Effect of Postextubation High-Flow Nasal Oxygen With Noninvasive Ventilation vs High-Flow Nasal Oxygen Alone on Reintubation Among Patients at High Risk of Extubation Failure: A Randomized Clinical Trial

Question  Among mechanically ventilated patients at high risk of extubation failure, does the use of high-flow nasal oxygen with noninvasive ventilation after extubation reduce the risk of reintubation compared with high-flow nasal oxygen alone?

Findings  In this randomized clinical trial that included 641 patients, high-flow nasal oxygen with noninvasive ventilation, compared with high-flow nasal oxygen alone, significantly decreased the rate of reintubation within the first 7 days after extubation (11.8% vs 18.2%).

Meaning  In patients at high risk of extubation failure, the use of high-flow nasal oxygen with noninvasive ventilation after extubation significantly decreased the risk of reintubation compared with high-flow nasal oxygen alone.

Importance  High-flow nasal oxygen may prevent postextubation respiratory failure in the intensive care unit (ICU). The combination of high-flow nasal oxygen with noninvasive ventilation (NIV) may be an optimal strategy of ventilation to avoid reintubation.

Objective  To determine whether high-flow nasal oxygen with prophylactic NIV applied immediately after extubation could reduce the rate of reintubation, compared with high-flow nasal oxygen alone, in patients at high risk of extubation failure in the ICU.

Design, Setting, and Participants  Multicenter randomized clinical trial conducted from April 2017 to January 2018 among 641 patients at high risk of extubation failure (ie, older than 65 years or with an underlying cardiac or respiratory disease) at 30 ICUs in France; follow-up was until April 2018.

Interventions  Patients were randomly assigned to high-flow nasal oxygen alone (n = 306) or high-flow nasal oxygen alternating with NIV (n = 342) immediately after extubation.

Main Outcomes and Measures  The primary outcome was the proportion of patients reintubated at day 7; secondary outcomes included postextubation respiratory failure at day 7, reintubation rates up until ICU discharge, and ICU mortality.

Results  Among 648 patients who were randomized (mean [SD] age, 70 [10] years; 219 women [34%]), 641 patients completed the trial. The reintubation rate at day 7 was 11.8% (95% CI, 8.4%-15.2%) (40/339) with high-flow nasal oxygen and NIV and 18.2% (95% CI, 13.9%-22.6%) (55/302) with high-flow nasal oxygen alone (difference, −6.4% [95% CI, −12.0% to −0.9%]; P = .02). Among the 11 prespecified secondary outcomes, 6 showed no significant difference. The proportion of patients with postextubation respiratory failure at day 7 (21% vs 29%; difference, −8.7% [95% CI, −15.2% to −1.8%]; P = .01) and reintubation rates up until ICU discharge (12% vs 20%, difference −7.4% [95% CI, −13.2% to −1.8%]; P = .009) were significantly lower with high-flow nasal oxygen and NIV than with high-flow nasal oxygen alone. ICU mortality rates were not significantly different: 6% with high-flow nasal oxygen and NIV and 9% with high-flow nasal oxygen alone (difference, −2.4% [95% CI, −6.7% to 1.7%]; P = .25).

Conclusions and Relevance  In mechanically ventilated patients at high risk of extubation failure, the use of high-flow nasal oxygen with NIV immediately after extubation significantly decreased the risk of reintubation compared with high-flow nasal oxygen alone.

Trial Registration  ClinicalTrials.gov Identifier: NCT03121482

In intensive care units (ICUs), approximately 10% to 15% of patients ready to be separated from a ventilator experience extubation failure leading to reintubation.¹ In patients considered at high risk, these rates can even exceed 20%.^1,2 Because reintubation is associated with particularly high mortality,^3,4 a strategy of oxygenation aimed at avoiding reintubation deserves consideration. Although noninvasive ventilation may prevent postextubation respiratory failure in patients at high risk,^5-9 only 2 small-scale randomized clinical trials (RCTs) have shown decreased reintubation rates compared with standard oxygen.^5,6 The most recent international clinical practice guidelines recommend the use of noninvasive ventilation to prevent postextubation respiratory failure in patients at high risk of extubation failure.¹⁰ However, up until now, no large-scale RCT has demonstrated a significant reduction of reintubation rates with noninvasive ventilation compared with standard oxygen. Thereby, most patients are treated with standard oxygen in clinical practice and only 10% of them receive noninvasive ventilation after extubation in the ICU.^2,11

High-flow nasal oxygen is an alternative strategy that may reduce the risk of reintubation in the ICU compared with standard oxygen.^12,13 A large-scale RCT has reported that high-flow nasal oxygen was noninferior to noninvasive ventilation in preventing reintubation in patients at high risk.¹⁴ Whereas high-flow nasal oxygen could be considered as a reference treatment after extubation, using high-flow nasal oxygen with noninvasive ventilation may further improve gas exchange and the work of breathing,¹⁵ thereby avoiding reintubation.

