Summary

Background and objectives: Antibiotic therapy alters the diagnostic value of protected brush specimens. With protected brush specimens alone, diagnosing pneumonia requires 24 or 48 h. Addition of direct Gram staining shortens this delay. Antibiotic prophylaxis, recommended after major surgery, may influence the contribution of Gram staining to diagnosing postoperative pneumonia. Methods: During a 1-yr period, we retrospectively studied all patients on mechanical ventilation suspected of having postoperative pneumonia who had undergone fibreoptic bronchoscopy with protected brush specimens. Postoperative pneumonia was diagnosed when quantitative protected brush specimens culture results yielded 103 colony-forming units mL−1. Results: Fifty patients were clinically suspected of having postoperative pneumonia after cardiac (n = 42), vascular (n = 5) or thoracic (n = 3) surgery. Eleven (22%) samples were obtained during antibiotic prophylaxis. Twenty-two (44%) episodes were microbiologically proven. Gram-stain sensitivity was 95.5%, with 82.1% specificity, 80.7% positive-predictive value and 95.8% negative-predictive value. Concordance between direct Gram-stain-identified pathogens and Gram stain of cultured pathogens was significantly less frequent during antibiotic prophylaxis (63.6%) than afterwards (94.9%) (P < 0.05). Conclusion: Antibiotic prophylaxis diminished the diagnostic value of Gram staining of protected brush specimens. When protected brush specimens was performed during antibiotic prophylaxis, Gram staining accurately enabled early exclusion of postoperative pneumonia because of its excellent negative-predictive value. After antibiotic prophylaxis, Gram staining permitted early diagnosis of postoperative pneumonia identification of the responsible pathogen.