease the PaCO 2 and RR of infants with severe bronchiolitis. Since 2004, numerous prospective or retrospective studies have been published and reported an increasing use of CPAP in this clinical setting (Table 27. 2 ) [ 8 -12 , 17 , 19 -21 ] . However, there is no clear consensus on clinical use of CPAP compared to intubation and invasive ventilation [ 22 ] .

Complications of intubation and mechanical ventilation are well known. In infants, endotracheal intubation can be complicated by subglottic edema with a risk of evolution to tracheal stenosis. Mechanical ventilation of children with severe airway obstruction is challenging and may expose the airways and the lungs to high pressures or high volumes, causing lung injury. Most of infants who are mechanically ventilated require sedative drugs and sometimes muscle paralysis. The need for central venous access is common. These invasive procedures are associated with blood loss and expose children to nosocomial infection (e.g., pneumonia, urinary tract infection, bacteremia). The safety of sedative drugs in immature brains is not completely established. Hence, NIV represents a good alternative because its use rarely requires sedative drugs. On the other hand, the main risks of NIV are pneumonia aspiration in a child with an altered level of consciousness and potentially delayed intubation.

Most of studies have confi rmed the results of early studies that CPAP and NIV improve gas exchange and the RR. Physiological studies suggested that application of CPAP improved gas exchange by decreasing the WOB and respiratory efforts, as assessed by the mWCAS. Median LOS was reduced in children in whom NIV succeeded compared to those with IV and those who failed NIV Few studies have compared this approach to the classic invasive ventilatory strategy on clinical outcome, such as the duration of ventilatory support, length of PICU stay, length of hospital stay (LOS), or ventilator-assisted pneumonia (VAP). Javouhey et al. in a pre/post study design showed that NIV as the primary ventilatory support was associated with a signifi cant decrease in the intubation rate: from 89 to 52 % [ 10 ] . The NIV failure rate was 33 %. This approach was associated with a decreased incidence of VAP and a decrease in the number of children with oxygen requirement for >8 days [ 10 ] .

Ganu et al. reported their 10-year experience of NIV for infants with severe bronchiolitis in their PICU from The Children's Hospital at Westmead in Sydney. Among the 520 infants admitted for bronchiolitis, 399 required ventilator assistance-285 with a trial of NIV, mainly CPAP [ 9 ] . They reported a signifi cant increase in the use of NIV (2.8 % increase per year) along with a decline in the intubation rate (1.9 % per year). The percentages of infants failing NIV decreased over the study period, from 31.8 to 13.5 %. This decline was also observed in centers in which NIV was widely used as the primary mode of ventilatory support [ 9 ] . In our center, for example, this percentage decreased from 33 to 5 % during the 2011-2012 epidemics (personal data). The median hospital LOS was longer for infants who were intubated and invasively ventilated than for those in whom NIV succeeded. The hospital LOS was also signifi cantly longer for children who failed NIV than for those with invasive ventilation. The same tendency had been found in a previous study [ 10 ] . Even with no control study, these results suggested that a strategy using NIV (mainly CPAP) as the primary ventilatory support was able to obviate the need for tracheal intubation.

More recently, a system of oxygen delivery was developed using heated and humidifi ed high-fl ow gases delivered via nasal cannulas that can generate PEEP. The level of PEEP provided depends of the fl ow and the leaks but can reach 3-5 cmH 2 O. This system has been used in children with severe bronchiolitis, with results similar to those achieved with CPAP, including improved alveolar ventilation and decreased RR, obviating the need for tracheal intubation [ 14 , 15 , 23 ] (Table 27. 3 ). Physiological studies showed that high-fl ow cannulas (HFCs) are able to improve lung mechanics and ventilatory function by a washout of the nasopharyngeal dead space, a decrease in airfl ow resistance, and improved mucociliary clearance [ 24 ] .

