Early Death in Patients with Disseminated Intravascular Coagulation during Induction Therapy for Acute Promyelocytic Leukemia: A Nationwide Analysis

Background: Real-world data studies showed poorer outcomes in patients with acute promyelocytic leukemia (APL) than randomized controlled trials, because elderly patients were excluded in such trials. Reportedly, the main cause of death was severe bleeding due to disseminated intravascular coagulation (DIC) during induction therapy for APL. The management of DIC was therefore crucially important especially in elderly patients. This study aimed to clarify factors associated with in-hospital death in all patients, and elderly patients with DIC during induction therapy for APL.

Study Design and Methods: We retrospectively identified 1,463 patients with newly diagnosed APL who received induction therapy including all-trans retinoic acid (ATRA) between July 2007 and March 2018 from a nationwide inpatient database in Japan. In-hospital death was evaluated with multivariable logistic regression models in all patients, and in ≥60 year-old patients. Anticoagulants included recombinant human soluble thrombomodulin, delteparin (low molecular weight heparin), danaparoid sodium, gabexate mesilate, and nafamostat mesilate which were administered within three days from admission. Patients who died within three days from the admission were excluded from the study to avoid immortal time bias.

Results: We identified a total of 1,138 (78%) patients who developed DIC. We excluded 23 patients who died within three days from the admission. The remaining 1,115 patients were analyzed. During hospitalization, 172 (15%) patients died at a median of 13 days (interquartile range: 7-30). Compared with younger patients (20 to 39 years old), elderly patients were significantly associated with higher in-hospital mortality (60 to 79 years old: odds ratio 5.58 [95% confidence interval 3.05-10.22], 80 years or older: 13.51 [6.07-30.08]). Patients who received ATRA monotherapy had significantly higher incidence of in-hospital death (2.48 [1.54-4.01]). Delayed initiation of ATRA was significantly associated with higher mortality (1.60 [1.11-2.30]). A total of 699 patients (63%) received anticoagulant therapies, but none of these were significantly associated with lower mortality. Use of multiple anticoagulants was significantly associated with higher in-hospital mortality (2.47 [1.16-5.26]). Subgroup analyses in patients ≥60 years old were then conducted. During hospitalization, 122 of 416 (29%) patients died at a median of 13 days (interquartile range: 7-29). Both late initiation of conventional chemotherapy and no conventional chemotherapy were significantly associated with higher in-hospital mortality (1.88 [1.01-3.49], 3.25 [1.74-6.06], respectively). Use of recombinant human soluble thrombomodulin and use of multiple anticoagulants were significantly associated with higher mortality (1.91 [1.09-3.35], 2.64 [1.01-6.90], respectively).

Conclusions: Elderly patients who developed DIC during induction therapy for APL were significantly associated with higher in-hospital mortality. Immediate initiation of ATRA and early initiation of conventional chemotherapy may have contributed to preferable outcomes.