Introduction
Difficult airway management is one of the most challenging tasks for anesthesiologists. Recent data from the American Society of Anesthesiologists (ASA) Management Closed Claims Project, specifically those findings related to the difficult airway, demonstrate that the percentage of claims resulting from adverse respiratory events, although on the decline (42% in the 1980s to 32% in the 1990s), continue to constitute a large source of injury. A closed claims analysis of the management of the difficult airway published in 2005 showed that of the 179 claims made between 1985 and 1999 ( n = 179), 87% ( n = 156) of claims came from the perioperative period. More recent closed claims analyses demonstrated that claims resulting in death and brain damage from difficult airway management were associated with induction of anesthesia but not other phases of anesthesia and decreased in the period between 1993 and 1999, as compared with the period between 1985 and 1992. In 2006, a closed claims analysis of trends in anesthesia-related death and brain damage showed an overall reduction in claims for death or brain damage between 1975 and 2000 (odds ratio [OR], 0.95 per year; 95% confidence interval [CI], 0.94 to 0.96, p < 0.01). Of all the respiratory events ( n = 503) responsible for death or brain damage, difficult intubation ( n = 115), inadequate oxygenation ( n = 111), and esophageal intubation ( n = 66) were the top three causes.
Of the three types of adverse respiratory events reported, claims for inadequate ventilation and esophageal intubation decreased significantly in the 1990s (9% as compared with 25% of claims for death and brain damage in the 1980s), possibly as a result of pulse oximetry and end-tidal carbon dioxide monitoring. Yet, the proportion of claims for difficult intubation (a technical act, uninfluenced by monitoring) and other respiratory events leading to death or brain damage remained relatively stable between the 1980s and 1990s (9% and 8%, respectively). Of the adverse respiratory events, three quarters were judged to be preventable. Thus it is possible that better prediction of and preparation for difficult airway management might lead to a reduction in these numbers.
Anesthesiologists are confronted daily with the task of determining whether endotracheal intubation will be of increased difficulty in a patient. Preoperative evaluation of the airway can be accomplished by a thorough history and physical examination, as related to the airway; in addition, various measurements of anatomic features and noninvasive clinical tests can be performed to enhance this assessment. Nonetheless, several reports have questioned whether true prediction is possible.
The recent National Audit Project, NAP4, conducted in the United Kingdom (2008-2009) gives a point estimate of one airway related death per 180,000 general anesthetic procedures and a 1 in 22,000 incidence of adverse airway events. O’Sullivan and colleagues suggest that the real incidence of a difficult airway is likely to be more common than 1 in 5500 and may thus be experienced on a “regular” basis. The data demonstrated that a formal airway assessment was conducted in only 35 of 133 cases of airway-related events occurring during anesthesia (26%). However, when an airway assessment was performed, difficulty was anticipated correctly in the majority (e.g, in 25 of 35 cases). This is suggestive that an airway examination is worthwhile. With an overall positive predictive value of 0.25, if the group identified as potentially difficult to intubate is regarded as having a “disease” and in need of some form of specialized “treatment” for airway management (e.g,. awake or sedated fiberoptic intubation), then this number needed to treat for preventing harm from failed intubation would be 4, which is acceptable.
Descriptions of Terms
Five terms are important to review and analyze in this area: failed intubation, difficult intubation, difficult laryngoscopy, difficult mask ventilation, and difficult laryngeal mask airway ventilation. The ASA Task Force on Management of Difficult Airway suggests the following descriptions:
Failed intubation, or the inability to place the endotracheal tube after multiple intubation attempts, is a clear-cut endpoint. Thus there is a fairly uniform reported incidence of approximately 0.05% of surgical patients or 1 : 2230 and approximately 0.13% to 0.35% of obstetric patients or 1 : 750 to 1 : 280.
Difficult tracheal intubation (DI) is described as intubation when tracheal intubation requires multiple attempts, in the presence or absence of tracheal pathology. The incidence of DI is higher than failed intubation and has been reported to be 1.2% to 3.8%.
Difficult laryngoscopy (DL) is described as not being able to visualize any portion of the vocal cords after multiple attempts at conventional laryngoscopy, and many investigators include grades III and IV or grade IV alone, according to the Cormack-Lehane original grading of the rigid laryngoscopic view ( Figure 15-1 ). According to these definitions, the incidence of difficult direct laryngoscopy varies from 1.5% to 13% in patients undergoing general surgery.
