Transfusion support after CAR-T cell therapy: Real-world patterns across commercial products and disease groups

Chimeric antigen receptor (CAR) T-cell therapy is frequently complicated by prolonged cytopenias requiring transfusion. Comparative real-world data on transfusion burden across commercial CAR-T products and post-therapy alloimmunization risk remain limited.

We retrospectively reviewed 102 patients treated with idecabtagene vicleucel (Abecma), ciltacabtagene autoleucel (Carvykti), lisocabtagene maraleucel (Breyanzi), axicabtagene ciloleucel (Yescarta), or brexucabtagene autoleucel (Tecartus) at a single center. Red blood cell (RBC) and platelet transfusions were quantified over 0-30, 30-90, and 90-180days post-infusion. Kruskal-Wallis tests and negative binomial regression with disease-stratified interpretation were used to estimate incidence rate ratios with 95% confidence intervals. Alloimmunization was assessed through routine, clinically driven antibody screening.

Adjusted exploratory models detected no product-specific differences in RBC support. Platelet transfusion distributions varied by product and disease group, but these findings were based on small absolute counts and were driven by a minority of high-need patients; therefore, they should be interpreted as hypothesis-generating rather than definitive product-level differences. No new RBC alloantibodies were detected through routine clinical antibody screening.

Transfusion needs after CAR-T therapy were concentrated in the first 30days and were modest overall, with most patients transfusion-free thereafter. This real-world descriptive analysis helps fill a current gap in understanding transfusion-support patterns across commercial CAR-T products and disease groups. Apparent product- and disease-related differences were driven by a minority of high-need patients and should be considered hypothesis-generating. No new RBC alloantibodies were detected through routine clinical screening, although standardized antibody surveillance was not performed.