Pulmonary Infiltrates (PI) in Acute Myeloid Leukemia (AML) During Induction Treatment (IT). How Much Do We Know?

During IT, AML patients may develop PI due to infection, leukostasis, pneumonitis, drug reaction or even acute respiratory distress syndrome. The risk association and the clinical outcome of such PI are poorly characterized in the literature. This study investigates the frequency, radiologic patterns, and outcome of PI during the IT of AML patients.

We retrospectively reviewed 363 cases of AML patients who received IT either cytarabine and anthracycline regimens, or hypomethylating agents between January, 2000-January, 2011 at William Beaumont Health System. Of 363 patients, 120 developed PI during IT, those patients were divided into 2 groups based on distribution of the infiltrate presenting whether localized or diffuse infiltrate. Distribution of the infiltrate was evaluated by dividing the thorax vertically in the midline, then horizontally by two lines originating above and below the hilus resulting in six areas. Localized infiltrate was defined as involvement of 1–3 areas, and diffuse infiltrate was defined as >3 areas were involved. Detected by either Bronchioalveolar lavage culture or sputum culture revealing a pathogen other than coagulase-negative staphylococci or corynebacteria was considered as pathogenic microorganisms. Treatment end points were defined as resolve or persistence of the infiltrate or death. Data on pts characteristics, leukemia subtype, cytogenetic risk, microorganism type, white blood cell count (WBC) at diagnosis, neutrophils count at the time of the infiltrate reported, response to antibiotic and/or antifungal therapy, using respiratory support, mortality rate, were retrieved from the records.

Our study involved 120 patients. Of these, 33% developed PI during their IT. This population was 53.3 % female, median age 66 years (range 23–93). We noted that 63 patients (52.5%) had localized infiltrate and 57 patients (47.5%) had diffuse infiltrate. Of the 120 patients, 46 (39.3%) had pathogenic microorganisms. These microorganisms were found to be 50% gram positive, 32% gram negative, 10.8% fungal, and 4.3% anaerobic. All 120 patients received two broad-spectrum antibiotics and/or an antifungal. Of the total number of patients, 58 (48.7%) required intubation and ventilatory support. The repeat computed tomography at the end of the hospital stay (after the treatment) showed resolution of the infiltrate in 67 patients (55.8%). During our study, 46 patients died from PI (38.3 %).

Further analysis of the data showed patients with localized PI (when compared to patients with diffuse infiltrate) were more likely to have positive pathogenic microorganisms (68.3% vs 8.8%, p <0.001), neutropenic (96.8% vs 21.1%, p <0.001), and tend to have potentially reversible infiltrates after treatment (87.3% vs 21%, p <0.001).

Patients with diffuse infiltrate (when compared to patients with localized infiltrate) were more like to require intubation (78.9% vs 21%, p <0.001), had leukocytosis (WBC >100 Bil/L) at diagnosis (54.4% vs 0%, p <0.001) and had a higher mortality rate (70.2% vs 9.5%, p <0.001).

During IT about one third of the AML patients developed PI. The radiological patterns of PI showed specific etiological and prognostic associations, may guide the clinical decision making processes. Diffuse PI is an unfavorable characteristic with overall dismal outcome.

No relevant conflicts of interest to declare.