Secondary Logo

Journal Logo

Critical Care and Resuscitation: Original Clinical Research Report

Comparison of the Mallampati Classification in Sitting and Supine Position to Predict Difficult Tracheal Intubation: A Prospective Observational Cohort Study

Hanouz, Jean-Luc MD, PhD*,†; Bonnet, Vincent MD*; Buléon, Clément MD*; Simonet, Thérèse MD*; Radenac, Dorothée MD*; Zamparini, Guillaume MD*; Fischer, Marc Olivier MD, PhD*; Gérard, Jean-Louis MD, PhD

Author Information
doi: 10.1213/ANE.0000000000002108
  • Free

Despite improvement in airway management guidelines and widespread use of airway devices such as video laryngoscopes and the gum elastic bougie, difficult airway management remains an important cause of hypoxic brain damage or anesthesia-related death.1,2 Consequently, preoperative detection of difficult airway management, ie, difficult mask ventilation and difficult tracheal intubation (DTI), is an important aspect of preoperative evaluation for anesthesiologists. The 2013 American Society of Anesthesiologists guidelines for evaluation and management of the difficult airway recommend that, whenever feasible, an airway history and physical examination be conducted in all patients before the initiation of anesthetic care.1 The airway physical examination prioritizes several clinical elements including the Mallampati classification (MLPT). The MLPT was first described with 3 classes, but a fourth class (inability to see the soft palate) was subsequently added.3,4

The MLPT remains the most popular bedside screening test and is included in nearly all multivariable scores aimed at predicting DTI.5–9 The MLPT depends on the visual inspection of pharyngeal structures seen in patients in the sitting position with the head in neutral position, mouth open as widely as possible, and the tongue extended to its maximum without phonation.5,9 However, the physical examination cannot always be performed in the sitting position. In emergency cases, where pain may limit mobilization, for example, or where immobilization in the supine position is required for potential neck injury, the MLPT often must be performed while the patient is supine.

To our knowledge, 6 studies have compared the MLPT in sitting and supine positions.10–15 Two studies showed that the MLPT changed toward a higher class when patient’s position switched from sitting to supine.10,11 Four studies have reported controversial results on the performance of the MLPT in sitting and supine positions to predict difficult laryngoscopy (ie, Cormack and Lehane grade).12–15 No study compared the diagnostic performance of the MLPT in sitting and supine positions to predict difficult intubation. Given the importance of preintubation airway risk assessment and the routine use of the MLPT, establishing its predictive power for difficult intubation in supine position is of clear relevance for anesthesiologists.

The aim of the current study was to compare the diagnostic accuracy of the MLPT in supine and sitting positions for the prediction of DTI in a large cohort of adult surgical patients. We hypothesized that the prediction of DTI would differ when the MLPT is performed in sitting and supine positions.

METHODS

This single-center prospective observational study was performed in the University Hospital of Caen from November 2012 to June 2013. This study conformed to the ethical principles for medical research involving human subjects written in the Declaration of Helsinki. Because we did not randomly assign patients and only routine care was performed, waived written informed consent was authorized by the local medical ethics committee (Comité de Protection des Personnes Nord Ouest III, Caen). Nevertheless, we sought oral informed consent from all patients. The report of the present study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) standards for observational studies.16

We prospectively included adult patients aged 18 years and above scheduled for surgery requiring general anesthesia and orotracheal intubation. Patients undergoing general anesthesia without orotracheal intubation, those scheduled for fiber-optic intubation or orotracheal intubation performed by residents or nurses less than 3 years of experience, and emergency cases were excluded. During the preanesthesia consultation, the anesthesiologist collected the physical examination of the airway using a standard form. During the preoperative anesthesia consultation, a senior anesthesiologist collected the following data: age, sex, weight, height, history of diabetes, history of snoring every night, presence of beard, complete edentulism, neck mobility, interincisor distance (or intergingival for edentulous patients), thyromental distance, MLPT observed in the sitting position with the head in neutral position, mouth fully open, tongue out, and without phonation. The MLPT was reported on the standard form, which was kept in the patient’s anesthesia record included in the patient’s medical record. The delay between the preoperative examination and surgery ranged from several days to several weeks.