An RCT pilot study performed in 19 infants with moderately severe bronchiolitis showed that heated/humidifi ed HFC therapy at 4-8 L/min improved the SpO 2 compared to a head-box oxygen group at 8 h (100 % vs. 96 %, p = 0.04) and 12 h (99 % vs. 96 %, p = 0.04) [ 13 ] . The stability of PEEP is not guaranteed and the level of PEEP reached can be insuffi cient to counterbalance the WOB. An in vitro study from Sivieri et al. showed that the airway pressure varied widely with the degree of nares occlusion by the prongs and by the amount of mouth leakage [ 25 ] . At 6 L/min HFC with the mouth open the airway pressure was <1.7 cmH 2 O. It was <10.0 cmH 2 O when the mouth was closed. Complete nares occlusion can generate high airway pressure (up to 20 cmH 2 O) when the mouth is closed. Further studies comparing CPAP and HFCs would be useful to understand which children should benefi t from CPAP rather than HFC. The latter system has the advantage of being simple to apply, usable in emergency units, and minimally expensive. Criteria used to initiate HFC or CPAP should be better defi ned and validated. A selection bias cannot completely be excluded because the level of severity of the infants treated is diffi cult to compare among studies. Moreover, as criteria to initiate ventilatory support are not well defi ned, those used in the various studies are likely to be different. Some authors have included children with severe respiratory distress and severe hypercapnic acidosis, whereas others have put children on ventilatory support considering only the signs of retraction or the level of tachypnea.

Criteria to initiate ventilatory support in children are not well defi ned and have not been validated. Most epidemiological studies have shown that infants with low weight and age < 42 days were more likely to be admitted to a PICU and ventilated.

Other factors predisposing to mechanical ventilation were factors linked to a medical history of lung and cardiac diseases, prematurity, and/or neuromuscular disease [ 2 , 9 , 20 ] . Evans et al. analyzed criteria for CPAP requirement in a retrospective cohort of 163 patients admitted to their center for severe bronchiolitis [ 20 ] . Among these 163 children, 28 required CPAP. The authors found seven predictors for CPAP requirement: young age, low gestational age, low SpO 2 , high level of oxygen requirement, respiratory and heart rates (RR, HR), and Glasgow Coma Score (GCS). Using receiver operator characteristic (ROC) curve analyses, they identifi ed several thresholds: age < 11 weeks, SpO 2 < 95 %, RR > 54, HR > 163, and GCS < 15. The strongest predictor was a low SpO 2 . The authors found a negative correlation between SpO 2 and O 2 requirement ( r = −0.656), a positive correlation between age and weight ( r = 0.836), and a positive correlation between gestational age and birth weight ( r = 0.824). They did not fi nd blood gas analyses as predictors of CPAP requirement [ 20 ] . Their results were limited by the retrospective nature of the study and by the small sample size.

Mansbach et al. identifi ed factors associated with CPAP and/or intubation requirement in a prospective multicenter study that included 161 children [ 26 ] . In the multivariate analysis, factors associated with CPAP and/or intubation requirement were age < 2 months [odds ratio (OR) 4.3, 95 % confi dence interval (CI) 1.7-11.5], maternal smoking during pregnancy (OR 1.4, 95 % CI 1.1-1.9), birth weight < 5 lb (OR 1.7, 95 % CI 1.0-2.6), breathing diffi culty began <1 day before admission (OR 1.6, 95 % CI 1.2-2.1), severe retractions (OR 11.1, 95 % CI 2.4-33.0), and room air SpO 2 < 85 % (OR 3.3, 95 % CI 2.0-4.8) [ 26 ] . Identifying patients at high risk of CPAP requirement is important because it can help the physician's decision about transferring the patient to the unit able to initiate the ventilatory support required.

Curiously, blood gas analyses have not been found to be good indicators of ventilatory support requirement except in the study of Campion et al., where a high level of CO 2 before CPAP was predictive of NIV failure defi ned as the need for invasive ventilation [ 8 ] . Similarly, composite scores of respiratory distress failed to identify the group of patients requiring ventilator support. In two French studies, high PRISM scores were predictors of the need for invasive ventilation. However, as this score is calculated 24 h after admission, it cannot help the physician make clinical decisions [ 8 , 11 ] .

The criteria for initiating CPAP should differ from those used to initiate invasive ventilation. Unfortunately, no reported studies have made such a comparison of these criteria. Therefore, the ventilatory strategy for children admitted with severe bronchiolitis is based on little evidence. CPAP and HFC can be proposed as fi rstline ventilatory support in most cases, although HFC is probably insuffi cient in children with severe hypercapnic acidosis. Response to this fi rst line of ventilatory support must be assessed within the fi rst 2 h following its initiation. Nonresponders are at high risk of complications and often require invasive ventilation. NIV in BiPAP or ASB mode or in pressure control mode can be attempted provided that rapid assessment is done and strict supervision is observed.

