When Life Gives You Lemons....

   If you've read an EMS-related journal or attended an EMS conference in the last couple of years, you're probably aware that prehospital airway management in general, and prehospital intubation specifically, is under close scrutiny.

   Multiple studies have indicated that paramedics may not have high intubation success rates, that prehospital intubation may worsen patient outcomes, and that airway management may receive less teaching time than it needs in EMS education programs.1 This body of evidence has led to changes in how EMS providers manage their patients' airways in emergency situations. At the same time, airway assessments aimed at predicting "difficult airways" have gained popularity; providers are now being encouraged to assess airways and make procedure choices based on their likelihood of success.

   However, many of the airway assessment techniques currently recommended have been adapted from other specialties, have practical limitations to their use in emergency airway management and have questionable validity from a statistical point of view. This article will look at the history of airway assessments, the popular mnemonic LEMON for predicting difficult laryngoscopy, and some of the limitations to its use in EMS, both practical and statistical.

History of Airway Assessment

   Airway assessment grew up in the OR as anesthesiologists sought ways to predict patients whom they would be unable to effectively ventilate. The development of alternative airway devices, such as the Combitube, offered choices in how to manage the operative patient's airway, but how do you choose which option is best for a particular patient? In the mid-1980s, Vijayalakshmi Patil, MD, proposed that the areas surrounding the head and superior neck had predictive value for difficult intubations, and Seshagiri Rao Mallampati, MD, published two articles describing a scale for judging ease of intubation based on pharyngeal structures visible when the patient opens his mouth and extends his tongue.2 Various combinations of these parameters, combined with other factors such as receding mandibles, buck teeth and x-ray findings, were also proposed.2 In 1993, the American Society of Anesthesiologists published a recommended Difficult Airway Algorithm, attempting to prevent airway-related deaths in the OR by standardizing an approach to airway management.

   In the meantime, rapid sequence intubation techniques were emerging in emergency rooms as a viable option for emergency airway management. This development raised the stakes for emergency providers. Previous methods of emergency airway management held the safety net that if the procedure was unsuccessful, the patient was still, for the most part, spontaneously breathing. The prospect of pharmacologically inducing respiratory arrest necessitated a careful risk/benefit analysis for every patient considered for RSI. Hence, airway assessment techniques were transplanted from the OR to the ED to help provide some guidance for practitioners. In much the same way, as RSI spread from the ED to the helicopter or ambulance, so did traditional airway assessments.

Current Airway Assessments: LEMON

   Recently, mnemonics for airway assessment that address all aspects of emergency airway management (laryngoscopy, BVM ventilation, extraglottic devices and surgical airway techniques) have been disseminated.3 The most commonly known mnemonic seems to be the one for laryngoscopy--LEMON:

  • Look externally
  • Evaluate 3-3-2
  • Mallampati scoring
  • Obstructions
  • Neck mobility.

   For those not familiar with the mnemonic, "Evaluate 3-3-2" refers to the geometry of the airway and oral access; the provider should be able to fit three of the patient's fingers into his mouth, fit three fingers underneath the chin, and fit two fingers from the top of the thyroid cartilage to the top of the neck. The 3-3-2 refers directly back to Patil's proposal of the importance of airway geometry and the thyromental distance (distance from the thyroid cartilage to the mentum of the chin).

Another commonly used measure of difficult laryngoscopy is the Cormack-Lehane scale. This system grades the difficulty of laryngoscopy based on the anatomical features visible after laryngoscopy has been performed. Complete visualization of the vocal cords receives a C/L Grade 1 score, whereas visualization of only the epiglottis receives a C/L Grade 4. The most obvious limitation of the Cormack-Lehane scoring system is that it offers no predictive abilities; the ease of laryngoscopy is not known until it has been performed. C/L scoring is very useful as a quality assurance benchmark, but of little use in prediction. As mentioned, the LEMON mnemonic is probably the most commonly used airway assessment for difficult laryngoscopy; however, it has some potentially serious limitations for use in emergency airway management. These limitations are described below.

