Observational data have consistently shown videolaryngoscopy (VL), compared to direct laryngoscopy (DL), to be associated with improved procedural outcomes such as increased first-attempt success and decreased esophageal intubations, difficult airways, and complications.^{2,3,4,5,6,7,8,9} Observational studies comparing VL to “optimized” DL augmented with maneuvers used to improve the success rate of DL (external laryngeal manipulation, positioning, bougie) still showed, in every comparison, that VL was associated with an increased adjusted odds of first attempt success.¹⁰ VL has been shown in many studies to improve the glottic view,^2,11 reduce the incidence of difficult intubations,^2,12 reduce failed airways when experiencing a difficult airway,¹¹ reduce airway trauma,^11,13 and lead to faster skill acquisition across both experienced and novice operators.^12,13 A 2022 Cochrane review, which occupies the highest position on the evidence pyramid, concluded that VL likely provides safer airway management than DL.¹⁴

So, it seems it should be straightforward for randomized controlled trials to confirm this association. Yet, not only has no trial shown a benefit (as of April 2023), one trial even showed potential harm with an increase in complications.^{15,16,17,18,19,20} Yet a more sophisticated mixed-effect analysis of two of the trials shows VL improves success rates in more difficult airways.²¹ It wasn’t until June 2023 that a multicenter trial showed a difference with VL.²² How can all of this be explained?

The bougie was developed as an adjunct to DL to facilitate tracheal intubation when the view of the glottis was compromised by the epiglottis. If the bougie improves success in these difficult situations, perhaps it should be used every time a standard geometry direct or video blade is used? The BEAM trial tested this hypothesis and found the routine use of a bougie with standard geometry VL had an astonishing first attempt success rate of 98%, compared to 87% when a styletted tracheal tube was used.²³ However, this trial was conducted at a single center where bougies were used routinely prior to the initiation of the study—a significant limitation. Due to this limitation, the trial was replicated in the 15-center BOUGIE trial where the bougie was uncommonly used, and the trial failed to show a benefit in the overall study population or in number of subgroup comparisons and sensitivity analyses.²⁴ What’s worse is that a secondary analysis using sophisticated machine learning approaches found that patients with difficult airway characteristics were less likely to be successfully intubated on first attempt if a bougie was used.^25,26 How can one trial show practice-changing results while the other shows no benefit whatsoever and even suggests the opposite?

The risk of cardiovascular collapse with rapid sequence intubation (RSI) makes the selection of induction agents an extremely important topic. The etomidate versus ketamine debate has been almost as controversial as the VL versus DL debate. The question revolves entirely around secondary pharmacologic effects that are only largely theoretically relevant—ketamine’s indirect sympathomimetic effect and etomidate’s temporary inhibition of cortisol synthesis. Meanwhile, ketamine’s direct myocardial depressant effect and etomidate’s negative effect on arterial elastance
are conveniently ignored.²⁷ The first clinical trial comparing these two agents for RSI, published in 2009, showed no difference in outcomes and it was reasonable to conclude at this point that both drugs were safe and effective.²⁸ Two changes then occurred around the same time, however, that reignited the debate. Hurricane Maria hit Puerto Rico which led to a global shortage in etomidate supply and ketamine exploded in popularity on social media.

Interestingly, however, observational data has been inconsistent. One study using the National Emergency Airway Registry (NEAR) database showed that ketamine was associated with more postintubation hypotension than etomidate, even when controlling for confounders with propensity matching,^29,30 while another study in Japan found the opposite.³¹ Still one more study showed that ketamine was used more often in the setting of an elevated shock index, but induction agent choice was not associated with postintubation hypotension.³² Another study in the ICU showed that etomidate and ketamine had similar outcomes, and both appeared to have higher mortality than when propofol was the sedative.³³ A prehospital study showed a progressive increase in the incidence of postintubation hypotension in patients with traumatic brain injuries after changing their protocol from etomidate to ketamine.³⁴ One randomized controlled trial compared etomidate and ketamine in ICU patients after an observational quality improvement project at their hospital suggested a difference in 7-day mortality. The trial showed a higher 7-day mortality with etomidate, but 7-day mortality is not a standard outcome, and 28-day mortality, which is the standard outcome, showed no difference.³⁵ The exploratory outcomes investigating the hypothesis that ketamine would have a better hemodynamic profile all trended worse for ketamine.³⁵

Observational studies on airway management in critically ill patients tend to look at all patients in a dataset and then adjust for confounders such as operator experience, and the presence or absence of difficult airway characteristics. However, the term “difficulty” presents a paradox. If your airway assessment indicates a potentially difficult airway, and you develop a strategy to overcome that difficulty, that results in successful first-attempt intubation without complications, is that a difficult airway? Yes and no, it’s a Schrödinger airway, both difficult and not difficult. You predicted difficulty and modified your strategy to overcome it (i.e., you did not wander into failure!) so the airway was not difficult, but it likely would have been had you not planned for it. Thus, the difficult airway presents a significant problem for research design as it is loaded with ambiguity.³⁶ Studies comparing devices, techniques, or methods often focus either on the anticipated difficult airway, which is poorly predicted,³⁷ or, more commonly, they deliberately exclude patients with the predicted difficult airway. Focusing inclusion criteria on anticipated difficulty or excluding anticipated difficulty can potentially sway the results of a trial significantly. Ideally, trials include all critically ill patients undergoing intubation, or select for potentially or actually difficult airways in a reproducible manner. Excluding patients with difficult airways is almost never the right approach to gaining accurate generalizable knowledge, but it would be ethically challenging to not allow exclusion if the treating clinician does not think there is equipoise for a given situation.