Incidence and Predictors of Difficult and Impossible Mask Ventilation

After obtaining institutional review board approval (University of Michigan, Ann Arbor, Michigan), all adult patients undergoing general anesthesia were prospectively included in this trial. Because no clinical interventions were studied and no patient-identifiable data were used, signed patient informed consent was waived per the institutional review board approval. For each anesthetic case, a preoperative history and physical and intraoperative record were documented using an electronic perioperative clinical information system (Centricity®; General Electric Healthcare, Waukesha, WI). Elements documented included a standard airway physical examination, physical features that may affect mask fit, patient history that may suggest airway anatomy pathology, and general patient and operation characteristics (tables 1 and 2).6,7 

The primary outcome measure was ease or difficulty of MV. A four-point scale ranging from grade 1 to 4 originally described by Han et al.  5is used at our facility (table 3). Grade 3 (DMV) is defined as MV that is inadequate to maintain oxygenation, unstable MV, or MV requiring two providers. Grade 4 MV is defined as IMV noted by absence of end-tidal carbon dioxide measurement and lack of perceptible chest wall movement during positive-pressure ventilation attempts despite airway adjuvants and additional personnel. Two incidence pilot studies were previously performed by Han et al. .5Previous studies suggest that the use of muscle relaxant does not alter the grade MV assigned.4,8Secondary outcomes measured were direct laryngoscopy (DL) view as defined by Cormack and Lehane,9a subjective assessment of difficult intubation (DI) due to more than three attempts by anesthesia attending staff, and the ability to perform successful tracheal intubation using DL. These data were collected using standardized pick-list choices with the option of free text entry if the choices did not offer the anesthesiologist the ability to fully document the clinical observation.

Anesthesia services were provided by anesthesiology attending staff with assistance from certified registered nurse anesthetists, anesthesia residents, and fellows in training. In general, both MV and intubation were attempted initially by the anesthesiology resident or certified registered nurse anesthetists present in the room. All clinical decisions regarding airway management (patient position, DL blade, use of thyroid pressure) were made by the attending staff. The attending could choose to perform an awake fiberoptic intubation at their discretion, thereby avoiding a MV attempt. MV was performed without a harness using a black rubber reusable mask (Rüsch; Teleflex Medical Inc., Research Triangle Park, NC) or clear disposable plastic mask (King Systems Corporation, Noblesville, IN). Laryngoscopy was performed using a fiberoptic DL handle and blade (Heine Inc., Dover, NH). DI was defined as grade III or IV DL view or more than three attempts at intubation by a staff anesthesiologist. Impossible intubation was defined as the inability to intubate the patient using DL technique despite more than three attempts.

Based on the work by Langeron et al.  4and Han et al.  5, we estimated that we would need to observe approximately 20,000 cases of MV to record 1,000 cases of difficult ventilation and 20 cases of impossible ventilation. We initially set our sample size and study duration to attain these sample sizes.

Univariate analysis was performed between patients with or without the following measured outcomes: grade 3 MV, grade 4 MV, and grade 3 or 4 MV and DI. Statistical significance was tested using Pearson chi-square or Fisher exact test. A P  value less than 0.05 was considered significant. All variables found to be significant in the univariate analysis were entered into a multivariate logistic regression model to identify independent predictors of the measured outcome.

If three or more risk factors were identified for an outcome, a risk factor scale was created to predict the outcome. Receiver operating characteristic curves and odds ratios were analyzed to assess the diagnostic value of the risk factor scale.

Of 61,252 anesthetic cases performed in adult patients during a 24-month period, 22,660 cases included an attempt at MV. Thirty-seven cases (0.16%) of grade 4 MV (impossible to ventilate) and 313 cases (1.4%) of grade 3 MV (difficult to ventilate) were recorded (table 3). Two cases of IMV were due to an existing patent tracheotomy site and were excluded from these data. No other patient exclusions were performed. Eighty-four cases (0.37%) of grade 3 or 4 MV and DI were observed.

During the first 9 months of the study period, the mandibular protrusion test was not recorded in the preoperative anesthesia history and physical form. Therefore, only 14,369 cases were included in the univariate and multivariate predictor analysis. All episodes of grade 3 or 4 MV were included in the analysis.

Univariate analysis demonstrated several risk factors associated with grade 3 MV, grade 4 MV, and grade 3 or 4 MV with DI (table 4). Body mass index (BMI) of 30 kg/m²or greater and age of 57 yr or older maximized the sum of sensitivity and specificity for each risk factor. Multivariate regression analysis identified the following independent predictors of grade 3 MV: BMI of 30 kg/m²or greater, presence of a beard, Mallampati classification III or IV, age of 57 yr or older, severely limited mandibular protrusion, and a history of snoring (table 5). These six indicators were used to create a prediction score. A patient was given one point if a preoperative predictor was noted. The area under the curve for the receiver operating characteristic curve was 0.75 (fig. 1). Weighting the factors did not improve the curve.

