PB1720 EARLY DEATHS IN ACUTE PROMYELOCYTIC LEUKEMIA: ANALYSIS OF PREDICTIVE FACTORS IN OUR INSTITUTION IN THE LAST 20 YEARS

Background: The outcome of acute promyelocytic leukemia (APL) has significantly improved in the last decades due to better understanding of the disease, introduction of effective targeted drugs and improvement in supportive care. However, induction mortality is approximately 30% and this important problem remains a disappointing and frustrating challenge for oncohemalogists. Aims: The aim of our study was to analyze which factors at diagnosis can be associated with an early death in APL. Methods: We performed a retrospective observational study on 49 consecutive patients diagnosed with APL in our hospital between 1998 and 2018. We collected a set of clinical, analytical, therapeutic and evolution data, with the following variables: age; sex; form of clinical presentation; hemoglobin, leukocytes and platelet count; severity of thrombocytopenia and presence or not of leukocytosis; standardized risk of APL according to the leukocyte and platelet count (Sanz score); serum lactate dehydrogenase (LDH) and uric acid levels; coagulation parameters (D‐dimer included); percentage of blasts in peripheral blood and bone marrow; subtype of APL (hyper‐ or hypogranular); immunophenotyping by flow cytometry (CD34 and HLA‐DR, among others); cytogenetic findings (specifying percentage of PML/RARA by FISH at diagnosis); initial chemotherapy scheme (ATRA + idarrubicin, ATRA, ATRA + arsenic trioxide) and idarrubicin starting day; the appearance of hemorrhage, thrombosis, infection and ATRA syndrome during induction treatment; finally, the death or not of the patient and, if it ocurred, at what time (before receiving any specific treatment, during the induction therapy or later). Statistical analysis of the data was performed using the updated SPSS version 17 for Windows, and included the X‐square or Fischer's exact test, or the Mann‐Whitney test, as appropriate. Results: Of the total of 49 patients, 13 died prematurely (26.5%) either before starting specific treatment (3) or during induction (10), between day + 1 and + 43 after diagnosis. When compared with those who survived, patients who died early were older (averages 60 versus 41 years; p = 0.008) and presented at diagnosis with a higher count of leukocytes (36,9 vs 13,3 x10 9 /L; p = 0.045), a higher percentage of blasts in peripheral blood (65 vs 35%; p = 0.014) and bone marrow (85 vs 75%; p = 0.03), a lower value of hemoglobin (8.8 vs 9.9 g/dL; p = 0.05), a serum LDH level above the normal value (92 vs 56%; p = 0.035), a higher incidence of hemorrhage (73 vs 35%; p = 0.041) and infection (46 vs 6%; p = 0.003) at induction, and a higher percentage of patients classified as “high risk” by Sanz score (62 vs 25%; p = 0.05). In the group of early death we also found a tendency to a higher incidence of leukocytosis (54 vs 35%; p = 0.081), a more frequent subtype of hypogranular variant APL (40 vs 12%; p = 0.064), a higher percentage of FISH positivity to PML/RARA fusion gen (79 vs 62%; p = 0.059) and a serum uric acid level above the normal value (27 vs 6%; p = 0.091). We did not find significant differences for the rest of the analyzed variables, including coagulation and immunophenotypic findings. Summary/Conclusion: Our early mortality rate is similar to that published by other authors. The knowledge of which factors can influence more decisively in an early death in patients with APL, may help to establish strategies directed to diminish this frustrating problem. We detected several interesting prognostic factors in our work, but studies that include a greater volume of patients are needed to try to improve population‐wide survival in this, otherwise, highly curable disease.