Before induction of anesthesia, the MLPT was evaluated in supine position on the operating room table (patient’s head on a 5-cm-high gel donut head pillow and neck in neutral position, mouth fully open, tongue out, and without phonation) by an independent observer (resident in anesthesiology with >3 years experience) placed at the patient’s head looking at oropharyngeal structures from above, with eyes aligned on a vertical axis starting from the middle of the opened mouth. The observer was not involved in the preoperative examination or in the patient’s anesthesia care and was blinded from the results of the preoperative MLPT. During induction of anesthesia performed by the senior anesthesiologist in charge of the patient and a nurse anesthesiologist, the independent observer also collected the following data: use of neuromuscular blocking agent, external laryngeal mobilization (backward, upward, and rightward pressure), characterization of mask ventilation (easy, difficult, and impossible), characterization of tracheal intubation (easy, difficult, and impossible), and the laryngoscopic view according to the Cormack and Lehane classification depicted on the standard form. Orotracheal intubation was performed by a senior anesthesiologist or nurse anesthetist (>3 years experience).

The standardized procedure for tracheal intubation used in our institution included patient’s head on a 5-cm-high gel donut head pillow, preoxygenation using the anesthesia ventilator (Aisys CS; GE Healthcare, Vélizy-Villacoublay, France), and positive inspiratory support (inspiratory pressure +4 to +8 cm H2O, positive end expiratory pressure + 5 cm H2O) until end-tidal O2 was superior to 90%. Mask ventilation was performed after placing an oropharyngeal cannula, using the anesthesia ventilator and controlled ventilation (tidal volume = 7 mL/kg, respiratory rate = 15/min, maximal airway pressure set at +20 cm H2O). During induction of anesthesia, patients were monitored with noninvasive blood pressure (repeated every 3 minutes), continuous 5-lead electrocardiography, and pulse oximetry (IntelliVue MP70; Philips HealthCare, Amsterdam, the Netherlands). If a neuromuscular blocking agent was administered, its effect was monitored by accelerometry at the thumb following Train-of-Four stimulations of the ulnar nerve repeated every 30 seconds (Philips Intellivue NMT; Philips HealthCare, Amsterdam, the Netherlands). During mask ventilation, end-tidal CO2, exhaled gas spirometry, and insufflation pressures were monitored.

All laryngoscopies were performed using metallic single-use laryngoscope blade (Callisto Single Use Fibre Optic Laryngoscope Macintosh Blade; Timesco, Essex, United Kingdom) to avoid differences in laryngoscopic visualization between plastic and metallic laryngoscope blades.17 After tracheal intubation, the correct positioning of the endotracheal tube was confirmed by the anesthesiologist via capnography and bilateral auscultation of lungs.

Definitions of Outcomes

The primary outcome was the discriminative power of the MLPT evaluated in the supine position for predicting DTI versus the MLPT evaluated in the sitting position. The discriminative power was assessed through the area under the receiver operating characteristic (ROC) curves.

Secondary outcomes were diagnostic accuracy (sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratio) of the MLPT evaluated in sitting and supine positions for predicting DTI and difficult mask ventilation (DMV), concordance between the MLPT in supine and sitting positions, independent variables associated with DTI, and the diagnostic accuracy of the prediction score for DTI described by Naguib et al18 using the MLPT in sitting and supine positions. This score was chosen because it was the best one to predict DTI and because it uses the MLPT.

DTI was defined as an orotracheal intubation requiring more than 2 laryngoscopies, lasting more than 10 minutes, or requiring an alternate device (gum elastic bougie, supraglottic device, or videolaryngoscope).8,18 Impossible intubation was included in the DTI group.

DMV was defined as the inability for the anesthesiologist to provide adequate ventilation because of one or more of the following problems: inability for the unassisted anesthesiologists to maintain oxygen saturation >92% using 100% oxygen, excessive gas leak requiring use of the oxygen flush valve more than twice, excessive insufflation pressure (>25 cm H2O), absence of spirometric measures of exhaled gas flow or a tidal volume <3 mL/kg, absence or inadequate exhaled CO2, and need to perform 2-handed mask ventilation.20

Statistical Analysis

Data are expressed as mean ± standard deviation or median (first quartile; third quartile) for nonnormally distributed continuous variables and numbers (percentages) for categorical variables. Normality was tested with the D’Agostino-Pearson test. The comparisons of variables between 2 groups (ie, easy and difficult tracheal intubation groups) were performed using Student t test or Mann-Whitney U test for quantitative variables and the Fisher exact test or χ2 for categorical variables.

