Low Albumin, High Ferritin, and Thrombocytopenia Before Transplant Predict Non-Relapse Mortality (NRM) Independent of the Hematopoietic Cell Transplantation Comorbidity Index (HCT-CI).

In 2005, the HCT-CI was reported as a weighted scoring system to evaluate pretransplant comorbidities and their effect on HCT outcomes. Among the general population or among patients (pts), serum albumin inversely correlates with age, smoking, obesity and hypertension; serves as a marker of inflammatory status; and predicts cardiovascular as well as all-cause mortalities. High serum ferritin concentration can indicate iron overload and is thought to predict morbidity and mortality after HCT. Cytopenia could indicate either progressive malignancy or extensively treated bone marrow. Here, we investigated 1) whether any of these three laboratory parameters independently impact NRM and 2) the possibility of incorporating these parameters into the HCT-CI to augment its predictive power. To this end, we collected data from 1448 pts with hematological malignancies treated with allogeneic HCT at a single institution. Proportional hazards models were used to estimate the hazard ratio (HR) for NRM associated with different cut-off values for serum albumin, ferritin, platelet count, hemoglobin, and absolute neutrophil count (Table 1). The models were adjusted for age (59% <50 vs 41% ≥50 years), disease type (69% myeloid vs 31% lymphoid), disease risk (47% low vs 53% high), recipient CMV sero-status (45% negative vs 55% positive), donor type (58% related vs 42% unrelated), conditioning regimen type (30% nonablative vs 70% ablative), and HCT-CI scores (36% “0” vs 16% “1” vs 18% “2” vs 16% “3” vs 15% “≥4”). All five laboratory parameters were predictive of NRM at most of the tested cutoff values (Table 1). We selected a single cutoff value for each of the five based on the highest statistical significance and prevalence and conducted another proportional hazard model where the effect of each parameter was adjusted for the effects of all others. In the final multivariate analysis, serum albumin <3.5 g/dl (HR 1.4, p=0.003), serum ferritin >1000 ng/ml (HR 1.4, p=0.03), and platelet count <100,000 /mm³ (HR 1.68, p<0.0001) independently predicted NRM among all pts, while neutrophil count <1000 /mm³ (HR 1.1, p=0.48) and hemoglobin <10 g/dl (HR 1.01, p=0.96) did not. The first 3 parameters were associated with statistically significantly higher NRM and worse survival among all pts and with strata based on the type of conditioning regimen (Table 2). Since the HR's for the first 3 parameters were less than 2, each parameter was assigned a weighted score of 1 for augmentation of the HCT-CI. At 2-years, overall survival rates were 79%, 64%, 52%, 47%, and 30% for the augmented HCT-CI scores of 0-1 vs 2 vs 3 vs 4 vs ≥5, respectively, suggesting preserved good linearity for outcome prediction. Serum albumin, ferritin, and platelets are simple, reliable, and important prognostic markers that should be assessed before allogeneic HCT, incorporated into risk-assessment, and prospectively validated for possibility of augmenting the performance of the HCT-CI.

No relevant conflicts of interest to declare.