Oral Abstract
Transfusion Service - Transfusion Practices
Shan Yuan, MD (she/her/hers)
Medical Director, Blood Bank and Transfusion Services
City of Hope
Duarte, California
Disclosure information not submitted.
The Hematopoietic Cell Transplantation Comorbidity Index (HCT-CI) is a pre-transplant scoring system used to predict the risk of non-relapse mortality (NRM) in patients after allogeneic hematopoietic cell transplantation (allo-HCT) by assessing the presence and severity of various comorbidities. Our previous work in acute myeloid leukemia (AML) patients showed that lack of transfusion independence by Day 30 post-HCT predicted decreased survival due to higher NRM but not relapse. This study was performed to evaluate the predictive value of prolonged transfusion dependence when adjusted for HCT-CI score in the same population.
Study
Design/Methods:
This single-center, retrospective study of AML patients evaluated the relationships between transfusion independence by Day 30, HCT-CI and NRM through multivariable Fine and Gray regression model.
Results/Findings:
A total of 676 of the 692 AML patients transplanted with graft from matched related (37.4%), unrelated (49.1%), haploidentical (8.2%) donors or with umbilical cord (5.3%) during the study period (2011-2017) were alive at Day 30 and included. There were 234 (35%) patients with pre-transplant HCT-CI value of 3 or higher. By Day 30 post HCT, 56.7% (95% confidence interval [CI]: 52.8-60.3%) and 86.1% (95% CI: 83.2-88.5%) had achieved RBC and PLT transfusion independence, respectively.
Median follow-up among survivors (n = 307) was 7.1 years (range: 2.7-11.8). In the multivariable regression model controlling for donor type and transfusion dependence, higher (≥3), but not lower HCT-CI score was significantly associated with increased NRM at 2 years; the incidences were 0.201 (95%CI: 0.152-0.255), 0.117 (95%CI: 0.078-0.165), and 0.141 (95%CI: 0.099-0.190) among patients with HCT-CI scores of ≥3, 1-2, and 0 respectively. The adjusted hazard ratio (HR) was 1.69 (95%CI: 1.16-2.47, p= 0.0066) when comparing HCT-CI of ≥3 vs. HCT-CI of 0, and 0.90 (95% CI 0.58-1.38, p=0.63) when comparing HCT-CI of 1-2 vs. 0.
After controlling for donor type and HCT-CI, the lack of RBC or PLT transfusion independence by Day 30 was strongly and independently associated with higher NRM. The 2-yr NRM incidences were 0.253 (95%CI: 0.204-0.350) vs. 0.078 (95%CI: 0.054-0.108) for those without or with RBC transfusion independence, with HR of 2.02 (95% CI: 1.44-2.84, p < .001). For those without and with platelet transfusion independence, the NRM incidences were 0.330 (95%CI: 0.237-0.426) and 0.125 (95%CI: 0.100-0.154), with HR of 2.07(1.30-3.29, p=0.002). The major causes of NRM were infections, organ failure, GVHD, and other reasons, which were not significantly different among patients with or without RBC or PLT independence (p=0.95 and 0.15, respectively).
Conclusions:
RBC or platelet transfusion dependence at Day 30 post-HCT predicted increased NRM independently from higher HCT-CI values, which can be used in conjunction with HCT-CI to identify patients at higher NRM risks.