Quiz Ref IDThis multicenter RCT involving patients at high risk of extubation failure in the ICU was conducted to determine whether high-flow nasal oxygen with noninvasive ventilation, compared with high-flow nasal oxygen alone, after extubation could reduce the rate of reintubation.

The study was conducted in 30 ICUs in France. For all the centers, the study protocol (Supplement 1) was approved by the central ethics committee (Ethics Committee Ouest III, Poitiers, France; registration No. 2016-A01078-43). Written informed consent was obtained from all patients or next of kin before inclusion in the study.

Quiz Ref IDAdult patients intubated more than 24 hours in the ICU and ready for extubation, after a successful spontaneous breathing trial performed according to the international conference consensus on weaning,¹⁶ were enrolled if they were at high risk of extubation failure (ie, older than 65 years or had any underlying chronic cardiac or lung disease).¹⁰ Underlying chronic cardiac diseases included left ventricular dysfunction, whatever the cause, defined by left ventricular ejection fraction equal to or below 45%; history of cardiogenic pulmonary edema; documented ischemic heart disease; or permanent atrial fibrillation. Underlying chronic lung diseases included chronic obstructive pulmonary disease, obesity-hypoventilation syndrome, or restrictive pulmonary disease.

The main exclusion criteria were long-term treatment with noninvasive ventilation or continuous positive airway pressure at home, contraindication to noninvasive ventilation, underlying chronic neuromuscular disease (myopathy or myasthenia gravis), or traumatic brain injury leading to intubation, as well as patients who underwent unplanned extubation (accidental or self-extubation) or with a do-not-reintubate order at time of extubation.

The trial was overseen by a steering committee that presented information regarding the progression and monitoring of the study at REVA (Réseau Européen de Recherche en Ventilation Artificielle) Network meetings every 6 months. No safety committee was required because the interventions used in the study were strategies of oxygenation typically used in clinical practice. Research assistants regularly monitored all the centers on site to check adherence to the protocol and the accuracy of the data recorded. An investigator at each center was responsible for daily patient screening, enrolling patients in the study, ensuring adherence to the protocol, and completing the electronic case-report form. Although the individual study assignments of the patients could not be masked, the coordinating center and all the investigators remained unaware of the study group outcomes until the data were locked in January 2019.

Randomization was computer-generated using a centralized web-based management system in permuted blocks of 4 participants (unknown to investigators), with stratification according to the center and arterial partial pressure of carbon dioxide (Paco₂) level (≤45 or >45 mm Hg) measured at the end of the spontaneous breathing trial (or under mechanical ventilation before the trial if this latter was not measured). Stratification was performed on Paco₂ level to include the same number of hypercapnic patients in the 2 groups because noninvasive ventilation may be more effective in these patients.^7,8 Patients were randomly assigned in a 1:1 ratio to receive high-flow nasal oxygen alone (control group) or with noninvasive ventilation (intervention group) immediately after extubation (Figure 1).

Patients assigned to the control group were continuously treated by high-flow nasal oxygen alone for at least 48 hours with a flow of 50 L/min and fraction of inspired oxygen (Fio₂) adjusted to obtain adequate oxygenation, with an oxygen saturation as measured by pulse oximetry (Spo₂) of at least 92%. To provide sufficient humidification, the temperature of the heated humidifier was set at 37°C as during invasive mechanical ventilation.

