Predictors of intubation vary by oxygen delivery method in patients with acute hypoxemic respiratory failure

Patients were treated with standard oxygen, high-flow nasal cannula oxygen, or noninvasive ventilation, and overall, 45% required intubation, with signs of persisting or worsening respiratory failure being the most common reason.


Predictors of intubation in patients with acute hypoxemic respiratory failure seem to vary according to the type of noninvasive oxygenation therapy received, according to a new study.

Researchers performed a post hoc analysis of randomized clinical trials involving 23 ICUs to identify factors associated with intubation in patients with acute hypoxemic respiratory failure. Included patients had a respiratory rate above 25 breaths per minute and a PaO2/FIO2 ratio of 300 mm Hg or lower and were treated with standard oxygen, high-flow nasal cannula oxygen, or noninvasive ventilation. The study's main outcome was to identify early factors—at baseline and at one hour—associated with intubation. Factors associated with 90-day mortality rates were also examined. The study results were published online Nov. 2 by Critical Care Medicine.

A total of 310 patients were included in the study, 72 (23%) with mild hypoxemia, 165 (53%) with moderate hypoxemia, and 73 (24%) with severe hypoxemia. Ninety-four patients were treated with standard oxygen, 106 were treated with high-flow nasal cannula oxygen, and 110 were treated with noninvasive ventilation. Overall, 45% of the study population (139 of 310 patients) required intubation, with signs of persisting or worsening respiratory failure being the most common reason. Bilateral pulmonary infiltrates were more likely in patients who needed intubation, and this group was also more likely to have a higher respiratory rate and a lower PaO2 at baseline compared with those who did not need intubation.

One hour after treatment began, intubation was more likely in patients receiving standard oxygen who had a respiratory rate of 30 breaths per minute or greater (odds ratio, 2.76; P=0.03). For patients receiving high-flow nasal cannula oxygen, increased heart rate one hour after the start of treatment was the only intubation predictor (odds ratio, 1.03; P<0.01), while for patients receiving noninvasive ventilation, independent predictors were a PaO2/FIO2 ratio less than or equal to 200 mm Hg (adjusted odds ratio, 4.26; P=0.003) and a tidal volume above 9 mL/kg of predicted body weight (adjusted odds ratio, 3.14; P=0.02). A tidal volume above 9 mL/kg of predicted body weight was also independently associated with mortality in the noninvasive ventilation group (adjusted odds ratio, 4.51; P=0.004).

The authors noted that their study had a post hoc design and that their results may apply only to patients with severe hypoxemic community-acquired pneumonia and no nonrespiratory organ dysfunction, among other limitations. However, they concluded that one hour after treatment initiation, a high respiratory rate in patients receiving standard oxygen and a high tidal volume in patients receiving noninvasive ventilation are associated with likelihood of intubation. They also noted an independent association between high tidal volume one hour after the start of treatment and 90-day mortality in patients receiving noninvasive ventilation. “When the tidal volume is so elevated, clinicians should be cautious concerning [noninvasive ventilation] prolongation, which could entail an increased risk of volutrauma,” the authors wrote.