Pulmonary Infections in Patients with Myelodysplastic Syndromes Receiving Azacytidine Treatment

Azacytidine (AZA) is a demethylating agent widely used in the treatment of patients with high-risk Myelodysplastic Syndromes (MDS). Pulmonary infections are often reported in these patients during AZA treatment, however, their incidence, etiology and impact in the overall outcome are still unclear. A major problem is represented by the low level of clinical and microbiological documentation, as most of patients treated with AZA are managed on an outpatient basis. This study aims to evaluate the incidence and clinical impact of pulmonary infections in MDS patients treated with AZA in 5 hematological Institutes in Rome, with a relatively homogeneous infective diagnostic work-up in the presence of fever or other infective clinical signs. We retrospectively evaluated 146 MDS patients [M/F 93/53, median age 69.5 years, interquartile range (IQR) 65.0 - 75.6] treated with AZA at our Institutions from 04/2009 to 01/2016. All patients received AZA cycles at standard dosage (75 mg/m² for 7 days every 28 days) as outpatients. In the event of febrile neutropenia or other infectious episodes patients underwent blood cultures, culture from other sites, and chest x-ray or (preferably) pulmonary CT-scan. Galactomannan assay from serum and from sputum/bronchoalveolar lavage was performed as indicated. The total number of AZA cycles was 1712, with a median per patient of 9 cycles (IQR 5 - 17 cycles). There were 75 episodes of lung infection (4.1% of AZA cycles), with 58 patients (39.7%) presenting at least 1 episode. Based on the above diagnostic work-up, pulmonary infiltrates were considered of fungal origin in 21 cases (28.0%), associated to bacteremia in 5 cases (6.7%) and of unknown origin in the remaining 49 cases (65.3%). As to the time of occurrence of lung infections, 29 episodes were documented in the first 4 cycles of AZA treatment (5.4% of 535 cycles) and the remaining 46 episodes beyond the 4^th cycle of AZA treatment (3.9% of 1177 cycles). Overall, a pulmonary fungal disease was documented in 10 of 535 (1.9%) cycles 1-4 and in 11 of 1177 (0.9%) cycles beyond the 4^th (p=0.001). Several clinical features (age, gender, Hb level, WBC and PLT counts, time from diagnosis) were evaluated as predictive factors for the occurrence of pulmonary infection, but none was significant. Out of 58 patients who developed at least one pulmonary infection, 39 (67.2%) definitively discontinued the AZA treatment within 3 months from the infectious episode due to deterioration of clinical conditions, hematologic disease progression and/or death. Attributable mortality rate of patients with pulmonary infection was 22.4% (13 of 58). The median Overall Survival (OS) of the whole cohort was 18.0 months (95%CI 14.4 - 21.5): the median OS of patients with pulmonary infection was 15.6 months (95%CI 13.1 - 18.0) compared with 21.5 months (95%CI 16.3 - 26.6) of patients without pulmonary infection (p=0.031). In conclusion, pulmonary infections are a common complication in MDS patients receiving AZA treatment, and are often associated to an interruption of AZA therapy. Pulmonary fungal infections are more frequently observed early during the first cycles of treatment. It should be defined if the poor outcome of patients who develop pulmonary infections during AZA therapy is an epiphenomenon of an immunologic deterioration associated to the hematologic disease progression or is independently related to the complication. If confirmed in other experiences, the results of our study raise the issue of the need of an antibacterial and/or antifungal prophylaxis particularly during the first months of AZA therapy.