Patients assigned to the intervention group (referred to here as the noninvasive ventilation group) were treated with high-flow nasal oxygen alternating with noninvasive ventilation. Noninvasive ventilation was initiated immediately after extubation with a first session of at least 4 hours and minimal duration of at least 12 hours a day during the 48 hours following extubation. Continuous application of noninvasive ventilation was promoted throughout the entire night period. Noninvasive ventilation was carried out with an ICU ventilator with noninvasive ventilation mode or dedicated bilevel ventilator in pressure-support mode with a minimal pressure-support level of 5 cm H₂O targeting a tidal volume around 6 to 8 mL/kg of predicted body weight, a positive end-expiratory pressure level between 5 and 10 cm H₂O, and a Fio₂ adjusted to obtain adequate oxygenation (Spo₂ ≥92%). Between noninvasive ventilation sessions, high-flow nasal oxygen was delivered as in the control group. Blood gases were performed 1 hour after treatment initiation under high-flow oxygen in the high-flow nasal oxygen alone group and under noninvasive ventilation in the noninvasive ventilation group. In the 2 groups, patients were treated for a minimum of 48 hours. When there were no signs of respiratory failure 48 hours after extubation, treatment was stopped and switched to standard oxygen. According to patient respiratory status, treatment could be continued until complete respiratory recovery. In case of established postextubation respiratory failure, the use of noninvasive ventilation was discouraged in accordance with the most recent international clinical practice guidelines,¹⁰ given that it has no proven benefit¹⁷ and can even increase the risk of death by delaying reintubation.¹⁸

The primary outcome was the proportion of patients who required reintubation within the 7 days following extubation. To ensure the consistency of indications across sites and reduce the risk of delayed intubation, patients were immediately reintubated if 1 of the following criteria was fulfilled: severe respiratory failure, hemodynamic failure with the need for vasopressors, neurological failure (altered consciousness with a Glasgow Coma Scale score <12), or cardiac or respiratory arrest. Severe respiratory failure leading to reintubation was defined by the presence of at least 2 criteria among the following: a respiratory rate greater than 35 breaths per minute, clinical signs suggesting respiratory distress with activation of accessory respiratory muscles, respiratory acidosis defined as a pH level below 7.25 units and Paco₂ greater than 45 mm Hg, hypoxemia defined as a need for Fio₂ at 80% or more to maintain an Spo₂ level at 92% or more, or a ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (Pao₂:Fio₂) equal to or below 100 mm Hg.

Secondary outcomes included reintubation at 48 hours, 72 hours, and up until ICU discharge; an episode of postextubation respiratory failure within 7 days following extubation; the proportion of patients in whom the treatment was continued beyond the first 48 hours following extubation; length of stay in the ICU and in the hospital; and mortality in the ICU, in the hospital, at day 28, and at day 90. An episode of postextubation respiratory failure was defined by the presence of at least 2 criteria among the following: a respiratory rate greater than 25 breaths per minute, clinical signs suggesting respiratory distress, respiratory acidosis defined as a pH level less than 7.35 units and Paco₂ level greater than 45 mm Hg, hypoxemia defined as a need for Fio₂ at least 50% to maintain Spo₂ level of at least 92%, or a Pao₂:Fio₂ ratio equal to or below 150 mm Hg.

Exploratory outcomes included blood gases 1 hour after treatment initiation, time to reintubation, the proportion of patients who met criteria for reintubation, reasons for reintubation, use of noninvasive ventilation as rescue therapy, the proportion of patients who were reintubated or died in the ICU, and mortality of reintubated patients.

Enrollment of 590 patients was determined to provide a power of 80% and to show an absolute difference of 8% in the rate of reintubation between the control group using high-flow nasal oxygen alone (rate of reintubation estimated to 18%) compared with the intervention group using high-flow nasal oxygen and noninvasive ventilation (rate of reintubation estimated to 10%) at a 2-sided α level of .05 (ie, exactly the same difference as that planned in a previous RCT comparing high-flow nasal oxygen vs standard oxygen on reintubation among patients at low risk of extubation failure¹³). To allow for the potential secondary exclusions, the number of patients to be enrolled was then inflated to 650 patients (increased by 10%).

All the analyses were performed by the study statistician according to a predefined statistical analysis plan. The analysis was performed on all randomized patients who were extubated and for whom the primary outcome was completed. Patients were analyzed according to their randomization group regardless of the treatment applied. The proportions of patients having needed reintubation within the 7 days following planned extubation were compared between the 2 groups by means of the χ² test.

Kaplan-Meier curves were plotted to assess time from extubation to reintubation and were compared by means of the log-rank test at day 7. Reintubation rates at the various predefined times, postextubation respiratory failure rates at day 7, and mortality rates in the ICU and hospital were compared between the 2 groups by means of the χ² test. Kaplan-Meier curves were plotted to assess the time from extubation to death and were compared by means of the log-rank test at day 90.