For better selection of patients who will respond to CPAP, some studies have assessed risk factors of NIV failure. Most of these studies were retrospective and compared patients whose ventilatory support was NIV alone versus those who were intubated after an NIV trial [ 8 , 10 , 20 , 27 -29 ] . Failure was defi ned as the need for intubation. Most of these studies included children with all types of respiratory distress, not bronchiolitis alone. The level of FiO 2 , ARDS and a high level of FiO 2 (over 80 %) 1h after starting NIV were found as factors associated with NIV failure in children with severe respiratory distress of various causes [ 27 , 28 ] . Larrar et al. identifi ed the absence of a reduction in PCO 2 as a predictive factor of NIV failure in a CPAP study. Abboud et al., in an HFC study, came to the same conclusion [ 11 , 23 ] . Neurological failure Altered level of consciousness with low reactivity or agitation not responding to oxygenation A French prospective multicenter study noted that a minimal reduction in CO 2 and a low increase in pH measured 2 or 4 h after NIV initiation were strong predictors of NIV failure [ 30 ] . In that study, the various centers had defi ned criteria for ventilatory support and absolute criteria for invasive ventilation (Tables 27.4 and 27 .5 ). The results suggested that early assessment of the response to NIV is crucial.

Noninvasive ventilatory supports include a number of systems that deliver pressure support to the patient via an interface. The CPAP delivery system has to be reliable, with good stability of the pressure during all the respiratory cycle length. It also has to be easy to use and install in children. Interfaces are chosen according to their ability to be connected to the CPAP delivery systems while minimizing air leaks, dead space, and discomfort.

To deliver heated/humidifi ed oxygen, an air-oxygen fl ow generator is required combined with a heated humidifi er. The circuit tubing and the size of the cannula differ according the age of the child. For infants weighing <10 kg, small-volume circuit tubing is required. Infant or pediatric cannulas can be used. Adult circuit tubing and cannulas are used for children weighing ≥10 kg.

As children with severe bronchiolitis requiring ventilatory support are young (<42 days) and of low weight, CPAP systems developed for neonates are used, such as Infant Flow (EME, Electro Medical Equipment, Brighton, UK), Infant Star 950 (Nellcor Puritan Bennett, San Diego, CA, USA), and Bubble CPAP (Fisher and Paykel Healthcare, Auckland, NZ). In the latter system, a water column delivers PEEP. In PICUs, an ICU ventilator or CPAP machine may be preferred. No study has compared the stability of PEEP in these delivery systems. It is well known that PEEP stability can be affect by the level of leaks, the degree of mouth opening, and the level of airfl ow in the circuit. The choice of the interface is crucial. The ideal interface is one that is easy to install, minimizes leaks, and does not cause skin or mucosal injury.

Low-resistance nasal prongs or cannulas are the interfaces most frequently used in infants with bronchiolitis. The nasal approach is preferred because infants predominantly breathe through the nose. In infants with bronchiolitis, the tolerance is reportedly good, although no study has specifi cally addressed skin or mucosal injuries in the context of bronchiolitis. As nasal breathing has to be preserved, nasal obstruction, which frequently occurs during RSV infection, should be systematically treated and monitored. Nasal obstruction is a source of discomfort and agitation for children treated by nasal CPAP. Consequently, nasal lavage with NaCl 0.9 % every 3 or 4 h is recommended. The choice of the cannula's size is important to limit leaks and avoid nose injuries. It is recommended that nasal prongs of different sizes with different inter-nostril distances be readily available. To limit mouth leaks, a dummy is frequently used and sometimes a chinstrap is required. To avoid skin irritation or ulceration and to improve the patient's comfort, colloid ulcer dressings (e.g., Comfeel, Coloplast) are applied to protect the nasal bridge as well as the nostrils. In our experience, nasal prongs or cannulas are well tolerated by infants weighing up to 5 kg. For larger infants, nasal masks are often better tolerated.

