Diagnostic tests for pneumonia in ventilated patients: prospective evaluation of diagnostic accuracy using histology as a diagnostic gold standard.

The diagnostic accuracy of protected-specimen brush (PSB), bronchoalveolar lavage (BAL), and endotracheal aspirates (EA) was prospectively evaluated in a series of 28 mechanically ventilated patients (MV patients) who died within 3 d of the bronchoscopic procedure, using postmortem lung examination as the gold standard for establishing the diagnosis of pneumonia. The entire fixed lungs were carefully dissected along the bronchovascular axes and each segment was cut into 5- to 10-mm thick sections, enabling gross examination of the lung parenchyma. Two tissue blocks were taken from each segment, including grossly abnormal areas whenever present. In several cases, two peripheral (subpleural) lung-tissue blocks were also taken from each lobe prior to systematic dissection of the lungs. Quantitative cultures (QC) and direct cytologic and microbiologic examination (DE) was performed on respiratory samples obtained within 72 h before death. Values of 10(3) cfu/ml of Ringer's solution, 10(4) cfu/ml of retrieved fluid, and 10(6) cfu/ml of respiratory secretions were used as cutoff points for quantitative PSB, BAL, and EA cultures, respectively. The main findings in this study were that: (1) Pneumonia was present in 67% of the patients. (2) Histologic lesions of pneumonia were mainly bilateral and predominated in the dependent lung segments. (3) Coexistence of a variety of noninfectious processes was a common finding in patients with pneumonia. (4) In several cases pneumonia was absent from peripheral lung samples while more central areas of the same segment displayed typical foci of pneumonia. (5) The sensitivity of quantitative cultures was 55%, 57%, and 47% for EA, PSB, and BAL, respectively, and the specificity was 85%, 88%, and 100%, respectively. Reducing the diagnostic threshold of EA to 10(5) cfu/ml of respiratory secretions instead of 10(6) cfu/ml resulted in a sensitivity of 63.1% and a specificity of 75% for EA. The sensitivity of direct examination (DE) was 50%, 47%, and 47%, respectively, and the specificity was 75%, 88%, and 87%. (6) The presence of intracellular organisms (ICO) in BAL had a 36.8% sensitivity and 100% specificity in establishing the diagnosis of pneumonia regardless of their percentage. (7) Although 15 patients (53%) were not on antibiotics or were off antibiotics for more than 48 h before testing, no relationship could be established between the patients' antibiotic status and the result of any diagnostic test. By using a recommended methodology for respiratory sampling techniques together with complete postmortem lung examination as a diagnostic "gold standard," this study provides a realistic insight into the diagnostic values of EA, PSB, and BAL in MV patients with suspected pneumonia.