The laryngeal mask airway (LMA)  maintains the airway under anesthesia using a spontaneous breathing technique without endotracheal intubation. However, the LMA does not always provide success in achieving an adequate airway. A 2%-6% incidence of an inadequate airway has been reported [2-4]. More than one attempt at insertion of the LMA is required in 10%-24% of patients [3-5]. Optimal seating of the LMA--with the laryngeal inlet totally within the rim of the LMA and the epiglottis and esophagus outside of the borders of the LMA--occurs in only 50%-66% of cases [6-8]. There is a large range of variation of seating of the LMA that still allows clinically adequate ventilation to be achieved. Yet the LMA is used in the management of a difficult airway [3,9]. This could be explained by the wide variation in final position of the LMA, which still allows ventilation. The Mallampati  classification is a means of predicting a difficult intubation. This study was undertaken to assess whether there was a relationship between the Mallampati classification and ease of seating of the LMA.
Hospital Ethics Committee and informed patient consent were obtained. One hundred adult patients ranging in age from 18 to 72 yr, ASA grades I or II, undergoing general anesthesia for a short elective surgical procedure using a spontaneous breathing technique were included. Patients with a history of reflux, hiatus hernia, obesity, or previous upper gastrointestinal tract surgery, patients taking H2 antagonists or antacids, and patients who had experienced a previous failed tracheal intubation were excluded. Mallampati grading was established at the preoperative visit by two post fellowship anesthesiologists. The second anesthesiologist was blinded to the Mallampati grade awarded by the first anesthesiologist. If both concurred on Mallampati grading, the patient was included. The establishment of a Mallampati grade was performed as described by Mallampati in his original paper  by noting the visibility of pharyngeal structures (faucial pillars, soft palate, base of uvula) with the patient sitting upright, mouth wide open, tongue protruding maximally, and not phonating. Each patient was directed to perform the maneuver twice to minimize the chance for error. Patients were divided into three grades according to the Mallampati classification: grade 1: faucial pillars, soft palate, and uvula visualized; grade 2: faucial pillars and soft palate visualized, but uvula masked by the base of the tongue; grade 3: only the soft palate could be visualized.
Monitoring included electrocardiography, pulse oximetry, noninvasive blood pressure, and capnography. After the patient breathed oxygen, alfentanil 5 micro gram/kg followed by propofol 2 mg/kg was given and an appropriately sized LMA was inserted when an adequate depth of anesthesia had been achieved by clinical estimation. A size 3 LMA was chosen for patients who weighed between 30 and 70 kg and 15 mL of air was injected, while a size 4 LMA was selected for patients weighing more than 70 kg and 25 mL of air was injected. All insertions of the LMA were performed by a single operator who was blinded to the Mallampati grading. The insertion was performed in the standard way as described by Brain . The head and neck were positioned as for laryngoscopy. Assisted ventilation was performed using 33% oxygen, 66% nitrous oxide, and 1% halothane (end-tidal), and a Portex bronchoscopic adaptor (Portex, London, UK) was applied to the end of the LMA. A fiberoptic bronchoscope (model Olympus LF1) was then passed down the LMA and through the grid at the laryngeal portion of the LMA. It was rotated through 360 degrees and each view graded according to an adapted classification that was previously described by Mizushima et al.  Table 1. The bronchoscopy was carried out by the same anesthesiologist who had inserted the LMA. If the first attempt at insertion was deemed unsatisfactory, only two more attempts were performed. An attempt at LMA insertion was said to have failed if there was clinically unacceptable chest expansion or an unacceptably large gas leak around the LMA during positive pressure ventilation via the reservoir bag beginning in the assisted ventilation phase of the anesthetic. However, a small leak with good chest expansion was deemed acceptable for spontaneous respiration. These decisions were made by the anesthesiologist who inserted the LMA. Between insertion attempts, the patient's lungs were ventilated using a face mask and Guedel airway technique. The LMA was deflated before removal and reinflated after the next insertion attempt. The volume of air injected into the LMA was changed after each failed attempt but never exceeded that described in the instruction manual. The size of the LMA was not changed. If the third attempt at insertion failed, the LMA was abandoned and a different airway technique used.
