Difficult tracheal intubation remains a major source of anaesthesia-related morbidity and mortality.1–5 Airway management remains one of the most critical responsibilities of anaesthesiologists and therefore the prediction of possible difficult airway situations is one of the most important items that need to be addressed during the preoperative anaesthesiological workup.6 This usually includes precise examination of the patient's head and neck anatomy including Mallampati classification,7 thyromental distance (also known as Patil's test),8 sternomental distance,9 body weight,10 mouth opening11 and head and neck movement.12 Several clinical studies have been conducted to identify predictive factors of difficult laryngoscopy.9,13–17 Only a few older studies used imaging techniques to find anatomical predictors of difficult laryngoscopy and all of these used conventional radiograph techniques with all the limitations of that technique, such as radiolucent soft tissue and cartilage structures.18–21 MRI may provide a better means of studying the complex airway anatomy and for identifying potential predictors of difficult laryngoscopy/tracheal intubation.
The aim of this study was to investigate whether there is a correlation between difficult laryngoscopy and the anatomical position of the vocal cords in relation to the cervical vertebrae as assessed by MRI.
Material and methods
In December 2008, prior to starting this study in January 2009, we contacted our local Ethics Committee (Ethics Committee of the Friedrich-Alexander Universität Erlangen, Erlangen, Germany; Chairperson Professor Peter Betz) and were told that based upon the design of the proposed study and according to German law, no formal Ethics Committee approval was required. Owing to the fact that we had not received a protocol number after this initial enquiry by telephone, we contacted our local Ethics Committee again in June 2015 and asked for a formal letter confirming that they had no objections to this study, and an official protocol number was received on 7 July 2015 (Nr. 171_15 Bc).
We enrolled 142 adult patients after obtaining written informed consent to this study between January 2009 and March 2012. Patients were undergoing neurosurgical procedures under general anaesthesia and scheduled for elective tracheal intubation. In a first step, we enrolled only patients with an existing preoperative MRI which included images of the neck and the cervical spine. In the second part of the study, we collected MRI data only from patients with difficult laryngoscopy, that is inclusion criteria for the second part were difficult laryngoscopy with Cormack and Lehane (CL) grade 3 or 4 (see further for definition) and existing imaging of the cervical spine. This adaptive enrichment study design was chosen because the incidence of difficult laryngoscopy is relatively low.22,23 Exclusion criteria were age below 18 years, patients with injuries of the cervical spine, restriction in reclining the head, acromegaly, history of radiation treatment of the neck, ENT tumours or restricted mouth opening of less than 3 cm.
We documented the patients’ age, sex, height, weight and American Society of Anesthesiologists’ physical status. Preoperative physical examination included known factors for prediction of a potential difficult airway. We recorded the modified Mallampati classification, the thyromental distance (TMD), the sternomental distance (SMD), mouth opening and the degree of retrognathia.
The modified Mallampati classification by Samsoon and Young24 describes the visibility of the pharyngeal structures. The patients were asked to hold the head in a neutral position, open the mouth as wide as possible and put out the tongue without phonation. Mallampati was graded in a standardised fashion24 by a single examiner.
TMD was measured along the upper edge of the thyroid cartilage to the chin with the head fully extended and mouth closed, and categorised as less than 6 cm (class 1), 6 to 6.5 cm (class 2) and more than 6.5 cm (class 3).25,26
SMD was measured as described by Al Ramadhani et al.9 on a line between the upper border of the manubrium sterni and the bony point of the mentum with the head in full extension and the mouth closed. It was categorised as 13.5 cm or less (class 1) and above 13.5 cm (class 2).
Mouth opening was defined as the space between the upper and lower incisors, with categorisation of 5 cm or less (class 1) and above 5 cm (class 2). In edentulous patients, the intergingival distance was measured.
Categories of retrognathia were none, moderate or strong in relation to a receding chin and rated by assessment of the hyoid-to-chin distance by a single study staff member for consistency.27
All laryngoscopies were performed by the same two skilled anaesthesiologists, who both had more than 10 years of clinical anaesthesia experience. For laryngoscopy, the patient's head was positioned in the ‘sniffing’ position and on a pillow elevating the head by 10 cm in a standardised fashion. A Macintosh laryngoscope blade size 3 was used and the view of the glottis was rated according to the classification of CL.28 The definitions of the different categories were as follows: CL 1, complete glottis visible; CL 2, anterior glottis not visible; CL 3, epiglottis visible, glottis not visible; CL 4, epiglottis not visible. We rated CL 1 and 2 as easy and CL 3 and 4 as difficult laryngoscopies.
