Influencing factors and prediction model construction of airway obstruction after cervical spine surgery

Objective To investigate the risk factors for airway obstruction after cervical spine surgery and to construct and validate a predictive risk model.

Methods A retrospective analysis of data from 250 patients who underwent cervical spinesurgery was performed. Patients were divided into modeling set (n=175) and validation set (n=75) in a ratio of 7 to 3. Patients in the modeling set were categorized into respiratory distress group and normal breathing group based on whether they experienced postoperative airway obstruction. General patient data were collected. Multivariate Logistic regression analysis was used to identify independent influencing factors for postoperative airway obstruction. Based on the identified factors, a risk prediction model was constructed. The R software was utilized to draw a nomogram and calibration curve for the risk prediction model. Receiver operating characteristic (ROC) curves were employed to assess the predictive value of the risk prediction model. The goodness-of-fit was evaluated using the Hosmer-Lemeshow test.

Results Among the patients in the modeling set, 28 cases developed postoperative airway obstruction, with an incidence rate of 16.00% (28/175). Univariate analysis results indicated that there were statistically significant differences in age, operation time, postoperative edema, surgical segments, preoperative tracheal shifting exercise and the number of involved surgical segments between the respiratory distress group and the normal breathing group (P < 0.05). Age, preoperative tracheal shifting exercise, postoperative edema and operation time were identified as independent influencing factors for postoperative airway obstruction (P < 0.05). Internal validation showed that the area under the ROC curve of this model was 0.888, with sensitivity of 75.1% and specificity of 92.1%. Hosmer-Lemeshow test showed that the prediction model had good fit and high prediction value (χ²=3.435, P=0.904). The area under the external validation ROC curve was 0.735, with sensitivity of 83.3% and specificity of 65.1%.

Conclusion Age, preoperative tracheal shifting exercise, postoperative edema and operation time are independent influencing factors for patients with airway obstruction after cervical spine surgery. The risk prediction model constructed based on these factors has good predictive performance and clinical application value.