SGEM#505: Close Enough for (ARF) Acute Respiratory Failure (HFNO vs NIV)

Date: March 11, 2026

Reference: RENOVATE Investigators and the BRICNet Authors; High-Flow Nasal Oxygen vs Noninvasive Ventilation in Patients With Acute Respiratory Failure: The RENOVATE Randomized Clinical Trial. JAMA March 2025

Guest Skeptic: Dr. Rory Spiegel is an emergency medicine and critical care physician known for his work in evidence-based medicine and critical care. He is widely recognized for translating emerging research into practical bedside insights through lectures, writing, and digital medical education. His work focuses on resuscitation science, airway management, and the critical appraisal of medical literature.

I’m in Maui at the Centre for Continuing Medical Education Year in Review Course. CCME has been doing courses for almost 40 years. The courses take place at amazing locations in the US, including Maui, Hilton Head, Key West, and NYC. CCME recruits four outstanding educators to review ~260 articles from the past year. It’s a unique course because there are no PowerPoint slides to get in the way of the attendees and the speakers. Two faculty members summarize a few articles on a topic in ½ hour with direct interaction with the speakers. You come to this course…you are up to date on the latest EM literature.

Case: A 64-year-old woman with a history of COPD (GOLD stage III) and hypertension presents to the emergency department (ED) with worsening shortness of breath over the past 24 hours. She reports increased sputum production and wheezing. On arrival, she is tachypneic and speaking in short phrases. Her vital signs are heart rate 104 beats per minute, blood pressure 148/86 mm Hg, respiratory rate 30 breaths per minute, and SpO₂ 88% on 4 L nasal cannula. She is using accessory muscles and has diffuse expiratory wheezes on auscultation. An arterial blood gas reveals pH 7.29, PaCO₂ 58 mm Hg, and PaO₂ 62 mm Hg. Chest X-ray shows hyperinflation without focal consolidation.

Background: Acute respiratory failure (ARF) is one of the most common serious respiratory problems managed in emergency medicine and critical care. For decades, noninvasive ventilation (NIV) has been a central part of therapy for selected patients. This is particularly true for those with COPD exacerbations and acute cardiogenic pulmonary edema. By delivering positive pressure, NIV reduces the work of breathing, improves oxygenation and ventilation. This intervention has been shown to reduce intubation rates and mortality in specific populations. However, NIV can be poorly tolerated, requires a tight mask seal and monitoring, and is resource-intensive [1-3]. These downsides can become more problematic in disease states that are not readily reversible over the first few hours.

High-flow nasal oxygen (HFNO) has emerged over the past decade as an attractive potential alternative. By delivering heated, humidified oxygen at high flow rates, HFNO improves oxygenation, improves ventilator efficiency by reducing dead space, and is often better tolerated than mask-based ventilation. Its physiologic appeal and ease of use have led to widespread adoption, particularly during the COVID-19 pandemic. Yet enthusiasm has at times outpaced evidence, and important clinical questions remain:

Is HFNO equivalent/non-inferior to NIV in preventing intubation or death?
How does it perform across different types of respiratory failure?
And when should clinicians choose one over the other?

Clinical Question: Is HFNO noninferior to NIV regarding the rates of endotracheal intubation or death at 7 days across five distinct patient groups with ARF?

Population: Hospitalized adults with ARF (hypoxemia plus respiratory effort or tachypnea) classified into 5 groups:
1. Nonimmunocompromised with hypoxemia
2. Immunocompromised with hypoxemia
3. COPD exacerbation with respiratory acidosis
4. Acute cardiogenic pulmonary edema (ACPE)
5. Hypoxemic COVID-19
  - Exclusions: The main exclusion criteria were if there was an urgent need for endotracheal intubation, hemodynamic instability or contraindications to NIV.
Intervention: High-flow nasal oxygen (HFNO) delivered continuously, titrated toward 60 L/min.
Comparison: Noninvasive ventilation (NIV) delivered through a face mask.
Outcome:
- Primary Outcome: Endotracheal intubation or death within 7 days.
- Secondary Outcomes: 28-day and 90-day mortality, mechanical ventilation-free days, and ICU-free days.
Type of Study: Multicenter, adaptive, noninferiority randomized clinical trial using a Bayesian hierarchical model with dynamic borrowing across patient groups.

