Association Between the Hematopoietic Cell Transplantation-Specific Comorbidity Index (CI) and Non-Relapse Mortality (NRM) After Reduced Intensity Conditioning (RIC) Allogeneic Stem Cell Transplantation (allo-SCT) for Acute Myeloid Leukemia (AML) in First Complete Remission (CR1).

Abstract 650

The Seattle group previously reported that the adapted Charlson CI (Sorror-CI) was useful for predicting NRM and survival in patients undergoing allo-SCT. However, the value of this CI index is still under considerable debate, since the cohort used by Sorror et al. (Blood 2005; 106:2912-9) included a large age range, a wide variety of diseases, with both myeloablative and RIC transplants. On the other hand, toxicities and NRM prediction is most likely needed in elderly and medically infirm patients who are increasingly offered RIC allo-SCT. Therefore, the current study was designed to test the performance of this CI and its association with outcomes among a cohort of 345 patients (i) aged >50 y.; (ii) diagnosed with a single disease entity, AML in CR1, and (iii) who underwent a RIC allo-SCT reported to the EBMT registry between 1999 and 2006.

In this series, the median year of allo-SCT was 2004, and the median age was 58 y. (range, 50-76). 32% of patients needed more than one induction course to achieve CR1. A fludarabine-based RIC regimen was used in 64% of patients, while 31% of patients received low-dose TBI as part of their RIC, and 6% received other non-specified RIC regimens. 76% of the patients received allo-SCT from an HLA-matched sibling donor.

Based on score calculated with hazard ratios (HR) estimated on the population studied, 161 patients (47%) had a CI score of 0, 96 patients (29%) had a score of 1, and 49 (14%) had a CI score of 2. The remaining 39 patients (11%) had CI scores of 3 or more. In this cohort, 2 years overall and leukemia-free survival rates were 64±3% and 54±3% and the 2 years relapse and -NRM cumulative incidences were 32±2%, and 15±2% respectively. The 2 years NRM incidences according to comorbidities score 0, 1, 2 and 3+ were 9±2%, 15±4%, 18±5% and 31±7%, respectively. In multivariate models (adjusted for recipient age, donor type, use of TBI or not, and cytogenetics risk group) comorbidities such as moderate active liver disease, obesity, prior history of renal dysfunction, and prior history of severe liver disease were associated with the highest HRs for 2 years NRM (varying from 2.11 to 2.76) and 2 years cumulative incidences of NRM varying from 22% to 44%, whereas previous solid tumor, diabetes, rheumatologic abnormalities, moderate pulmonary diseases, cardiac abnormalities (other than arrhythmia and valve disease) were associated with the lowest HRs (varying from 0.2 to 1.0) with 2 years cumulative incidences of NRM varying from 5 to 17%.

Results from this large study performed in a single disease entity and homogeneous allo-SCT setting, suggest that the hematopoietic cell transplantation-specific CI is a simple, informative and useful tool for capturing pre-transplant comorbidities, and for predicting NRM after RIC allo-SCT for AML in CR1 in patients aged >50 y. Such index may be used for clinical trials and patient counselling before RIC allo-SCT.