High-flow nasal cannula (HFNC) is a relatively new device for oxygenation that has become a popular tool in the treatment of hypoxic respiratory failure [1]. Recently, there has been a growing interest in its use in hypercapnia. Physiological studies have shown that HFNC reduces rebreathing of carbon dioxide in the upper and lower anatomical dead space [2,3]. HFNC users generate less carbon dioxide by reduced work of breathing [4]. Some studies have also suggested that HFNC may even improve alveolar ventilation [5]. However, the effectiveness of HFNC in hypercapnic respiratory failure has not been fully demonstrated as relevant studies have been mostly retrospective with some contradictory results [6-8].

To clearly understand the effects of HFNC in hypercapnic respiratory failure, we conducted a prospective observational study with frequent arterial blood gas analysis (ABGA) in the first 24 hours after starting HFNC. For safety reasons, we recruited patients with mild to moderate hypercapnia without severe acidosis. The aim of this study was to examine the competence of HFNC in hypercapnia correction and to investigate the clinical use of HFNC in patients with hypercapnia.

A multicenter prospective observational study was conducted in four tertiary care hospitals in South Korea from November 2017 to October 2018. Inclusion criteria for the subjects were patients with mild to moderate hypercapnia (43 ≤ arterial partial pressure of carbon dioxide [PaCO₂] ≤ 70 mm Hg) in ABGA. Recruitment was conducted from the emergency room, general ward, and intensive care unit. Exclusion criteria were: under 18 years of age; respiratory rate more than 35 breaths per minute; severe acidosis (pH < 7.30); using accessory muscle of respiration; any other signs of progressive respiratory failure requiring imminent intubation and mechanical ventilation; unstable hemodynamics; post-cardiopulmonary resuscitation status or on extracorporeal membrane oxygenation; consciousness disorder that would require intubation (e.g., not responsive, agitated, uncooperative, high risk for aspiration); facial surgery, trauma or deformity and airway obstruction that precluded application of nasal cannula; and those receiving end-of-life care.

HFNC was initiated with a pulse oximeter oxygen saturation target of 92%. The initial flow was set at 35 L/min and adjusted to suit the best comfort of the subject. ABGA was checked at 1, 3, 6, and 24 hours after starting HFNC. When the 24-hour data collecting period ended, the attending physician decided on the continuation and duration of the HFNC. Medications such as bronchodilators and antibiotics were administered to treat the underlying medical condition and precipitating factors.

HFNC was discontinued if the subject met any of the following criteria: symptoms and signs of worsening respiratory distress, such as retractions or sweating; increased respiratory rate to more than 35 breaths per minute; elevation of PaCO₂ greater than 10 mm Hg from the previous ABGA results; pH less than 7.2 in any of the ABGA results; pH decrease more than 0.2 from previous ABGA results; or deteriorating mental status.

In this prospective observational study, we report our findings on changes in ABGA parameters after applying HFNC to patients with mild to moderate hypercapnia. A significant decrease in PaCO₂ and pH correction were observed in the first hour of HFNC application. The effect of PaCO₂ reduction was more prominent in patients who did not have underlying obstructive lung disease. In terms of clinical outcomes of the 45 study subjects, two subjects required noninvasive mechanical ventilation for worsening hypercapnia, and one subject required invasive mechanical ventilation after intubation for worsening respiratory distress. There were no deaths or serious adverse outcomes in the study population.

With its unique physiological benefits, HFNC has become a popular device for oxygen delivery. HFNC provides a constant FiO₂, improves mucociliary clearance, reduces the effort required to breathe, and decreases anatomical dead space [1]. With acute hypoxic respiratory failure, multiple prospective randomized trials have shown the clinical advantages of HFNC over conventional oxygen delivery systems [9,10]. Therefore, HFNC is widely being used in various conditions with hypoxia.

However, the effects of HFNC on hypercapnic respiratory failure remain unclear; therefore, more evidence is needed to support its use in patients with hypercapnia. Although most clinical studies have reported a reduction in PaCO₂ after applying HFN, there have been some studies with conflicting results. In a study by Sztrymf et al. [8], involving intensive care unit patients with acute respiratory failure, a significant increase in PaCO₂ was observed. In a study by Nilius et al. [11], the use of HFNC resulted in different individual responses in respiratory rates and PaCO₂ levels.

Because hypercapnia has complex pathophysiological mechanisms and multiple etiologies, there is wide heterogeneity in patients [12]. This could explain the mixed results in clinical studies and becomes a challenge to establishing clear indications for the use of HFNC in hypercapnic respiratory failure. However, even with various causes, there are commonalities in these patients. Whatever the diagnosis, mechanical ventilation (invasive and noninvasive) is a standard treatment in severe cases of hypercapnia [13]. Therefore, the potential CO₂ removal capacity and physiologic benefits make HFNC an attractive alternative to conventional oxygen therapy in patients with hypercapnia.

With our study, an overall decrease in PaCO₂ with a correction of pH was observed after applying HFNC in patients with mild to moderate hypercapnia. Additionally, 93% of the study population was successfully treated without progressing to mechanical ventilation. In other studies investigating clinical implications of HFNC on hypercapnic respiratory failure, intubation rate ranged from 0% to 25% [14,15]. A recent metaanalysis suggested that HFNC was not inferior to noninvasive ventilation with respect to intubation rate and mortality [16]. Considering all current evidence, HFNC may be a feasible alternative option for oxygen supplementation in patients with mild to moderate hypercapnia.

However, our study found a different response to HFNC depending on underlying lung disease. We analyzed changes in PaCO₂ after grouping the patients according to the presence or absence of an obstructive lung pathology. Contrary to our anticipation, the effects of HFNC were more pronounced in patients who did not have underlying obstructive lung disease. This finding suggests that the decrease in PaCO₂ during HFNC was possibly related to the decreased effort required to breathe rather than reduction in dead space. Further studies are needed to better understand the effects of HFNC on hypercapnia based on underlying lung pathology.

We recognize the limitations of this study. We acknowledge that the findings in this study cannot be fully accepted without a control group comparison. Also, because of the nature of this study design, selection bias is possible. Last, the small number of subjects also limits the ability to generalize the study’s findings.

In conclusion, we prospectively investigated changes in ABGA parameters and clinical implications of HFNC in patients with mild to moderate hypercapnia. After applying HFNC, an overall decrease in PaCO₂ with pH correction was seen. Most of our subjects were treated successfully without progressing to mechanical ventilation. Therefore, we suggest that HFNC could be a feasible alternative option for oxygen delivery in patients with mild to moderate hypercapnia. However, further controlled studies with different etiologies of hypercapnia are needed to establish the clear role of HFNC in patients with hypercapnic respiratory failure.