Much has been written regarding the benefits of apneic oxygenation. The physiological underpinnings are sound with minimal logistical and resource based costs. As such it has enjoyed widespread adoption throughout the Emergency Medicine and Critical Care world. Despite its popularity, the evidence supporting its use has been less inspirational.
Recently, a meta-analysis published in Annals of Emergency Medicine found the use of apneic oxygenation was associated with a decreased rate of hypoxia when compared to standard care (1). And while this appears promising, when the authors included only the highest quality RCT data these benefits vanished. We are now faced with yet another RCT, the ENDAO Trial, examining the use of apneic oxygenation during RSI in an Emergency Department population. Once again the results are impressively negative.
Published in Academic Emergency Medicine, Caputo et al randomized 200 patients presenting to a single Emergency Department requiring endotracheal intubation with the intention of undergoing rapid sequence intubation (RSI) (2). Patients were excluded if they presented in arrest or the treating clinician planned to perform an awake intubation. All patients received a fairly rigorous pre-oxygenation strategy using either flush-rate oxygen via an NRB-mask, a BVM with PEEP valve, or a BiPAP circuit with an FiO2 of 100%. Patients in the apneic O2 group received an additional 15 L/min of flow using a standard nasal cannula as well as an ETCO2 nasal cannula set at 15 L/min, both of which were initiated during the pre-oxygenation phase and continued throughout the apneic phase. To minimize bias, outcome variables were collected by independent observers.
The authors found no benefit in their primary outcome, the median lowest oxygen saturation, between the two groups (92% vs 93%, p=0.08). Nor did the authors find a statistical difference in the number of patients who desaturated to less than 90% (17% vs 15%) or less than 80% (3% vs 4%). Once again, a high quality randomized control trial found no benefit to the use of apneic oxygenation in patients requiring emergent intubations. But before completely abandoning apneic oxygenation let us for a moment put this study in proper perspective.
Given recent failures, the question becomes whether apneic oxygenation is a quantifiable reality or just a physiological delusion that has not been proven in the clinical arena? But apneic oxygenation is more than just a bedtime story used to comfort novice intubators, it has proven an effective means of maintaining appropriate oxygen saturations in multiple studies. The most infamous, by Frumin et al, in which healthy subjects scheduled for minor elective surgeries were induced using thiopental and succinylcholine and endotracheally intubated (3). Paralysis and sedation were continued and patients were left apneic supported only by low flow oxygen down the endotracheal tube. Patients were left apneic for up to 50 minutes without dropping their oxygen saturations. Recent data using high-flow oxygen delivery systems in the OR have demonstrated similar results (4). There have even been data supporting the use of low-flow oxygen strategies delivered through a simple nasal cannula. So how do we reconcile the studies supporting the physiological underpinnings of apneic oxygenation with the recent slew of clinically oriented negative trials? As always the answer given is based primarily on the question asked.
Modern-day evidence supporting the use of low-flow apneic oxygenation is based off two RCTs performed in the OR which demonstrated that the use of apneic O2 extended the time it took patients to desaturate. In both these trials patients were randomized to either undergo an apneic period with or without the use of low-flow oxygen via a nasal cannula. Both sets of authors reported that patients randomized to the apneic oxygenation group experienced significantly longer times to desaturation. And yet compared to the ENDAO Trial these cohorts were diminutively small, 30 and 40 patients in each respective trial. Despite these shortcomings in statistical power these studies demonstrated a far greater signal of benefit than what was demonstrated in the ENDAO cohort.
The first of these trials was published in Anesthesia in 2006 by Taha et al (5). Patients were randomized to either standard preoxygenation followed by an apneic period with no oxygenation or a period in which 5 l/min was applied through a standard nasal cannula. The authors allowed patients to remain apneic until the SO2 dropped below 95% or 6 minutes whichever occurred first. In the control group the SO2 fell below 95% at a mean time of 3.65 min. In contrast, 100% of patients in the apneic oxygenation group maintained a SO2 greater than 95% for 6 minutes.