Independent Variables Associated With Difficult Tracheal Intubation

All variables recorded in the study were possible predictors for DTI. We performed 2 multiple forward stepwise logistic regressions to determine independent variables associated with DTI including the MLPT in either the sitting or the supine position. These models were constructed to examine if independent variables previously described as associated with DTI were observed in the present study.1,7,8 Besides the MLPT, independent variables included in the models were age, American Society of Anesthesiologists class, body mass index, history of snoring, complete edentulism, neck mobility, interincisor distance, and thyromental distance. The calibration of the final models was assessed by the Cessie-van Houwelingen statistic. The discrimination of the final models was assessed by the C statistic. A 10-fold cross-validation was performed to calculate the optimism and the optimism-corrected C statistic.

Diagnostic Accuracy of the Mallampati Classification

The sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratio of the MLPT in the sitting and supine positions were calculated. The discriminative power of the MLPT in sitting and supine positions was assessed using the ROC curve obtained by averaging 1000 samples bootstrapped (sampling with replacement) from the original study population. Bootstrapping limits the impact of outliers and provides robust representation including 95% confidence interval (CI) depicted using box plots. The comparison of areas under the 2 ROC curves was performed by the nonparametric method described by DeLong et al.21

Reclassification tables were constructed to evaluate the net reclassification improvement and the integrated discrimination improvement in cases with DTI and in cases with easy intubation. Two classes of risk were defined: low risk of DTI (MLPT classifications 1 and 2) and high risk of DTI (MLPT classifications 3 and 4).

Agreement Between the Mallampati Classification in the Sitting and Supine Positions

The strength of the relationship between MLPT in the sitting and supine positions was evaluated by the Spearman rank correlation. The agreement between 2 ordinal scales, ie, MLPT in sitting and supine positions, was examined by the weighted Cohen κ statistics.

Naguib Score Accuracy Including Mallampati in Sitting and Supine Positions

We tested the predictive performance for DTI of the score described by Naguib et al18 with Mallampati either in the sitting or in the supine position according the equation:

where TMD is the thyromental distance in centimeters, ID is the interincisor distance in centimeters, and Mallampati score is dichotomized into 0 (MLPT 1 and 2) and 1 (MLPT 3 and 4).12 According to this equation, intubation would be easy if the numerical value (l) is less than zero and difficult if it is greater than zero. The Naguib score was reported as the best one to predict unanticipated DTI. It enabled us to compare its diagnostic accuracy using the MLPT in the sitting and supine positions.18 The discriminative power of the 2 scores was assessed using average ROC curves obtained from 1000 samples bootstrapped (sampling with replacement) from the original study population. The comparison of areas under the 2 ROC curves was performed by the nonparametric method described by DeLong et al.

Sample Size Justification

The number of patients to be included was estimated based on a previously reported summarized area under the ROC curve at 0.75 for the MLPT in the sitting position to predict DTI.19 Based on available literature, we hypothesized that the MLPT in the supine position would modify the area under the ROC curve by 10%.10–15 The ratio of positive cases (ie, DTI) to negative cases was estimated at 1/17 (6% incidence of DTI according to the chosen definition).8,20 The number of positive and negative cases required were calculated at 106 and 1806, respectively (α risk = 5%, β risk = 10%). However, we planned to record 15 variables possibly associated with DTI and to analyze those who were independent predictors of DTI. To limit the risk of overfitting, 10 events per variable are suggested in multivariable analysis. Consequently, we decided to increase the number of positive cases at 150 (2550 negative cases).21 Taking 10% of missing data for outcome into account, we planned to include 2800 consecutive patients.

All P values were 2-tailed, and a P value of less than .05 was required to reject the null hypothesis. Statistical analysis was performed with R: A Language and Environment for Statistical Computing (R Core Team; R Foundation for Statistical Computing, Vienna, Austria) and specific packages.

RESULTS

A total of 3036 patients were included in the diagnostic performance analysis and in the concordance analysis between the Mallampati classification in the supine and sitting. Multivariate analysis included 2982 patients because of missing data for dependent variables in 54 patients.

According to the chosen definitions, there were 157 (5.1%) DTIs including 8 (0.3%) impossible intubations and 195 (6.4%) DMVs including 6 (0.2%) impossible facemask ventilation. Fifty-one (1.6%) patients presented both difficult DMV and difficult DTI. Table 1 shows the characteristics of patients with and without DTI.