Multivariate regression identified history of snoring (P = 0.004) and thyromental distance of less than 6 cm (P = 0.040) as independent predictors of grade 4 MV.

Multivariate regression analysis identified the following independent predictors of grade 3 or 4 MV combined with DI: limited or severely limited mandibular protrusion, thick/obese neck anatomy, history of sleep apnea, history of snoring, and BMI of 30 kg/m²or greater (table 5). The receiver operating characteristic curve demonstrated an area under the curve of 0.78 (fig. 2).

Of the 37 cases of grade 4 MV, only 1 patient could not be intubated and required emergent cricothyrotomy. Ten patients had a DI, and 26 were intubated without difficulty.

The results of this study confirm the incidence of grade 3 MV (1.4%) to be similar to the 1.6% incidence reported in the review of Han et al.  5of 1,405 patients using the same MV scale. This is less than the 5% DMV incidence reported in the study of Langeron et al.  4using a different MV scale. Grade 4 MV (IMV) has an incidence of 0.16% in the studied tertiary care center surgical patient population. Abnormalities in the mandibular protrusion test may be associated with grade 3 MV. Although grade 4 MV is associated with DI, an overwhelming majority of patients can still be intubated.

Langeron et al.  4reported the incidence and predictors of DMV in a study of 1,502 adult patients designed explicitly for this endpoint. The 1.4% observed incidence of grade 3 MV in our population was markedly lower than that reported by Langeron et al.  4(5%) but similar to that reported by Han et al.  5(1.5%), Asai et al.  10(1.4%), and Rose and Cohen11(0.9%). This is most likely due to the different definitions of DMV. Historically, only three categories of MV have been used (easy, difficult, and impossible).1,4,12Of note, our inclusion of a grade 2 definition (ventilated by mask with oral airway/adjuvant with or without muscle relaxant) is applicable to attempts that are neither easy nor difficult. This may be the most important explanation for why our incidence of grade 3 MV was lower than Langeron’s incidence of DMV. Our study participants were able to describe an airway that was not “easy” but nevertheless presented some challenges. Clinically, neither grade 1 nor grade 2 MV raises significant clinical concern for the experienced anesthetist. This is similar to the four-point scales used to describe DL view (Cormack and Lehane) and oropharyngeal anatomy (Mallampati examination as modified by Samsoon and Young).7,9The American Society of Anesthesiologists Task Force on Management of the Difficult Airway succinctly defined DI as intubation requiring “multiple attempts,”1whereas the definition of DMV was a list of signs and symptoms ranging from objective monitoring abnormalities to subjective assessments of adequacy of air movement.1Although Han’s MV scale is also limited by definitions including multiple signs and symptoms, the four-point scale may be superior at discriminating clinically significant MV challenges. Although other small studies have identified similar DMV rates, Langeron’s three categories and 5% DMV incidence may overstate the incidence of clinically significant DMV.12,13 

We were able to confirm Langeron’s observation that increased BMI, presence of a beard, history of snoring, and advanced age are independent predictors of grade 3 MV. Our observations indicated an optimal sensitivity and specificity at a BMI of 30 rather than 26 kg/m²as previously reported.4Although significantly increased BMI has been found to be a risk factor for DI,14these data demonstrate that even moderately increased BMI is the most important risk factor for grade 3 MV. We have confirmed presence of a beard as an important risk factor for grade 3 MV. Poor mask fit and gas leak are the intuitive anatomical pathology relating presence of a beard and grade 3 MV.15A beard is the only easily modifiable risk factor for DMV. Now that it has been confirmed that a beard is a significant risk factor, we are obligated to inform patients of this risk. We may need to recommend that they shave their beard before the procedure, especially in patients with other risk factors for DMV. Further investigation in this area is necessary.

We were able to confirm that a history of snoring is associated with grade 3 MV as reported previously by Langeron et al.  4and Yildiz et al.  12Snoring has been shown to be related to upper airway collapse.16A history of obstructive sleep apnea requiring surgical or positive airway pressure treatment was not found to be related to grade 3 MV. Given the high prevalence of snoring, we had hoped to find a more specific historical or physical examination element that may provide improved positive predictive value.17In contrast to Langeron’s findings, we were unable to identify lack of teeth as an independent predictor of grade 3 MV. Although age of 57 yr or older may seem to be the independent predictor responsible for eliminating edentulous dentition from the multivariate model, interaction analysis did not support this theory. This deviation from the results of Langeron et al.  warrants further study.

Our data do identify a possible relation between abnormalities in the mandibular protrusion test and grade 3 MV. The jaw-thrust maneuver as a tool in restoring patency of the upper airway is a mainstay of anesthetic practice and has been described for more than 100 yr.18Although evaluation of the mandibular protrusion test is a part of the American Society of Anesthesiologists Task Force on Management of the Difficult Airway’s standard physical examination,1many institutions, including ours, have not historically performed the test, been aware of its significance, or documented its findings.19Calder et al.  6and Takenaka et al.  20indicated the need to further study the value of this quick and simple test. The inability to protrude the mandible, particularly in patients with characteristics predisposing them to upper airway collapse, may be an important risk factor. Our data do indicate a role for this test as part of the standard airway examination and its inclusion as a risk factor for grade 3 MV.

