Don't extubate patients to conventional O2; use HFNC or NIV, says new guideline
Is it care improvement, or overkill? Probably a bit of both
When extubating patients with respiratory failure, applying conventional nasal cannula oxygen has been first-line, with high-flow nasal oxygen (HFNO) or noninvasive ventilation (NIV) generally reserved for patients seen at higher risk or who are failing conventional oxygen.
A major American thoracic society has released a new official clinical practice guideline for the use of noninvasive respiratory support in adults with respiratory failure in the ICU and medical wards.
According to their expert panel, we should be doing this the other way around: extubating to HFNO or NIV in most patients, with conventional oxygen considered in the occasional low-risk patient. Most patients at the highest risk for extubation failure should receive NIV, not HFNO, according to the panel. (PulmCCM is not affiliated with any specialty society.)
Does the advisement hold up to scrutiny?
Parts 1 and 2 here:
Conventional Oxygen Is Insufficient For Most Patients Post-Extubation: Panel
Between 15 and 20% of extubated patients require reintubation within 72 hours, often due to respiratory fatigue resulting from high work of breathing. It’s possible that increased use of prophylactic advanced noninvasive respiratory support (NIV or HFNO), applied at the time of extubation, would prevent some of these reintubations.
The panel reviewed over 50 randomized trials testing some version of this question, in most of which patients were extubated directly to either HFNO or NIV, with conventional O2 as the control. That produced a solid signal of benefit for both modalities in pooled analyses:
NIV (35 RCTs) reduced reintubation rates by 25% (RR 0.75, 95% CI 0.63 to 0.88)
HFNO (17 RCTs) reduced reintubation rates by 20% (RR 0.80, 95% CI 0.64 to 1.00)
The authors mention that of the 53 trials, “nearly half were conducted in Europe.” In our cursory AI-supported analysis of the trials, about 70% of patients were enrolled in developed/high-income nations (Europe/North America/Australasia).
The pooled effect sizes from trials conducted in developing countries were larger than those from trials in developed countries, but only moderately so.
NIV is better for high-risk patients, probably
The panel concluded (from an unspecified ‘subgroup analysis’) that for high-risk patients, NIV was more effective than HFNO at preventing reintubation, when prophylactically applied at the time of extubation.
High-risk in this context was defined as the presence of:
Age ≥ 65 years
Obesity (BMI ≥ 30)
Cardiopulmonary disease
APACHE II score ≥ 12
Laryngeal edema or other airway patency concerns
Secretion concerns (weak cough, frequent suctioning)
≥7 days mechanical ventilation
“More than two comorbidities”
This aligns with our ongoing evidence review, some of which we touched on here in 2025:
Does prophylactic noninvasive ventilation prevent reintubation in obese patients?
After liberation from mechanical ventilation, obese patients may be at increased risk for recurrent respiratory failure and reintubation.
Can HFNO/NIV, Applied Broadly, Really Reduce 20-25% of Reintubations?
A broader criticism is that all such trials are compromised by the potential influence of the intervention on clinician psychology: a patient on conventional oxygen may seem more at risk and be more likely to be intubated, e.g., while a clinician may allow a patient at similar risk to continue for longer wearing NIV or HFNO.
Research teams, and their academic colleagues caring for trial subjects, may have a subtle “home team” bias to reintubate control arm patients sooner, exaggerating any actual treatment benefit (or producing a false positive, if no real benefit exists).
Two relevant studies can help put this skeptical criticism in context.
PROPER trial
This pragmatic crossover RCT performed at Vanderbilt (AJRCCM 2021) explicitly tested the question: should all-comers undergoing extubation in a medical ICU receive noninvasive respiratory support?
A protocol was put in place. In the intervention arm, upon extubation, patients who had suspected hypercapnia received NIV. If they had no suspected hypercapnia, they received HFNO. Usual care patients could receive any modality. Two halves of a single ICU switched back and forth periodically en bloc (intervention/usual care) to generate balanced data.
Most patients in the intervention group (73%) were extubated to HFNO, and only ~18% to NIV; most control patients were extubated to conventional oxygen (97%), sometimes with NIV (14%).
So this trial could be seen as testing universal HFNO after extubation vs usual care (as NIV rates did not differ meaningfully between intervention/control).
The results? Among 751 extubated patients, numerically more patients in the protocolized “HFNO group” were reintubated within 96 hours (57, or ~16%) compared with 52 patients (~13%) in the usual care group (odds ratio, 1.23; 95% confidence interval, 0.82 to 1.84; P = 0.32).
So that didn’t work, there.
But NIV for higher risk patients did:
NIV Prevented Reintubations When Applied Routinely To Higher-Risk Patients (Thille et al 2016)
This wasn’t an RCT, but a prospective before/after study, testing a protocol to apply prophylactic NIV to all extubated patients ≥65 years old or with cardiorespiratory disease at a French teaching hospital’s ICU.
Before the protocol, 28% of higher-risk patients were reintubated. After the protocol to extubate higher-risk patients to NIV was begun, 15% of higher-risk patients were reintubated.
Reintubation rates for the excluded (average risk) patients remained the same (10.2% vs 10.7%).
Other trials further support the use of prophylactic NIV post-extubation in high-risk and obese patients:
Even Low Risk Patients Should Get HFNO After Extubation: Panel
The panel suggests HFNC even for low-risk patients, and NIV for high-risk patients, “to reduce the need for re-intubation following extubation after critical illness.”
Of course, applying high-flow oxygen to hundreds of thousands of stable, low-risk patients would seem to be wasteful and expensive as well as a bit ridiculous.
The panel suggested that only “very low risk” patients should be extubated to conventional oxygen or room air. “Very low risk” is a stringent category indeed, including post-op patients from non-thoracic or abdominal surgery, those without any heart or lung disease, who have not been long on the vent, with good coughs and minimal secretions.
Conclusion
Evidence suggests that either high-flow oxygen or noninvasive ventilation, applied prophylactically, could prevent reintubation in some patients.
In response, the major American thoracic society now advises that all but the lowest-risk patients should receive these advanced modalities at the time of extubation, with HFNO generally for lower-risk and NIV for higher-risk patients.
Most experienced clinicians already extubate high-risk patients to an advanced modality. Applying the guideline uniformly to the hundreds of thousands of patients extubated from mechanical ventilation in U.S. ICUs each year (only a fraction of whom would be considered “very low risk” and appropriate for conventional O2) would create a large burden of additional care, prolonging ICU and hospital stays, without adequate proof of benefit.
It makes sense to try to nudge clinicians toward greater use of these modalities, which are probably underused in higher-risk patients. Inexperienced clinicians who might miscategorize risk for reintubation in particular should follow the guideline.
Rather than the blanket application of a guideline, though, clinical judgment by an experienced intensivist remains the best determinant of which oxygen therapy is indicated in a patient after extubation.
Reference
Goel NN, Ferreyro BL, Pitre T, Lewis K, Homer-Bouthiette C, Angriman F, Kawano Dourado L, Drake M, Eaton TL, Esbrook CL, Frat JP, Grieco DL, LeTourneau W, Mosier J, Patel BK, Qadir N, Riviello E, Roca O, Twagirumugabe T, Vranas KC, Munshi L, Rochwerg B; American Thoracic Society Assembly on Critical Care. Noninvasive Respiratory Support for Adult Patients with Acute Respiratory Failure. An Official American Thoracic Society Clinical Practice Guideline. Am J Respir Crit Care Med. 2026 Jun 29:aamag302. doi: 10.1093/ajrccm/aamag302. Epub ahead of print. PMID: 42371750.