Bedside tests for predicting difficult airways: an abridged Cochrane diagnostic test accuracy systematic review

Although bedside screening tests are routinely used to identify people at high risk of having a difficult airway, their clinical utility is unclear. We estimated the diagnostic accuracy of commonly used bedside examination tests for assessing the airway in adult patients without apparent anatomical abnormalities scheduled to undergo general anaesthesia. We searched for studies that reported our pre-specified bedside index screening tests against a reference standard, published in any language, from date of inception to 16 December 2016, in seven bibliographic databases. We included 133 studies (127 cohort type and 6 case-control) involving 844,206 participants. Overall, their methodological quality (according to QUADAS-2, a standard tool for assessing quality of diagnostic accuracy studies) was moderate to high. Our pre-specified tests were: the Mallampati test (6 studies); modified Mallampati test (105 studies); Wilson risk score (6 studies); thyromental distance (52 studies); sternomental distance (18 studies); mouth opening test (34 studies); and the upper lip bite test (30 studies). Difficult facemask ventilation, difficult laryngoscopy, difficult intubation and failed intubation were the reference standards in seven, 92, 50 and two studies, respectively. Across all reference standards, we found all index tests had relatively low sensitivities, with high variability, but specificities were consistently and markedly higher than sensitivities. For difficult laryngoscopy, the sensitivity and specificity (95%CI) of the upper lip bite test were 0.67 (0.45-0.83) and 0.92 (0.86-0.95), respectively; upper lip bite test sensitivity (95%CI) was significantly higher than that for the mouth opening test (0.22, 0.13-0.33; p < 0.001). For difficult tracheal intubation, the modified Mallampati test had a significantly higher sensitivity (95%CI) at 0.51 (0.40-0.61) compared with mouth opening (0.27, 0.16-0.41; p < 0.001) and thyromental distance (0.24, 0.12-0.43; p < 0.001). Although the upper lip bite test showed the most favourable diagnostic test accuracy properties, none of the common bedside screening tests is well suited for detecting unanticipated difficult airways, as many of them are missed.