Patients with three or more points in the predictor scale had a grade 3 MV incidence of 5%, nearly 20 times the baseline incidence of 0.26% for patients with zero points (fig. 3). Some providers may wish to use a risk factor cutoff of three to guide their MV preparation. Given the morbidity associated with airway difficulty, our predictive factor system may serve to help the practitioner prepare for a possible episode of grade 3 MV by having the patient shave his beard, ensuring the presence of another anesthesia provider in the room, or preparation for alternate methods of MV.1 

The study of IMV has been limited to anecdotal reports of its occurrence.21,22No previous study has been able to comment on its incidence or predictors.13Our series of 37 patients demonstrates the largest group of IMV patients reported thus far and may offer some insight. The incidence of grade 4 MV is rare at 0.16%, but still more frequent than other dreaded anesthesia complications such as malignant hyperthermia or homozygous atypical pseudocholinesterase.23,24Given that MV is an important rescue technique in cases of DI, the inability to mask ventilate represents an event with significant potential morbidity and mortality.13We were only able to identify two predictors of grade 4 MV: a history of snoring and thyromental distance of less than 6 cm. This is almost certainly due to limited statistical power given the relatively small number of cases available for study, masking true relations. However, it may actually reveal an underlying variation in etiology between grade 3 MV and grade 4 MV.

The most important result from our series of 37 grade 4 MV patients is the fact that only 1 patient required surgical airway access. Given the overlap of conditions that predispose to grade 3 MV and DI, a valid a priori  concern would be that IMV cases may have a high incidence of impossible intubation via  DL. Although a disproportionate share of these patients had poor views on DL, they were successfully intubated. Unfortunately, provider preoperative concern for DI may be markedly skewing our results. Out of concern for impossible ventilation and intubation, the anesthesia provider may have chosen elective awake fiberoptic intubations and thereby excluded these patients from our data set. Table 6does demonstrate that patients undergoing elective awake fiberoptic intubation had much higher rates of the risk factors for grade 3 MV than the general population studied (P < 0.01). Clearly, despite our large overall sample size, we were unable to detect a large number of IMV cases and struggle to provide conclusions regarding IMV risk factors. Further studies assessing incidence, predictors, and impact of IMV are essential.

Analysis of patients with grade 3 or 4 MV and DI represents a fruitful and important effort because of the frequency observed in our population (0.37%) and a large enough series to provide meaningful data (84 cases). Limited or severely limited mandibular protrusion, thick/obese neck anatomy, a history of sleep apnea, a history of snoring, and BMI of 30 kg/m²or greater were identified as independent predictors. This clinical situation represents among the most feared airway outcomes: a patient in whom establishing endotracheal ventilation is difficult and the primary rescue technique, MV, is also challenging. The ability to predict this situation would offer the clinician the ability to prepare for it with alternative airway tools before engaging in anesthesia induction: laryngeal mask airway, fiberoptic intubation cart, Bullard laryngoscope, and so forth. Intuitively, the presence of a beard should not be a shared anatomical abnormality for both grade 3 MV and DI, and the data are consistent with this hypothesis. The mandibular protrusion test was the most important predictor for this outcome. This supports the theory that defects in mandibular protrusion may be a shared abnormality between DI and DMV as suggested by Takenaka et al.  20 

There are several limitations to our conclusions. To garner a large enough sample size, we could not introduce a data collection process or care protocol that interfered with delivery of clinical care. Despite general standardization of MV and intubation technique at our institution, we cannot guarantee that controlled and uniform conditions were applied across all the MV attempts. In addition, both the possible predictors and outcomes were recorded by providers as part of their clinical documentation responsibilities. Although the format and specificity of some elements were prospectively altered to provide more detailed data for analysis, we did not use a distinct data collection form with diagrams and extensive definitions to assist providers in accurate selection as recommended in other studies.25A more consistent reported incidence of grade 3 MV is a first step to predicting its occurrence. To that end, we recommend the validation of an MV scale as described by Han et al.  5Our definition was more stringent than that used by Langeron and may be underestimating the incidence of clinically significant grade 3 MV as a result. Our analysis of grade 4 MV is limited by the rarity of the event more so than ambiguity in its definition. Despite reviewing more than 20,000 cases, we were able to identify only 37 occurrences and were unable to derive reliable predictors of the event. Multicenter trials combining patient populations or detailed retrospective studies of patients exhibiting IMV may be warranted.

In conclusion, we have been able to demonstrate the value of the mandibular protrusion test in predicting DMV and DMV combined with DI. We have provided confirmation of previous studies indicating the predictive value of advanced age, increased BMI, presence of a beard, and a history of snoring. Furthermore, we have been able to comment on the incidence and predictors of two more rare yet clinical worrisome situations: IMV and DMV combined with DI. We hope our data can serve to help anesthesia providers prepare for possible DMV with greater accuracy.