A multiple logistic regression analysis was performed for the primary outcome to adjust on the stratification variable (Paco₂ level) and on potential baseline unbalanced variables. Lack of balance was defined as P < .05, and the only variable ultimately included in the model was underlying chronic lung disease. The results were presented as odds ratios with 95% CIs. A post hoc random-effects multilevel logistic regression model was used to take into account the effect of the hospital. Treatment group was introduced in the model as a fixed effect and hospital was introduced in the model as a random effect. A subgroup analysis was performed for the primary and secondary outcomes according to the Paco₂ level (≤45 or >45 mm Hg) prior to extubation after an interaction test carried out to detect heterogeneity of treatment effect between hypercapnic and nonhypercapnic patients. Because of the potential for type I error due to multiple comparisons, findings for analyses of secondary end points should be interpreted as exploratory. A 2-tailed P value of less than .05 was considered to indicate statistical significance. We used SAS software version 9.4 (SAS Institute) for all analyses.

From April 2017 through January 2018, 3121 patients were extubated in the 30 participating units, 969 were eligible for inclusion in the study, and 648 underwent randomization (mean [SD] age, 70 [10] years; 219 women [34%]) (Figure 1). Seven patients were secondarily excluded because they were protected under French law, which was not known at randomization (n = 3), died before extubation (n = 1), or were missing data for the primary outcome (n = 3), leaving 641 patients included in the analysis: 302 patients were assigned to high-flow nasal oxygen alone and 339 to high-flow nasal oxygen with non-invasive ventilation.

The characteristics of the patients at inclusion were similar in the 2 groups except for a higher proportion of patients with underlying chronic lung disease in the noninvasive ventilation group (Table 1). The median duration of mechanical ventilation prior to extubation was 5 days (interquartile range [IQR], 3-10), and weaning was considered difficult or prolonged in 32% of patients (206 of 641 patients). At the time of extubation, 111 patients (17%) had hypercapnia (Paco₂ >45 mm Hg).

Initial mean (SD) settings were as follows: in the high-flow nasal oxygen alone group, the gas flow rate was 50 (5) L/min with Fio₂ of 0.41 (0.13); in the noninvasive ventilation group, the pressure-support level was 7.8 (2.5) cm H₂O, PEEP was 5.3 (1.1) cm H₂O, and Fio₂ was 0.34 (0.10), resulting in a tidal volume of 8.6 (2.9) mL/kg of predicted body weight.

In the noninvasive ventilation group, noninvasive ventilation was delivered for a mean (SD) of 22 (9) hours within the first 48 hours following extubation (mean of 13 hours within the first 24 hours) and was delivered for 4 hours or less due to intolerance in 20 patients (6%). In the high-flow nasal oxygen alone group, high flow nasal oxygen was delivered for a mean (SD) of 42 (11) hours within the first 48 hours.

Quiz Ref IDThe reintubation rate at day 7 was 11.8% (95% CI, 8.4%-15.2%) with noninvasive ventilation and 18.2% (95% CI, 13.9%-22.6%) with high-flow nasal oxygen alone (difference, −6.4% [95% CI, −12.0 to −0.9]; P = .02) (Figure 2).

Reintubation rates were also significantly lower with noninvasive ventilation than with high-flow nasal oxygen alone at 48 hours, 72 hours, and until ICU discharge (Table 2). The proportion of patients with postextubation respiratory failure at day 7 was significantly lower with noninvasive ventilation than with high-flow nasal oxygen alone (21% vs 29%; difference, −8.7% [95% CI, −15.2% to −1.8%]; P = .01). In the noninvasive ventilation group, noninvasive ventilation was continued beyond the first 48 hours for incomplete recovery of respiratory status in 86 patients (25%), whereas in the high-flow nasal oxygen alone group, high-flow nasal oxygen was continued in 106 patients (35%) (difference, −9.7% [95% CI, −16.8% to −2.6%]; P < .01). Quiz Ref IDMortality in the ICU, in the hospital, and at day 90 were not significantly different between groups (Table 2; eFigure in Supplement 2).

One hour after treatment initiation, Pao₂:FiO₂ was higher with noninvasive ventilation than with high-flow nasal oxygen alone (mean [SD], 291 [97] mm Hg vs 254 [113] mm Hg; difference, 37.0 [95% CI, 19.7 to 54.3]; P < .001), whereas the proportion of patients with hypercapnia did not differ (21% vs 19%, respectively; difference, 1.2% [95% CI, −5.4% to 7.7%]; P = .72).

