General Articles Craniocervical Extension Improves the Specificity and Predictive Value of the Mallampati Airway Evaluation George A. Mashour, MD, PhD*† Warren S. Sandberg, MD, PhD*† BACKGROUND: The modified Mallampati (MMP) classification is a standard airway examination that assesses mouth opening and structures within the oral cavity. Recent data suggest that maximal mouth opening (as measured by interdental distance) is possible only with extension of the craniocervical junction. Because the MMP examination is performed with the head in a neutral position, the airway may appear worse because of submaximal interdental distance. We hypothesized that adding craniocervical extension to the MMP would allow for greater mouth opening, lower scores, and less false positives than the tradi- tional MMP examination. METHODS: Multiple clinicians with at least 1 yr of airway experience evaluated adult airways (n = 60) with the MMP examination (with head in neutral position). The same examination was then repeated with the addition of craniocervical extension (Extended Mallampati Score, EMS). RESULTS: On average, craniocervical extension decreased the MMP class (P  0.002). The EMS improved specificity from 70% to 80% and positive predictive value from 24% to 31% when compared with the traditional MMP. The sensitivity (83%) was the same for MMP and EMS. CONCLUSIONS: Craniocervical extension improves the specificity and positive pre- dictive value of the MMP airway evaluation while retaining sensitivity of the traditional MMP examination. The introduction of the EMS into clinical practice should be considered. (Anesth Analg 2006;103:1256 –9) The Mallampati evaluation is a standard method of assessing the airway for potentially difficult intuba- tion. First formulated as a hypothesis (1), it was later demonstrated in a prospective study to be predictive of difficult laryngoscopy (2). Samsoon and Young (3) modified the original Mallampati classification system by including a fourth class of airway, in which the soft palate could not be visualized. The modified Mallam- pati (MMP) Class 4 airway was shown to be associated with difficult tracheal intubation in a retrospective study of the obstetric population. The proposed ana- tomic basis for this examination is the relationship of the tongue to the oral cavity: when the base of the tongue is disproportionately large, the glottis may be obscured during laryngoscopy. The MMP examination is performed with the pa- tient sitting upright, the head in neutral position, and the tongue maximally protruded (3). Lewis et al. (4) demonstrated that the predictive value of the MMP is dependent on the position of the cervical spine and recommended that the MMP be performed with ex- tension of the craniocervical junction. It has been demonstrated more recently that mouth opening is also dependent on cervical spine positioning: to achieve maximal interdental distance, the craniocervi- cal junction must be extended (5). When patients were kept in the neutral position and prevented from full extension, mouth opening was limited by an average of 12 mm. Similarly, when craniocervical extension is limited with soft cervical collars, mouth opening is also limited (6). These biomechanical studies suggest that when the head is in the neutral position with respect to the cervical spine, mouth opening is limited and submaximal. Because the MMP airway evaluation is performed with the head in the neutral position, the degree of mouth opening seen by the examiner is less than what could be achieved with craniocervical extension. This creates the possibility of the airway appearing worse than it actually is, as the relationship of the tongue to the apparent size of the oral cavity is also dependent on the degree of mouth opening. The consequent result of a false positive examination (where a “posi- tive” is predictive of difficult laryngoscopy) may decrease the specificity and predictive value of the MMP. From the *Department of Anesthesia and Critical Care, Massa- chusetts General Hospital; and †Department of Anaesthesia, Har- vard Medical School, Boston, Massachusetts. Accepted for publication June 30, 2006. Supported by departmental and institutional funds. Address correspondence and reprint requests to George A. Mashour, MD, PhD, 55 Fruit St., Clinics 309, Boston, MA 02114. Address e-mail to gmashour@partners.org. Copyright © 2006 International Anesthesia Research Society DOI: 10.1213/01.ane.0000237402.03329.3b Vol. 103, No. 5, November 2006 1256