Practical Limitations

   Both Mallampati scoring and evaluation of 3-3-2 require significant patient cooperation to accurately assess. This is a non-issue in most perioperative patients who routinely undergo airway assessments in the pre-op period while they are still alert. This is probably a much different patient than the one encountered by emergency providers. Our patients are typically hypoxic, scared, possibly combative and somewhere along the spectrum of "altered mental status" ranging from verbally responsive to completely obtunded. Some have recommended the technique of using the laryngoscope as a tongue blade, opening the patient's mouth and extending their tongue for them to gauge the Mallampati score; however, the effectiveness of this technique compared to patient cooperation is unknown.

   Neck mobility also plays an important role in airway assessment. The ET tube must pass through the mouth, pharynx and larynx, all of which have different visual axes in the supine patient. Some degree of neck mobility is helpful to align these axes through use of the "sniffing position." However, many patients requiring emergency airway management also require spine immobilization, which significantly hinders neck mobility, as well as jaw extension.

   In 2004, a retrospective study of 850 ED patients requiring emergency airway management looked at the ability of emergency providers to predict difficult laryngoscopy using techniques of Mallampati scoring, thyromental distance and neck mobility.4 The assessments described above could not have been used collectively in two-thirds of the patients. However, the vast majority (n=839) were successfully intubated using laryngoscopy, either with or without RSI (laryngoscopy failed in 3 RSI patients, 8 underwent nasotracheal intubation). The implications of the study are clear: Emergency airway providers cannot reliably depend on the level of patient cooperation needed to perform Mallampati and 3-3-2 evaluations, and a significant number of these patients have limited neck mobility that make alignments of the visual axes difficult to assess.

Statistical Limitations

   Like all other clinical tests, airway assessments are subject to statistical values and properties. The two most-commonly utilized tests for assessments and tests are sensitivity and specificity:

  • Sensitivity is an "inclusion" value; a test with 90% sensitivity will correctly identify 90 out of 100 patients who have the condition being tested for.
  • Specificity is an "exclusion" value; a test with 90% specificity will correctly identify 90 out of 100 patients who do NOT have the condition being tested for.

   Sensitivity and specificity are typically mutually exclusive. The more generic you make test criteria to improve sensitivity, the more "false positives" will slip through the cracks, lowering specificity and vice versa.

   Testing of the various common airway assessments (Mallampati scoring, thyromental distance and neck mobility) through the years has yielded conflicting results with regard to sensitivity and specificity. A meta-analysis of this research published in 20025 found a wide variety of values for the three tests:

  • Mallampati scoring: sensitivity 42%-60%, specificity 53%-82%
  • Thyromental distance: sensitivity 7%-91%, specificity 25%-99%
  • Neck mobility: sensitivity 10%-17%, specificity 98%.

   It should be noted that the studies in this analysis tested only a single predictive test, whereas the LEMON mnemonic incorporates all three. The effect of combining these predictive tests is unclear.

   Another statistical value to be considered when evaluating a test is the positive predictive value. This value is similar to sensitivity (measures the likelihood of having the condition and a positive test result); however, it also incorporates the actual incidence of the condition. This is an important consideration for the use of airway assessments because the incidence of failed laryngoscopy with optimal muscular relaxation through RSI and performance by experienced, skilled providers is very low, ranging in studies from 0.4% to 1%.

   Applying an incidence of 0.4% to the predictive tests mentioned in the meta-analysis, and using the highest sensitivity and specificity ratings for each test (which is unlikely due to the mutually exclusive relationship between the two), the positive predictive value for each test is:

  • Mallampati scoring: 1.32%
  • Thyromental distance: 26.76%
  • Neck mobility: 3.3%.

   The low positive predictive values of these airway assessment methods, even when given the highest sensitivity and specificity values, further limits their utility in emergency airway management.


