Predictive models for difficult laryngoscopy and intubation. A clinical, radiologic and three-dimensional computer imaging study

Purpose: To identify the variables most useful in predicting difficult laryngoscopy and intubation from various clinical, skeletal (lateral x-rays) and soft tissue (three-dimensional computed tomography imaging) measurements. Methods: Twenty-four adult patients in whom an unanticipated difficult tracheal intubation was identified according to established criteria were evaluated. Further, a control group of 32 patients in whom tracheal intubation was easily accomplished was studied. We applied multivariate discriminant analysis to clinical and radiological data of all patients to select those variables most useful in predicting difficult laryngoscopy and intubation, The receiver operating characteristic (ROC) curve was used to describe the discrimination abilities and to explore the trade-offs between sensitivity and specificity of the model. Results: With the clinical data alone, discriminant analysis identified four risk factors that correlated with the prediction of difficult laryngoscopy and intubation: thyrosternal distance, thyromental distance, neck circumference and Mallampati classification. With both clinical and radiological data, discriminant analysis identified five risk factors: thyrosternal distance, thyromental distance, Mallampati classification, depth of spine C2 and angle A (the most antero-inferior point of the upper central incisor tooth). The positive predictive value of this combined (clinical and radiological) model was greater than that of the clinical model alone (95.8% vs 87.5%, respectively). The areas under the ROC curves, that measure the probability of the correct prediction of the clinical and the combined models, were found to be 0.933 and 0.973, respectively. Conclusions: These models can be used for predicting difficult laryngoscopy and intubation in clinical practice.