MRI (1.5 Tesla Magnetom Sonata Maestro class, Siemens, Germany) was performed in advance or after the scheduled surgery. The patient's head was positioned on a pillow elevating the head by 10 cm, thus mimicking the intubation position. In some patients, however, pillows with a lower height had to be used to fit the patient's head below the MRI coil. Images were then screened for the slices in which the vocal cords could be identified and a projection line was plotted to determine the anatomical localisation in relation to the cervical vertebrae (Fig. 1). This localisation was independently categorised by three anaesthesiologists who were blinded with respect to difficulty of laryngoscopy, and a consensus had to be found. In case of disagreement, images were reviewed by the three anaesthesiologists together and discussed until a consensus was found. Categories were as follows: cervical disc between third and fourth cervical vertebrae; vertebral body four; cervical disc between fourth and fifth cervical vertebrae; vertebral body five; cervical disc between fifth and sixth cervical vertebrae; and vertebral body six.
Data were tested for normal distribution using the Shapiro–Wilk test. Continuous data were tested for differences between the two groups (easy vs. difficult laryngoscopy) using the t test for unpaired samples or the Mann–Whitney test. Categorical data were tested for differences between the groups using the χ2 or Fisher's exact tests. To identify possible predictors of difficult laryngoscopy, we used, in a first step, Spearman's rank correlation coefficient. Variables which showed a significant correlation with difficult laryngoscopy were then analysed further using a generalised linear model with binary logistical regression. Starting with a full model with all variables selected by correlation analysis, we removed in a stepwise procedure those which did not show significance in the generalised linear model. The ability of the model to correctly classify and predict difficult laryngoscopy was assessed by sensitivity, specificity and positive and negative predictive value. Statistical significance was defined as P less than 0.05. Statistical analysis was performed using SPSS (IBM, SPSS Statistics, Version 21.0, New York, USA). The primary outcome measure was to compare the position of the vocal cords relative to the cervical vertebrae as measured by MRI in patients with easy vs. difficult laryngoscopy. The secondary outcome measures were whether inclusion of vocal cord position relative to the cervical vertebrae would improve performance of a multivariate model to predict difficult intubation.
We enrolled 142 patients in this study, 91 with easy (CL 1 or 2) and 51 with difficult (CL 3 or 4) laryngoscopy. In the first part of the study, 94 patients were included (91 with CL 1 or 2, and three with CL 3 or 4). In the second part of the study, only patients with difficult intubation and MRI were included (48 with CL 3 or 4). Nine other patients with difficult intubation either did not have MRI of the cervical spine or were lost to follow-up, and were therefore excluded. Patients’ demographic data are shown in Table 1. There was no significant difference between the groups in relation to American Society of Anesthesiologists’ physical status, age, height, weight and BMI. The difficult laryngoscopy group contained significantly more male patients than the easy laryngoscopy group (61 vs. 40%, P = 0.02).
The distribution of the anatomical position of the vocal cords in relation to the cervical vertebrae differed significantly between the two groups (Fig. 2, P < 0.0001). In the difficult laryngoscopy group (CL classification 3 or 4), 37% of the patients had a more cranial vertebral body position of the vocal cords. In contrast, in the easy laryngoscopy group (CL classification 1 or 2), the majority of patients (55%) had a vocal cord position projecting to vertebral body five.
Some of the physical examination variables showed significantly different distributions in patients with difficult laryngoscopy (Table 2). The correlation analysis revealed the following variables as possible predictors of difficult laryngoscopy: male sex (P = 0.015); high Mallampati score (P = 0.001); retrognathia (P < 0.0001); prominent upper incisors (P < 0.0001) and position of the vocal cords (P = 0.036). As mandibular retrognathia and maxillary prognathism were highly intercorrelated (P < 0.0001), only mandibular retrognathia was included in the generalised linear model. During further analysis using the generalised linear model, only four of these variables were maintained in the final model: sex (P = 0.058); Mallampati score (P = 0.064); mandibular retrognathia (P < 0.001) and position of the vocal cords (P = 0.033). Using these four variables, the model revealed acceptable values for classification and predictivity (Table 3). Without the factor ‘vocal cord position’, the model showed a markedly reduced sensitivity. When the parameters sex and/or Mallampati score were removed, the model predictivity appeared to decrease, but did not reach significance in the model.
