Real-world analysis of safety and efficacy of CAR T-cell therapy in lymphoma patients with decreased kidney function

To the Editor, Eligibility criteria for clinical trials, including CD19 chimeric antigen receptor T-cell (CAR-T) therapy trials, require patients to have adequate major organ function, and with respect to kidney function, this is defined as a creatinine clearance of more than 60 mL/min/1.73 m2.1, 2 This resulted in patients with reduced kidney function (RKF) including those with chronic kidney disease (CKD) being excluded from CAR T trials despite its prevalence in lymphoma patients, with approximately one third having CKD.3, 4 While real-world studies have shown that CAR T-cell therapy can be effective even among large B-cell lymphoma (LBCL) patients not meeting trial eligibility, there is a lack of data regarding the impact of CKD on toxicity and disease outcomes.2, 5, 6 The objective of this retrospective single centre study was to evaluate the impact of glomerular filtration rate on toxicity and survival outcomes in real-world adult patients with relapsed/refractory (r/r) LBCL and RKF treated with commercial CAR T-cell therapy in the third-line setting between 1 January 2018 and 15 March 2021. Normal kidney function (NKF) and RKF were defined as estimated glomerular filtration rate (eGFR) of ≥ and <60 mL/min/1.73 m2, respectively, at time of lymphodepleting chemotherapy (LDC).7 We identified 210 patients, 41 (19.5%) had RKF at the time of LDC and median eGFR at LDC was 86 mL/min/1.73 m2 (15–163). None of the patients were dialysis dependent at the time of LDC. Most had mild to moderate decrease in kidney function with 29 patients (70%) having eGFR of 45–59 mL/min and 9 patients (22%) having eGFR of 30–44 mL/min. Only three patients (7%) had a severe decrease in kidney function (eGFR 15–29 mL/min). Out of the 41 patients, 30 had CKD (RKF ≥3 months) and 11 had subacute kidney disease (RKF for <3 months) that was predominantly (9 out of 11 patients) persistent post CAR T-cell therapy. Axi-cel was the main (94%) CAR T-cell product administered. Per discretion of the treating physician, 15 out of 41 RKF patients received a reduced dose of fludarabine versus 8 out of 169 in the NKF group (p < 0.0001). The dose reduction varied between 20%–50% and 20%–33.3% in the RKF and NKF groups respectively. Median age of the overall study population was 60 years (range 18–88), while 63% of patients in the RKF group were older than 60 in comparison with 45% in the NKF group (p = 0.034). In addition, median haematopoietic cell transplantation-specific comorbidity index (HCT-CI) score was higher (3 vs. 1, p = 0.0049) in the RKF versus NKF groups respectively (Table 1). Seventy-eight patients (37%) of the overall cohort had intensive care unit (ICU) event(s) and with median ICU stay of 3 days (IQR = 4). Number of ICU events was comparable between the RKF and NKF groups, 18 (44%) vs. 60 (36%), respectively, p = 0.318; however, RKF patients had longer ICU stay in comparison with NKF patients: median ICU stay of 6 vs. 3 days (p = 0.0173) respectively. Thirty-three patients experienced grade ≥3 cytokine release syndrome (CRS) (15.7%) of the overall cohort; 88 patients (42%) had grade ≥3 immune effector cell-associated neurotoxicity syndrome (ICANS) with 37 patients (42%) having longer than 2 days of grade ≥3 ICANS (Table 2). No statistically significant difference was noted in the incidence of grade ≥3 ICANS (46% vs. 41%, p = 0.521) and grade ≥3 CRS (17% vs. 15%, p = 0.790) between the RKF and NKF groups, respectively, but interestingly grade ≥3 ICANS was noted to have a prolonged duration in the RKF group, median of 3 days vs. 2 days for the NKF group (p = 0.04). Moreover, older age (>60) was not associated with prolonged