Objective To investigate the risk factors of delayed extubation after non-cardiac surgery and establish a predictive model. Methods Retrospective analysis of 1 009 patients who underwent non-cardiac surgery and underwent postoperative anesthesia awakening in PACU from September to October 2020. The patients were divided into two groups according to whether extubation was delayed during anesthesia recovery: delayed extubation group (extubation time exceeded 1 hour) and non-delayed extubation group (extubation time did not exceed 1 hour). LASSO and multivariate logistic regression were used to establish a predictive model. Receiver operating characteristic (ROC) curve, area under the curve (AUC) and decision curve were analyzed to evaluate the predictive value of the predictive model for delayed extubation after non-cardiac surgery. Results There were 253 patients (25.1%) in the delayed extubation group. The Results of multivariate Logistic regression analysis showed that ACCI score (≥3 points), BMI (≤22.66 kg/m2), intraoperative use of rocuronium bromide, intraoperative blood transfusion, operative time (≥166 minutes), indwelling catheter, use of propofol in PACU, and use of vasoactive drugs in PACU were independent predictors of delayed postoperative extubation. The AUC of predictive model was 0.730 (95% CI 0.695-0.765, P < 0.001), and sensitivity was 814%, and specificity was 55.4%. The analysis of decision curve showed that this prediction model has important clinical value for the prediction of delayed extubation. Conclusion ACCI score (≥3 points), BMI (≤22.66 kg/m2), intraoperative use of rocuronium bromide, intraoperative blood transfusion, operation time (≥166 minutes), indwelling catheter, use of propofol in PACU, and use of vasoactive drugs in PACU have significant predictive value for delayed extubation after general anesthesia. |