Difficulty in performing endotracheal intubation is the end result of difficulty in performing laryngoscopy, which depends on the operator’s level of expertise, patient characteristics, and circumstances. Thus it has been suggested that the definition of DI be based on a uniform understanding of the best attempt at performing laryngoscopy/intubation and should use the number of attempts and time as boundaries only. The best attempt should incorporate the effect of changing the patient’s position; the effect of changing the length or type of laryngoscope blade; and the effect of simple maneuvers, such as conventional cricoid pressure, backward, upward, rightward pressure (BURP), and optimal external laryngeal manipulation (OELM).
Difficult mask ventilation (DMV) is a condition in which it is not possible for the anesthesiologist to provide adequate face mask ventilation because of one or more of the following problems: an inadequate mask seal, excessive gas leakage, or excessive resistance to the ingress or egress of gas. It is clear from clinical experience that there are grades of difficulty, similar to DI. The incidence of DMV also varies in the literature from 0.01% to 5%.
Difficult laryngeal mask airway ventilation is a situation in which providing ventilation and oxygenation to a patient with a laryngeal mask airway (LMA) is difficult. Even though not defined by ASA, researchers have defined this as an inability to place the LMA in a satisfactory position within three attempts to allow adequate ventilation and airway patency. Indices of clinically adequate ventilation are generally expired tidal volume > 7 mL/kg and leak pressure > 15 to 20 cm H 2 O. Verghese and Brimacombe, in their study of more than 11,000 patients, had a failure rate of 0.16%.
Descriptive Terms Used for Predicting a Difficult Airway
The following terms are commonly used to analyze the usefulness of predictive tests.
Specificity: Identifies all normal intubations as being normal. A sensitivity of 90% indicates that 90% of normal intubations will be identified as normal and 10% will be falsely identified as difficult. Ideally, specificity should be 100%.
Positive predictive value (PPV): The percentage of procedures that are true DIs from all those predicted by the test to be DIs. If the test predicts 20 DIs and only four are actually difficult, the PPV for the test is 20%. Even though PPV is a useful test, it is limited by the fact that it is dependent on the prevalence of DI in the sample group.
Likelihood ratio (LR): This is a useful term and can be calculated very quickly using sensitivity and specificity only. It is the chance of a positive test if the procedure is a DI divided by the chance of a positive test if the procedure was normal. LR is sensitivity/1 − specificity. It can be seen as a factor that links pretest probability to post-test probability of a DI with the use of a nomogram.
Receiver operating characteristic (ROC) curves: These help in determining the best predictive scores. The ROC has sensitivity on the y axis and 1 − specificity on the x axis. The test with the greatest area under the curve is the better one.
Prediction of the Difficult Airway: the Problem
There has been a heightened awareness of and a steady rise in the amount of literature being published on the recognition and prediction of the difficult airway. Evaluation of the evidence supporting the various methods of prediction of the difficult airway involves understanding the actual endpoints and their effect on the patient outcomes of mortality or brain death. The frequency of airway difficulty varies according to the population studied and the definition of DI used. There is no universally accepted definition of DI. Most of the larger studies concentrate on DI, broadly defined by difficult rigid laryngoscopic view (Cormack-Lehane grades III and IV or grade IV only), without the best attempt used. To be useful, a classification of laryngeal view should predict difficulty (or ease) of tracheal intubation, which requires the views to be associated with increasing degrees of intubation difficulty. Nonetheless, in a study of 1200 patients, Arne and colleagues found a significant difference between the incidence of Cormack-Lehane grades III and IV laryngoscopic views and the occurrence of DI in the general population, as many of the grades III and IV views were actually easy intubations. Thus one of the problems in the prediction of the difficult airway is that a DI is often not identified until laryngoscopy is performed and, as mentioned previously, there are discrepancies in the literature as to what defines difficulty.
