Preoperative Airway ultrasonography for predicting difficult intubation: comparative performance of skin-to-epiglottis distance and bedside tests – a prospective observational study

Unanticipated difficult airway management remains a major contributor to perioperative morbidity and mortality. Conventional bedside tests have shown limited and variable predictive accuracy, and no single test reliably identifies patients at risk. Point-of-care airway ultrasonography has emerged as a promising adjunct, but the diagnostic performance of specific sonographic parameters — particularly the skin-to-epiglottis distance (SED) — requires further evaluation. In this prospective, single-center observational study, 400 adults (18–70 years; ASA I–III) scheduled for elective surgery under general anesthesia with orotracheal intubation were enrolled. Before induction, the attending anesthesiologist performed a standardized bedside assessment (modified Mallampati classification, upper lip bite test, and thyromental, sternomental, interincisor, and mentohyoid distances). A second anesthesiologist experienced in airway ultrasonography, blinded to the bedside findings and intraoperative outcomes, obtained sonographic measurements including SED, skin-to-thyroid cartilage and thyroid cartilage-to-epiglottis distances, pre-epiglottic space depth, tongue thickness, and hyomental distance in neutral and extended positions. The prespecified primary outcome was difficult intubation (two or more intubation attempts); difficult mask ventilation (DMV; requiring two-handed two-provider technique or airway adjuncts) was the prespecified secondary outcome. Between-group comparisons and receiver operating characteristic (ROC) curve analysis were performed, with optimal cutoffs identified by the Youden index. Difficult intubation occurred in 20 patients (5.0%) and DMV in 27 (6.8%). Among bedside tests, the upper lip bite test and modified Mallampati classification were significantly associated with difficult intubation. SED yielded the highest point-estimate discrimination (AUC 0.912; 95% CI 0.811–1.000; optimal cutoff 2.14 cm; 85% sensitivity, 91% specificity); given the limited number of events, these estimates are subject to imprecision and should be interpreted as exploratory. Predictors of DMV followed a distinct pattern, dominated by demographic, comorbid, and broader anatomical factors rather than by any single ultrasound parameter. Preoperative airway ultrasonography — and in particular SED — provided supplementary anatomical information beyond conventional bedside assessment for difficult intubation, whereas predictors of difficult mask ventilation followed a distinct pattern, supporting the view that these are related but non-identical airway phenotypes. Given the single-center design and the limited number of events, these findings require external validation in multicenter studies incorporating multiparametric ultrasound-and-clinical models and contemporary videolaryngoscopy-based airway management. The study was retrospectively registered at ClinicalTrials.gov (NCT07383610; registered 26 January 2026) after study initiation. This timing is reported transparently and does not affect data integrity or the prespecified analyses.