During the last decade, manufacturers have designed small nasal masks specifi cally intended not to leak. They allow us to put small infants on CPAP with standard ventilators in the PICU. These masks are also available for the Bubble CPAP system and the Infant Flow CPAP generators. It is also possible to use a nasal mask with intentional leakage to connect infants to a CPAP or BiPAP machine. The Resmed Sullivan Infant Bubble mask (ResMed, Waterloo, Australia) is used for the smallest infants and the Small Child Profi l Lite mask (Philips Respironics, Murrysville, PA, USA) for the others. Skin protection can be used to minimize skin irritation. The choice of the headgear or bonnet fi tted to the head's form and size is important to avoid mask displacement, which can increase leaks, and fastening the mask too tightly on the face, which could increase the risk of skin injuries.

Bucconasal masks are used only when the leaks are interfering with synchronization of the infant with the ventilator for NIV. When the mouth is open or if the nose is obstructed, application of nasal CPAP becomes ineffective. A major concern is the absence of specifi c bucconasal masks for infants. Most often, anesthesia masks or adult nasal masks are used. However, in these cases, the risk of skin injury is much higher than with nasal masks, particularly laceration or ulceration of the nasal bridge. Skin protection with colloid dressings must be used to prevent these injuries. Progress in the design of bucconasal masks is needed to enable NIV in infants and young children. Helmets represent an alternative in children weighing >5 kg. They cannot be used in smaller children because the helmet compresses the chest, reducing its effi cacy. Some experiences with helmets have been reported even in small children. They report rather good tolerance and improvement of alveolar ventilation [ 31 , 32 ] .

Essouri and colleagues showed that a CPAP level of 7 cmH 2 O was better than either 4 or 10 cmH 2 O in 10 children admitted to a PICU for severe bronchiolitis [ 16 ] . The decrease WOB, assessed by PTP di and PTP es , was more signifi cant with 7 cmH 2 O.

This level was closest to the auto-PEEP level (6.3 cmH 2 O) and is consistent with the level of CPAP used in the main clinical studies (Table 27. 2 ). Based on these results, starting with a level between 6 and 8 cmH 2 O is recommended. For HFC, the recommended fl ow by the manufacturer is 1-2 L/kg/min. In a retrospective study, Schibler et al. used a fi xed fl ow of 8 L/min but did not explain the reason for this choice [ 15 ] . In practice, we start with a fl ow of 1 L/kg/min and increase it to 2 L/kg/min according to the tolerance of the child. As already noted, the optimal level of fl ow is unknown and depends of the degree of nares obstruction and mouth leakage.

Concerning the BiPAP or ASB ventilation modes, as no study has been performed comparing different ventilatory settings we are not able to make any recommendations. The studies that have reported the use of BiPAP or ASB in patients with bronchiolitis used a level of PEEP varying from 4 to 8 cmH 2 O and inspiratory pressures between 10 and 20 cmH 2 O. The level of pressure support varied from 4 to 12 cmH 2 O. NIV pressures >20 cmH 2 O are associated with a high risk of gastric dilatation by gas. To minimize this phenomenon, a nasogastric tube is routinely inserted to defl ate the stomach when necessary. In our clinical practice, when an infant is switched from CPAP to NIV on pressure support, we start with a level of PEEP equal to the level of CPAP used and then add pressure support of 6 cmH 2 O above PEEP or an inspiratory pressure of 6 plus PEEP. Then, after assessing the effi cacy and tolerance, we adapt the ventilator setting or the interface, avoiding exceeding 20 cmH 2 O. The inspiratory pressure is titrated by 2 cmH 2 O increments to a level where the RR, signs of WOB, and blood gases are improved.

The main problem with NIV in pressure support mode, such as BiPAP, is asynchrony. The sensitivity of ventilatory triggers is sometimes insuffi cient for young infants, who are unable to trigger a ventilatory cycle. On the contrary, when the sensitivity of the trigger is too high, and when the leaks are important, autotriggering may appear, generating discomfort and asynchrony. Control of leakage is another factor contributing to synchrony: If the ventilator is unable to compensate for the leaks, the inspiratory time can be prolonged into the period when the child wants to expire. These causes of asynchrony are a source of discomfort and poor tolerance of NIV in infants.

