ABO identical and washed blood transfusions as candidate strategies to reduce early mortality in acute promyelocytic leukemia.

Background: Despite dramatically improved long term outcomes seen with all-trans retinoic acid therapy, and now arsenic trioxide, in acute promyelocytic leukemia (APL), early mortality remains a substantial challenge. Recent data from a single center study and the Surveillance, Epidemiology and End Results (SEER) registry report 30day mortality rates of 26% (n=18 of 70) and 17% (n=238 of 1400), respectively. Early deaths are predominately due to hemorrhage. Patients with APL invariably have abnormal laboratory hemostasis tests. The standard of practice is to prophylactically transfuse platelets, plasma and cryoprecipitate to mitigate abnormal platelet counts, PT/PTT and fibrinogen levels. Standard blood bank practice is to transfuse platelets, plasma and cryoprecipitate largely without regard to ABO blood group (platelets, cryoprecipitate), and, in some centers, transfusing ABO non-identical universal donor group AB plasma. Evidence from observational studies suggests that use of ABO non-identical blood components may be associated with increased bleeding. We hypothesized that use of ABO identical blood components and saline washed transfusions (red cells and platelets) would be associated with reduced early mortality in APL by avoidance of transfusion induced hemostatic dysfunction.

Methods: This is a single center cohort study of APL patients treated in an 800 bed university community and referral hospital. Novel approaches to transfusion support, based upon randomized trials, include implementation of ABO identical platelet transfusions for all patients with acute leukemia in 1990, use of only ABO identical cryoprecipitate in 2005, and washed transfusions of red cells and platelets for all patients with acute leukemia <50years of age beginning in 2006. Plasma transfusion has always been ABO identical. Two comparison populations were recent literature reports and the New York State Cancer Registry. We characterized 30 day mortality in APL patients seen in our institution since 2000 as a convenience sample comparable to literature reports, beginning approxcimately when ATRA therapy became uniform for induction therapy. Only patients receiving their induction therapy in our hospital were included.

Results: Of 41 patients there were 2 early (30 day) deaths (5%; a 71-81% reduction from expected). Early mortality at 100 days was 7% (n=3). The 30 day mortality in the younger cohort <50years of age (n=16) receiving washed transfusions was 0%. Restricting the analysis to patients treated since 2006 (ABO identical transfusions, mostly washed) (n=27; mean age 43 years; median 41 years; range 12-79), the early mortality rate at 30days was 3.7%. Long-term survival (5 years) of our APL patients was similar to New York State Cancer Registry and literature reports (80-83%).

Discussion: APL patients supported with transfusion regimens including ABO identical blood components, with or without washing, experienced early mortality at 30 days that was strikingly improved (71% to 86% lower) compared with that reported in the recent literature (3.7% to 5% vs. 17% to 26%). If these observed low rates of early mortality are related to transfusion practices, avoidance of ABO immune complex formation, and subsequent interference with hemostasis, is a plausible contributing mechanism. These favorable results provide a rationale for randomized trials of relatively simple and inexpensive approaches to reducing early hemorrhagic mortality in APL: use of ABO identical transfusions and washing to remove supernatant plasma.