Development and validation of a score for prediction of postoperative respiratory complications.

Abstract

Postoperative respiratory failure is associated with increased morbidity and mortality, as well as high costs of hospital care. Using electronic anesthesia records, billing data, and chart review, the authors developed and validated a score predicting reintubation in the hospital after primary extubation in the operating room, leading to unplanned mechanical ventilation within the first 3 postoperative days. Using multivariable logistic regression analysis, independent predictors were determined and a score postulated and validated. In the entire cohort (n = 33,769 surgical cases within 29,924 patients), reintubation occurred in 137 cases (0.41%). Of those, 16%, (n = 22) died subsequently, whereas the mortality in patients who were not reintubated was 0.26% (P < 0.0001). Independent predictors for reintubation were: American Society of Anesthesiologist Score 3 or more, emergency surgery, high-risk surgical service, history of congestive heart failure, and chronic pulmonary disease. A point value of 3, 3, 2, 2, and 1 were assigned to these predictors, respectively, based on their β coefficient in the predictive model. The score yielded a calculated area under the curve of 0.81, whereas each point increment was associated with a 1.7-fold (odds ratio: 1.72 [95% CI, 1.55-1.91]) increase in the odds for reintubation in the training dataset. Using the validation dataset (n = 16,884), the score had an area under the curve of 0.80 and similar estimated probabilities for reintubation. The authors developed and validated a score for the prediction of postoperative respiratory complications, a simple, 11-point score that can be used preoperatively by anesthesiologists to predict severe postoperative respiratory complications.