Change in Clinical Management Associated with Flexible Bronchoscopy and Bronchoalveolar Lavage in Hematopoietic Stem Cell Transplant Recipients with New Pulmonary Infiltrates

BACKGROUND: Pulmonary complications occur in 60% of hematopoietic stem cell transplant (HSCT) recipients with significant morbidity and mortality. Bronchoscopy with bronchoalveolar lavage (BAL) is an important diagnostic tool, but yield can be variable. The aim of this study was to investigate the utility of BAL in HSCT patients with new pulmonary infiltrates and determine factors associated with higher BAL yield. METHODS: Retrospective review of BAL results from January 2014 to July 2016. Included first bronchoscopy for HSCT patients over 18 years of age. Positive BAL determined by positive culture (bacterial, fungal, mycobacteria), viral PCR, elevated aspergillus galactomannan antigen (AGA), and cytology. Logistic regression analysis was performed. RESULTS: 54 HSCT recipients were included: 93% allogeneic, 34% neutropenic, 39% on prednisone, 59% on supplemental oxygen, 8% receiving mechanical ventilation (MV), 85% with multilobar infiltrates, 92% on antimicrobials (antibacterial 83%, antifungal 92%, antiviral 13%). 65% had positive BAL (23/54 bacterial, 16/54 elevated AGA, 6/54 fungal, 14/54 viral PCR, 1/54 mycobacteria). Median time to BAL from HSCT was 7.5 months. Not on levofloxacin prophylaxis (P = 0.004), not on MV (P = 0.037), or unrelated donor were associated with positive BAL results. Positive bacterial BAL culture was predictive of antibacterial escalation (P < 0.001). Elevated BAL AGA was associated with antifungal escalation and antibacterial de-escalation (P = 0.012). Antiviral initiation was more likely with positive BAL PCR (P = 0.002). Antifungal de-escalation (P = 0.047) and steroid initiation (P = 0.010) were more likely with negative BAL. 6 month mortality was 41%, and more likely with positive bacterial BAL culture (P = 0.046). Overall positive BAL was not predictive of mortality. CONCLUSION: Bronchoscopy with BAL assisted in making critical management changes: elevated AGA with antifungal escalation and antibacterial de-escalation; and negative BAL with prednisone initiation and antifungal de-escalation. Unrelated donors, patients not on levofloxacin, or not on MV were more likely to have a positive BAL. Antimicrobial use likely reduced diagnostic yield for bacterial pathogens. Overall 6 month mortality was high, especially among those with bacterial infections. DISCLOSURES: All authors: No reported disclosures.