Authors’ Conclusions: “Compared with NIV, HFNO met prespecified criteria for noninferiority for the primary outcome of endotracheal intubation or death within 7 days in 4 of the 5 patient groups with ARF. However, the small sample sizes in some patient groups and the sensitivity of the findings to the choice of analysis model suggests the need for further study in patients with COPD, immunocompromised patients, and patients with ACPE.”

Quality Checklist for Randomized Clinical Trials:

The study population included or focused on those in the emergency department. Yes
The patients were adequately randomized. Yes
The randomization process was concealed. Yes
The patients were analyzed in the groups to which they were randomized. Yes
The study patients were recruited consecutively (i.e. no selection bias). Yes
The patients in both groups were similar with respect to prognostic factors. Unsure
All participants (patients, clinicians, outcome assessors) were unaware of group allocation. No
All groups were treated equally except for the intervention. Unsure
Follow-up was complete (i.e. at least 80% for both groups). Yes
All patient-important outcomes were considered. Yes
The treatment effect was large enough and precise enough to be clinically significant. Unsure
Funding: Supported by a grant from the Brazilian Ministry of Health. Fisher & Paykel Healthcare provided the high-flow nasal oxygen equipment and associated disposables. The trial coordinating center and sponsor were the Hcor Research Institute. “The funders/sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.”
Financial conflicts of interest. Multiple authors reported multiple COIs, with one author receiving personal fees from Fisher & Paykel Healthcare.

Results: RENOVATE randomized 1,800 hospitalized adults with ARF across 33 hospitals in Brazil, with 1,766 completing the trial. The mean age was 64 years, and 40% of participants were women. The population was primarily older adults with moderate to severe respiratory failure. The largest of the five pre-defined subgroups (almost half) consisted of patients with hypoxemic COVID-19.

Key Result: HFNO was noninferior to NIV for the composite outcome of endotracheal intubation or death within 7 days in four out of five subgroups.

Primary Outcome: Endotracheal intubation or death
- Overall, 39% HFNO vs 38% NIV
- Noninferiority: Met in four of the pre-specified groups (Nonimmunocompromised, COPD, ACPE, and COVID-19).
- Immunocompromised Group: Stopped for futility (57.1% HFNO vs 36.4% NIV).
Secondary Outcomes: There were no statistically or clinically meaningful differences in 28-day mortality, 90-day mortality, mechanical ventilation-free days, or ICU-free days overall. However, subgroup-specific secondary outcome estimates were imprecise and should be interpreted cautiously.

1. Diverse Etiology: The authors enrolled all adult patients presenting with hypoxic respiratory failure to non-invasive support with either HFNO or NIV. Using such a broad enrollment criterion led them to enroll a wide variety of clinical etiologies in this trial.

The advantages of such broad inclusion criteria mean that the results can be applied broadly to patients presenting with hypoxic respiratory failure. In this case, it is likely accurate to say that neither NIV nor HFNO is superior when treating an undifferentiating population of patients with hypoxic respiratory failure.

A disadvantage is that the 5 subtypes of respiratory failure represent very different physiological causes of ARF, which may respond differently to different forms of respiratory support. There may, in fact, exist potential benefits for either NIV or HFNO in one or more of these specific subgroups that are obscured when looking at this greater population. The authors address this by performing subgroup analyses for each of the 5 predefined subgroups. Unfortunately, due to the small sample sizes in each subgroup, there were too few patients to confidently demonstrate that one form of non-invasive support is preferred over the other.

Therefore, the broad inclusion criteria used by these authors make it very difficult to identify a potential benefit of NIV or HFNO in any one subtype of hypoxic respiratory failure. For example, among immunocompromised patients with hypoxemia, the primary outcome (endotracheal intubation and death within 7 days) occurred in 57.1% of patients in the HFNO group vs 36.4% of patients in the NIV group. Due to the small sample sizes (28 and 22 patients in the two groups, respectively), the 20.7% difference met criteria for futility, as the confidence intervals were too wide to demonstrate non-inferiority or superiority.

