A Predictive Scoring Model for Postoperative Tracheostomy in Patients Who Underwent Cardiac Surgery

BACKGROUND: A subset of patients require a tracheostomy as respiratory support in a severe state after cardiac surgery. There are limited data to assess the predictors for requiring postoperative tracheostomy (POT) in cardiac surgical patients. METHODS: The records of adult patients who underwent cardiac surgery from 2016 to 2019 at our institution were reviewed. Univariable analysis was used to assess the possible risk factors for POT. Then multivariable logistic regression analysis was performed to identify independent predictors. A predictive scoring model was established with predictor assigned scores derived from each regression coefficient divided by the smallest one. The area under the receiver operating characteristic curve and the Hosmer-Lemeshow goodness-of-fit test were used to evaluate the discrimination and calibration of the risk score, respectively. RESULTS: A total of 5,323 cardiac surgical patients were included, with 128 (2.4%) patients treated with tracheostomy after cardiac surgery. Patients with POT had a higher frequency of readmission to the intensive care unit (ICU), longer stay, and higher mortality (p < 0.001). Mixed valve surgery and coronary artery bypass grafting (CABG), aortic surgery, renal insufficiency, diabetes mellitus, chronic obstructive pulmonary disease (COPD), pulmonary edema, age >60 years, and emergent surgery were independent predictors. A 9-point risk score was generated based on the multivariable model, showing good discrimination [the concordance index (c-index): 0.837] and was well-calibrated. CONCLUSIONS: We established and verified a predictive scoring model for POT in patients who underwent cardiac surgery. The scoring model was conducive to risk stratification and may provide meaningful information for clinical decision-making.