Several authors have suggested the modification of the four-grade Cormack-Lehane scoring system (see Figure 15-1 ), which classifies the laryngeal view during laryngoscopy. This widely adopted classification system was described to allow simulated DI, yet it is applied inaccurately by the majority. Yentis and Lee modified this scoring system by subdividing a grade II laryngoscope view into IIa (partial view of glottis visible) and IIb (only arytenoids visible). This five-grade classification is referred to as the modified Cormack-Lehane system (MCLS) and allows refining the definition of DL as including IIb, III, and IV (see Figure 15-1 ). Koh and colleagues found that this system better delineated the difficulty experienced during laryngoscopy and intubation than the four-grade Cormack-Lehane system. Thus the true incidence of DL may be underestimated because it excludes a subgroup of the original grade II (IIb), which may be difficult to manage.
Cook further divided the Yentis and Lee modified systems into 3a (epiglottis can be seen and lifted) and 3b (epiglottis visualized but cannot be lifted); thus it consists of six grades, divided into three functional classes: easy, restricted, and difficult. Easy views were defined as when the laryngeal inlet is visible and thus suitable for intubation under direct vision (grades 1 and 2a). Restricted views were defined as when the posterior glottic structures (posterior commissure or any arytenoid cartilages) are visible or the epiglottis is visible and can be lifted (grades 2b and 3a). These views are likely to benefit from indirect intubation methods (e.g., gum elastic bougie). Difficult views were defined as when the epiglottis cannot be lifted or when no laryngeal structures are visible, which are likely to need specialist methods for intubation and may need to be performed blindly (grades 3b and 4). Cook proposed that this three-category classification system is of more practical value and had greater discrimination than Cormack-Lehane’s. He found that an easy view predicts easy intubation in 95% of cases and has less than 3% need of any intubation adjuncts. A difficult view is associated with DI in three quarters of cases, and specialist intubation techniques are likely to be required. Between these extremes, a restricted view is likely to require the use of a gum bougie but no other adjuncts.
It would be useful to predict DI before it occurs, but no preoperative test has adequate sensitivity to identify most cases without substantial false-positive results. Several prospective studies have identified various individual characteristics, which have significant association with laryngoscopic or intubation difficulties. Sensitivity and PPVs of these individual variables are low, ranging from 33% to 71% for specificity. Several combinations of these variables have been shown to be more effective predictors of DI.
A meaningful evaluation of the available literature requires an assumption about a reasonable level of expectancy in terms of sensitivity and specificity of the tests used for prediction of DI. Thus if at least 9 of 10 DIs are to be predicted, a sensitivity of 90% will be required. In addition, if one assumes that one false alarm a week is acceptable, in a hypothetical practice of 10,000 cases a year, it would correspond to a specificity of 99.5%. A number of investigators have attempted to achieve the goal of predicting DL or DI, or both, by combining different predictors and deriving multivariate indices so that the occurrence of false-negative results is decreased and the PPVs are increased. However, to date, no single multifactorial index can be applied to all of the various surgical populations. In addition, most, with the exception of Wilson’s index, have not been validated prospectively.
New investigative modalities, including x-ray, ultrasound, and three-dimensional computed tomography (CT) scans of the airway, have been proposed to help predict a difficult airway. A recent review performed by Sustic suggests that ultrasound can be used to assess anatomy of the upper respiratory organs and possibly assist in various applications of airway management.
The Upper Lip Bite Test (ULBT), a new, simple clinical bedside test performed by having the patient attempt to bite his or her own upper lip, has recently been suggested to aid in the prediction of difficulty with intubation. It is classified as follows: lower incisors can bite the upper lip above the vermilion line—Class 1; below the vermilion line—Class II; and cannot bite the upper lip—Class III. A recent external prospective evaluation of the reliability and validity of ULBT demonstrated that the interobserver reliability was better than the Modified Mallampati (MMP) score (Mallampati classification [MPT], as modified by Samsoon and Young). They also found that they could not use the test on edentulous patients (11% of 1425 patients), and concluded that, like the MMP score, the ULBT was a poor predictor when used as a single screening test.
Additionally, advanced computing techniques over the last decade have improved statistical analysis, allowing improved testing of variables for successful prediction of the difficult airway. Nonetheless, given the low incidence of DI and the wide variation in acceptable definitions of airway terms, it is difficult to compare different studies and perform a meta-analysis of the predictors of difficult airway management.
