Red blood cell transfusion and associated outcomes in patients referred for palliative care: A retrospective cohort study.

Abstract

We aim to describe the occurrence of red blood cell transfusion and associated predictive factors and outcomes among patients referred for palliative care. This retrospective cohort study used linked health administrative data of adults referred for palliative care at an academic hospital from 2014 to 2018. Multivariable regression models were employed to evaluate patient characteristics associated with transfusion and the relationship between transfusion status and location of death. Survival analyses were performed using log-rank tests and Cox proportional hazards modeling. Of 6980 evaluated patients, 885 (12.7%) were transfused following palliative care consultation. Covariate factors associated with transfusion included younger age, higher performance status, lower baseline hemoglobin, and a diagnosis of hematologic malignancy (OR=2.97, 95% CI 2.20-4.01) or solid organ tumor (OR=1.37, 95% CI 1.10-1.71) vs. noncancer diagnosis. Median survival from palliative care consultation was 19 (IQR 5-75) days; 83 (32-305) days in those transfused and 15 (4-57) days in the nontransfused group (p < .0001). Median survival following transfusion was 56 (19-200) days. Solid organ tumor diagnosis was independently associated with poor survival (HR=1.7, 95% CI 1.39-2.09 vs. non-cancer diagnosis). Among individuals who survived ≥30 days, transfusion was associated with a higher likelihood of death in hospital (OR=2.15, 95% CI 1.71-2.70 vs. home/subacute setting). Transfusions commonly occurred in patients receiving palliative care, associated with cancer diagnoses and favorable baseline prognostic factors. Poor survival following transfusion, particularly in solid organ tumor patients, and the twofold likelihood of death in hospital associated with this intervention have important implications in prescribing transfusion for this population.