grade ≥3 ICANS (p = 0.54). No statistical significance difference was observed in the incidence of Day 30 post CAR T grade ≥3 CTCAE v5 haematological toxicity or sepsis between the RKF and NKF groups (Table 2). Similarly, there was no significant difference in overall survival (OS), progression-free survival (PFS) or overall response (OR) rate between the two groups (Supporting information). Nonrelapse mortality (NRM) rate was 13% (27 patients) for the entire cohort with a median follow-up of 22.9 months (95% CI: 21.1–25.6 months). The NRM rate decreased to 8.5% (18 patients) after the exclusion of deaths related to a second malignancy and unknown aetiology, limiting the mortality to CAR T-cell-associated toxicity and infection (Table S1). Patients in the RKF group had a higher rate of NRM than those in the NKF; 22% vs. 10.7%, p = 0.0525 respectively. Nonetheless, in multivariable analysis to evaluate the effect of RKF on NRM with adjustment of age and LDH in the model, the association between RKF and NRM was shown to be non-significant with OR of 2.3 (95% CI: 0.89–5.59; p = 0.087). Regarding kidney related complication, 35 patients (17%) of the total cohort developed acute kidney injury (AKI) during the first 30 days of CAR T-cell therapy; 77% (27/35) with Stage 1 and 6% (2/35) with severe AKI that required haemodialysis. Twenty-five patients out of 35 (71%) had kidney recovery and 9 of the 10 (90%) patients who did not have kidney recovery died with persistent AKI, including the two patients who required haemodialysis. Notably, there was no statistically significant difference between the NKF and RKF groups in AKI incidence (15% vs. 24%, p = 0.144), AKI severity (Stage 1; 76% vs. 80%, p = 1.0), and AKI full recovery (64% vs. 80%, p = 0.774), respectively. These findings support the growing evidence that post CAR T-cell AKI, irrespective of baseline kidney function, is a prevalent complication that is predominantly mild in severity and has favourable kidney outcome.8-11 To our knowledge, this is the largest study reporting the outcomes of patients who would be ineligible to participate in CAR T clinical trials on the basis of kidney function. We observed comparable outcomes related to toxicity, disease response and survival between the groups with RKF and NKF irrespective of eGFR value. Similar findings highlighting the impact of RKF on CAR T-cell toxicity and survival were reported in three other studies involving axi-cel (with/without tisa-cel and brexu-cel) and in one study involving liso-cel5, 6, 8, 12 (Table 3). CRS≥3 0% vs. 7%, p = 0.610 ICANS ≥3 44% vs. 30%, p = 0.250 PFS rate at 12 months 38 vs. 47, p = 0.360 OS rate at 12 months 50 vs. 69, p = 0.170 Best CR at 12 months 56% vs. 64%, p = 0.530 CRS≥3 12% vs. 8%, p = 0.639 ICANS ≥3 35% vs. 23%, p = 0.372 AKI 41% vs. 21%, p = 0.078 Median PFS 8.9 vs. 9.9 months, p = 0.75 Median OS 10.4 months versus not reached, p = 0.14 DR <20% in two patients 20%–39% in 23 patients 40%–50% in two patients DR versus SD survival Median PFS 8.9 vs. 8.4 months, p = 0.99 Median OS 26.6 versus not reached DR versus SD toxicity CRS ≥3 11% vs. 8%, p = 0.703 ICANS ≥3 30% vs. 24%, p = 0.626 Axi-cel Tisa-cel CRS grade 3 4.5% vs. 0%, p = 0.77 ICANS ≥3 57% vs. 20%, p = 0.21 AKI within 100 days 29 vs. 28%, p = 0.99 Median PFS 8.8 vs. 2.9 months, p = 0.78 Median OS 10 vs. 7 months, p = 0.64 Axi-cel Tisa-cel Brexu-cel 51a (19%) TEAE grade 3–5 92% vs. 76% CRS ≥3 2% vs. 2% NE. ≥3 20% vs. 8% DRb 19 patient (7%) Eliminated dose Two patient (0.7%) 41 (19.5%) CRS ≥3 17% vs. 15%, p = 0.790 ICANS ≥3 46% vs. 41%, p = 0.521 Median PFS 4.3 vs. 4.3 months, p = 0.921 Median OS 18.7 vs. 13.3 months, p = 0.388 OR 89.2% vs. 80.7%, p = 0.33 DR in RKF, 15 (36.5%) 37.5%–50%, two patients 33.3%–25%, 12 patients 