The number of attempts at LMA insertion were recorded. An analysis of the relationship between Mallampati grade and the number of attempts at LMA placement was prepared using chi squared analysis with Yates continuity correction.
There was no patient morbidity as a result of this study. Adequate ventilation was possible in 98 patients, but in two patients the LMA was abandoned after the third attempt at insertion. Both of these patients were Mallampati grade 3. A face mask and Guedel airway technique were used for the remainder of the anesthetic in both cases. Table 1 illustrates the relationship between Mallampati grading, with a breakdown of the number of patients in each Mallampati grade, and the view at fiberoptic bronchoscopy. There was more laryngeal inlet occlusion with worsening Mallampati grade resulting in P = 0.0001 by chi squared analysis. Seating of the LMA was suboptimal in 30 patients, and in seven cases there was no view of the vocal cords at all. These seven patients were all Mallampati grade 3.
(Table 2) illustrates the number of patients in each Mallampati grade referenced to the number of attempts required for successful insertion of the LMA. chi squared analysis was performed, and this demonstrated a strong association between Mallampati grading and ease of insertion of the LMA (P = 0.0001). Increasing occlusion of the laryngeal inlet and increasing difficulty of LMA insertion occurred with Mallampati grades 2 and 3.
Unexpected difficulties in endotracheal intubation are an etiologic factor in patient morbidity and mortality. The LMA may be a way to overcome this problem. Many reports in the literature testify to the use of the LMA in the cannot ventilate, cannot intubate the trachea scenario [9,13-17]. It may be used as a conduit to introduce a gum elastic bougie into the trachea, followed by removal of the LMA and intubation over the gum elastic bougie , or as an aid to both blind endotracheal intubation  and fiberoptic intubation .
Current American Society of Anesthesiologists practice guidelines  suggest that the LMA has a role in the management of the difficult airway. On the other hand, there have been reports of difficulty in maintaining the airway with the LMA in a significant percentage of patients [2-4] and suboptimal seating of the LMA occurs in 30%-66% of cases [6-8], suggesting that its role may be limited.
The actual position of the LMA is variable by radiologic assessment [7,8]. Blind tracheal intubation via the LMA has a failure rate on the first attempt of 26% with a 10% overall failure rate, while placement of an intubating stylet via the LMA has an 18% failure rate [18,22]. Other authors allege that LMA insertion is not compromised in patients with a difficult intubation [23,24]; however, in neither study was reference made to standardization of the Mallampati grading. When using solely the Mallampati classification to predict a difficult intubation, we found that there was a strong association between predicted difficult intubation and ease of seating of these LMAs. These anatomic factors may also be the cause of the variation in position of the LMA previously described.
The Mallampati classification and other similar classifications are based on the concept that the size of the base of the tongue is an important factor in determining the degree of difficulty of laryngoscopy. The Mallampati classification was modified by Samsoon and Young  to incorporate a fourth class when only the tongue is visible at mouth opening. If the base of the tongue is large and occludes the uvula, palate, and faucial pillars, laryngoscopy will probably be difficult. If it is difficult to see these structures, then it follows that blindly passing a LMA over the base of the tongue to reach the larynx will also be difficult. The LMA may glide past the larynx because of the large tongue or the large tongue may make the larynx sit relatively more anteriorly, again causing the LMA to glide past the optimal position into the pharynx. Cormack and Lehane  cited a large or poorly mobile tongue and an anteriorly cited larynx as possible causes of difficult laryngoscopy. This would agree with Mallampati's and our findings and explains why LMA seating is poor with increasing Mallampati grading. A Mallampati grade 3 should indicate to the anesthesiologist that correct seating of the LMA may not always be possible. The use of the LMA in patients who are classed as Mallampati grade 3 should be assessed critically.
This study is limited by the small number of Mallampati grade 3 patients (n = 10). However, in seven patients there was no view of the vocal cords at fiberoptic bronchoscopy and in two patients there was failure to achieve an adequate airway with the LMA. Although the precise risk cannot be quantified based on this data, it does suggest that a substantial percentage of Mallampati grade 3 patients will encounter difficulty with placement of the LMA.
Experience with LMA use remains limited in patients in whom difficulties are anticipated in securing the airway. Our data suggest that when the LMA is used electively as the airway of choice, there is a significant incidence of suboptimal placement.
