Objectives: In our retrospective study we investigated the optimal timing of bronchoalveolar lavage (BAL) and the comparison of galactomannan antigen (GM) determined on BAL and serum in patients (pts) with pneumonia. Methods: From January 2007 to June 2009, 130 consecutive pts with haematological disease were investigated for the presence of 175 pneumonias in our Unit. At the onset of fever, blood cultures and assay GM were perfomed for 3 days. A chest computerized tomography (CT) was performed if pt was still febrile after 4 days. Pts began antifungal drugs if there was a possible or probable fungal infection. The decision to perform a BAL was made on the basis of CT images and response to therapy (tp). Pts characteristics and results are summarized in Table 1. Results: The time between the onset of neutropenia and fever, the onset of fever and CT, the CT and BAL and between the start of antifungal tp and the BAL was 8, 5.5, 5, 5.38 days respectively. Among 148 serum GM, 31 were positive with a mean value of 1.35 (range 0.6-4.8). We performed 75 BAL (average 11 days after the onset of fever). At the time of procedure, 25 pts were already receiving broad-spectrum antibiotic tp and 32 antibiotic and antifungal tp at therapeutic doses (76% patients were in tp). 25 BAL demonstrated a pathogen by culture, immunofluorescence or RT-PCR or by direct microscopic examination. GM in BAL was performed in 52 samples. GM was positive on BAL in 15 pts: mean value of 2.06 (range 0.54-4.47); in 4 cultures were positive for fungus. 9/15 were on antifungal tp from an average of 7 days. In 9/15 also serum GM was positive at the time of BAL. Of the 31 events with positive serum GM, BAL was performed in 16: among them, 9 confirmed GM BAL positivity, while 3 were negative for GM. 2 showed Aspergillus culture positive without GM. We found 6 pts with BAL GM positive and serum GM negative. 36 BAL have provided an etiological diagnosis. The diagnostic power was 48%. We changed tp in 38% of patients thanks to BAL. However, there was no difference in mortality rate between patients with a positive and negative BAL. Conclusion: In our experience, when BAL is used late, pt's clinical conditions and the great number of antibacterial and antifungal therapies may negatively influence its results. We noticed that if we perform it early during the course of disease we could improve the outcome and avoid unnecessary therapies. We must keep on comparing the sensibility of GM on serum and on BAL.
Full conference title:
20th European Congress of Clinical Microbiology and Infectious Diseases
- ECCMID 20th (2010)