To the Editor:—
I would like to congratulate Hubert Schmitt et al. 1on their study of difficult intubation in acromegalic patients. Their experience is unrivaled, and it is reassuring to know that serious difficulty with the airway is not common.
I would like to make two pleas to future investigators. The first is for the abandonment of the four-grade Mallampati classification. A fourth grade was added by Samsoon and Young 2in their retrospective study.
I have been able to discuss the reason for this change with Dr. Young, the senior author. He confirmed my suspicion that they made an entirely reasonable assumption that the four grades of glottic visibility described by Cormack and Lehane 3would be predicted by four grades of oropharyngeal visibility. We know now that this is not the case and should cease pretending that we can describe oropharyngeal appearances with such precision.
My second plea is for investigators to report likelihood ratios because they provide an understandable estimate of risk when one is confronted by a patient with a test result. 4The likelihood ratio is calculated by dividing the sensitivity by 1-specificity, and the result is the number of times more likely it is that a positive result will be seen in someone with the condition being sought than in one without. The likelihood ratio for a positive Mallampati test in the series of Schmitt et al. 1thus would be 44/24 = 1.8. I find that this gives me a better understanding of the poor performance of the Mallampati. With the use of nomograms, 4a knowledge of the likelihood ratio also allows estimation of the positive predictive value in populations with a different risk because of a higher or lower incidence of difficulty. In this case, I estimate that had Schmitt et al. 1chosen to define difficult laryngoscopy as grade 3 after the application of external laryngeal pressure and use of an appropriate laryngoscope blade (three patients), then the positive predictive value of the Mallampati would be 3%.
The only report in which the Mallampati has performed well was of a series of patients with cervical spine disease (likelihood ratio = 14). 5The failure of the Mallampati in the study of Schmitt et al. 1tends to support our contention that the Mallampati owes much of its predictive power to the effect of cranio-cervical rigidity on mouth opening ability, malalignment of oropharyngeal axes, and oropharyngeal visibility. The Mallampati can be regarded as essentially a test of cranio-cervical extension.