S1637 A BAL‐DRIVEN ANTIMICROBIAL TREATMENT IMPROVES CLINICAL OUTCOME IN HEMATOLOGIC MALIGNANCIES PATIENTS WITH LUNG INFILTRATES DETECTION: A PROSPECTIVE MULTICENTER STUDY OF THE SEIFEM GROUP

• Published in: HemaSphere 2019-06, Vol.3 (S1), p.756-757
• Main Authors: Marchesi, F., Cattaneo, C., Criscuolo, M., Delia, M., Dargenio, M., Del Principe, M.I., Spadea, A., Fracchiolla, N. S., Melillo, L., Perruccio, K., Alati, C., Russo, D., Garzia, M., Zappasodi, P., Cefalo, M.G., Armiento, D., Cesaro, S., Decembrino, N., Mengarelli, A., Busca, A., Pagano, L.
• Format: Article
• Language: English
• ISSN: 2572-9241; 2572-9241
• DOI: 10.1097/01.HS9.0000564796.22734.06

Background: The best diagnostic approach of lung infiltrates (LI) in patients with hematologic malignancies (HM) remains to be well established. Bronchoscopy with bronchoalveolar lavage (BAL) is useful but prospective studies evaluating the efficacy of BAL in response to antimicrobial treatment and survival are lacking. Aims: To evaluate the efficacy of BAL in patients with HM. Primary end‐points were: 1) rate of detection of a putative causal agent; 2) evaluating radiologic response following BAL‐driven antimicrobial therapy at day +30 from LI detection. Secondary end‐points were to evaluate overall survival (OS) and attributable mortality (AM) at day +120 from LI detection comparing a BAL‐driven with an empiric approach. Methods: We conducted a prospective multicenter observational study in 18 hematology institutes participating to SEIFEM group from January 1 to September 30, 2018. Adult and pediatric severely immunocompromised patients with HM with LI detection at CT scan were enrolled according to the following criteria: 1) presence of a host risk factor for invasive fungal disease (EORTC/MSG 2008); 2) absence of any microbiological and/or serological documentation at standard diagnostic work‐up (including serum galactomannan, GM); 3) fever or respiratory distress not responding to broad‐spectrum antimicrobial therapy. According to the investigators, the patients fulfilling these criteria underwent a BAL. Results: Out of 3.055 patients admitted on wards of the 18 institutions in the study period, 434 (14%) had LI and 177 of them (6%) were enrolled into the study and evaluable for analysis. Median age was of 60 years (1–83). Most patients had a AML/MDS (81), followed by lymphoma (41), ALL (27), MM (15), CLL (15) and other HM (6). Median neutrophilis count was of 460/mm3, and 106 patients (60%) had experienced a prolonged and severe neutropenia (neutrophilis under 100/mm3 for a median of 12 days). BAL was feasible in 145 cases (82%) in a median time of 4 days from LI detection, whereas in the remaining 32 cases BAL was not performed for clinical decision. A putative causal agent was detected in 111 cases (76%): in 41 cases BAL was positive for GM (median value: 2,93), in 36 cases for bacteria, in 5 for Pneumocystis and in the remaining cases for viruses. In 87 cases (61%) the results of BAL allowed to modify the ongoing empiric antimicrobial therapy, whereas in 56 cases BAL was not informative. We observed a significant better outcome of LI at day +30 fr