Validation of a Preoperative Risk Model for Pneumonia in Patients undergoing CABG Surgery

Background and Purpose: Pulmonary problems are among the most frequently reported complications of Coronary Artery Bypass Graft (CABG) surgery. However, the risk of postoperative pulmonary complications (PPCs) is not the same for all patients. The aim of this study was to validate and simplify a previously developed preoperative risk model for postoperative pulmonary complications, particular pneumonia, in patients undergoing CABG surgery. Subjects and Methods: Prospectively collected data for 421 adult patients who had undergone elective CABG surgery, in a university medical center in the Netherlands, were used to validate the preoperative risk model. The model assesses six risk factors (age ≥ 70 years, productive cough, diabetes mellitus, a positive smoking history, predicted inspiratory vital capacity < 75%, and predicted maximal expiratory pressure < 75%), with scores ranging from -4 (minimum risk) to 10 points (maximum risk). Patients with a score of -4 to -2 were classified as being at low pulmonary risk and patients with a score of -1 to 10 points were classified as being at high pulmonary risk. The accuracy of the model was tested by comparing the expected and observed incidence of pneumonia for each patient. The prospectively collected data were also used to develop a simple, accurate prognostic model for pneumonia. Results: Of the 421 patients, 227 (53.9%) were classified as being at high pulmonary risk, 24 (10.6%) of whom developed pneumonia, whereas only 4 of the 194 (2.1%) patients classified as being at low pulmonary risk developed pneumonia (OR= 5.62; 95% CI, 1.91 to 16.48). The area under curve (AUC) of the receiver-operating characteristics (ROC) curve was 0.84 and the model had a negative predictive value of 97.9%. Logistic regression analysis showed that the six-factor preoperative risk model could be simplified to a three-factor risk model, containing preoperative productive coughing, diabetes mellitus, and decreased lung function, to distinguish between patients with a high risk of developing pneumonia and those with a low risk (AUC: 0.79; negative 104 Model validation predictive value: 99.2%). Discussion and Conclusion: The easy-to-use six-factor risk model is accurate in identifying patients at low risk of developing PPCs. Three easy-to-measure preoperative variables (productive cough, diabetes mellitus, and decreased lung function) can be used to help clinicians tailor preoperative and postoperative interventions according to a patient’s risk. 105