One common thread in all studies that look at intubation-related harm is that there is often a failure to assess for difficulty, instead relying on a habitual practice (e.g., rapid sequence intubation and a size 3 Macintosh blade). The first of the three core actions in this book is to assess for potential difficulty or danger. This chapter will describe how to assess for difficulty in completing the technical aspects of laryngoscopy, intubation, and rescue techniques. This requires a reliable and reproducible method for identifying the anatomically difficult airway. This evaluation must be expeditious, easy to remember, and complete.

In clinical practice, we identify the potentially difficulty airway by assessing four aspects of airway management to inform our ability to:

There is overlap among the dimensions of airway management. Some aspects that may make intubation difficult may make mask ventilation difficult, etc. However, a distinct evaluation is required for difficult laryngoscopy, difficult BMV, difficult EGD, and difficult surgical access; and each evaluation must be applied to each patient, even if time pressure requires abbreviated assessments, before any airway management (Fig. 5.1).

The concept of difficult laryngoscopy and intubation is inextricably linked to poor glottic view; the less adequate the glottic view, the more challenging the intubation. This concept developed during an era when almost all intubations were done by direct laryngoscopy (DL), is still relevant even in the era of video laryngoscopy (VL). Nearly all research relating specific patient characteristics to difficult or impossible intubation is based on studies of DL. VL, especially hyperangulated VL (HA-VL), is much less affected than DL by the presence or number of difficult airway attributes. For example, the acute curvature of HA-VL was designed to overcome challenges viewing the airway in the presence of a large tongue or cervical immobility—two components of LEMON. However, despite a design to reduce the incidence and effect of difficulty, extreme abnormalities of some elements of LEMON will affect both DL and VL alike. Severely reduced mouth opening, for example, makes it impossible for any laryngoscope blade to be inserted. So, although abnormal
LEMON findings will differentially impact DL more, we recommend performing a difficult laryngoscopy assessment, using the LEMON mnemonic, on all patients for whom intubation is planned, including for planned VL. When able to be inserted, it is rare for VL, particularly HA-VL, to yield a Cormack-Lehane (C-L) grade III (or worse) glottic view.^1,2,3 VL accomplishes this independently of the need to align the various airway axes, as must occur during DL.

It follows that evidence-based guidelines for prediction of difficult VL may be challenging, or even impossible, to create. A mnemonic for difficult VL, “CRANE,” has been developed, based on limited evidence from the anesthesia literature, but its utility in airway management is unclear. If orotracheal intubation is planned and the airway is not overwhelmingly soiled or obliterated by large obstructing airway pathology, then VL provides the best opportunity for intubation success despite the presence of other “CRANE” abnormalities. Providers should proceed with using VL unless patient characteristics are present that would make both VL and DL virtually impossible, in which case, another approach may be required (i.e., awake nasotracheal flexible VL for advanced angioedema of the tongue).

C-L introduced the most widely used system of categorizing the degree of visualization of the larynx during DL, in which a complete laryngoscopic view is designated grade 1 and the worst possible view, grade 4 (Fig. 5.2). C-L grade 3 view (only the epiglottis is visible) and grade 4 view (no glottic structures are visible) are highly correlated with difficult or failed intubation. C-L grade 1 (visualization of virtually the entire glottic aperture) and grade 2 (visualization of the posterior portion of the cords or the arytenoids) are not typically associated with difficult intubation. The C-L grading system does not differentiate precisely the degree to which the laryngeal aperture is visible during laryngoscopy: A grade 2 view may reveal little view of the vocal cords, or none if only the arytenoids are visible. This led to adoption of a grade 2a/2b system, wherein a 2a shows any portion of the cords and a 2b shows only the arytenoids. Grade 2a airways perform comparably to those scored as grade 1, whereas grade 2b airways behave more like grade 3 airways. When DL is used, grade 2b accounts for only about 20% of grade 2 views. However, when a grade 2b view occurs, two-thirds of patients are difficult to intubate, whereas only about 4% of patients with grade 2a views are characterized as difficult intubations. A grade 1 view reveals virtually the entire glottis and is associated with nearly universal intubation success.