Evidence
History
After thorough review of the literature, the published evidence is not sufficient to evaluate the effect of either a bedside medical history or a review of prior medical records on predicting the presence of a difficult airway. According to the ASA task force, there is suggestive evidence (which is defined by the ASA as enough information from case reports and descriptive studies to provide a directional assessment of the relationship between a clinical intervention and a clinical outcome) that some features of both may be related to the likelihood of encountering a difficult airway.
Many congenital and acquired syndromes are associated with difficult airway management. Also, certain disease states, such as obstructive sleep apnea and diabetes, have been suggested to correlate with an increased risk of DI. Trauma to the airway, either caused by external forces or iatrogenic from routine endotracheal intubation, may also be associated with difficult airway management. Recently, Tanaka and colleagues demonstrated increased airflow resistance attributable to intraoperative swelling of the laryngeal soft tissues in patients whose airways were predicted to be normal (or easy to intubate) and who underwent routine tracheal intubation. Others have observed serious laryngeal injuries (e.g., vocal cord paralysis, arytenoid cartilage subluxation, laryngeal granulomas, and scars) after short-term intubation and anesthesia. Additionally, the ASA task force found that a previous history of difficult airway management offers clinically suggestive evidence that difficulty may recur.
Physical Examination
Single Predictors of Difficult Laryngoscopy/Intubation
The ability of a specific test to predict a DI is decreased by the variability of definitions of DL and DI and the inherent inaccuracy of numeric grading systems. Nonetheless, several investigations have identified anatomic features that have unfavorable influences on the mechanics of direct laryngoscopy and endotracheal intubation ( Table 15-1 ). The majority of anesthesiologists rely on predicting DI mainly as a result of several preoperative bedside screening tests.
| Predictors | Study | No. of Patients | Incidence (%) | Sensitivity (%) | Specificity (%) | Positive Predictive Value (%) | Negative Predictive Value (%) | Definition of Difficult Intubation * | Best Attempt | Population |
|---|---|---|---|---|---|---|---|---|---|---|
| Mallampatti III or IV | Arne et al, 1998 | 1200 | 4 | 78 | 85 | 19 | 99 | 4 | + | General + ENT |
| Savva, 1994 | 355 | 1.14 | 64.7 | 66.1 | 8.9 | 1, 3, and 4 | + | General + OB (10%) | ||
| Oates et al, 1991 | 675 | 1.8 | 42 | 84 | 4 | − | General | |||
| Butler and Dhara, 1992 | 220 | 8.2 | 56 | 81 | 21 | − | General | |||
| Frerk, 1991 | 244 | 4.5 | 81 | 82 | 17 | − | General | |||
| Rose and Cohen, 1994 | 18,558 | 1.8 | Relative risk, −4.5 | 3 > 2 attempts | − | General | ||||
| Voyagis et al, 1998 | 1833 | 8.3 | 88.1 | 37.2 | Original † | 1 | − | Obese General | ||
| 86.8 | 50 | Modified † | ||||||||
| Bergler et al, 1997 | 91 | 10 | 60 | 72 | − | General + ENT | ||||
| Brodsky et al, 2002 | 100 | 12 | 58.3 | 70.5 | 1 and 3 | − | Morbidly Obese | |||
| Khan et al, 2003 | 300 | 5.7 | 82.4 | 66.8 | 13 | 98.4 | 1 | − | General | |
| Yamamoto et al, 1997 | 3680 | 1.3 | 67.9 | 52.5 | 2.2 | 1 | + | General | ||
| El-Ganzouri et al, 1996 | 10,507 | 1 | 44.7 | 89 | 21 | 96.1 | 1 | + | General | |
| 59.8 | 4.4 | 2 | ||||||||
| Wong and Hung, 1999 | 411 | 1.99 | 85.7 | 62.6 | 3.8 | 99.6 | 1 | − | Chinese ♀ | |
| IV Only | Savva, 1994 | 355 | 1.14 | 52.9 | 87 | 1, 3, 4 | + | General + OB (10%) | ||
| Wong and Hung, 1999 | 28.6 | 98.3 | 22.2 | 98.8 | 1 | − | Chinese ♀ | |||
| Thyromental Distance | ||||||||||
| <6 | Butler and Dhara, 1992 | 220 | 8.2 | 62 | 25 | 16 | − | General | ||
| <6 | El-Ganzouri et al, 1996 | 10,507 | 1 | 7 | 99.2 | 38.5 | 94.3 | 1 | + | General |
| 16.8 | 99 | 15.4 | 99.1 | 2 | ||||||
