Editorial 

December 10, 2024

Is High-Flow Oxygen the Standard for All Patients With Acute Respiratory Failure?

Jean-Pierre Frat, Sylvain Le Pape, Arnaud W. Thille

JAMA. Published online December 10, 2024. doi:10.1001/jama.2024.25906

Hypoxemic acute respiratory failure denotes oxygenation failure attributable to various mechanisms, including ventilation-perfusion mismatch, increased shunt, diffusion impairment, and alveolar hypoventilation. The most common etiology of hypoxemic acute respiratory failure is pneumonia.¹ A definition of hypoxemic acute respiratory failure was recently proposed, emphasizing signs of respiratory distress and Pao₂ level less than 60 mm Hg in room air or level of Pao₂:Fio₂ (fraction of inspired oxygen) less than 300 under oxygen.^2,3 By contrast, hypercapnic acute respiratory failure is characterized by ventilation failure defined by respiratory acidosis (pH level ≤7.35 with Paco₂ >45 mm Hg) and is mainly caused by exacerbation of chronic obstructive pulmonary disease (COPD).²

In this issue of JAMA, the RENOVATE trial, which was conducted by the RENOVATE Investigators and the BRICNet Authors,⁴ addresses this fundamental question in acute respiratory failure management of whether to initiate noninvasive ventilation vs high-flow oxygen, regardless of the underlying cause of acute respiratory failure. Among the 1766 patients with a range of acute respiratory failure etiologies, high-flow oxygen was found to be noninferior to noninvasive ventilation for preventing endotracheal intubation or death within 7 days for most acute respiratory failure etiologies, including hypoxemic acute respiratory failure (nonimmunocompromised patients), COPD exacerbation, cardiogenic pulmonary edema, and acute respiratory failure due to COVID-19, with the exception of immunocompromised patients with hypoxemic acute respiratory failure.

In previous trials, among patients with severe acute respiratory failure requiring intensive care unit admission, risks for endotracheal intubation and mortality varied significantly according to the underlying cause. Hypoxemic acute respiratory failure presents the highest risk profile, with endotracheal intubation rates of 30% to 50% (particularly in immunocompromised patients and in those with COVID-19) and mortality rates of 20% to 35% (though lower in patients with COVID-19).² By contrast, hypercapnic acute respiratory failure and cardiogenic pulmonary edema show more favorable outcomes, with endotracheal intubation rates approximating 10% and mortality rates below 15%.⁵

Oxygen therapy through a nonrebreathing mask (standard oxygen) is the first-line therapy for hypoxemia in the most severe cases of acute respiratory failure.² However, even while correcting hypoxemia, this approach does not alleviate dyspnea or the clinical signs of increased work of breathing. Noninvasive respiratory support, including high-flow nasal cannula oxygen therapy, continuous positive airway pressure, and noninvasive ventilation with bilevel positive pressure support are used to reduce inspiratory effort and, subsequently, to decrease the probability of endotracheal intubation due to respiratory exhaustion as well as the incidence of adverse events related to invasive mechanical ventilation, including mortality. However, the benefit of these oxygen strategies varies depending on the underlying causes of acute respiratory failure, and, as a consequence, the specific recommendations regarding their use have been based on the type or cause of acute respiratory failure.^2,5

Noninvasive ventilation is strongly recommended in patients with hypercapnic acute respiratory failure or cardiogenic pulmonary edema because it reduces the risk of endotracheal intubation and mortality.⁵ However, its benefits remain controversial in hypoxemic acute respiratory failure and in immunocompromised patients. Small studies of 30 to 100 patients conducted in the 1990s and early 2000s suggested noninvasive ventilation reduced endotracheal intubation risk compared with standard oxygen in hypoxemic acute respiratory failure.^2,3 This perspective shifted after 2015 when larger trials (100-300 patients) comparing noninvasive ventilation with high-flow oxygen found no superiority for noninvasive ventilation.^1,6 For immunocompromised patients, research in the early 2000s initially showed promising results, with 2 small, randomized trials in heterogeneous populations reporting benefits with noninvasive ventilation vs standard oxygen regarding endotracheal intubation and mortality outcomes.^7,8

However, subsequent larger trials challenged these findings. For example, the study by Lemiale et al⁹ found no benefit with noninvasive ventilation vs standard oxygen even though nearly 40% of patients received high-flow oxygen therapy, which may have offset the benefit obtained using noninvasive ventilation. The trial by Coudroy et al¹⁰ similarly found no benefit with noninvasive ventilation vs high-flow oxygen in an immunocompromised patient population. The COVID-19 pandemic prompted numerous trials of oxygen support strategies.