A poor glottic view is associated with low intubation success. The highest first-attempt success has historically been seen with a C-L grade I and II airways and is largely dependent on device. Recent data from the National Emergency Airway Registry (NEAR), a cohort of patients like ICU patients have provided insight into the C-L view during intubations in critically ill ED patients.
In an analysis of nearly 12,000 patients intubated with either HA-VL or SG-VL, a C-L grade I or II view was obtained in 94.1% and 87.4%, respectively.³ A second NEAR study, assessing the intubating conditions during RSI with either DL or VL observed a C-L grade I or II view in approximately 89% of all encounters. In this same study, the first-attempt success rate was 87% to 88% with an ultimate success rate >99%.⁴ In a single-center prospective evaluation of 750 ED intubations over a 2-year period, during which 255 intubations were performed with a C-MAC and the rest with a conventional laryngoscope, the C-MAC yielded grade I/II views in 94% of cases compared with 83% for DL.⁵ In the ICU, a large single-center registry study showed a C-L grade I or II view in an average of 80% to 87% of all intubations, with a dramatic difference between VL and DL (86% vs. 62%).⁶

Despite scores of clinical studies, no evidence to date has identified a foolproof set of patient attributes that, when absent, always predicts successful intubation and, when present, predicts certain intubation failure. In the absence of a proven and validated system that can predict intubation difficulty with 100% sensitivity and specificity, it is important to develop an approach that will enable a clinician to quickly and simply identify those patients who might be difficult to intubate to inform an appropriate strategy. In other words, when asking the question, “If I induce this patient, will I have difficulty seeing the airway and placing an endotracheal tube?” we value sensitivity (i.e., identifying all those who might be difficult) more than specificity (i.e., always being correct when identifying a patient as difficult).

The mnemonic LEMON is a useful guide to identify as many of the anatomic risks of difficulty as quickly and reliably as possible and can be done urgently or in patients with poor mental status. The elements of the mnemonic are assembled from an analysis of the difficult airway prediction instruments in the anesthesia, emergency medicine, and critical care literature. The mnemonic, which we developed for The Difficult Airway Course has been externally validated in emergency department (ED) patients.⁷ The modified LEMON (all aspects of LEMON except the Mallampati score and thyromental distance) has undergone additional external validation and has been found to have very high negative predictive value for both conventional laryngoscopy and VL.⁸ LEMON has now been adopted as a recommended airway assessment tool in the Advanced Trauma Life Support (ATLS) manual.

The mnemonic is as follows:

The first “3,” therefore, assesses mouth opening. A normal patient can open their mouth sufficiently to accommodate three of their own fingers between the upper and lower incisors (Fig. 5.3A). This is an approximate measurement as it would be unusual to ask an acutely ill or injured patient to stick three fingers in their mouth. If a patient can comply, ask if they can open the mouth as wide as possible. This will give a meaningful sense of whether the patient is able to open fully, partially, or not at all. The second “3” evaluates the length of the mandibular space by ensuring the patient’s ability to accommodate three of their own fingers between the tip of the mentum and chin-neck junction (hyoid bone) (Fig. 5.3B). The “2” assesses the position of the glottis in relation to the base of the tongue. The space between the chin-neck junction (hyoid bone) and the thyroid notch should accommodate two of the patient’s fingers (Fig. 5.3C). Thus, in the 3-3-2 rule, the first 3 assesses the adequacy of oral access, and the second 3 addresses the dimensions of the mandibular space to accommodate the tongue on DL. The ability to accommodate fewer than three fingers is associated with greater degrees of difficulty in visualizing the larynx at laryngoscopy: The former because the length of the oral axis is elongated, and the latter because the mandibular space may be too small to accommodate the tongue, requiring it to remain in the oral cavity or move posteriorly, obscuring the view of the glottis. Encroachment on the submandibular space by infiltrative conditions (e.g., Ludwig angina) is identified during this evaluation. The final 2 identifies the location of the larynx in relation to the base of the tongue. If significantly more than two fingers are accommodated, meaning the larynx is distant from the base of the tongue, it may be difficult to reach or visualize the glottis, particularly if a smaller blade size is used initially. Fewer than two fingers may mean that the larynx is tucked up under the base of the tongue and may be difficult to expose. This condition is often imprecisely called an “anterior airway.”