Table 1.
Table 1.:
Univariate Comparison of Patients With or Without Difficult Tracheal Intubation
Table 2.
Table 2.:
Multivariate Analysis for Difficult Tracheal Intubation Including the Mallampati Classification in Sitting and Supine Positions

Logistic regression with MLPT in sitting and supine positions are summarized in Table 2. The common independent predictors of difficult intubation identified in the 2 logistic regression models were the MLPT 3 and 4, snoring and stiff cervical spine (Table 2). In contrast, complete edentulism was associated with a decreased risk of DTI. The le Cessie-van Houwelingen test confirmed the goodness of fit of the model with MLPT in sitting (P = .36) and supine (P = .73) positions. The C statistics of the model with MLPT in sitting and supine positions were 0.80 (0.76–0.84) and 0.86 (0.83–0.89), respectively. The optimism-corrected C statistic was 0.78 (optimism = 0.02) for the model in the sitting position and 0.85 (optimism = 0.01) for the model with the MLPT in supine position.

Performance of the Mallampati Classification Evaluated in Supine Position for the Prediction of Difficult Tracheal Intubation

Table 3.
Table 3.:
Diagnostic Accuracy of the Mallampati Classification in Supine and Sitting Positions for the Prediction of Difficult Tracheal Intubation
Figure 1.
Figure 1.:
Discriminative power of the Mallampati classification in sitting and supine positions to predict difficult tracheal intubation represented by receiver operating characteristic (ROC) curves. Distribution of the 1000 bootstrapped samples from the original study population is represented by box plots. Lines represent the average ROC curve.

Table 3 shows the diagnostic characteristics of the MLPT in supine and sitting positions for predicting DTI. The area under the ROC curve for the MLPT in supine position (0.82 [0.78–0.84]) was greater than that for the MLPT in the sitting position (0.70 [0.66–0.75]; P < .001; Figure 1).

Performance of the Mallampati Classification Evaluated in Supine Position for the Prediction of Difficult Mask Ventilation

Table 4.
Table 4.:
Diagnostic Accuracy of the Mallampati Classification in Supine and Sitting Positions for the Prediction of Difficult Mask Ventilation

Table 4 shows the diagnostic characteristics of the MLPT in supine and sitting positions for predicting DMV. The area under the ROC curve for the MLPT in supine position (0.74 [0.70–0.77]) was greater than that for the MLPT in the sitting position (0.65 [0.61–0.69]; P < .001).

Agreement Between the Mallampati Classification Evaluated in Supine and Sitting Positions

The Spearman rank correlation coefficient ρ was 0.50 (P < .001), suggesting a moderate monotonic relationship. Figure 2 summarizes the concordance between MLPTs evaluated in sitting and supine positions. The weighted Cohen κ statistic measuring the agreement between the sitting and supine MLPTs was 0.40 (0.38–0.43), which was considered fair.

Figure 2.
Figure 2.:
Bubble plot showing the distribution of the Mallampati classification observed in the sitting and supine positions. Area of circle is proportional to the number of case.

Among the 2664 patients in whom the MLPT in the sitting position was 1 and 2, 257 (13%) had an MLPT 3 and 4 in the supine position. Among the 372 patients in whom the MLPT in the sitting position was 3 and 4, 120 (32%) had an MLPT 1 and 2 in the supine position.

Table 5.
Table 5.:
Reclassification Tables Using the Mallampati Classification as the Classifier of the Risk of Difficult Intubation

The net reclassification improvement for events (ie, DTI) was 0.28 and the net reclassification improvement for nonevents (ie, no DTI) was 0.07. As compared to the MLPT in the sitting position, the MLPT in the supine position improved the proportion of correctly classified patients (net reclassification improvement = 0.21; Table 5). The integrated discrimination improvement for events (ie, DTI) was 0.28 and the integrated discrimination improvement for nonevents (ie, no DTI) was 0.06.

Evaluation of the Diagnostic Performance of the Naguib Score Including Mallampati Classification in Sitting and Supine Positions

Table 6.
Table 6.:
Diagnostic Characteristic of the Naguib Score Calculated With the Mallampati Classification in Sitting and Supine Positions for the Prediction of Difficult Tracheal Intubation
Figure 3.
Figure 3.:
Discriminative power of the Naguib score calculated with the Mallampati in sitting or supine position to predict difficult tracheal intubation represented by receiver operating characteristic (ROC) curves. Distribution of the 1000 bootstrapped samples from the original study population is represented by box plots. Lines represent the average ROC curve.