| <6.5 | Savva, 1994 | 355 | 1.14 | 65 | 81 | 15 | 1, 3, 4 | + | General + OB (10%) | |
| <6.5 | Arne et al, 1998 | 1200 | 4 | 16 | 95 | 12 | 96 | 4 | + | General + ENT |
| <7 | Frerk, 1991 | 244 | 4.5 | 91 | 82 | 19 | − | General | ||
| <7 | Schmitt, 2002 | 270 | 5.9 | 81 | 73 | 1 | + | General | ||
| Ratio of Height to Thyromental Distance | ||||||||||
| 25 | Schmitt et al, 2002 | 270 | 5.9 | 81 | 91 | 1 | + | General | ||
| Sternomental Distance | ||||||||||
| <12.5 | Savva, 1994 | 355 | 1.14 | 82.4 | 88.6 | 26.9 | 1, 3, 4 | + | General + OB (10%) | |
| Neck Movement | ||||||||||
| <80° | El-Ganzouri et al, 1996 | 10,137 | 1 | 10.4 | 98.4 | 29.5 | 94.4 | 1 | General | |
| 16.78 | 7.9 | 2 | ||||||||
| <90° | Arne et al, 1998 | 1200 | 54 | 85 | 14 | 98 | General + ENT | |||
| Atlanto-Occipital Extension | ||||||||||
| <35° | Wong and Hung, 1999 | 411 | 85 | 70 | 4.8 | 1 | − | Chinese ♀ | ||
| Obesity | ||||||||||
| BMI > 30 kg/m 2 | Voyagis et al, 1998 | 1833 | 8.3 | 88.9 | 66.7 | Obese | ||||
* Definition of difficult intubation: (1) Cormack and Lehane grade III or IV; (2) Cormack and Lehane grade IV only; (3) No. of attempts; (4) special techniques and others.
† Original indicates tongue protruded by the patient; Modified indicates tongue actively pulled out by the anesthesiologist.
Mallampati Classification.
The MPT focuses on the relative visibility of oropharyngeal structures when the patient is examined in the sitting position with the mouth fully opened, the tongue fully extended, and without phonation. Samsoon and Young proposed the modified MPT (MMP) in which there are four oropharyngeal classes instead of the original three ( Figure 15-2 ), yet Ezri et al and Maleck et al further suggest adding a fifth class, class 0, defined as the ability to visualize any part of the epiglottis on mouth opening and tongue protrusion. Samsoon and Young’s method is by far the most widely investigated method of airway evaluation. The practical value of this method lies in its ease of application, yet practitioners often perform this examination in the supine position with or without phonation. A wide range of observations shows that this method is subject to significant interobserver variability. Overall, the literature suggests that the true sensitivity of the MMP, as modified by Samsoon and Young, is most likely between 60% and 80% and the true specificity is between 53% and 80%; the PPV is approximately 20%. A recent meta-analysis of the accuracy of MPT/MMP found substantial differences and variability in reported sensitivity and specificity values. Overall accuracy of the test was poor to good and depended on which version of the test and reference tests were used. The meta-analysis also suggested that the MPT/MMP was a poor predictor of DMV. Krobbuaban and colleagues found that MMP Classes III and IV had a sensitivity of 70% and specificity of 60% with a PPV of 20%.
Additionally, a recent study suggested that the best way to perform MPT was by placing the patient in the sitting position, with the patient’s head in full extension, tongue protruded, and with phonation, yet phonation did not influence the overall accuracy of this classification. Mashour and Sandberg evaluated 60 patients first with the MMP test and then repeated the examination with craniocervical extension. They found that by including craniocervical extension, the MMP scores were reduced. Class II MMP became Class 1.6, Class III became 2.6, and Class IV became 3.5. The sensitivity remained the same but the specificity improved from 70% to 80%. The PPV increased from 24% to 31%, and the negative predictive value (NPV) increased marginally from 97% to 98%. A recent meta-analysis of 55 studies involving 177,088 patients concluded that the prognostic value of the MMP was worse than earlier estimates with a pooled sensitivity and specificity of 0.35 and 0.91 and an OR of 5.89. Another recent but smaller study of 1956 patients determined that MPT is insufficient for predicting DI on its own.
