Objective: To determine the factors associated with ventilator-associated pneumonia recurrence in patients alive after 8 days of treatment for a first episode.
Design: A 16-month, prospective, observational cohort study of patients diagnosed with a first ventilator-associated pneumonia episode. Predictors of recurrence were assessed by logistic regression analysis.
Setting: Two intensive care units in a university hospital.
Patients: Bronchoscopy was performed in 124 patients with clinically or radiologically suspected ventilator-associated pneumonia. Ventilator-associated pneumonia was confirmed by the presence of at least two of the following criteria: >/=2% of cells with intracellular bacteria found on direct examination of bronchoalveolar lavage fluid, protected specimen brush sample culture >/=103 colony-forming units/mL, or bronchoalveolar lavage culture >/=104 colony-forming units/mL. Ventilator-associated pneumonia recurrence was confirmed using the same microbiological criteria. Antibiotic treatment for ventilator-associated pneumonia lasted 14 days.
Measurements and main results: Clinical, radiologic, and biological data at intensive care unit admission, on the day of bronchoscopy (D1) and on D8, and outcome variables were prospectively recorded. Ventilator-associated pneumonia recurred in 28 patients (all of them still on mechanical ventilation on D8), 21 +/- 9 days after the first episode (82% after D14). Factors significantly associated with recurrence were: acute respiratory failure as initial reason for mechanical ventilation, D1 radiologic score >7, D8 radiologic score >8, adult respiratory distress syndrome on D8, mechanical ventilation persistence on D8, D8 temperature >38 degrees C, and D8 temperature >D1 temperature, but not disease-severity scores at inclusion and D8, or first-episode pathogen(s). Multivariate analysis identified D1 radiologic score >7 (odds ratio = 3.9; 95% confidence interval, 1.3-11.6), D8 temperature >38 degrees C (odds ratio = 4.4; 95% confidence interval, 1.4-13.4), and adult respiratory distress syndrome on D8 (odds ratio = 14.6; 95% confidence interval, 1.5-143.5) as predictors of recurrence.
Conclusions: Factors of ventilator-associated pneumonia recurrence evaluated on D8 of a 14-day course of antibiotics are linked to the severity of lung injury and persistence of fever, but not to first-episode pathogen(s).