The median time to reintubation was not significantly different between groups: 33 hours (IQR, 7-81) with noninvasive ventilation and 39 hours (IQR, 12-67) with high-flow nasal oxygen alone (difference, −5.0 [95% CI, −42.0 to 32.0]; P = .76).

Among the 100 patients who were reintubated in the ICU, 96% met prespecified criteria for reintubation: 95% (39/41) in the noninvasive ventilation group vs 97% (57/59) in the high-flow nasal oxygen alone group (difference, −1.5% [95% CI, −13.0% to 7.4%]; P = .99). The reason for reintubation was severe respiratory failure in 88 patients, neurological failure in 37 patients, hemodynamic failure in 16 patients, and respiratory or cardiac arrest in 10 patients (eTable 1 in Supplement 2).

Among the 88 patients who had postextubation respiratory failure with high-flow nasal oxygen alone, 28 patients (32%) were treated with noninvasive ventilation as rescue therapy delivered for a mean (SD) of 20 (18) hours, of whom 12 patients (43%) needed reintubation. Results for additional exploratory outcomes are shown in Table 2.

No significant interaction was noted between Paco₂ at enrollment and treatment group with respect to the primary outcome (P for interaction = .25). Among the 111 patients with Paco₂ greater than 45 mm Hg before extubation, the reintubation rate at day 7 was significantly lower with noninvasive ventilation than with high-flow nasal oxygen alone (8% vs 21%; difference, −12.9% [95% CI, −27.1% to −0.1%]; P = .049) (Figure 3). Among the 530 patients with Paco₂ of 45 mm Hg or less, reintubation rates at day 7 were not significantly different between groups (13% with noninvasive ventilation vs 18% with high-flow nasal oxygen alone; difference, −5.0% [95% CI, −11.2% to 1.1%]; P = .10) (eTables 2, 3, and 4 in Supplement 2).

After adjustment for Paco₂ level at enrollment (≤45 or >45 mm Hg, stratification randomization variable) and underlying chronic lung disease (variable unbalanced between both groups indicated in Table 1), the odds ratio for reintubation at day 7 was significantly lower with noninvasive ventilation than with high-flow nasal oxygen alone (adjusted odds ratio, 0.60 [95% CI, 0.38-0.93]; P = .02). The post hoc analysis showed a lower reintubation rate with noninvasive ventilation than with high-flow nasal oxygen alone after adjustment for the hospital random effect (P = .02 without hospital random effect, and P = .02 after adjustment for the hospital random effect) (eTable 5 in Supplement 2).

During the study, there were no severe adverse events attributable to the randomization group.

In this multicenter, randomized, open-label trial, high-flow nasal oxygen with noninvasive ventilation, compared with high-flow nasal oxygen alone, decreased the rate of reintubation within the first 7 days after extubation in the ICU.

This study was designed to assess noninvasive ventilation in a large population of patients who are particularly easy to identify and extubated daily in the different ICUs. Although noninvasive ventilation may be beneficial on outcomes of hypercapnic patients,^7,8 these patients account for only about 20% to 30% of patients at high risk of extubation failure in the ICU.^7,19 Patients older than 65 years or with underlying chronic cardiac or respiratory disease are also at high risk of reintubation²⁰ and could benefit from noninvasive ventilation.²¹

To our knowledge, the combination of high-flow nasal oxygen with noninvasive ventilation had not been previously assessed after extubation in the ICU. A preliminary study observed a reintubation rate of 15% at day 7 with noninvasive ventilation and standard oxygen in exactly the same population.²¹ Therefore, the study hypothesized that a new strategy combining high-flow nasal oxygen with noninvasive ventilation could further reduce the rate of reintubation, whereas the estimated rate would exceed 15% in the control group.^13,14 For an overall reintubation rate around 15% in the ICU,²² an absolute difference of at least 5% (relative reduction of one-third) would be considered clinically significant and in the range of previous large multicenter RCTs assessing reintubation as the main outcome.^13,23 Reintubation rates were almost exactly the expected rates in the 2 groups (18.2% and 11.8%), reinforcing the external validity of the study.

