To the Editor: I read with interest the article by Tse et al. [1] concerning the prediction of difficult endotracheal intubation in 471 adult patients presenting for routine elective surgery. Their findings prompt comments on several accounts. First, 62 patients were found to be difficult to intubate [laryngoscopy grade III or IV, Cormack and Lehane classification [2]], which is an incidence of 13%. This is an unusually high incidence of difficult intubation, using the definition described previously, in patients who had no malformations of the airway and no disease of the cervical spine; previous studies have shown this incidence to be around 1% [3-5]. Could this be accounted for by the relative inexperience of the anesthesiology residents performing the endotracheal intubation? Second, the authors found that oropharyngeal Class 3 done according to Mallampati criteria [6] and a thyromental distance <or=to7 cm, when used as predictors of difficult intubation either alone or in combination, had low sensitivity and specificity. These findings confirm those of previous studies [3,5,7] that these predictors are of limited use as clinical tests and contradict the results produced by Frerk [4]. Third, the authors found that use of a head extension angle <or=to80 degrees to predict difficult intubation had a very low sensitivity. Sternomental distance (the distance from the mentum to the upper border of the manubrium sterni with the head fully extended and the mouth closed) has been described by Savva [5] as an alternative method of assessing head extension. When used to predict difficult intubation, a sternomental distance <or=to12.5 cm was shown to have a sensitivity of approximately 80% [5,8]. I agree with the author's comment that designing a good predictive test for difficult intubation is problematic since there are many factors that can cause difficulty. It has been shown that mandibular protrusion Class C as described by Calder et al. [9] is always associated with difficult intubation [8,9]. It has also been shown that there is no correlation between maximum interincisor gap (IG) and difficulty with intubation [5] when the IG is >or=to2.0 cm. It has accordingly been suggested that an IG <2.0 cm should be used to predict difficult intubation. It appears that a sternomental distance <or=to12.5 cm, mandibular protrusion Class C, and IG <2.0 cm should be used as a composite examination having >80% sensitivity in the preoperative screening for difficult intubation. Dino Savva, MB, ChB, DA, FRCA M. Maroof, MD Department of Anesthesiology King Fahd National Guard Hospital Riyadh 11426, Saudi Arabia