Open Label

2. Lack of Masking: This was an open-label trial where both clinicians and patients knew which treatment was assigned. This design can lead to performance bias if clinicians adjust co-interventions, supportive care, or monitoring based on treatment knowledge. It may also cause detection bias, especially with outcomes that involve subjective judgment. For example, the decision to intubate may depend in part on the clinician’s assessment. If clinicians believe one treatment is less effective, they might escalate care earlier, impacting outcome rates. Risk-of-bias assessment tools emphasize concerns about deviations from planned interventions and bias in outcome measurement in unmasked trials. This can be more important when outcomes are not strictly objective.

3. Dynamic Borrowing: The trial used dynamic borrowing, which allows data from one patient group to influence the results of another if their treatment effects are similar. The authors utilized dynamic borrowing across the five defined subgroups of acute respiratory failure. This means that instead of analyzing each group separately, the model allows “borrowing” from other subgroups when the treatment effects look similar across groups.

Dynamic borrowing is a statistical approach used in Bayesian hierarchical models where multiple related patient groups are analyzed together rather than completely separately. Instead of estimating the treatment effect in each subgroup independently, the model allows groups to share data. If subgroup results look consistent, the model blends them slightly to improve precision, especially for smaller groups with less data. If the results look very different, the model reduces the amount of sharing. This improves statistical efficiency and stabilizes estimates in small subgroups, but the trade-off is that if true differences between groups exist, dynamic borrowing can partially smooth them out and make subgroup effects appear more similar than they really are.

In the RENOVATE Trial, dynamic borrowing artificially increased the size of each subgroup, leading to more precise results. It is important to note that when the authors speak of groups being similar or dissimilar, they refer to statistical similarity, or a similarity in measured outcomes, rather than clinical similarity, or a similarity in the physiological mechanisms underlying respiratory compromise. And so, while a technique like borrowing may improve statistical efficiency, especially in smaller subgroups, it may not always be appropriate in heterogeneous clinical populations and may, in fact, obscure real differences that are simply underpowered in the original subgroup. The authors also showed how using this statistical model can shift clinical interpretations. A post hoc analysis without borrowing changed the results for ACPE (showing potential superiority of HFNO) and the immunocompromised/COPD groups (showing potential favouring of NIV).

4. Noninferiority Trials. As our friend Justin Morgenstern from First10EM says, “You don’t understand non-inferiority trials (and neither do I)”. These trials differ fundamentally from superiority trials because their goal is to demonstrate that a new intervention is not significantly worse than a standard treatment by more than a specific margin (delta). This margin must be clinically justified, and choosing the right value is crucial. If delta is too large, it can lead to a clinically important loss of efficacy being considered acceptable. On the other hand, if delta is too small, minor, clinically insignificant differences might lead the trial to be viewed as a failure. Additionally, analytical choices in noninferiority trials can bias results toward the null hypothesis, often leading to noninferiority conclusions. Therefore, it’s good to be skeptical of this study design.

Most pertinent to this trial is that statistical power can drastically affect the interpretation of its results. As the number of patients in the cohort drops, the confidence intervals, or in this case, the credibility intervals, widen. What can occur is you have a point estimate that demonstrates equivalency or even superiority of one treatment over another, but due to the wide confidence/credibility intervals, the non-inferiority margin may still be crossed.

5. Patient Comfort: This was a secondary, supportive outcome. Comfort was measured using a patient-reported scale reflecting how tolerable the respiratory support felt to the patient. This matters because NIV requires a tight-fitting mask that can cause claustrophobia, pressure discomfort, air leaks, and difficulty speaking or eating. In contrast, HFNO uses nasal prongs and is generally perceived as less intrusive. In the RENOVATE trial, HFNO was associated with higher comfort scores than NIV, consistent with prior physiologic and observational studies suggesting improved tolerance. While comfort was not a primary or secondary efficacy endpoint, it is important to consider patients’ preferences and values within an EBM framework.

This goes directly to the non-inferiority trial design. The concept of HFNO offers other advantages, most specifically comfort and logistical simplicity. If it is found to be no worse than the NIV as defined by the non-inferior margin, then these other advantages may outweigh the small clinical benefit NIV provides.

Comment on the Authors’ Conclusion Compared to the SGEM Conclusion: We generally agree with the authors’ conclusions.

