Ventilator-associated pneumonia after heart surgery: a prospective analysis and the value of surveillance.

OBJECTIVE

To determine the frequency, etiology, and risk factors of ventilator-associated pneumonia (VAP) and purulent tracheobronchitis (TBX) in patients who have undergone heart surgery. To study the predictive role of systematic surveillance cultures.

DESIGN

Prospective study.

SETTING

Heart surgery intensive care unit.

PATIENTS: Intubated heart surgical patients.

INTERVENTIONS: Systematic tracheal aspirate and protected brush catheter cultures of all intubated patients.

MEASUREMENTS AND MAIN

RESULTS

Studied were the frequency of lower respiratory tract infection in ventilated patients and the role of surveillance cultures. The frequency of VAP was 7.87% (34.5 per 1,000 days of mechanical ventilation), and the criteria for purulent tracheobronchitis was fulfilled by 8.15% of patients (31.13 per 1,000 days of mechanical ventilation). After multivariate analysis, the variables independently associated with the development of respiratory tract infection were central nervous system disorder (relative risk [RR] = 4.7), ulcer disease (RR = 3.6), New York Heart Association score >/=3 (RR = 4), need for mechanical circulatory support (RR = 6.8), duration of mechanical ventilation >96 hrs (RR = 12.3), and reintubation (RR = 63.7). Mortality in our study was as follows: VAP patients, 57.1%; purulent tracheobronchitis patients, 20.7%; colonized patients, 11.5%; and noncolonized patients, 1.6%. Regular surveillance cultures were taken from all ventilated patients to assess the anticipative value of the cultures in predicting respiratory tract infection. A total of 1,626 respiratory surveillance samples were obtained. Surveillance cultures effectively predicted only one episode of VAP and one of tracheobronchitis.

CONCLUSIONS

Patients undergoing heart surgery have a high frequency of VAP. VAP is associated with a poor prognosis. In this study, surveillance cultures failed as an anticipative diagnostic method.