Thyromental Distance.
The concept of thyromental distance (TMD), noted as the distance between the chin and the notch of the thyroid cartilage, was described by Patil and associates in 1983. They proposed that this distance should be 6.5 cm in the healthy adult, and that if this distance is less than 6 cm, there may be intubation difficulties. Of all the morphometric measurements, TMD has been questioned the most for its value in predicting DI. The sensitivity of this test is between 60% and 80% and has a specificity of 80% to 90% in some studies. Arne and colleagues and El-Ganzouri and colleagues found the test to be highly insensitive (sensitivity, 16% to 17%) but very specific (specificity, 95% to 99%) with a PPV of 12% to 16%, if a more stringent definition of DI involving best attempt (with OELM) is applied.
Recently, the role of TMD has been challenged by some authors. Chou and Wu suggest that the receding mandible, one of the two components of a micrognathic mandible, is not the real cause for DL in these patients, thus TMD is irrelevant. Qudaisat and Al-Ghanem suggest that TMD is a surrogate for inadequate head extension. They found that among the factors that determine TMD only the degree of head extension was significantly different between the two laryngoscopy groups. The two other factors (sagittal angulomental distance, representing mandibular growth and sagittal angulothyroid distance, representing laryngeal descent in the neck) did not differ between the two laryngoscopy groups. Wong and Hung studied TMD, along with MMP and atlanto-occipital extension (AOE) and demonstrated the limitation of absolute anatomic measurements in their study involving Chinese women. The optimal TMD criterion was 5.5 cm in this study, which achieved a sensitivity of 71% and a specificity of 83%, yet the PPV was only 7.5%. Schmitt and colleagues attempted to adjust this measurement to the patient’s size and proposed the ratio of the patient’s height to thyromental distance (RHTMD). Using the ROC curve, they found a cutoff value to be 25 or greater for this ratio to predict DL with a reasonable degree of sensitivity (81%) and specificity (90%).
A recent meta-analysis performed by Shiga and colleagues stated, “the diagnostic value of TMD proved unsatisfactory in their analysis.” They determined that there was a wide range in the sensitivity, which could possibly be due to different cutoff points (4.0 to 7.0 cm). They also found that the positive LR of TMD improved from 3.4 to 4.1 when a more strict cutoff criterion (<6.0 cm) was applied.
Recently, Krobbuaban and colleagues conducted a prospective randomized study of 550 consecutive Thai patients. They found that the RHTMD had higher sensitivity (77%), higher PPV (24%), and fewer false-negative results (16%). They also found that RHTMD of 23.5 or greater, neck movement less than 80 degrees, and MMP Classes III and IV were major predictors of DL. Rosenstock and colleagues found that the interobserver agreement for TMD and neck mobility was low.
Hyomental Distance.
Hyomental distance (HMD), a measurement from the tip of the chin to the hyoid cartilage, has also been considered as one of the predictors of DI. Both TMD and HMD give an idea of the available space for the tongue during laryngoscopy. In an investigation involving 12 cadavers and 334 patients, Turkan and colleagues, using cervical spine radiographs of patients in the neutral position, found that mean HMDs were less than the stated limit of 7 cm and that HMD was the only objective variable not affected by age. However, both McIntyre and Randall demonstrated that radiologic measurements have not been capable of providing sensitive criteria for prediction of DI and that radiographic studies were, at best, regarded as valuable in understanding problems encountered during laryngoscopy.
Sternomental Distance.
Sternomental distance (SMD), a measurement from the tip of the chin to the sternal notch, normally greater than 12.5 cm, was suggested by Savva to predict DI if less than 12 cm with maximal head extension. Savva found that this measurement was both more sensitive and more specific than TMD and that it may give a more accurate estimate of head extension. This measure functionally “added” the atlanto-occipital joint into the physical evaluation of the airway. Ramadhani and colleagues suggested that SMD was a superior measurement, compared with others, by showing that SMD had an increased sensitivity (71.1%) and specificity (66.7%) for predicting subsequent DL and that it was unaffected by age. However, the patient group in their study was limited to women of childbearing age only. Turkan and colleagues, on the other hand, demonstrated that SMD measurements were affected both by age and gender, as both younger (20- to 30-year olds) and male patients had longer SMD measurements.
