A Comparison of the Mallampati evaluation in neutral or extended cervical spine positions: a retrospective observational study of >80 000 patients

Abstract

Background: The Mallampati examination is a standard component of an airway risk assessment. Existing evidence suggests that cervical spine extension improves the predictive power of the Mallampati examination for detecting difficult laryngoscopy and tracheal intubation, but a comparative effectiveness study has not been conducted.

Methods: The extended Mallampati examination (EMS) was introduced to the standard preoperative airway assessment, in addition to the standard Modified Mallampati examination (MMP). This study compared the accuracy of both Mallampati examinations on the prediction of difficult laryngoscopy, tracheal intubation, and bag mask ventilation. Univariate and adjusted analyses were performed.

Results: 80 801 patients with recorded MMP and EMS, and subsequent glottic view obtained during direct laryngoscopy, were examined. There was increased specificity (88.7% cf. 81.9%) but reduced sensitivity (33.3% cf. 45.7%) in the detection of difficult direct laryngoscopy with use of the EMS. The area under the receiver operating characteristic curve of each test performed in combination with other airway predictors for the models predicting difficult laryngoscopy was 0.740 (95% CI 0.731-0.753) for MMP and 0.739 (95% CI 0.729-0.752) for EMS. The area under the receiver operating characteristic curve of each test, performed in combination with other airway predictors for the models predicting difficult intubation was 0.699 (95% CI 0.688-0.711) for MMP and 0.695 (95% CI 0.683-0.707) for EMS.

Conclusions: This retrospective observational study demonstrates that cervical extension improves the specificity but decreases sensitivity of Mallampati examination. The Mallampati evaluation should be performed with the cervical spine in the neutral position to maximize test sensitivity.

Keywords: glottis; laryngoscopy; physical examination; risk assessment; sensitivity and specificity.