Table 6 shows the diagnostic characteristics of the Naguib score calculated with the MLPT in sitting or supine position for the prediction of DTI. As shown in Figure 3, the area under the ROC curve for the prediction of DTI by the Naguib score calculated with MLPT in the supine position (0.78 [95% CI, 0.74–0.82]) was significantly greater than that for the Naguib score calculated with MLPT in the sitting position (0.69 [95% CI, 0.63–0.74]; P < .001).

DISCUSSION

In this prospective study of 3036 patients, we found a fair agreement between the MLPT evaluated in sitting and supine positions related to correct reclassification of 21% of patients with DTI, that the performance of the supine MLPT for predicting DTI and DMV was superior to the sitting MLPT, and that use of the supine MLPT improved the diagnostic performance of the Naguib score.

As emphasized in recent guidelines, airway physical examination is a key aspect of preanesthetic airway assessment.1 During the examination, assessment of several clinical signs are recommended including the MLPT, interincisor and thyromental distances, relationship of maxillary and mandibular incisors, cervical spine mobility, and neck anatomy. The MLPT is the most popular morphologic test used to evaluate the airway.23 It was defined and studied with the patient in the sitting position, head in neutral position, mouth widely open, tongue protruded, and without phonation.3,4 Although well described, this definition raises several issues including the sitting position and the head, mouth, and tongue relationships. With respect to the sitting position, it is not possible in patients with traumatic spine fractures, hip and lower limb fractures, pain, and mandatory immobilization in supine position. With respect to the head, mouth, and tongue relationships, it has been shown that when the patients were asked to perform maximal mouth opening, a spontaneous 26° craniocervical extension from neutral position occurred resulting in a 10-mm increase in the interincisor distance.24 Furthermore, the response of patients to instructions varies widely with 25% of patients phonating spontaneously.10,25 Finally, it has been highlighted that no standard method was used in the literature to evaluate the MLPT, limiting the interpretation of the results and precluding between-study comparisons.12,19

The present study, based on a large prospective cohort of patients, focused on the patient’s position. Although we found a fair agreement between the MLPT in sitting and supine positions, we also showed that MLPT in the supine position could improve diagnostic performance for predicting DTI.

Studying 64 and 80 patients, respectively, Tham et al10 and Singhal et al11 suggested that the MLPT shifted toward a higher grade when the patient was turned to the supine position from the sitting position.9 In contrast, the present study showed that the supine position worsened the view of the pharyngeal structures in 13% of patients but improved it in 32% of patients. Nevertheless, the supine MLPT was associated with an improved reclassification of patients. The fair agreement between the MLPT in the sitting and supine positions may be explained, at least in part, by complex differences in upper airway anatomy between sitting and supine positions. Thus, in awake volunteers, the supine position resulted in a lower oropharyngeal cross-sectional area as compared to the sitting position.26 In contrast, neck extension increased oropharyngeal cross-sectional area and improved mouth opening.24,26 Because we did not measure the craniocervical angle in sitting and supine positions, we cannot rule out a difference that may have modified the oropharyngeal view. Finally, the effect of anesthesiologists’ level of experience and interobserver variability may have played a role. The interobserver agreement of the MLPT remains unclear since it has been evaluated in small sample size studies and reported to be poor to good.27,28 However, in a large prospective cohort study, it has been suggested that interobserver variability should have no significant effect on the MLPT.29

The diagnostic characteristics of the sitting MLPT we reported was comparable to that reported in meta-analyses and large cohort studies.7,19,29 We found that the supine MLPT substantially increased sensitivity but slightly decreased specificity in predicting DTI. Nevertheless, the positive likelihood ratios were the same for both positions and consistent with those previously reported.19 Previously published studies reported no change in the diagnostic characteristics of the MLPT between sitting and supine positions to predict difficult laryngoscopy.12–15 The present study showed that the performance of the supine MLPT for predicting DTI was superior to the sitting MLPT and increased the proportion of correctly classified patients. Additionally, we found that replacing the sitting MLPT with the supine version improved the diagnostic accuracy of the Naguib score. Although unclear, it can be hypothesized that a combination of anatomical differences between sitting and supine positions may modify visualization of oropharyngeal structures.17,22 Further studies are required to elucidate these points.

Our study showed that the performance of the supine MLPT for predicting DMV was superior to the sitting MLPT. Limited data are available on DMV but the MLPT 3 or 4 was identified as an independent predictor in several studies.30 Specific studies on DMV are required.