To date, only 2 RCTs have observed lower reintubation rates with noninvasive ventilation than with standard oxygen.^5,6 To our knowledge, the present study is the largest RCT showing a reduced risk of reintubation after extubation in the ICU with noninvasive ventilation. Unlike a previous RCT that reported similar reintubation rates between noninvasive ventilation and high-flow nasal oxygen applied 24 hours after extubation in nonhypercapnic patients,¹⁴ this study combined noninvasive ventilation and high-flow nasal oxygen for at least 48 hours and treatment was prolonged if necessary. Although the beneficial effects of noninvasive ventilation on oxygenation, alveolar ventilation, and work of breathing are well demonstrated,^24,25 continuation of high-flow nasal oxygen between sessions of noninvasive ventilation may further provide clinical improvement by decreasing work of breathing.^15,26

Quiz Ref IDApproximately one-third of patients treated with high-flow nasal oxygen alone received noninvasive ventilation as rescue therapy in case of postextubation respiratory failure. Although noninvasive ventilation as rescue therapy may avoid reintubation in a number of cases, it has been shown to possibly be deleterious and increase mortality in this setting.¹⁸ Moreover, international clinical practice guidelines suggest that noninvasive ventilation should not be used in the treatment of patients with established postextubation respiratory failure.¹⁰

This study has several limitations. First, high-flow nasal oxygen rather than standard oxygen was used in the control group. However, it has been shown that reintubation rates with high-flow nasal oxygen were reduced as compared with standard oxygen.^12,13 According to clinical practice in participating centers, the use of standard oxygen alone would have been considered a suboptimal strategy for patients at high risk. Therefore, high-flow nasal oxygen was used in the control group to promote equipoise and facilitate inclusions in different centers.

Second, attending physicians could not be blinded to the study group and this could have modified the decision of reintubation by promoting early reintubation in patients treated with high-flow nasal oxygen alone. However, almost all reintubated patients met prespecified criteria for reintubation, and they had particularly high mortality (exceeding 30%), confirming high severity of patients who were reintubated.

Third, the weaning protocol and the type of spontaneous breathing trial performed before extubation may have influenced the results.²⁷ In addition, inclusion criteria identifying patients at high risk were different from previous studies.^6,13,14 However, international clinical practice guidelines specify that patients at high risk who may benefit from noninvasive ventilation are those older than 65 years or who have any underlying cardiac or respiratory disease.¹⁰

In mechanically ventilated patients at high risk of extubation failure, the use of high-flow nasal oxygen with noninvasive ventilation immediately after extubation significantly decreased the risk of reintubation compared with high-flow nasal oxygen alone.

Corresponding Author: Arnaud W. Thille, MD, PhD, Médecine Intensive Réanimation, CHU de Poitiers, 2 rue la Milétrie, 86021 Poitiers Cedex, France (aw.thille@gmail.com).

Accepted for Publication: September 9, 2019.

Published Online: October 2, 2019. doi:10.1001/jama.2019.14901

Correction: This article was corrected online on February 25, 2020, to clarify the intervention description in the Abstract and Methods section; to correct data reported for exploratory outcomes in the Results section; and to fix the x-axis labels in Figure 2 and Figure 3.

Author Contributions: Dr Thille had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors give their agreement to be accountable for all aspects of the work, and ensure the accuracy and integrity of any part of the work.

Concept and design: Thille, Girault, Dellamonica, Lacave, Zanre, Ragot, Frat.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Thille, Coudroy, Girault, Cabasson, Nanadoumgar, Zanre, Demoule, Ragot, Frat.

Critical revision of the manuscript for important intellectual content: Thille, Muller, Gacouin, Coudroy, Decavèle, Sonneville, Beloncle, Girault, Dangers, Lautrette, Rouzé, Vivier, Le Meur, Ricard, Razazi, Barberet, Lebert, Ehrmann, Sabatier, Bourenne, Pradel, Bailly, Terzi, Dellamonica, Lacave, Danin, Gibelin, Zanre, Deye, Demoule, Maamar, Nay, Robert, Ragot, Frat.

Statistical analysis: Thille, Zanre, Ragot.

Obtained funding: Thille, Pradel, Zanre.

Administrative, technical, or material support: Thille, Gacouin, Dangers, Rouzé, Le Meur, Razazi, Lebert, Lacave, Zanre, Demoule, Robert.

Supervision: Thille, Zanre, Robert, Ragot.

