CORRESPONDENCE 505 PREDICTING DIFFICULT INTUBATION

Sir,—Oates and colleagues [1] reported a useful study, comparing two published methods of predicting difficult intubation, and have added to the considerable recent literature encouraging preoperative attempts to predict difficult in-tubation as part of a routine to increase patient safety. The authors acknowledged how important it is to perform the Mallampati test [2] correctly and, because observer 4 was not doing this during the study, subtracted the results for this observer. This gave a sensitivity to the Mallampati test slightly better than that for the Wilson test [3] and specificity and positive predictive value (PPV) slightly less. After the authors removed the data of observer 4 from their results, nine difficult laryngoscopies (grades 4 and 5) remained. This makes it hard for the reader to obtain sensitivities of "0.40 " and " 0.50 " in table III for the Wilson and Mallampat