No study has been performed comparing different ventilatory modes with different ventilators in infants who suffer from severe bronchiolitis. Neurally adjusted ventilatory assistance (NAVA) is a promising mode that would limit the incidence of asynchrony. Liet and colleagues reported three cases of infants with severe bronchiolitis treated with this mode during invasive mechanical ventilation and showed that NAVA was able to improve synchrony, decrease the oxygen requirement, and decrease peak airway pressure from 28 ± 3 to 15 ± 5 cmH 2 O [ 33 ] .

It has been shown In 15 neonates and children that NAVA decreased patient-ventilator asynchrony and the peak inspiratory pressure [ 34 ] . The percentage of time in asynchrony was lower in the NAVA group (8.8 %) than in the pressure (33.4 %) and fl ow (30.8 %) trigger groups (ventilated either in pressure control or pressure regulated volume controlled). Moreover, the peak inspiratory pressure was 1.9-2.0 cmH 2 O lower in NAVA than in the pressure and fl ow groups, respectively ( p < 0.05 for both) [ 34 ] .

We reported our experience of NAVA in NIV mode in 18 infants with severe bronchiolitis. The tolerance and the feasibility were good, and 16 of 18 infants had NAVA mode success, thereby avoiding invasive ventilation (personal communication). As no study comparing classic NIV to NAVA NIV has been reported, this technique cannot be recommended but represents a new mode to be considered when asynchrony is detected frequently with classic NIV.

During the last decade, ventilatory support for children with severe bronchiolitis has radically changed. Nasal CPAP is become the fi rst mode of NIV for children who meet the criteria for ventilatory support. Numerous studies have suggested that this strategy is associated with a decreased need for intubation and invasive ventilation. Although the level of evidence of improved outcomes related to this strategy is low, in the absence of prospective controlled studies the data published have shown that children can be safely managed less invasively without prolonging the PICU stay. Some studies suggested that responders to nasal CPAP had a lower length of PICU stay than those who were intubated. Physiological studies have provided some evidence that a CPAP level of 6-7 cmH 2 O is able to decrease the WOB and improve alveolar ventilation in children with obstructive bronchiolitis [ 7 , 16 , 17 ] . The application of extrinsic PEEP to the airways at a level greater than the level of auto-PEEP generated by dynamic obstruction of small airways allowed reduction of efforts made by the child to initiate the next inspiratory cycle. This mechanism is responsible for the clinical improvement observed in the children after initiation of nasal CPAP. The responders are those whose RRs are reduced and CO 2 levels and heart rates are decreased within 2-4 h of starting NIV with CPAP. Early identifi cation of those who will respond is crucial so as not to delay applying NIV with two levels of pressure or intubation with invasive ventilation.

More recently, the HFC, which is able to deliver humidifi ed/heated oxygen, has been reported to be another alternative to nasal CPAP [ 14 , 15 , 23 ] . This system has been shown to generate a low level of PEEP, induce washout of nasopharyngeal dead space, match inspiratory fl ow rates in infants, and improve mucociliary clearance [ 24 ] . Children with apnea and those with severe hypercapnic acidosis are more likely to fail HFC and can be treated by nasal CPAP. As no study has been conducted comparing HFC to nasal CPAP, no recommendation can be drawn. There is a crucial need of studies to better distinguish groups of children who will respond to HFC, to CPAP, or to NIV because the level of expertise and equipment differ significantly between these modes. HFC can be initiated in the emergency or intermediate care units, whereas CPAP and NIV should be reserved for use in an intermediate care unit or an ICU according the level of the teams' experience. Stratifi cation of respiratory distress severity is required for better patient selection at admission. We know that patients with a medical history of chronic lung, heart, or neuromuscular diseases are at higher risk of complications and failure of HFC or CPAP. Young infants, particularly those born prematurely and those with low weight, are more likely to require ventilatory support [ 1 , 6 ] . However, the clinical score, biological markers, and blood gas criteria associated with ventilatory support and with CPAP failure, are not well defi ned and require further study. Moreover, as no study has been performed on NIV at two pressure levels in bronchiolitis, there is no evidence that NIV after CPAP or HFC failure can obviate the need for intubation and invasive ventilation. Only a multicenter prospective study comparing different ventilatory strategies would be able to determine the best ventilatory support treatment.