Two clinical trials demonstrated superiority with high-flow oxygen vs standard oxygen in preventing endotracheal intubation, but the mortality rates remained similar.^11,12 Only one small trial showed benefits for noninvasive ventilation delivered through a helmet interface.¹³ Accordingly, recent recommendations have suggested using of high-flow oxygen rather than standard oxygen and noninvasive ventilation as first-line therapy in patients with hypoxemic acute respiratory failure, even though high-flow oxygen reduces endotracheal intubation risk but not mortality.² For immunocompromised patients with acute respiratory failure, experts were unable to make any recommendation for noninvasive ventilation.²

One might question whether equipoise was present in the study design of the RENOVATE trial,⁴ given international guidelines favoring high-flow oxygen for hypoxemic acute respiratory failure and noninvasive ventilation for COPD exacerbation and cardiogenic pulmonary edema. However, this concern is mitigated by several factors. First, only 2 prior clinical trials have shown benefit of high-flow oxygen vs noninvasive ventilation in hypoxemic acute respiratory failure, indicating a need for additional validation.³

Second, the authors of the RENOVATE trial,⁴ proposed protective settings of noninvasive ventilation with positive end-expiratory pressure levels of 8 to 12 cm H₂O, which is higher than those applied in most of the previous trials. As the authors argue,⁴ while the superiority of noninvasive ventilation has been established vs standard oxygen in COPD exacerbation and cardiogenic pulmonary edema, it has not been directly compared with high-flow oxygen, which offers beneficial physiological effects in these conditions. Moreover, the study protocol also allowed for treatment escalation from high-flow oxygen to noninvasive ventilation, which ultimately occurred in 23% of patients with COPD exacerbation and 5% of those with cardiogenic pulmonary edema. The authors chose a composite primary outcome (endotracheal intubation or death at day 7)⁴ that effectively encompassed all critical events, including death without prior endotracheal intubation. The inclusion of 30- and 90-day mortality as secondary outcomes provided additional insights into longer-term treatment effects.

What is the clinical interpretation of the results of the RENOVATE trial? The bayesian adaptive statistical methods used in the trial and interim analyses render clinical interpretation complex, especially given significant disparity in group sizes across the different patient populations. Specifically, even though the results demonstrated noninferiority of high-flow oxygen vs noninvasive ventilation in the nonimmunocompromised patients with hypoxemic acute respiratory failure (n = 485) and COVID-19 acute respiratory failure (n = 882), which were the 2 groups with the most patients, the same findings for COPD exacerbation with a smaller number of patients (n = 77) warrants careful consideration and cautious interpretation.

The broader applicability of the results to patients with COPD exacerbation is also questionable due to the unusually high rates of endotracheal intubation (24%-29%) in this group of patients in the trial⁴ compared with typical clinical experience. Discontinuation of the group including immunocompromised patients based on futility criteria resulted in a small number in this trial group⁴ as well (n = 50), and meant a low likelihood of finding noninferiority between noninvasive ventilation and high-flow oxygen. This outcome stands in contrast to 2 previous trials^9,10 that failed to demonstrate superiority with noninvasive ventilation vs standard or high-flow oxygen, and included large numbers of immunocompromised patients.

The findings from the RENOVATE trial⁴ should be interpreted cautiously. Rather than suggesting that high-flow oxygen can universally replace noninvasive ventilation for all causes of acute respiratory failure, the results are best interpreted as indicating that initiating treatment with high-flow oxygen is generally not harmful. This interpretation of the trial results provides useful clinical guidance for the initial care settings (such as admission to the intensive care unit or emergency department) because it allows time for clinicians to establish the diagnosis and the cause of acute respiratory failure before applying the most appropriate ventilatory support. However, the key to optimal patient care should remain individualization—moving beyond a standardized approach to identify specific risk factors for poor outcomes, particularly endotracheal intubation and mortality risks.

The results from the RENOVATE trial⁴ demonstrate that high-flow oxygen is noninferior to noninvasive ventilation across most acute respiratory failure etiologies. These results suggest that high-flow oxygen can serve as a safe bridge therapy while the underlying cause of acute respiratory failure is determined and the most appropriate respiratory support is ultimately implemented. Further studies are required to better define which specific patient populations might benefit most from alternative first-line treatments such as noninvasive ventilation or continuous positive airway pressure.