Relationship of Microbiologic Diagnostic Criteria to Morbidity and Mortality in Patients With Ventilator-Associated Pneumonia

Abstract

To evaluate whether the mortality and the morbidity of ventilator-associated pneumonia, defined by positive result of protected specimen brush culture, was different from that defined by other methods. Matched-cohort study. All patients with clinical suspicion of pneumonia were investigated with protected specimen brush, bronchoalveolar lavage, and blind bronchial samplings. Two groups were defined: brush-positive patients (positive culture of the protected specimen brush) and brush-negative patients (negative culture of the protected specimen brush, but positive culture with another method). A 14-bed medicosurgical intensive care unit (ICU) in an 850-bed teaching hospital. All patients with documented ventilator-associated pneumonia over 4 years 9 months. A total of 102 cases documented by protected specimen brush culture and 223 documented by another sampling procedure. Patients were matched according to diagnosis on admission, age, sex, date of admission, Acute Physiology and Chronic Health Evaluation II (APACHE II) score, and date of onset of pneumonia. Mortality rate, duration of mechanical ventilation, duration of ICU stay, duration of hospital stay, sampling methods, and microbiologic cultures. A total of 76 pairs were submitted for analysis. The effectiveness of matching was 81.85%. There was no difference in mortality between brush-positive patients and brush-negative patients. The ICU fatality rate was 38% in the brush-positive group and 39.4% in the brush-negative group (odds ratio [OR], 0.95; 95% confidence interval [CI], 0.46-1.93). The hospital fatality rate was 41% (OR, 1; 95% CI, 0.5-2.01). The mean (SD) duration of ventilation was 26 (23) days in the 2 groups (range, 3-132 days). The duration of ICU stay was 33 (27.4) days in the 2 groups (range, 3-152 days). When confounding factors are controlled, patient outcome is the same if ventilator-associated pneumonia has been diagnosed by protected specimen brush or by another sampling method.