Predictors of prolonged mechanical ventilation in a cohort of 3,269 CABG patients.

Background. Prolonged mechanical ventilation after heart surgery is associated with increased patient morbidity and mortality (4.9% vs 22-38%). A prospective observational cohort study was carried out to assess the predictors of prolonged mechanical ventilation and its impact on hospital survival in a cardiac surgical patient cohort admitted to our 8 bed postoperative ICU from January 1997 through June 2004. Methods. All of the patient perioperative and ICU variables were input into an electronic database. Patients were divided into: 1) an Early Extubation group, undergoing a successful extubation within 12 h and 2) a Delayed Extubation group, needing mechanical ventilation longer than 12 h. Results. A total of 3,269 patients undergoing a coronary artery bypass graft operation were admitted. A multivariate Logistic Regression model allowed us to identify: 1) redo surgery (OR = 3.090, 95% CI = 1.655-5.780); 2) cardiopulmonary bypass time longer than 91' (OR= 1.390, 95% CI = 1.013-1.908); 3) intraoperative transfusions of more than 4 units of red blood cells (OR = 3.144, 95% CI = 2.331-4.255) or fresh frozen plasma (OR = 2.976, 95% CI = 1.984-4.830); and 4) left ventricular ejection fraction < 30% (OR = 2.444, 95% CI 1.291-3.205) as independent predictors of prolonged mechanical ventilation. The Early Extubation group showed a significantly higher cumulative survival 180 days after the ICU admission (Log-Rank = 16.617, p=0.000). Conclusion. This audit allowed us to assess a predictive model identifying a priori coronary artery bypass graft patients that are more likely to undergo prolonged mechanical ventilation.