In their meta-analysis, Shiga and colleagues found that SMD yielded moderate sensitivity and specificity. It also yielded a high positive LR and diagnostic OR. The negative LR for SMD was the lowest, suggesting that it could be the best single test for ruling out DI. Nonetheless, their study was based on only three studies that included SMD.
Neck Movement and Mouth Opening.
Neck movement and mouth opening have also been considered as variables in predicting DI. El-Ganzouri and colleagues demonstrated that three single variables, that is, restricted head and neck movement, including flexion and especially extension capability (<80 degrees or <90 degrees ), restricted mouth opening (<4 cm or <5 cm ), and inability to protrude the mandible, have a significant association with DI. The accuracy of the estimation of AOE with use of the Bellhouse test has been questioned and, similar to other clinical methods, is subject to wide interobserver variability.
Individual examinations and tests are subject to wide interobserver variability; thus any evidence needs to be evaluated accordingly. In a study involving 59 patients, Karkouti and colleagues determined that mouth opening and chin protrusion had excellent interobserver reliability, whereas seven tests (i.e., TMD, mandible subluxation, AOE and angle, profile classification, ramus length, and oropharyngeal best view) were only moderately reliable between observers. In addition, the MMP technique of assessing the oropharyngeal view has poor interobserver reliability.
Rosenstock and colleagues evaluated the interobserver reliability of the Simplified Airway Risk Index (SARI). The parameters used in SARI include mouth opening, TMD, ability to protrude the mandible, MMP score, head and neck mobility, and body weight. Two pairs of assessors (two specialists and two residents) performed the assessment. They used five of seven tests from SARI and evaluated 120 patients with normal airways and 16 patients documented to have airways that were difficult to intubate. They found good interobserver agreement with mouth opening, MMP class, and mandibular protrusion, whereas TMD and neck movement had low levels of interobserver agreement.
In the Yildiz and colleagues multicenter study, the most sensitive criterion when used alone was mouth opening (sensitivity, 43%). In their study, the incidence of DI was significantly higher in patients with MMP Classes III and IV, a decreased average TMD and SMD, decreased mouth opening, or decreased protrusion of the mandible ( p < 0.05). Combination of the tests did not improve their results.
Rose and Cohen analyzed the data regarding problems and prediction of difficult airway management in 18,500 patients and found that although the most common single abnormalities noted were restricted neck movement (3%) and decreased visualization of the hypopharynx (2.2%), with a relative risk of 3.2 and 4.5, respectively, decreased mouth opening (<2 fingers; relative risk, 10.3) and shortened TMD (<3 fingers; relative risk, 9.7) were the best predictive factors of DI.
Weight.
Obesity has been studied as isolated body weight (>110 kg) or body mass index (BMI; >30 kg/m 2 ) and shown to be associated with DL, especially when accompanied with a large tongue (as assessed by MMP). Juvin and colleagues, in a study involving 134 lean (BMI < 30 kg/m 2 ) and 129 obese patients (BMI ≥ 35 kg/m 2 ), determined that DI is more common among obese than nonobese patients by using the Intubation Difficulty Scale (IDS) developed by Adnet and colleagues, which includes both qualitative and quantitative dimensions of DI. It is an objective scoring system involving seven variables: number of intubation attempts, skill and experience of the operators, alternative intubation techniques, glottic exposure (Cormack-Lehane), lifting force applied to the laryngoscope, application of external laryngeal pressure, and position of the vocal cords at intubation. In this study, they defined two groups of patients according to the IDS values: those with an IDS score of less than 5 (easy and slightly difficult) or 5 or greater (difficult). They found that among the classic risk factors for DI, only an MMP score of III or IV is a risk factor for DI in obese patients (OR, 12.51; specificity, 62%; PPV, 29%). They also determined that the risk of hypoxemia is higher in obese patients during anesthesia induction and that further investigation is necessary to identify the risk factors for DI in this population.