Our study demonstrated that in patients with difficult laryngoscopy, the vocal cords were positioned significantly more cranially compared with patients with easy laryngoscopy. This might be explained by the curvature of Macintosh blades, which does not allow viewing around the corner – a disadvantage that is overcome by using video laryngoscopes.
In addition to this new predictor for difficult intubation, the classical predictors of difficult laryngoscopy investigated in our study such as Mallampati score, TMD and mandibular retrognathia showed a significant correlation with difficult laryngoscopy, as did sex. In contrast, SMD and mouth opening did not reach significance, which might have been in part secondary to the arbitrarily chosen categories. Our results regarding the classical predictors do not contradict the published literature. Many studies have demonstrated the varying impact of these physical variables. Kheterpal et al.13 demonstrated that age over 46 years, BMI more than 30 kg m−2, male sex, Mallampati classification 3 or 4, neck mass or radiation, limited TMD, sleep apnoea, presence of teeth, beard, thick neck, limited cervical spine mobility and limited jaw protrusion were independent predictors of difficult mask ventilation and difficult laryngoscopy. These results emphasise the fact that there is no single predictor of difficult laryngoscopy. In the same way, several authors argued that Mallampati classification alone is not a reliable predictor of difficult laryngoscopy.29,30 The same is true for SMD9 and TMD.31
Our linear generalised model included sex, Mallampati score, retrognathia and position of the vocal cords with the best values of sensitivity (60%), specificity (96%) and positive (75%) and negative (92%) predictive value. Leaving vocal cord position out of the model led to a markedly reduced sensitivity with unchanged specificity. These results range within the better corridor of the values reported by other groups.11,16,30,32 Lundstrom et al.29 reported ranges from 0 to 100% for sensitivity and 44 to 100% for specificity in their meta-analysis. Only one publication33 reported 100% for sensitivity, specificity and positive predictive value when measuring complicated angles in lateral radiographs. In that study, Kamalipour et al.33 compared the Mallampati classification with several anatomical angles obtained from preoperative radiographs. In their prospective triple-blind study, 100 patients (aged 18 to 89 years), scheduled for elective surgery were randomly selected. Lateral neck radiographs were obtained from every patient prior to surgery. Several angles and variables on the radiograph were proposed to illustrate a relationship with easy or difficult intubation. Compared with the Mallampati classification, the analysis of the lateral radiograph proved to be a suitable method for predicting difficult intubation. Whether the knowledge gained about the patient's airway justifies the exposure to ionising radiation remains to be determined.
Chou and Wu19 evaluated lateral cervical radiographs and measured the vertical distance between the mandible and the hyoid bone. They reviewed the positions of the mandibular angle and hyoid bone in relation to the cervical vertebrae. They found that a short mandibular ramus or a relatively caudal position of the larynx can cause difficult laryngoscopy. Unfortunately, they studied only 11 patients in whom direct laryngoscopy was difficult.
A radiograph analysis of 13 patients who were rated as presenting difficult laryngoscopy was published by White and Kander.21 After examining the upper jaw, mandible and cervical spine, they found that the most important factor determining the ease of laryngoscopic examination was the posterior depth of the mandible (which means the distance between the alveolus immediately behind the third molar tooth and the lower border of the mandible). An increase in this measurement was thought to hinder displacement of the soft tissues by the laryngoscope blade.
Bellhouse and Doré18 recruited 33 patients, 19 of whom presented with difficult laryngoscopy, who were evaluated with lateral radiographs. The variables which, when combined, were most reliable in predicting likely difficult intubation were reduced atlanto-occipital extension, the proportion of mandibular space and the thickness of the tongue. Unfortunately, they did not comment on the level of the vocal cords in relation to the cervical vertebrae.
