Randomized Comparison of Helmet CPAP Versus High-Flow Nasal Cannula Oxygen in Pediatric Respiratory Distress

Discussion

This study showed that both helmet CPAP and HFNC methods were efficient in improving the clinical conditions of subjects with mild-to-moderate respiratory distress when compared with a control group, although clinical response to helmet CPAP was more efficient and rapid compared with that of HFNC-treated children. Moreover, whereas helmet CPAP was responsible for a statistically significant improvement of pH, ABG P_aCO2, and P_aO2/F_IO2, HFNC was more effective in improving S_pO2%, ABG P_aO2, and P_aO2/F_IO2.

HFNC therapy delivers a high flow and heated humidified oxygen into the nasal passages. Humidified delivery prevents mucous plugging and improves comfort.¹² In detail, HFNC exceeds the peak inspiratory flow so that a precise F_IO2 can be delivered without dilution of the administered oxygen through entrainment of room air into the nasal passages during inspiration. By exceeding peak inspiratory flow, HFNC therapy provides some of the driving pressure in the upper airway during inspiration, thereby reducing the work that the patient is required to perform. The flow during expiration washes out the dead space of the upper airway, reducing the ventilatory requirement for the patient. The expiratory resistance provides some PEEP, thereby increasing functional residual capacity and oxygenation, and the small amount of PEEP delays small airway closure during expiration, improving tidal ventilation. Nevertheless, the flow must be humidified for nasal comfort, and this humidification might improve mucus clearance and prevent mucous plugging; however, this is a minor benefit compared with the other physiologic benefits.

However, the clinical efficacy of HFNC as a therapeutic approach to bronchiolitis has only been reported in non-experimental observational studies, which indicates lack of appropriately powered randomized studies with well-defined clinically important outcomes.^13,14 Some clinical studies using HFNC in a non-randomized design have shown a reduction in intubation rates in critically ill infants in intensive care settings.^15–17 Pilot studies have also investigated the use of HFNC in general pediatric ward settings, suggesting that this method is safe and efficient.¹⁸ Their results indicate that pediatric ICU admission rates could be significantly reduced by 2.5 times compared with a control group.¹⁸ Franklin et al have proposed an extensive prospective randomized, controlled trial on the use of HFNC in children with bronchiolitis to assess the efficiency of this method in resolving respiratory distress and avoiding admission to the pediatric ICU.²

The current study highlights the efficiency of HFNC in improving clinical conditions and blood gas exchanges of children affected by bronchiolitis. This treatment revealed a statistically significant improvement of S_pO2% (P = .009), ABG P_aO2 (P = .009), and P_aO2/F_IO2 (P = .009) when compared with parameters recorded at baseline (see Table 3), which are consistent with the mechanism of action of HFNC. Nevertheless, HFNC failed in one subject, who required endotracheal ventilation due to gradual deterioration of clinical conditions and lack of response to noninvasive support after the first 24 h. Future studies should clarify risk factors for eventual HFNC failure in pediatric patients with respiratory distress.

Our study showed a more efficient response in helmet CPAP-treated subjects compared with the HFNC group. During CPAP administration, the patient's airway is maintained throughout the respiratory cycle at a selected constant pressure (CPAP), which is higher than the atmospheric pressure. This method is used to improve respiratory mechanics and gas exchange in patients with intact neuromuscular function, representing a supportive therapy in patients with various forms of respiratory distress. CPAP acts through improving arterial oxygenation and respiratory mechanics and reducing the patient's respiratory drive and effort.^19–21 Because the inspiratory effort creates a negative pressure inside the thorax, the ventricle afterload is decreased. Accordingly, a decrease in inspiratory effort implies a reduction in the left ventricle afterload. Therefore, venous return and ventricle sizes are decreased with a parallel drop in the wall tension and myocardial oxygen consumption.^19–21 In patients with non-hydrostatic pulmonary edema, CPAP could improve gas exchange and respiratory mechanics, thereby increasing the end-expiratory lung volume and preventing alveolar collapse.^19–21 The alveolar extension provides a greater gas exchange surface, which improves the respiratory mechanics of ventilation and results in a consequent decrease of P_aCO2 in blood gas analysis.^19–21 Moreover, CPAP has been demonstrated to be efficient in subjects with respiratory distress associated with neurological conditions, such as autoimmune encephalitis.²²

In the current study, CPAP was delivered by helmet, since the lack of pressure points on the face helps to avoid skin necrosis and pain. It also reduces discomfort and improves the patient's tolerance. Similar to HFNC, helmet CPAP was found to be efficient in improving respiratory distress after the first few hours of treatment, and its influence was proved to be more effective on ABG pH (P = .043), ABG P_aCO2 (P < .001), and P_aO2/F_IO2 (P < .001) changes (see Table 2). None of the subjects treated with CPAP required transfer to the pediatric ICU in this study. Compared with HFNC, helmet CPAP showed a higher efficiency in resolving respiratory distress (see Table 4). This was mostly due to the effect of CPAP on alveolar extension, which causes an increase of the alveolar surface responsible for blood gas exchange. Nevertheless, we have to consider that our CPAP group was younger than the HFNC subjects, and contained more females and fewer asthma causes. Their blood gases at baseline were also wildly different, although we have to consider that younger age is a risk factor for respiratory instability, respiratory distress worsening, and treatment failure. On the other hand, the amount of extrathoracic dead space is different in different-sized children. If one of the theories for HFNC “improving” ventilation is dead space washout, then smaller children would be more likely to benefit, but also larger children with higher flows can benefit from this method of respiratory support. This assumption highlights the need for further stratification studies on the efficiency of HFNC and CPAP in different age groups.