Our study has limitations that deserve consideration. First, this study was performed on adult patients undergoing elective surgery and thus cannot be generalized to pediatric patients or obstetric patients. Furthermore, the generalizability of the study is also limited by its single-center design. Second, we excluded patients scheduled for fiber-optic intubation because of a high likelihood of DTI. Consequently, we excluded patients in whom history and clinical examination indicated a high probability of difficult airway management. Nevertheless, we believe that improving the diagnostic performance for DTI is useful in patients in whom DTI is not anticipated. Third, the incidence of DTI depends on the chosen definition. We used the definition from French guidelines and obtained a 5.1% DTI incidence, which is comparable to that reported in other studies.8,17 Importantly, orotracheal intubation was only performed by experienced anesthesiologists with metallic laryngoscope blade. Fourth, we lack data on the interobserver variability of the MLPT within the anesthesiologist pool. However, previous studies have found little effect of interobserver variability on the MLPT score.28 In addition, it is likely that the interobserver variability in this study is representative of that existing in daily clinical practice. Fifth, the MLPT in the supine position was evaluated by residents in anesthesiology (>3 years of experience), whereas the MLPT in the sitting position was evaluated by senior anesthesiologists. Thus, differences in the evaluation of MLPT and different interobserver variability may have occurred. To limit this bias, we used a standard form with specific figures to guide the evaluation.

In conclusion, the MLPT performed in the supine position is possibly superior to that performed in the sitting position for predicting difficult intubation in adults. Thus, the MLPT evaluated in the supine position can be used in preoperative airway examination in patients in whom the sitting position is impossible. In patients in whom both sitting and supine positions are possible, evaluating the MLPT in the supine position may improve the ability to predict DTI.

DISCLOSURES

Name: Jean-Luc Hanouz, MD, PhD.

Contribution: This author helped design, supervise, and participate in the study, helped perform data management, analyze data, and write the manuscript.

Name: Vincent Bonnet, MD.

Contribution: This author helped participate in the study and read the manuscript.

Name: Clément Buléon, MD.

Contribution: This author helped participate in the study and read the manuscript.

Name: Thérèse Simonet, MD.

Contribution: This author helped participate in the study and read the manuscript.

Name: Dorothée Radenac, MD.

Contribution: This author helped participate in the study and read the manuscript.

Name: Guillaume Zamparini, MD.

Contribution: This author helped participate in the study and read the manuscript.

Name: Marc Olivier Fischer, MD, PhD.

Contribution: This author helped participate in the study and read the manuscript.

Name: Jean-Louis Gérard, MD, PhD.

Contribution: This author helped read the manuscript.

This manuscript was handled by: Avery Tung, MD, FCCM.