Technically, manufacturers have improved their products to facilitate CPAP application. Nasal masks and nasal prongs of different sizes are now available, allowing us to fi t the equipment to the child's facial and head morphology. The objectives of these interfaces are to facilitate setup, limit the dead space, and reduce air leaks. Experience and the use of specifi c nursing protocols are factors associated with a high success rate of NIV techniques, suggesting that only teams with a high level of training and experience should apply NIV.

Review of epidemiology and clinical risk factors for severe respiratory syncytial virus (RSV) infection

Epidemiology and prevention of respiratory syncytial virus infections among infants and young children

Risk factors for respiratory syncytial virus associated apnoea

Incidence of apnea in infants hospitalized with respiratory syncytial virus bronchiolitis: a systematic review

Key Major Recommendations • Nasal CPAP and HFCs are the best fi rst option for ventilatory support of children with severe bronchiolitis. Their use may avoid intubation and invasive mechanical ventilation

• There is an insuffi cient level of evidence of the effi cacy of NIV on mortality or morbidity criteria

• NIV with pressure support is an option when CPAP fails

• Early assessment (within the fi rst 2 h) of responders to CPAP or HFC is required to prevent secondary critical deterioration

• Improved blood gas levels after CPAP is a good indicator of response to CPAP

• Infection, apnea, and young age are associated with NIV failure. • Nasal cannulas are the most appropriate interface for infants weighing <5 kg

Acute respiratory distress syndrome caused by respiratory syncytial virus

Hospitalizations for respiratory syncytial virus bronchiolitis in preterm infants at <33 weeks gestation without bronchopulmonary dysplasia: the CASTOR study

Nasal continuous positive airway pressure decreases respiratory muscles overload in young infants with severe acute viral bronchiolitis

Non-invasive ventilation in infants with severe infection presumably due to respiratory syncytial virus: feasibility and failure criteria

Increase in use of non-invasive ventilation for infants with severe bronchiolitis is associated with decline in intubation rates over a decade

Non-invasive ventilation as primary ventilatory support for infants with severe bronchiolitis

Effects of nasal continuous positive airway pressure ventilation in infants with severe acute bronchiolitis

Non-invasive ventilation for severe bronchiolitis: analysis and evidence

Pilot study of vapotherm oxygen delivery in moderately severe bronchiolitis

High fl ow nasal cannulae therapy in infants with bronchiolitis

Reduced intubation rates for infants after introduction of high-fl ow nasal prong oxygen delivery

Optimal level of nasal continuous positive airway pressure in severe viral bronchiolitis

6 cmH(2) O continuous positive airway pressure versus conventional oxygen therapy in severe viral bronchiolitis: a randomized trial

Continuous positive airway pressure in bronchiolitis

Continuous positive airway pressure by nasal prongs in bronchiolitis

Clinical predictors of nasal continuous positive airway pressure requirement in acute bronchiolitis

Randomised controlled trial of nasal continuous positive airways pressure (CPAP) in bronchiolitis

Use of continuous positive airway pressure (CPAP) in acute viral bronchiolitis: a systematic review

Predictors of failure in infants with viral bronchiolitis treated with high-fl ow, high-humidity nasal cannula therapy

Research in high fl ow therapy: mechanisms of action

Effect of HFNC fl ow rate, cannula size, and nares diameter on generated airway pressures: An in vitro study

Prospective multicenter study of children with bronchiolitis requiring mechanical ventilation

Predictive factors for the success of noninvasive mask ventilation in infants and children with acute respiratory failure

Noninvasive positive pressure ventilation: fi ve years of experience in a pediatric intensive care unit

Predictive factors of non invasive ventilation failure in critically ill children: a prospective epidemiological study

Non invasive ventilation in children with acute respiratory failure during RSV epidemic period: a prospective multicenter study

Air-oxygen helmet-delivered continuous positive airway pressure to manage respiratory failure due to bronchiolitis

Helmet-delivered continuous positive airway pressure with heliox in respiratory syncytial virus bronchiolitis

Respiratory support by neurally adjusted ventilatory assist (NAVA) in severe RSV-related bronchiolitis: a case series report

Comparison of pressure-, fl ow-, and NAVA-triggering in pediatric and neonatal ventilatory care