Shiga and colleagues found that the incidence of DI in obese patients (BMI > 30 kg/m 2 ) was more than three times higher than in nonobese patients. Also, Cattano and colleagues found that obesity had the highest sensitivity (32%) and a PPV of 16% for predicting difficulty of intubation. The same sensitivity (32%) was found with an MMP score of Classes III and IV. Brodsky and colleagues, on the other hand, studied 100 consecutive morbidly obese subjects (BMI > 40 kg/m 2 ) and concluded that neither absolute body weight (obesity) nor BMI is associated with intubation difficulties. Rather, they found that a large neck circumference (NC; measured at the level of the superior border of the cricothyroid cartilage) of 40 cm showed a 5% probability and an NC of 60 cm showed a 35% probability of problematic intubation; high (III or greater) MMP scores are the only predictors of potential intubation problems in this patient population. Thus whether tracheal intubation is more difficult in obese patients is debatable. Lundstrom and colleagues, in a cohort study of 91,332 concluded that a BMI greater than 35 kg/m 2 is a weak (sensitivity, 7.5%; PPV, 6.4%) but statistically significant predictor of difficult and failed intubation with an OR of 1.031.
Kim and colleagues demonstrated a 13.8% versus 4.8% ( p = 0.016) incidence of DI in 123 obese patients (BMI > 27.5 kg/m 2 ) compared with 125 nonobese patients. Multivariate analysis showed that the MMP score, the Wilson score, and the ratio of the NC and TMD independently predicted DI (defined as intubation difficulty scale > 5) in obese patients. NC/TMD had the highest sensitivity and NPV.
NAP4 data suggest that patients with a BMI of more than 30 kg/m 2 were at least twice as likely to have serious complications of airway management as those with a BMI of 30 kg/m 2 or less. A BMI of more than 40 kg/m 2 increased the risk fourfold.
Increasing knowledge of the sonoanatomy of the upper airway could potentially play a significant role in predicting difficult airways. Komatsu and colleagues used ultrasound to quantify anterior neck soft tissue thickness and predict DL in 64 morbidly obese patients (BMI ≥ 35 kg/m 2 ). They performed an ultrasound scan of the anterior neck soft tissue and measured the distance from the skin to the anterior aspect of the airway at the level of the vocal cords. In contrast to Brodsky’s findings, they concluded that the thickness of pretracheal soft tissue at the level of the vocal cords is not a good predictor of DL in either white or black obese patients. In contrast, Ezri and colleagues studied Middle Eastern patients and determined that soft tissue in the neck did influence difficulty in intubation. Adhikari and colleagues, in their pilot study with 5 of 51 patients having DL, found that sonographic measurements of anterior neck soft tissue thickness at the level of hyoid bone and thyrohyoid membrane can be used to distinguish difficult and easy laryngoscopies.
Additionally, Siegel and colleagues demonstrated that ultrasound of the airway was a reliable, simple, and comfortable method of identifying the mechanism of airway obstruction. The role of preintubation ultrasound assessment elsewhere in the upper airway for the detection of pharyngeal or laryngeal pathology, such as tumors, abscesses, or epiglottitis, has also been studied. Because of these discrepancies in the literature, convincing evidence to correlate soft tissue thickness of the neck with DI does not exist.
Combined Predictors of Difficult Laryngoscopy and Intubation
Although no single factor has been shown to be a predictor of DI on its own, it has been widely suggested that combinations of factors improve predictability of DI. Various combinations of individual predictors have been studied, and several multivariate indices have been proposed ( Table 15-2 ); however, very few have been prospectively evaluated for their efficacy. In his editorial, Wilson concluded that no single test is likely to be a perfect predictor of DI, and Bainton suggests that the most satisfactory solution would be the “best algebraic sum” of several tests.
Related posts:
Is a Preoperative Screening Clinic Cost-Effective?
Which Patient Should Have a Preoperative Cardiac Evaluation (Stress Test)?
Is There a Best Approach for Patients with Difficult Airways: Regional versus General Anesthesia?
What Is the Best Technique for Hip Surgery?
Should Thoracic Epidural/Spinal Analgesia be Used for Coronary Artery Bypass Grafting?
Optimal Postoperative Analgesia
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