All of these publications used radiographs as the imaging technique, recruited a rather small number of patients and found different predictors of difficult laryngoscopy. Our study used MRI instead of a radiograph technique and focused on the position of the vocal cords in relation to the cervical vertebrae. This has been described in the literature only twice before, as single case reports. Schneider et al.34 presented a case of a 58-year-old man with a malignant non-Hodgkin's lymphoma of the base of the tongue. According to the authors, preoperative MRI proved to be extremely useful in the evaluation of the upper aerodigestive tract by providing sagittal sections of the hypopharynx and the inlet of the larynx. Thus, a potentially difficult intubation was confirmed in a patient in whom other investigations would not have produced comparable conclusions. In the same year, Hotchkiss et al.35 used MRI for retrospectively identifying an anatomical variant of the epiglottis, a large elongated epiglottis, as the cause of a previously encountered difficult laryngoscopy and tracheal intubation.
For the first time, we have undertaken a study using MRI to identify potential anatomical predictors for difficult laryngoscopy in a larger study population. We were able to demonstrate that difficult laryngoscopy coincides with significant cranial position of the vocal cords in relation to the cervical spine. In part, this agrees with the TMD which predicts a difficult laryngoscopy if it is less than 6.5 cm, implying a cranial position of the larynx and therefore of the vocal cords.8
A limitation of our study is that the position of some of our patients during MRI scan differed from the position during laryngoscopy. However, technical needs of the MRI technique make this a limitation that will be difficult to overcome in future studies. Our study population was relatively small and this may explain why some of the previously identified predictors of difficult laryngoscopy did not reach levels of statistical significance in our study. Another limitation might be the fact that our ratio of difficult to easy laryngoscopy does not reflect reality because we did not include consecutive patients but chose to use an enrichment study design.22,23 The rationale for this approach was that we had not been able to examine hundreds of patients to include a sufficient number of patients with difficult laryngoscopy. This does not influence the result of the study concerning the anatomical position of the vocal cords, but emphasises that we do not consider MRI to be a general screening instrument for possible difficult laryngoscopy.
In conclusion, we found that a cranial position of the vocal cords is related to difficult laryngoscopy. Furthermore, we believe that evaluation of MRI scans can have significant impact for predicting difficult airway. We are aware that imaging techniques such as MRI are time consuming and expensive and hence do not advocate their use for routine MRI imaging to assess the airway for possible predictors of difficult intubation. However, in the presence of existing preoperative images, we believe that it is absolutely necessary to include MRI scan findings in the preoperative evaluation of potential difficult laryngoscopy alongside the established bedside tests. For carefully selected patients, MRI imaging of the neck might provide useful information for providing safer anaesthesia.
Acknowledgements relating to this article
Assistance with the study: none.
Financial support and sponsorship: none.
Conflict of interest: none.
Presentation: this work was in part presented at the annual meeting of the German Society for Anaesthesia and Intensive Care in 2012 and at the Euroanaesthesia meeting in Barcelona in 2013.
1. Caplan RA, Posner KL, Ward RJ, Cheney FW. Adverse respiratory events in anesthesia: a closed claims analysis. Anesthesiology
2. Peterson GN, Domino KB, Caplan RA, et al. Management of the difficult airway: a closed claims analysis. Anesthesiology
3. Cook TM, Woodall N, Frerk C. Fourth National Audit Project. Major complications of airway management in the UK: results of the Fourth National Audit Project of the Royal College of Anaesthetists and the Difficult Airway Society. Part 1: anaesthesia. Br J Anaesth
4. Gannon K. Mortality associated with anaesthesia. A case review study. Anaesthesia
