Predictors of Long-Term Mortality in Patients with Ventilator-Associated Pneumonia

Purpose To determine the long-term outcome of ventilator-associated pneumonia (VAP) and identify factors associated with increased mortality. Methods We retrospectively studied 671 patients with VAP admitted to an intensive care unit between 1994 and 2000. We determined long-term and out-of-hospital mortality for these patients. Results The in-hospital mortality was 42.3%; 19.8% of patients had concomitant bacteremia, the mortality was 59.7% versus 38.0% for those without bacteremia (P<.001). The factors associated with increased hospital mortality by univariable analysis were: diagnosis on admission, the need of vasopressors during the stay in the intensive care unit, not undergoing a tracheostomy, the absence of fever, the presence of concomitant bacteremia, and renal failure or the need for dialysis. Patients transferred from an outside hospital and patients with normal serum bicarbonate, serum total bilirubin <2 mg/dL, and platelets >120 × 4> 103/μL had a lower in-hospital mortality. All of these factors except bilirubin level, platelet count, transfer from outside hospital, and serum bicarbonate remained significant on multivariable analysis. The estimated mortality at 1, 3 and 5 years is 25.9% (95% confidence interval [CI], 20.2-30.1%), 33.6% (95% CI, 27.4-39.2%) and 44.7% (95% CI, 38.1-50.6%), respectively. Conclusions VAP is associated with a high rate of hospital and long-term mortality. The presence of bacteremia is associated with a high mortality. The 5-year estimated mortality of the survivors is less than 50%.