The efficiency of CPAP in pediatric respiratory distress has been investigated since 1981, when Beasley and Jones²³ retrospectively described 23 children with bronchiolitis, showing a positive outcome after treatment with CPAP. Nevertheless, few data have since been published in this regard, and the experiences on pediatric wards outside of neonatal ICU and pediatric ICU settings are scarce. To our knowledge, the first study on the use of noninvasive respiratory support in a non-intensive care setting was published in 2005 by Prado et al,²⁴ who prospectively evaluated the medical records of 14 subjects (between 1 month and 13 y of age). In that study, the presence of pulmonary restrictive diseases (80% of cases) was found to be the main indication for bi-level positive airway pressure. In the other cases, NIV was performed in subjects who used NIV at home and had exacerbations of their neuromuscular respiratory disease plus subjects with hypoxic and hypercapnic respiratory distress.²⁴ In this study, only one subject required intubation for mechanical ventilation, while all the other subjects improved.

In 2008, Thia et al²⁵ published a randomized controlled trial investigating the use of nasal CPAP in bronchiolitis in a pediatric intermediate care setting, which included 29 children treated randomly by CPAP or standard treatment. The authors found that CPAP significantly decreased P_aCO2 levels compared with standard treatment. Moreover, CPAP demonstrated a significantly stronger reduction in P_aCO2 when it was used at first, compared with when it was used afterward. CPAP was well-tolerated, and no complications were detected.

A systematic review of Jat and Mathew²⁶ included 2 studies of 50 subjects < 12 months of age. In one study, the authors found a high risk of bias for incomplete outcome data and selective reporting, whereas both studies had an unclear risk of bias for several domains, including random sequence generation. The effect of CPAP in children with acute bronchiolitis was found indeterminate due to imprecision around the effect estimate. One trial found that CPAP significantly improved breathing frequency compared with no CPAP application, whereas the other study reported no difference between groups with no numerical data to pool. Changes in arterial oxygen saturation were measured in only one trial, and the results were imprecise. The effect of CPAP on the change in P_aCO2 was also imprecise. Change in P_aO2, hospital admission rate (from emergency department to hospital), duration of emergency department stay, requirement for ICU admission, local nasal effects, and shock were not assessed in either study. Therefore, it was concluded that the effect of CPAP in children with acute bronchiolitis is uncertain due to limited evidence available, which underscores the need for further studies and more extensive trials.²⁷

Reviewing the literature on the 2 aforementioned noninvasive support modalities, very few studies that compare the efficacy of CPAP versus HFNC in resolving respiratory distress and avoiding pediatric ICU admissions have been published. In fact, there is conflicting evidence about the ability of HFNC devices to produce clinically important PEEP, since some research studies have demonstrated that HFNC levels of PEEP are too low to be clinically important.^27–30

One retrospective case review compared 2 methods of noninvasive support in 34 infants with bronchiolitis admitted to pediatric ICU during bronchiolitis seasons of 2 consecutive years.³⁰ Infants admitted during the first year received CPAP, whereas those in the subsequent year received HFNC with oxygen flow of up to 3 L/kg/min (up to a maximum of 8 L/min). The authors described a significant mean maximum breathing frequency in F_IO2 and a significant increase in the noninvasive respiratory support, but they found no differences between the 2 groups for both of these parameters.³¹ To our knowledge, this is the only study comparing the efficiency of HFNC and CPAP in children with respiratory distress. Since the aforementioned study was performed in a pediatric ICU setting, we can confirm that the current study is the first ever performed on CPAP versus HFNC efficiency in a pediatric intermediate care setting.

Previous literature has already underscored the need to perform further clinical studies to evaluate the effect of CPAP and HFNC on important clinical outcomes and with various levels of care.^32,33 In the current study, we found a decrease of pediatric ICU admission of 97.5% (only one HFNC-treated case required endotracheal intubation in the pediatric ICU). Helmet CPAP was found to be more efficient than HFNC in resolving respiratory distress. This correlation was observed regarding duration of treatment (from admission to evidence of clinical and ABG improvement) and duration of hospitalization and pharmaceutical drug administration (see Tables 1 and 4). Moreover, it was found that supported subjects had a better clinical course compared with the control group (20 children affected by mild-to-moderate respiratory distress for bronchiolitis), since their total length of hospital stay and pharmaceutical drug consumption were statistically and significantly lower than the control group (see Table 4). It should be noted that routine blood analyses were carried out for all of the children at admission, and for those with signs of bacterial infections, broad-spectrum antibiotic therapy was started at the onset of hospitalization. There was no septic shock reported.

We should mention that our study also presents some weaknesses, such as treatment allocation that was not blinded and the choice for heterogeneous disease groups, due to the heterogeneity of the population admitted to our complex O.U. during the study period. Moreover, the different age and size of the 2 studied groups may represent an important influence on the final results, as far as blood gas differences at admission between the studied groups, before any treatment was performed. Therefore, this may encourage further study with a larger sample size and a more uniform distribution of demographic parameters of the included subjects.

The significance of the current findings is due to vast economic burden of bronchiolitis in children, since the economic burden of infants hospitalized with bronchiolitis is estimated to be 1.73 billion United States dollars/y. In Australia, approximately 11,000 infants with bronchiolitis require hospital admission per year. In 2013, 1,254 of these 11,000 cases (12%) were admitted to a pediatric ICU.^5,34 Therefore, it is crucial to establish new protocols for the treatment of respiratory distress in pediatric emergency departments to avoid high rates of pediatric ICU admissions.