Conflict of Interest Disclosures: Dr Thille reported receiving grants from the French Ministry of Health and personal fees and nonfinancial support from Fisher & Paykel Healthcare during the conduct of the study and personal fees from Maquet-Getinge, GE Healthcare, and Covidien outside the submitted work. Dr Sonneville reported receiving grants from the French Ministry of Health, the European Society of Intensive Care Medicine, and the French Society of Intensive Care Medicine and personal fees from Baxter outside the submitted work. Dr Beloncle reported receiving personal fees from Lowenstein Medical and nonfinancial support from GE Healthcare, Getinge Group, and Covidien outside the submitted work. Dr Girault reported receiving grants, personal fees, and nonfinancial support from Fisher & Paykel Healthcare during the conduct of the study and grants and nonfinancial support from ResMed outside the submitted work. Dr Ricard reported receiving travel and accommodation expenses from Fisher & Paykel Healthcare outside the submitted work. Dr Ehrmann reported receiving grants, nonfinancial support, and other funding from Fisher & Paykel Healthcare during the conduct of the study; grants, personal fees, nonfinancial support, and other funding from Aerogen; grants from Hamilton; personal fees from La Diffusion Technique Française; and personal fees from Baxter outside the submitted work. In addition, Dr Ehrmann had a patent to EP17305015 issued. Dr Terzi reported receiving personal fees from Boehringer Ingelheim and Pfizer outside the submitted work. Dr Danin reported receiving fees for lectures from Fisher and Paykel during the conduct of the study. Dr Deye reported receiving lecture and travel fees from Zoll and Bard outside the submitted work. Dr Demoule reported receiving personal fees from Medtronic, Baxter, Hamilton, and Getinge; grants, personal fees, and nonfinancial support from Philips and Lungpacer; personal fees and nonfinancial support from Fisher & Paykel Healthcare; and grants from the French Ministry of Health and Respinor outside the submitted work. Dr Frat reported receiving personal fees and nonfinancial support from Fisher & Paykel Healthcare during the conduct of the study and personal fees and nonfinancial support from SOS Oxygen outside the submitted work. No other disclosures were reported.

Funding/Support: The study was funded by the “Programme Hospitalier de Recherche Clinique National 2015” of the French Ministry of Health through the University Hospital of Poitiers, Poitiers, France.

Role of the Funder/Sponsor: The funder 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. Fisher & Paykel provided to all the participating centers the high-flow nasal oxygen equipment and masks for noninvasive ventilation but had no other involvement in the study.