Dr. Rory Spiegel

SGEM Bottom Line: This trial reinforces that HFNO and NIV are both seem reasonable first-line strategies for many patients with acute hypoxemic respiratory failure in the ED. It suggests that choosing HFNO for comfort and ease of application is unlikely to worsen mortality compared with NIV. Although the small sample sizes of the individual subgroups make more precise conclusions fraught. In patient populations with prior evidence supporting the use of NIV, like hypercapnia or cardiogenic pulmonary edema, these results are not robust enough to change your first-line choice for non-invasive respiratory support. The same can be said for HFNO in patients with HFNO due to pneumonia or ARDS. But these results support a flexible, phenotype-driven approach rather than a one-size-fits-all respiratory support algorithm and the use of HFNO as an alternative tool in patients who are not tolerating NIV.

Case Resolution: This patient has an acute COPD exacerbation with hypercapnic respiratory acidosis (pH 7.29, PaCO₂ 58 mm Hg), which places her in a group where noninvasive ventilation (NIV) has strong physiologic and evidence-based support as first-line therapy. NIV (BiPAP) reduces the work of breathing, improves alveolar ventilation, corrects hypercapnia, and decreases the need for intubation and mortality in COPD exacerbations with acidosis. While the RENOVATE trial included COPD patients and suggested HFNO may be noninferior overall, the evidence still favours NIV in hypercapnic respiratory failure. Therefore, this patient should receive NIV promptly along with bronchodilators, systemic corticosteroids, and appropriate treatment of any precipitating infection, reserving HFNO only if NIV is not tolerated or contraindicated.

Clinical Application: Consider HFNO as a first-line alternative to NIV for many patients with ARF, while understanding the physiological causes of each patient’s respiratory distress and the specific benefits of both HFNO and NIV.

What Do I Tell the Patient? We are going to use a special mask to help prevent having to put a large tube down your throat. If you cannot tolerate the mask, we can switch to another device that blows humidified oxygen up your nose.

Keener Kontest: Last episode’s winner was Ryker Kiel from Wyoming. He knew that according to the study from Alberta, the factor associated with the largest relative increase in ED length of stay for admitted patients was inpatient capacity issues.

Other FOAMed:

SGEM#7: Every Breath You Take (Non-Invasive Ventilation)
SGEM#135: The Answer My Friend is Blowin’ in your Nose – High Flow Nasal Oxygen
SGEM#186: Apneic and the O, O, O2 for Rapid Sequence Intubation
SGEM#228: Winds of Change – High Flow Nasal Oxygen for Acute Bronchiolitis?
SGEM #379: Heigh Ho High Flow versus CPAP in Acutely Ill Children
SGEM #401: Hey Ho! High Flow vs Standard Oxygen Therapy for Hospitalized Children with Respiratory Failure
SGEM#447: Just What I Needed – Preoxygenation Prior To Intubation
SGEM#477: I Can Feel It Coming In the Air Tonight…But By Which Pre-Oxygenation Strategy

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.

References:

Rochwerg B, Brochard L, Elliott MW, Hess D, Hill NS, Nava S, Navalesi P Members Of The Steering Committee, Antonelli M, Brozek J, Conti G, Ferrer M, Guntupalli K, Jaber S, Keenan S, Mancebo J, Mehta S, Raoof S Members Of The Task Force. Official ERS/ATS clinical practice guidelines: noninvasive ventilation for acute respiratory failure. Eur Respir J. 2017 Aug 31;50(2):1602426. doi: 10.1183/13993003.02426-2016. PMID: 28860265.
Osadnik CR, Tee VS, Carson-Chahhoud KV, Picot J, Wedzicha JA, Smith BJ. Non-invasive ventilation for the management of acute hypercapnic respiratory failure due to exacerbation of chronic obstructive pulmonary disease. Cochrane Database Syst Rev. 2017 Jul 13;7(7):CD004104. doi: 10.1002/14651858.CD004104.pub4. PMID: 28702957; PMCID: PMC6483555.
Berbenetz N, Wang Y, Brown J, Godfrey C, Ahmad M, Vital FM, Lambiase P, Banerjee A, Bakhai A, Chong M. Non-invasive positive pressure ventilation (CPAP or bilevel NPPV) for cardiogenic pulmonary oedema. Cochrane Database Syst Rev. 2019 Apr 5;4(4):CD005351. doi: 10.1002/14651858.CD005351.pub4. PMID: 30950507; PMCID: PMC6449889.