The impact of granulocyte colony stimulating factor on patients receiving chimeric antigen receptor <scp>T</scp>‐cell therapy

Abstract

Infectious complications following immunosuppressive therapy increase morbidity and mortality in patients combatting malignancy.1 The inverse relationship between circulating leukocyte levels and infection potential places patients with prolonged and profound neutropenia at greatest infection risk.2 Granulocyte colony stimulating factor (GCSF) after aggressive chemotherapy is a supportive measure recommended to reduce neutropenia intensity and ultimately abrogate infection. Chimeric antigen receptor T-cell (CAR-T) therapy represents a major advancement in the management of refractory or relapsed malignancies.3 Patients receiving CAR-T therapy have multiple factors that impair their immunity and increase infection risk.4 The theoretical potential for GCSF to exacerbate cytokine release syndrome (CRS) prompted guidelines to recommend against administering GCSF within 14 days after CAR-T infusion.5 However, few studies exist that characterize outcomes after CAR-T therapy and the ultimate impact of concurrent GCSF remains unknown.6 Our primary objective was the comparison of neutrophil kinetics between patients who did and did not receive GCSF after axicabtagene ciloleucel. Secondary outcomes included comparing CAR-T therapy-related toxicity characteristics and infection rates. This Mayo Foundation Institutional Review Board approved retrospective analysis evaluated consecutive, adult cancer patients who received commercial axicabtagene ciloleucel between January 2018 and December 2020. Patients actively participating in a CAR-T clinical trial, considered a vulnerable population, or refusing research participation were excluded. All patients provided written informed consent or had consent provided for them by their legal power of attorney prior to data collection. The electronic health record (EHR) data was compiled utilizing a standardized form developed by the investigators. Baseline criteria were established immediately prior to lymphodepleting chemotherapy. Lymphoma diagnosis with subtyping was reported according to World Health Organization classification criteria.7 Relevant laboratory values, medications, and microbiology were abstracted from the time of CAR-T infusion through disease relapse, death, or a maximum of 100 days after infusion. Neutropenia and severe neutropenia were defined as an absolute neutrophil cell (ANC) count below 0.5 × 109 cells/L and 0.1 × 109 cells/L, respectively.1 The ANC recovery occurred when the count exceeded and stayed above 0.5 × 109 cells/L cells for two consecutive days. Also, GCSF was originally prescribed and administered throughout the period of neutropenia until ANC recovery. In April 2019, our institutional cellular therapy practice committee changed from ANC-based GCSF administration, where prescribing occurred at ANC of 0.5 × 109 cells cells/L and GCSF was discontinued after two consecutive days where an ANC measured above 0.5 × 109 cells/L, to an infection-based indication with GCSF prescribing reserved for patients presenting after CAR-T infusion with febrile neutropenia and an increased concern for infection after a thorough evaluation. Both CRS and Immune effector cell-associated neurotoxicity syndrome (ICANS) events were graded according to published criteria.8 An infectious episode was defined as any positive culture from any site. Catheter-related infections were defined in accordance with previously published guidelines. Patients were considered positive for oral candidiasis based on clinical diagnosis after oral examination and symptom assessment. Invasive fungal and viral infections met consensus definitions and institutional methods of detection. Baseline patient characteristics and laboratory values were summarized with counts and percentages for categorical variables, and with medians and interquartile ranges (IQR) for continuous variables. Distribution of baseline variables and outcomes according to GCSF use were assessed with t tests, Kruskal-Wallis, chi-square test, and Fisher Exact Test as appropriate. Kaplan-Maier estimation was used to calculate the cumulative incidence of infection with a 95% confidence interval. Univariate Cox proportional hazard models, with time dependent covariates when appropriate, were used to demonstrate association between the independent variables and outcomes. Analyses were conducted using R (Version 3.6.3, R Foundation for Statistical Computing, Vienna, Austria). Note, p values less than 0.05 were considered statistically significant. The 70 included patients had a median (IQR) age of 59.4 (44.3, 63.9) years at the beginning of lymphodepletion, 45 (64%) were male, and 63 (90%) were Caucasian. A diagnosis of diffuse large B-cell lymphoma was most common (n = 44, 64%) and transformed follicular lymphoma was the second most common (n = 14, 20%). Baseline characteristics and laboratory parameters were similar between groups with the exception of median body surface area (BSA) which was greater in the patients who received GCSF (2.0 vs. 1.9, p = 0.04). Lymphodepletion was universally cyclophosphamide 500 mg/m2 with fludarabine 30 mg/m2. Median duration of follow up after CAR T-cell infusion was 5.3 (3.8, 12.7) months with 22 patients still in remission at day 100 post CAR T-cell infusion. There were 32 patients who received axicabtagene ciloleucel prior to April, 2019 when GCSF was prescribed for the ANC-based indication and 38 patients who received axicabtagene ciloleucel after the practice change to an infection-based indication for prescribing GCSF. Thirty (94%) of patients received GCSF prior to the change in practice with five (13%) patients prescribed GCSF after the change was implemented. All GCSF prescribing occurred in accordance with our institutional protocol before and after the practice change. With too few patients prescribed GCSF according to an infection-based indication to provide any meaningful analysis, we report our results simply based on GCSF use. So, GCSF was prescribed to 35 (50%) patients for a median of eight (IQR: 6, 13) doses with a median cumulative dosage of 3840 mcg (IQR: 1920–6120) and median time to first dose of 3 days (IQR: 1, 4.4) post CAR T-cell infusion. There were significantly fewer total days of neutropenia (6.5 vs. 10, p = 0.05) and median days of the first neutropenia episode (4.5 vs. 9, p < 0.01) with GCSF use. However, total days of severe neutropenia was similar between groups (4 vs. 6, p = 0.20) and median days of severe neutropenia in the first episode was not statistically significantly reduced (4 vs. 5, p = 0.06). Additional neutropenia-related outcome details are presented in Table 1. Patients had a similar incidence of CRS following GCSF compared to those without GCSF use, 86% versus 88%, respectively. There was no significant difference in CRS severity between groups (median grade: 2 vs. 2); however, duration of CRS was significantly longer for those prescribed GCSF compared to no GCSF, 8 versus 4.5 days (p < 0.01), respectively. Additionally, the incidence, severity, and duration of ICANS were similar between those groups. The use of toxicity management medications, particularly tocilizumab and corticosteroids, was similar between those who were and were not prescribed GCSF. Further information on toxicity-related outcomes are summarized in Table 1. There was a lower overall risk of infection (HR 0.80, 95% CI: 0.41–1.59, p = 0.53) and lower risk of bacterial infection (HR: 0.61, 95% CI: 0.24–1.57, p = 0.30) with GCSF use; but these were not statistically significant (Figure S1). The patients prescribed GCSF in our cohort of adult patients receiving axi-cel for lymphoma treatment had neutropenia rates of any grade that were slightly higher than reported in previous studies; however, this may reflect the differences between clinical trial enrollment criteria and our real-world population.3 Interestingly, those with GCSF use were significantly more likely to experience more neutropenic episodes; however, the total days of neutropenia and median duration of neutropenia during the first neutropenic episode were significantly reduced. Unfortunately, total days of severe neutropenia and median days of severe neutropenia in the first episode were similar between groups. This raises the possibility that ANC values of more than 0.5 × 109 cells/L for two consecutive days is a less than optimal criteria for discontinuing GCSF support after CAR-T infusion. Both CRS and ICANS median severity grade as well as grade 3 or higher rates were similar between those prescribed and not prescribed GCSF. This is concordant with other similar evidence demonstrating no significantly increased risk of CAR T-cell related CRS or ICANS when prescribing GCSF.6 Notably, the time to CRS recovery was longer for patients who received GCSF. The significance of this finding is unknown as corticosteroids and tocilizumab use for CAR-T associated toxicity management was similar between those who were and were not prescribed GCSF. Risk of overall and bacterial infection was lower with ANC-based initiation of GCSF, although non-significant possibly due to small sample size. Ultimately, we believe that GCSF use is reasonable in the CAR T-cell patient population and the potential benefit for using GCSF as well as the optimal timing after CAR T-cell infusion requires further investigation. This publication was supported by CTSA Grant Number UL1 TR002377 from the National Center for Advancing Translational Science (NCATS). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Dr. N. Nora Bennani: Has participated on advisory boards for Purdue Pharma, Verastem Oncology, Acrotech Biopharma, and Sea Gen, Inc. Dr. Yi Lin: Has served as a consultant or received research funding from Kite/Gilead, Celgene, JUNO, Bluebird Bio, Janssen, Legend BioTech, Gamida Cells, Novartis, Merck, Takeda. Has participated on an advisory board for Sorrento: Data and Safety Monitoring Board (DSMB). Jason N. Barreto participated in protocol concept and design, data collection, data analysis, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Radhika Bansal participated in protocol concept and design, data collection, data analysis, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Matthew A. Hathcock participated in protocol concept and design, data collection, data analysis, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Corina J. Doleski participated in data collection, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Justin R. Hayne participated in data collection, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Tuan A. Truong participated in data collection, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Adrienne N. Nedved participated in protocol concept and design, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Stephen M. Ansell participated in protocol concept and design, data analysis, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. N. Nora Bennani participated in protocol concept and design, data analysis, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Jonas Paludo participated in protocol concept and design, data analysis, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Jose C. Villasboas participated in protocol concept and design, data analysis, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Yi Lin participated in protocol concept and design, data collection, data analysis, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Dr. Lin contributed equally w/ Dr. Johnston as co-senior authors. Patrick B. Johnston participated in protocol concept and design, data analysis, data interpretation, and manuscript creation involving critical writing and revising of the intellectual content. Dr. Johnston contributed equally w/ Dr. Lin as co-senior authors. The deidentified data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. Figure S1. Cumulative risk of infection in CAR T-cell patients according to colony stimulating factor use. (A). Cumulative risk of any infection. (B) Cumulative risk of bacterial infection. 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.

Topics & Concepts