The second study by Ramachandran et al, published in Journal of Clinical Anesthesia in 2010, found fairly similar results (6). Patients were randomized to either receive standard preoxygenation with or without the addition of a nasal cannula at 15 L/min applied during the apneic period. And again patients were kept apneic until a saturation less than 95% was reached or a total time of 6 minutes had elapsed. Similar to the previous study, the patients randomized to the apneic oxygenation arm experienced much longer safe apneic times. Time to desaturation less than 95% was 5.29 min vs 3.49 min. 87.7% of the patients in the apneic oxygenation arm had an SO2 greater than 95% at 6-minutes vs only 4.4% in the control arm.
Despite adequate pre-oxygenation every subject enrolled in these trials were given adequate time to desaturate during the apneic period, allowing for a noticeable difference in SO2 levels between the groups. This is in direct contrast to the ENDAO Trial, which was performed in a very high functioning Emergency Department that employed a robust preoxygenation strategy and demonstrated a high level of skill-based competency. As a result, patients were intubated fairly quickly, allowing minimal time for desaturation. The large majority of patients were successfully intubated by 60 seconds, 80% by 80 seconds, 90% by 100 seconds, and 100% by 195 seconds. None of the patients were allowed the 3.49 minutes that it required the control group in the Krishna et al study to reach an SO2 of less than 95%. While, this was a pragmatic trial, performed on sick patients in an Emergency Department, and it would be unrealistic and unethical to allow patients to remain apneic until an SO2 of less than 95% was achieved, it is essentially impossible for this study to demonstrate a benefit in extending safe apnea time, or decrease the rate of severe desaturation as so few patients experienced such events.
Essentially there is a conflict between what we associate with a statistically negative trial and the reality of what this trial can actually disprove. The question we hope to answer is, does the use of apneic oxygenation in Emergency Department RSI decrease the rate of clinically important desaturation, leading to a decrease in peri-intubation complications and death? But what the authors powered their study to detect was a difference in the lowest mean oxygen saturation. On first glance this seems like an appropriate outcome, but in reality it fails to address the clinically important question at the heart of this discussion.
We have discussed the issues surrounding the use of lowest mean oxygen saturation in a previous post, but in brief it both lacks statistical and clinical validity and results in a trial with insufficient power to evaluate the effect of the treatment in question on any clinically important variables. In the case of the ENDAO trial so few patients experienced clinically important desaturation (only 16% had SO2 levels <90% and 3.5% < 80%), that it would be impossible to detect any potential benefits apneic oxygenation may provide.
In both the FELLOW trial and the PREOXYFLOW trial, which utilized apneic oxygenation in RSIs conducted in the ICU setting, lowest SO2 was again used as the trials’ primary outcome (7,8). And like ENDAO, neither of these trials found a difference. But unlike ENDAO, both these trials observed clinically important outcomes at a high enough rate to note the potential benefits of apneic oxygenation. The FELLOW trial observed a 9.2% absolute difference in the rate of patients who experienced severe desaturation (defined as an SO2<80%), a 2.8% difference in the rate of peri-intubation arrest, and a 14.2% difference in hospital mortality. Similarly, PREOXYFLOW reported an 8.5% increase in the rate of severe complications (defined as desaturation <80 % or cardiovascular collapse), a 1.8% difference in peri-intubation arrest, and 6.7% difference in 28-day mortality. None of these difference reached statistical significance and this may very well be because apneic oxygenation provides no benefit to standard pre-oxygenation techniques. Or it could be because these studies were not powered to detect a difference in dichotomous outcomes due to a misconceived power calculation based off a clinically questionable continuous variable.
While not definitive, the ENDAO Trial did demonstrate in an unselected Emergency Department cohort, when appropriate pre-oxygenation techniques are utilized, in a high functioning system, the added benefits of apneic oxygenation are minimal. But we do not utilize apneic oxygenation for the unselected cohort. Because if we were guaranteed successful endotracheal intubation with minimal delay in all our patients, then the majority of our preparation would be rendered superfluous. As with much of our preparation, apneic oxygenation is employed universally in the event of the rare case of unexpected difficulty. And while its utility in these patients is unknown, an underpowered trial, in which so few patients experienced the outcome in question, does not disprove the potential benefits of apneic oxygenation. For this a far more statistically robust endeavor is required.
Additional FOAM resources:
–EM Literature of Note
–St. Emlyn's Journal Club
-Fabulous Editorial by John Sakles published along side the ENDAO Trial