Caution should be exercised in selection of the LMA as a means of securing the airway in patients who are Mallampati grade 3.
We gratefully acknowledge the contribution made to the results section of this paper by Dr. F. J. Boylan, MB, FFARCSI, Consultant Anesthesiologist.
1. Brain AIJ. The laryngeal mask: a new concept in airway management. Br J Anaesth 1983;55:801-5.
2. Smith I, White PF. Use of the LMA as an alternative to face mask during outpatient arthroscopy. Anesthesiology 1992;77:850-5.
3. Maltby JR, Loken RG, Watson NC. The LMA: clinical appraisal in 250 patients. Can J Anesth 1990;37:509-13.
4. Broderick PM, Webster NR, Nunn JF. The LMA: a study of 100 patients during spontaneous breathing. Anaesthesia 1989;44:238-41.
5. Sorma VJ. The use of the LMA in spontaneously breathing patients. Acta Anesthesiol Scand 1990;34:669-72.
6. Fuller B, Pothmann W. The LMA: fiberoptic detection of positioning and measurement of gas leakage. Anesth Analg 1992;74(Suppl):S101.
7. Nandi PR, Nunn JF, Charleswith CH, Taylor SJ. Radiological study of the LMA. Eur J Anesthesiol 1991;Suppl 4:33-9.
8. Denman W, Goudsouzian NG, Shorten G. The position of the LMA as determined by magnetic resonance imaging [abstract]. Anesthesiology 1991;75:1045.
9. Brain AIJ. Three cases of difficult intubation overcome by the LMA. Anaesthesia 1985;40:353-5.
10. Mallampati SR, Gatt SP, Laverne D, Gugino PD. A clinical sign to predict difficult tracheal intubation: a prospective study. Can Anaesth Soc J 1985;32:429-34.
11. Brain AIJ. Proper technique for insertion of LMA. Anesthesiology 1990;73:1053.
12. Mizushima A, Wardell GJ, Simpson JL. The LMA in infants. Anaesthesia 1992;47:849-51.
13. Brain AIJ. The LMA: a possible new solution to airway problems in the emergency situation. Arch Emerg Med 1984;1:229-32.
14. Calder I, Ordman AJ, Jackowski A, Crocheral HA. The brain LMA: an alternative to emergency tracheal intubation. Anaesthesia 1990;45:137-9.
15. Baraka A. LMA in the cannot intubate, cannot ventilate, situation. Anesthesiology 1993;79:1151.
16. Bennmof JC. Management of the difficult airway: with special emphasis on the awake tracheal intubation. Anesthesiology 1991;75:1087-110.
17. The confidential enquiry into maternal deaths in the United Kingdom 1985-1987. London: HMSO, 1991.
18. Allison A, McCrory J. Tracheal placement of a gum elastic bougie using the laryngeal mask airway [letter]. Anaesthesia 1990;45:419-20.
19. Brain AIJ. The development of the LMA: a brief history of the invention, early clinical studies and experimental work from which the LMA evolved. Eur J Anesthesiol 1991;Suppl 4:5-17.
20. Beumof JL. Use of LMA to facilitate fiberscope aided tracheal intubation [letter]. Anesth Analg 1992;74:313-9.
21. American Society of Anesthesiologists: practice guidelines for management of the difficult airway: a report by the American Society of Anesthesiologists Task Force into Management of the Difficult Airway. Anesthesiology 1993;78:597-602.
22. Heath ML. Endotracheal intubation through the LMA: helpful when laryngoscopy is difficult or dangerous. Eur J Anesthesiol 1991;Suppl 4:41-5.
23. Brimacombe J, Berry A. Mallampati classification and LMA insertion. Anaesthesia 1993;48:347.
24. Mahion P, Narchi P, Reyrac P. Is the LMA easy to use in the care of difficult intubation? [abstract]. Anesthesiology 1992;77:A1228.
25. Samsoon GLT, Young JRB. Difficult tracheal intubation: a retrospective study. Anaesthesia 1987;47:487-90.
26. Cormack RS, Lehane J. Difficult tracheal intubation in obstetrics. Anaesthesia 1984;39:1105-11.