   Careful consideration of the components of LEMON yields some serious problems with applying them in the prehospital setting; the patients we deal with are unlikely to give us reliable test results. Moreover, given the low incidence of failed laryngoscopy due to patient anatomy, even reliable test results from your patient may not actually indicate difficult laryngoscopy, which might seem to lead to a clinical rock and hard place. We're encouraged to use airway assessments to form an airway management plan, but the assessments don't offer much concrete advice.

   Even if the specifics of LEMON aren't particularly effective, the ideas and principles they're based on are sound. Oral access, the geometry of the airway, and ability to align the oral, pharyngeal and laryngeal axes are all important aspects of laryngoscopy that depend to some degree on your patient's airway anatomy. Although no airway assessment can look at all aspects of an individual's airway anatomy, completely discounting the anatomical obstacles to laryngoscopy and intubation would be unwise.

It would be similarly unwise, however, to judge the likelihood of success based solely on the patient's anatomy. The technique, skill and competence of the provider are also important factors and collectively probably override anatomical concerns in most situations. Poor technique can turn a predicted non-difficult airway into a failed laryngoscopy attempt, and good technique can turn a predicted difficult airway into a first-pass success. Ignoring other determinants of laryngoscopy success to focus on the patient's anatomy is not a recipe for success.

   Yet another determinant in airway management success lies in the mechanical limitations of the airway tools at your disposal. Not all situations can be easily managed by one tool or procedure. Inability to fit an extraglottic device into a patient's airway makes for just as difficult an airway as an inability to identify anatomical landmarks for laryngoscopy and intubation. Consideration of how your airway devices work can help guide you further in formulating an airway management plan.

   Airway management is a collection of infinite situations linked by common procedures. The variables are the patient's anatomy, the provider's skill and competence, the tools and procedures available and the patient's clinical needs. Assessing the airway, whether with a stepwise LEMON approach or a "gut feeling" of difficulty, is only the beginning of a dynamic process that will continue until you turn over patient care. Arming yourself for successful airway management includes impeccable technique for whatever airway management options you have (curved and straight blade laryngoscopy, use of the bougie, video laryngoscopy, extraglottic devices, BVM ventilation, and so forth), an understanding of the mechanics of how they work and what they require to work effectively, and the ability to use critical thinking to choose the best options for the situation at hand. At the end of the day, the final goal is oxygenation and ventilation of the patient, not successful completion of a procedure.

   Thanks to Geoffrey Breeden, NREMT-P, CCEMT-P, for his help and insight in the preparation of this article.


1. Bledsoe BE, Gandy WE. The disappearing endotracheal tube: historic skill threatened by lack of practice and new devices. JEMS 34:3, pp. 88-99.
2. Murphy MF, Doyle DJ. Airway Evaluation. In Murphy MF, Hung O, eds. Management of the Difficult and Failed Airway. New York, NY: McGraw-Hill, pp. 6-7, 2008.
3. Murphy MF, Walls RM. Identification of the Difficult Airway. In Walls RM, Murphy MF, eds. Manual of Emergency Airway Management. Philadelphia, PA: Lippincott Williams & Wilkins, pp. 81-93, 2008.
4. Levitan RM, Everett WW, Ochroch AE. Limitations of difficult airway prediction in emergency department intubated patients. Ann Emerg Med 44: 307-13, 2004.
5. Yentis SM. Predicting difficult intubation: Worthwhile exercise or pointless ritual? Anaesthesia 57:105-109, 2002.

   Benjamin Dowdy, BS, NREMT-P, WALS, graduated from Eastern Kentucky University with a BS in Emergency Medical Care and is a paramedic with Georgetown-Scott County (KY) EMS. He also works with Wolfe County (KY) SAR in the Red River Gorge area and is an instructor for Wilderness Medical Associates, where he indulges his hobbies of hiking, backpacking and climbing. Contact him at b.dowdy@gscems.com.