M29 How effective is bronchoscopy in aiding microbiological diagnosis in the immunocompromised patient?

Abstract

Introduction and objectives Bronchoscopy is frequently used to guide anti-microbial treatment in immunocompromised patients with possible respiratory infection. The causative organism in this group can be varied and early identification can improve clinical outcome. Our aim is to assess the effectiveness of diagnostic bronchoscopy in aiding microbiological diagnosis and its impact on patient management. Methods This was a single centre retrospective analysis of the case notes of immunocompromised patients undergoing bronchoscopy over the last two years. A database of their pre-procedural neutrophil and lymphocyte counts, evidence of infection on imaging, pre-procedural use of antibiotics and the number of positive lavage samples was collected. Details of organisms identified at lavage was recorded. Results 97 immunocompromised patients underwent bronchoscopy in our centre over the last two years. The majority (n=64) were immunosuppressed due to haematological disease. Overall 62 (64%) patients had either positive BAL culture or viral PCR. 18 of these patients had both. 39 patients had organisms identified in lavage culture (40%). 35 were bacterial and 4 were fungal. The most common bacteria were Stenotrophomonas maltophilia and Escherichia coli. 41 patients were found to have a positive lavage viral PCR. In 21 patients with positive viral PCR this was information not known prior to the bronchoscopy. 23 patients had no radiological changes on imaging prior to the procedure but high clinical suspicion of infection. In this group 8 patients isolated identifiable organisms after culture (35%) and 9 had a positive respiratory viral PCR, of which 5 provided new clinical information. Conclusion We found the rate of organism identification in this patient group was significant and likely to have influenced the patient management. The majority of patients were immunosuppressed due to haematological disease, however there was no difference in organism identification based on the cause of immunosuppression. The respiratory viral PCR results were frequently positive (42%) but the information this provided was often already available from previous viral throat swabs (51%) and therefore less likely to impact patient management. The awareness of our diagnostic rate will aid informed consent in future discussion with immunocompromised patients.