Study objectives: To determine the clinical significance of Candida sp isolated from bronchoscopic samples in patients with suspected pneumonia.
Design: A retrospective chart review was performed in all nonneutropenic adult patients with Candida sp isolates from respiratory secretions obtained by bronchoscopy over a 5-year period (1991 to 1995). Potential risk factors, therapeutic decisions, and outcome were recorded. Microbiological findings, chest radiograph reports, and pathologic material were reviewed. Isolates were classified as definite, probable, or indeterminate contamination, or as definite pulmonary candidiasis, on the basis of histologic findings, therapeutic decisions, and outcome.
Setting: A 600-bed teaching hospital with 16 beds in a medical-surgical ICU.
Patients: Thirty-seven consecutive patients with positive cultures for Candida sp from respiratory samples obtained by bronchoscopy were evaluated. Thirty-two of these 37 patients (86.5%) received antibiotic therapy prior to sampling, and 23 (62.2%) were intubated.
Results: Contamination was classified as definite in 3 patients (8.1%) and probable in 30 others (81.0%). Contamination was indeterminate in two cases (5.4%). Two additional patients (5.4%) received antifungal agents for systemic candidiasis. No cases of pulmonary candidiasis could be demonstrated, although 24 of 28 patients showed protected specimen brush cultures > or = 10(3) cfu/mL.
Conclusions: Nonneutropenic patients with isolation of Candida sp from bronchoscopic samples, even in high concentrations, are unlikely to have invasive candidiasis. Indication for initiation of antifungal therapy should be based on histologic evidence or identification from sterile specimens.