5. O'Sullivan E, Laffey J, Pandit JJ. A rude awakening after our fourth ‘NAP’: lessons for airway management. Anaesthesia
6. American Society of Anesthesiologists Task Force on Management of the Difficult Airway. 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
7. Mallampati SR, Gatt SP, Gugino LD, et al. A clinical sign to predict difficult tracheal intubation: a prospective study. Can Anaesth Soc J
8. Patil VU, Stehling LC, Zauder HL. Predicting the difficulty of intubation utilizing an intubation gauge. Anaesthesiol Rev
9. Al Ramadhani S, Mohamed LA, Rocke DA, Gouws E. Sternomental distance as the sole predictor of difficult laryngoscopy in obstetric anaesthesia. Br J Anaesth
10. Lavi R, Segal D, Ziser A. Predicting difficult airways using the intubation difficulty scale: a study comparing obese and nonobese patients. J Clin Anesth
11. Karkouti K, Rose DK, Wigglesworth D, Cohen MM. Predicting difficult intubation: a multivariable analysis. Can J Anaesth
12. Nichol HC, Zuck D. Difficult laryngoscopy – the ‘anterior’ larynx and the atlanto-occipital gap. Br J Anaesth
13. Kheterpal S, Healy D, Aziz MF, et al. Incidence, predictors, and outcome of difficult mask ventilation combined with difficult laryngoscopy: a report from the multicenter perioperative outcomes group. Anesthesiology
14. Kim WH, Ahn HJ, Lee CJ, et al. Neck circumference to thyromental distance ratio: a new predictor of difficult intubation in obese patients. Br J Anaesth
15. Knill RL. Difficult laryngoscopy made easy with a ‘BURP’. Can J Anaesth
16. Shah PJ, Dubey KP, Yadav JP. Predictive value of upper lip bite test and ratio of height to thyromental distance compared to other multivariate airway assessment tests for difficult laryngoscopy in apparently normal patients. J Anaesthesiol Clin Pharmacol
17. Schmitt HJ, Kirmse M, Radespiel-Troger M. Ratio of patient's height to thyromental distance improves prediction of difficult laryngoscopy. Anaesth Intensive Care
18. Bellhouse CP, Doré C. Criteria for estimating likelihood of difficulty of endotracheal intubation with the Macintosh laryngoscope. Anaesth Intensive Care
19. Chou HC, Wu TL. Mandibulohyoid distance in difficult laryngoscopy. Br J Anaesth
20. Samra SK, Schork MA, Guinto FC Jr. A study of radiologic imaging techniques and airway grading to predict a difficult endotracheal intubation. J Clin Anesth
21. White A, Kander PL. Anatomical factors in difficult direct laryngoscopy. Br J Anaesth
22. Pinsky PF, Gallas B. Enriched designs for assessing discriminatory performance – analysis of bias and variance. Stat Med
23. Simon N, Simon R. Adaptive enrichment designs for clinical trials. Biostatistics
24. Samsoon GL, Young JR. Difficult tracheal intubation: a retrospective study. Anaesthesia
25. Frerk CM. Predicting difficult intubation. Anaesthesia
26. King TA, Adams AP. Predicting difficult intubation: what factors influence the thyromental distance? Anaesthesia
27. Butler PJ, Dhara SS. Prediction of difficult laryngoscopy: an assessment of the thyromental distance and Mallampati predictive tests. Anaesth Intensive Care
28. Cormack RS, Lehane J. Difficult tracheal intubation in obstetrics. Anaesthesia
29. Lundstrom LH, Vester-Andersen M, Moller AM, et al. Poor prognostic value of the modified Mallampati score: a meta-analysis involving 177 088 patients. Br J Anaesth
30. Mashour GA, Kheterpal S, Vanaharam V, et al. The extended Mallampati score and a diagnosis of diabetes mellitus are predictors of difficult laryngoscopy in the morbidly obese. Anesth Analg
31. Qudaisat IY, Al-Ghanem SM. Short thyromental distance is a surrogate for inadequate head extension, rather than small submandibular space, when indicating possible difficult direct laryngoscopy. Eur J Anaesthesiol
32. Lundstrom LH, Moller AM, Rosenstock C, et al. High body mass index is a weak predictor for difficult and failed tracheal intubation: a cohort study of 91,332 consecutive patients scheduled for direct laryngoscopy registered in the Danish Anesthesia Database. Anesthesiology
33. Kamalipour H, Bagheri M, Kamali K, et al. Lateral neck radiography for prediction of difficult orotracheal intubation. Eur J Anaesthesiol
34. Schneider M, Probst R, Wey W. Magnetic-resonance imaging – a useful tool for airway assessment. Acta Anaesthesiol Scand
35. Hotchkiss RS, Hall JR, Braun IF, Schisler JQ. An abnormal epiglottis as a cause of difficult intubation – airway assessment using magnetic resonance imaging. Anesthesiology