20%, one patient DR versus SD toxicity (RKF) CRS ≥3 13.3% vs. 19.2%, p = 1.00 ICANS ≥3 66.7% vs. 34.6%, p = 0.06 Grade ≥3 ICANS >2 days 60% vs. 66.7% (1.00) Axi-cel Tisa-cel In our study, the NRM rate was higher in the RKF group than in the NKF group. Upon assessment of the timing of NRM, we observed the RKF group experienced NRM earlier; NRM occurred with 30–90 days post CAR T cell infusion for 28% of the NKF group versus 78% of the RKF group. This tendency could be related to longer ICU hospitalization for the RKF group, longer grade ≥3 ICANS, and subsequently, death occurring at a later point compared to the NKF group. Considering the direct association of RKF with aging, the higher NRM might be in part related to old age. Furthermore, old age has been identified in other studies as a possible risk factor for NRM and ICANS. However, in our study, we did not observe an association between age and longer grade ≥3 ICANS.13, 14 Recent data suggest fludarabine overexposure is associated with higher ICANS incidence in B-NHL patients treated with CAR T-cell therapy and having a decreased kidney function can contribute to fludarabine overexposure.15, 16 When evaluating the association between LDC dose and longer grade ≥3 ICANS in the RKF group, we did not observe lower ICANS duration in the LDC dose reduction subgroup (Table 3). Nonetheless, as the number of patients who received reduced LDC was small in our cohort and the percentage of dose reduction was heterogeneous, further study of dose reduction effect on outcomes, including ICANS duration, in this population is warranted. There are several other limitations of this study, the main being its retrospective nature. Additionally, the association that we observed between RKF with ICANS and NRM and disease response might be attributed to chance due to the small number of patients with RKF included in the study or due to co-variables for which we did not control. OM and SA prepared the manuscript. PS and SM contributed to designing the study and writing the manuscript. LF and RS provided data analysis and interpretation. AA, RES, RN, CF, JLR, NS, SAS, REC, PK, LJN, MAR, EJS, YN, JW. and SSN provided study materials or data and reviewed the manuscript. None. SA received a research support to institution for clinical trials from Seattle Genetics, Merck, Xencor, Chimagen and Tessa Therapeutics, has membership on Tessa Therapeutic's and Chimagen scientific advisory committee, she serves on Data Safety Monitoring Board for Myeloid Therapeutics; she is a consultant for ADC therapeutics and KITE/Gilead. RES received a research funding from Seagen, BMS, GSK and Rafael Pharmaceuticals. EJS served as a consultant for Partner Therapeutics, Mesoblast, Synthego Corporation, Bayer, ASC Therapeutics, Novartis, Magenta, Cimeio Therapeutics AG, NY Blood Center, Adaptimmune, Navan, Celaid Therapeutics, Zelluna Immunotherapy, FibroBiologics, and Axio, and she has licence agreement with Takeda, Affimed and Syena. SSN received research support from Kite/Gilead, BMS, Allogene, Precision Biosciences and Adicet Bio; served as Advisory Board Member/Consultant for Kite/Gilead, Merck, Sellas Life Sciences, Athenex, Allogene, Incyte, Adicet Bio, BMS, Bluebird Bio, Fosun Kite, Sana Biotechnology, Caribou, Astellas Pharma, MorphoSys, Janssen, Chimagen, ImmunoACT, Orna Therapeutics, Takeda and Synthekine; has stock options from Longbow Immunotherapy, Inc; and has intellectual property related to cell therapy. The study was approved by MDACC Institutional Review Boards and performed in accordance with the Declaration of Helsinki. Patient informed consent was not required given the study was retrospective and posed a minimal risk to participants. Appendix S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.