Reply: Apneic Oxygenation Has Not Been Disproven

From the Authors:

We appreciate the interest demonstrated by Dr. Pavlov, and the airway management community in general (1), in our randomized trial of apneic oxygenation during endotracheal intubation of critically ill adults (2). Dr. Pavlov’s primary concern is the power of our trial. Our sample size (150 patients) was selected using the same primary endpoint (lowest arterial oxygen saturation) and minimum clinically meaningful difference between groups (5%) as prior high-quality trials targeting desaturation during endotracheal intubation (3, 4). We observed a numerical difference in lowest oxygen saturation between the apneic oxygenation and usual care arms of just 2%, well short of clinical or statistical significance. This was true despite nearly half of the patients in each arm experiencing lowest oxygen saturations in the 60s, 70s, and 80s, which are sufficient rates of desaturation for apneic oxygenation to have conceivably had an effect. As Dr. Pavlov correctly asserts, however, our trial was not powered to make inferences regarding less common secondary outcomes, such as the incidence of lowest oxygen saturation less than 80%. Although we respectfully resist Dr. Pavlov’s proclamation of “a huge difference (15.8% vs. 25.0%) in the incidence of saturation lower than 80% between the two groups” (referring to six total patients), and his assertion that “statistical significance was not attained … simply because the sample was too small” (reliant on the flawed assumption that small differences in one of many secondary outcomes would persist in a larger sample), we agree with his overall point. The relationship between immediate complications of endotracheal intubation and long-term, patient-centered outcomes remains incompletely understood. If only the most extreme desaturations are of clinical importance, then our trial (and most other emergent intubation trials) would be vastly underpowered. Future research should empirically examine which surrogate endpoints most closely relate to patient-centered outcomes such as cardiac arrest and death. Future trials should also consider larger sample sizes to target increasingly robust surrogate endpoints or clinical outcomes.

Finally, Dr. Pavlov titles his letter “Apneic Oxygenation Has Not Been Disproven.” The interventions we apply to our patients should be proven to be effective, not presumed effective until proven otherwise. Despite being promoted by experts for half a decade (5) and administered to thousands of patients across the world, before our trial apneic oxygenation during intubation outside the operating room had never even been tested, much less proven. We applied apneic oxygenation in the manner recommended (nasal cannula set at 15 L/min [6]) to the patient population recommended (all patients being intubated, including those receiving noninvasive ventilation or bag-valve-mask ventilation [6]) and did not find it to be effective. This suggests, at a minimum, that the effect of apneic oxygenation on desaturation during emergent intubation is not as great as we had previously hoped. We readily acknowledge that our trial was just one trial of one method of delivering apneic oxygenation to one patient group. Although our results agree with another recently published trial that failed to demonstrate a benefit of apneic oxygenation at 60 L/min among patients with hypoxic respiratory failure (4), different patient or operator populations may produce different results. If, in the future, a high-quality trial (large, randomized, concealed-allocation, high compliance with the assigned intervention, minimal missing data and loss to follow up, and objective data collection of a clinically meaningful outcome) demonstrates apneic oxygenation to be effective in a specific context, we will eagerly employ it in that context. Until then, we will shift focus away from apneic oxygenation and toward airway management interventions proven to help patients.