Patterns and Safety of Glucocorticosteroids Use Following CD19 CAR-T Cell Therapy for B-Cell Lymphoma

Introduction: CAR T-cell therapy is a well-established therapy for relapsed and/or refractory B cell non-Hodgkin's lymphomas (R/R B-NHL). However, CAR-T cell therapy may cause significant toxicities such as CRS and ICANS. Glucocorticosteroids play a pivotal role in managing these toxicities but can also suppress T cell function and may impede the persistence and effectiveness of CAR-T cells. The impact of glucocorticosteroids has resulted in conflicting conclusions in the literature.

Objective: This single-center, retrospective, observational study aimed to compare the outcomes of patients with R/R B-NHL who received steroids to those who did not after CAR-T cell therapy.

Methods: Patients' records were reviewed to identify any steroid administration within 30 days of CAR-T cell infusion. Patients receiving steroids within 7 days before CAR-T infusion were excluded to focus on CAR-T impact without confounding baseline steroid use. Information on the type of steroids, indication, timing and duration was collected. Steroids were expressed as dexamethasone equivalent. Steroids-associated variables were analyzed as both continuous and categorical variables, using median as the threshold. Cox regression models and Kaplan-Meier analysis was employed for PFS and OS estimation using steroids as a time-dependent covariate and a 30-day landmark analysis, with events as disease progression or death for PFS and death only for OS.

Results: 292 patients with R/R B-NHL received commercial CAR-T cell therapy (48% axi-cel, 24% tisa-cel, 22% liso-cel, and 6% Brexu-cel), primarily for large B cell cell lymphoma (85%), mantle cell (11%) or follicular (4%) lymphoma. Only 49 of the 292 patients (17%) received steroids between the day of infusion (day 0) to 30 days after. The median cumulative dexamethasone-equivalent dose at 30 days was 60 mg (range: 1-1982), the median day from CAR-T cells infusion to steroids initiation was 7 days (range 0-26) and the median duration of corticosteroid use was 3 days (range, 1-27). Patients receiving steroids had a significantly higher number of lines of prior therapies (p=0.036), more stage 3-4 disease (p=0.027), and elevated pre-lymphodepletion LDH levels(p=0.004). However, there were no other significant differences, including CD19 CAR-T product type, between the groups.

The rates of grade 2 or higher CRS and ICANS were documented in 42% and 18%, respectively. ICANS (47%), CRS (14%), or both (21%) were the leading indications for steroid administration. Other indications for administration of steroids were progression of disease (n=4), adrenal insufficency (n=3), graft vs host disease (n=1), and premedication with intravenous immune globulin (n=1).

With a median follow-up of 20 months, median PFS and OS were 6.8 months (95% CI 6-11) and 27 months (95% CI 18-not reached), respectively. The use, duration, and cumulative dose of steroids, whether analyzed as continuous or categorial variables in univariable Cox regression models, were not associated with PFS and OS in a 30 landmark analysis ( Figure). The analysis of the adminstration of steroids as a time-dependent covariate did not show significant differences for either PFS or OS. Importantly, patients exposed to steroids were also not at increased risk for bloodstream infections (p=0.68).

Conclusion: In this largest analysis to date on the impact of steroids on CAR-T outcomes, steroid exposure following CAR-T cell infusion was not associated with a greater likelihood of CAR-T cell therapy failure. Our results deviate from previous observations that associate using steroids with shorter PFS and OS (Starti et al, Blood 2021, Terao et al, JTCT 2023). The discrepancy is likely due to our study's exclusion of steroid use prior to CAR-T and statistical analysis. Overall, our findings suggest that steroid administration for CAR-T cell therapy toxicity management is safe and does not appear to have a negative impact on efficacy.