REFERENCES

Apfelbaum JL, Hagberg CA, Caplan RA, et al. Practice guidelines for management of the difficult airway: An updated report by the American Society of Anesthesiologists Task Force on Management of the Difficult Airway. Anesthesiology. 2013;118:251–270.
2. Peterson GN, Domino KB, Caplan RA, Posner KL, Lee LA, Cheney FW. Management of the difficult airway: a closed claims analysis. Anesthesiology. 2005;103:33–39.
3. Mallampati SR, Gatt SP, Gugino LD, et al. A clinical sign to predict difficult tracheal intubation: a prospective study. Can Anaesth Soc J. 1985;32:429–434.
4. Samsoon GL, Young JR. Difficult tracheal intubation: a retrospective study. Anaesthesia. 1987;42:487–490.
5. Wilson ME, Spiegelhalter D, Robertson JA, Lesser P. Predicting difficult intubation. Br J Anaesth. 1988;61:211–216.
6. el-Ganzouri AR, McCarthy RJ, Tuman KJ, Tanck EN, Ivankovich AD. Preoperative airway assessment: predictive value of a multivariate risk index. Anesth Analg. 1996;82:1197–1204.
7. Shiga T, Wajima Z, Inoue T, Sakamoto A. Predicting difficult intubation in apparently normal patients: a meta-analysis of bedside screening test performance. Anesthesiology. 2005;103:429–437.
8. Langeron O, Cuvillon P, Ibanez-Esteve C, Lenfant F, Riou B, Le Manach Y. Prediction of difficult tracheal intubation: time for a paradigm change. Anesthesiology. 2012;117:1223–1233.
9. Arné J, Descoins P, Fusciardi J, et al. Preoperative assessment for difficult intubation in general and ENT surgery: predictive value of a clinical multivariate risk index. Br J Anaesth. 1998;80:140–146.
10. Tham EJ, Gildersleve CD, Sanders LD, Mapleson WW, Vaughan RS. Effects of posture, phonation and observer on Mallampati classification. Br J Anaesth. 1992;68:32–38.
11. Singhal V, Sharma M, Prabhakar H, Ali Z, Singh GP. Effect of posture on mouth opening and modified Mallampati classification for airway assessment. J Anesth. 2009;23:463–465.
12. Lewis M, Keramati S, Benumof JL, Berry CC. What is the best way to determine oropharyngeal classification and mandibular space length to predict difficult laryngoscopy? Anesthesiology. 1994;81:69–75.
13. Bindra A, Prabhakar H, Singh GP, Ali Z, Singhal V. Is the modified Mallampati test performed in supine position a reliable predictor of difficult tracheal intubation? J Anesth. 2010;24:482–485.
14. Amadasun FE, Adudu OP, Sadiq A. Effects of position and phonation on oropharyngeal view and correlation with laryngoscopic view. Niger J Clin Pract. 2010;13:417–420.
15. Khan ZH, Eskandari S, Yekaninejad MS. A comparison of the Mallampati test in supine and upright positions with and without phonation in predicting difficult laryngoscopy and intubation: A prospective study. J Anaesthesiol Clin Pharmacol. 2015;31:207–211.
16. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370:1453–1457.
17. Buléon C, Parienti JJ, Lesage A, et al. Comparison of plastic and metallic single-use and metallic reusable laryngoscope blades: a randomised controlled trial. Eur J Anaesthesiol. 2013;30:163–169.
18. Naguib M, Scamman FL, O’Sullivan C, et al. Predictive performance of three multivariate difficult tracheal intubation models: a double-blind, case-controlled study. Anesth Analg. 2006;102:818–824.
19. Lundstrøm LH, Vester-Andersen M, Møller AM, Charuluxananan S, L’hermite J, Wetterslev J; Danish Anaesthesia Database. Poor prognostic value of the modified Mallampati score: a meta-analysis involving 177 088 patients. Br J Anaesth. 2011;107:659–667.
20. Diemunsch P, Langeron O, Richard M, Lenfant F. [Prediction and definition of difficult mask ventilation and difficult intubation: question 1. Société Française d’Anesthésie et de Réanimation]. Ann Fr Anesth Reanim. 2008;27:3–14.
21. DeLong ER, DeLong DM, Clarcke-Pearson DL. Comparing the area under two or more correlated receiver operating characteristic curves: A nonparametric approach.Biometrics. 1988;44:837.
22. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49:1373–1379.
23. Hung O, Law JA, Morris I, Murphy M. Airway assessment before intervention: what we know and what we do. Anesth Analg. 2016;122:1752–1754.
24. Calder I, Picard J, Chapman M, O’Sullivan C, Crockard HA. Mouth opening: a new angle. Anesthesiology. 2003;99:799–801.
25. Wilson ME, John R. Problems with the Mallampati sign. Anaesthesia. 1990;45:486–487.
26. Jan MA, Marshall I, Douglas NJ. Effect of posture on upper airway dimensions in normal human. Am J Respir Crit Care Med. 1994;149:145–148.
27. Karkouti K, Rose DK, Ferris LE, Wigglesworth DF, Meisami-Fard T, Lee H. Inter-observer reliability of ten tests used for predicting difficult tracheal intubation. Can J Anesth. 1996;43:554–559.
28. Rosenstock C, Gillesberg I, Gätke MR, Levin D, Kristensen MS, Rasmussen LS. Inter-observer agreement of tests used for prediction of difficult laryngoscopy/tracheal intubation. Acta Anaesthesiol Scand. 2005;49:1057–1062.
29. Healy DW, LaHart EJ, Peoples EE, Jewell ES, Bettendorf RJ Jr, Ramachandran SK. A comparison of the Mallampati evaluation in neutral or extended cervical spine positions: a retrospective observational study of >80 000 patients. Br J Anaesth. 2016;116:690–698.
30. El-Orbany M, Woehlck HJ. Difficult mask ventilation. Anesth Analg. 2009;109:1870–1880.
Copyright © 2017 International Anesthesia Research Society