Group Information: Members of the HIGH-WEAN Study Group and REVA Research Network include the following: Florence Boissier (Centre Hospitalier Universitaire de Poitiers, Médecine Intensive Réanimation, Poitiers), Delphine Chatellier (Centre Hospitalier Universitaire de Poitiers, Médecine Intensive Réanimation, Poitiers), Céline Deletage (Centre Hospitalier Universitaire de Poitiers, Médecine Intensive Réanimation, Poitiers), Carole Guignon (Centre Hospitalier Universitaire de Poitiers, Médecine Intensive Réanimation, Poitiers), Florent Joly (Centre Hospitalier Universitaire de Poitiers, Médecine Intensive Réanimation, Poitiers), Morgane Olivry (Centre Hospitalier Universitaire de Poitiers, Médecine Intensive Réanimation, Poitiers), Anne Veinstein (Centre Hospitalier Universitaire de Poitiers, Médecine Intensive Réanimation, Poitiers), Dalila Benzekri-Lefevre (Groupe Hospitalier Régional d’Orléans, Médecine Intensive Réanimation, Orléans), Thierry Boulain (Groupe Hospitalier Régional d’Orléans, Médecine Intensive Réanimation, Orléans), Yves Le Tulzo (Centre Hospitalier Universitaire de Rennes, Hôpital Ponchaillou, Service des Maladies Infectieuses et Réanimation Médicale, Rennes), Jean-Marc Tadié (Centre Hospitalier Universitaire de Rennes, Hôpital Ponchaillou, Service des Maladies Infectieuses et Réanimation Médicale, Rennes), Suela Demiri (Groupe Hospitalier Pitié-Salpêtrière Charles Foix, Service de Pneumologie et Réanimation Médicale, AP-HP, Paris), Julien Mayaux (Groupe Hospitalier Pitié-Salpêtrière Charles Foix, Service de Pneumologie et Réanimation Médicale, Paris), Lila Bouadma (Hôpital Bichat–Claude Bernard, Médecine Intensive Réanimation, Paris), Claire Dupuis (Hôpital Bichat–Claude Bernard, Médecine Intensive Réanimation, Paris), Pierre Asfar (Centre Hospitalier Universitaire d’Angers, Département de Médecine Intensive Réanimation, Angers), Marc Pierrot (Centre Hospitalier Universitaire d’Angers, Département de Médecine Intensive Réanimation, Angers), Gaëtan Béduneau (Centre Hospitalier Universitaire de Rouen, Hôpital Charles Nicolle, Département de Réanimation Médicale, Rouen), Déborah Boyer (Centre Hospitalier Universitaire de Rouen, Hôpital Charles Nicolle, Département de Réanimation Médicale, Rouen), Benjamin Delmas (Centre Hospitalier Universitaire Félix Guyon, Service de Réanimation Polyvalente, Saint Denis de la Réunion), Bérénice Puech (Centre Hospitalier Universitaire Félix Guyon, Service de Réanimation Polyvalente, Saint Denis de la Réunion), Konstantinos Bachoumas (Centre Hospitalier Universitaire de Clermont-Ferrand, Hôpital Gabriel Montpied, Clermont-Ferrand), Edouard Soum (Centre Hospitalier Universitaire de Clermont–Ferrand, Hôpital Gabriel Montpied, Clermont-Ferrand), Marie-Anne Hoppe (Centre Hospitalier de La Rochelle, Service de Réanimation, La Rochelle), Quentin Levrat (Centre Hospitalier de La Rochelle, Service de Réanimation, La Rochelle), Saad Nseir (Centre Hospitalier Universitaire de Lille, Center de Réanimation, Lille), Olivier Pouly (Centre Hospitalier Universitaire de Lille, Center de Réanimation, Lille), Gaël Bourdin (Hôpital Saint-Joseph Saint-Luc, Réanimation Polyvalente, Lyon), Sylvène Rosselli (Hôpital Saint-Joseph Saint-Luc, Réanimation Polyvalente, Lyon), Charlotte Garret (Centre Hospitalier Universitaire de Nantes, Médecine Intensive Réanimation, Nantes), Maelle Martin (Centre Hospitalier Universitaire de Nantes, Médecine Intensive Réanimation, Nantes), Guillaume Berquier (Hôpital Louis Mourier, Réanimation Médico-Chirurgicale, Colombes) Abirami Thiagarajah (Hôpital Louis Mourier, Réanimation Médico-Chirurgicale, Colombes), Guillaume Carteaux (Hôpitaux Universitaires Henri Mondor, Service de Réanimation Médicale, Créteil), Armand Mekontso-Dessap (Hôpitaux Universitaires Henri Mondor, Service de Réanimation Médicale, Créteil), Antoine Poidevin (Groupe Hospitalier Régional Mulhouse Sud Alsace, site Emile Muller, Service de Réanimation Médicale, Mulhouse), Anne-Florence Dureau (Groupe Hospitalier Régional Mulhouse Sud Alsace, site Emile Muller, Service de Réanimation Médicale, Mulhouse), Marie-Ange Azais (Centre Hospitalier Départemental de Vendée, Service de Médecine Intensive Réanimation, La Roche Sur Yon), Gwenhaël Colin (Centre Hospitalier Départemental de Vendée, Service de Médecine Intensive Réanimation, La Roche Sur Yon), Emmanuelle Mercier (Centre Hospitalier Régional Universitaire de Tours, Médecine Intensive Réanimation, Tours), Marlène Morisseau (Centre Hospitalier Régional Universitaire de Tours, Médecine Intensive Réanimation, Tours), Alexandre Massri (Centre Hospitalier de Pau, Service de Réanimation, Pau), Walter Picard (Centre Hospitalier de Pau, Service de Réanimation, Pau), Marc Gainnier (CHU La Timone 2, Médecine Intensive Réanimation, Marseille), Thi-My-Hue Nguyen (Centre Hospitalier Henri Mondor d’Aurillac, Service de Réanimation, Aurillac), Gwenaël Prat (Centre Hospitalier Universitaire de Brest, Médecine Intensive Réanimation, Brest), Carole Schwebel (Centre Hospitalier Universitaire Grenoble Alpes, Médecine Intensive Réanimation, Grenoble), and Matthieu Buscot (Centre Hospitalier Universitaire de Nice, Réanimation Médicale Archet 1, Université Cote d’Azur, Nice).

Meeting Presentation: Presented at the annual congress of the European Society of Intensive Care Medicine, October 2, 2019, Berlin, Germany.

Additional Contributions: We thank Jeffrey Arsham (a translator employed by CHU de Poitiers, Poitiers, France) for reviewing and editing the original English-language manuscript.

Data Sharing Statement: See Supplement 3.