Variables predicting reintubation after cardiac surgical procedures

Background. This study sought to determine patient characteristics, processes of care, and intermediate outcomes as predictors of reintubation after cardiac surgical procedures. Methods. We performed a retrospective case-control study that included all patients undergoing cardiac surgical intervention who required reintubation and an equal number of control patients not requiring reintubation. Putative risk factors were analyzed univariately by chi(2), Fisher exact, Student's t, or Mann-Whitney tests. A logistic regression model was developed using data from patients requiring reintubation for cardiorespiratory reasons. Results. Of the 1,000 consecutive patients reviewed, 41 (4.1%) required reintubation (30 [3%] for cardiorespiratory reasons and 11 [1.1] for unplanned operations). Univariate predictors of reintubation (p < 0.05) were older age, chronic obstructive pulmonary disease, New York Heart Association functional class IV, preoperative renal failure, lower arterial oxygen tension, insertion of intraaortic balloon pump, longer time in the operating room, longer duration of cardiopulmonary bypass times, positive fluid balance, postoperative renal failure, and worse pulmonary mechanics. Patients requiring reintubation also required a longer initial period of mechanical ventilation (median, 16.3 versus 6.0 hours; p < 0.05). Excellent prediction was found with a model consisting of four variables: operating room time, respiratory rate, vital capacity, and chronic obstructive pulmonary disease. Conclusions. Patients who required reintubation were sicker and had worse respiratory function and more comorbidity. Prompt extubation did not contribute to reintubation. Patients identified as having a high risk for reintubation should be followed up closely, and interventions should be directed to treating the problems leading to reintubation. (Ann Thorac Surg 1999;67:661-5) (C) 1999 by The Society of Thoracic Surgeons.