Objectives: To check on the accuracy of a new protected blind brush (BB) inserted through an endotracheal tube to collect respiratory secretions to be used in the diagnosis of nosocomial pneumonia (NP) in ventilated patients.
Design: Prospective study of patients who had undergone both BB and plugged telescoping catheter via fiberoptic bronchoscopy (PTC-FB) sample collection sessions.
Setting: Intensive Care Unit of a referral-based University Hospital.
Patients: All patients (n = 37) mechanically ventilated for more than 3 days with clinical and radiological criteria of NP between July 1990 and March 1991.
Interventions: Randomized BB and PTC-FB sample collection sessions carried out less than 30 min apart.
Measurements and main results: The two sampling procedures resulted in similar findings with both cultures either negative or positive and identified the same organism and colonies in 31 patients (83.7%). Agreement was 90% when the patients with right or bilateral pulmonary infiltrates were grouped together and 100% when only the right field was considered. Complications arising from BB sampling were much lower than those from the conventional PTC-FB technique.
Conclusions: Our results, pending confirmation by other prospective studies, indicate that BB sampling is useful in the diagnosis of NP in ventilated patients with radiological evidence of either right or bilateral pulmonary infiltrates and that it could stand in for PTC-FB in ICU settings where this procedure is not available.