de Carvalho CC, et al. - Researchers compared voice sounds to modified Mallampati test for the prediction of a difficult airway. Participants were 453 patients scheduled for elective surgery under general anaesthesia with tracheal intubation. They recorded five phonemes and analyzed their formants. A modified Mallampati classification, first formant of /i/(iF1), and second formant of /i/(iF2) were included in the model that achieved better performance, compared to five regression models assessed, to predict difficult laryngoscopy. An area under curve of 0.918 was generated with a regression model that included both formants and Mallampati; this was higher as compared with formants alone (area under curve 0.761) and modified Mallampati alone (area under curve 0.874). According to the findings, voice presented a significant link with difficult laryngoscopy during general anaesthesia demonstrating a 76.1% chance of correctly categorizing a randomly selected patient.