High-dose anakinra to treat refractory immune effector cell-associated neurotoxicity syndrome in CAR T-cell therapy recipients

Abstract BACKGROUND Our group (Gazeau et al., TCT 2023) and others showed the safety and potential efficacy of treatment with the recombinant IL-1 receptor antagonist anakinra in patients with refractory cytokine release syndrome (CRS) and/or immune effector cell-associated neurotoxicity syndrome (ICANS) after chimeric antigen receptor (CAR) T-cell therapy. While anakinra is now routinely used, its efficacy has not been robustly investigated, and the optimal timing of anakinra initiation and factors associated with its efficacy are unknown. Here, we report the outcomes of a large multicenter cohort of patients treated with anakinra for refractory CRS and/or ICANS and analyze factors associated with the efficacy of anakinra to treat refractory ICANS. METHODS We included adults undergoing CAR T-cell therapy for hematologic malignancies at Fred Hutch Cancer Center (FHCC; n = 82) or Oregon Health n = 42) between 2017 to February 2025 who received anakinra for refractory CRS and/or ICANS per institutional guidelines (persistent or worsening despite corticosteroids; N = 124). Anakinra was initiated IV (n = 113) or SC (n = 11) at the dose of 8-10 mg/kg/day in 3 divided doses (FHCC) or 200 mg every 8 hours (OHSU). Cumulative incidence (CI) estimates of time to ICANS resolution were computed with death as the competing risk. RESULTS The most common disease types were large B-cell lymphoma (LBCL; n = 53, 43%), mantle cell lymphoma (n = 25, 20%), and multiple myeloma (n = 21, 17%). The most common CAR T-cell products were axicabtagene ciloleucel (n = 31, 25%), brexucabtagene autoleucel (n = 26, 21%), and tisagenlecleucel (n = 21, 17%). Complete or partial anti-tumor responses at day +28 occurred in 94 patients (89%). The CI of early death at 28 days was 7% (95% CI, 4-13%). Anakinra was initiated for persistent and/or severe ICANS in most patients (n = 106, 85%; grade ≥3: n = 82, 76%). Nearly all (n = 123; 99%) patients received concurrent dexamethasone (dex) with a median total dose of 234 mg (range, 30-690). More than half (n = 67; 54%) received concurrent methylprednisolone (MP) with a median total dose of 3,000 mg (range, 140-9,450). Anakinra was initiated after lack of response to dex in 106 patients (85%) and lack of response to both dex and MP in 14 patients (11%). In 21 patients (17%), anakinra and MP were initiated simultaneously after dex failure. Sixteen patients (13%) received additional therapies, including ruxolitinib (n = 1), siltuximab (n = 3), cetuximab (n = 1), intrathecal chemotherapy (n = 11), and dasatinib (n = 1). ICE scores significantly improved by 48 hours after anakinra initiation (Wilcoxon signed-rank p 0.001). The median time to ICANS resolution from anakinra initiation was 10 days. In patients without ICANS resolution (n = 12), the most common cause of death was refractory toxicity (n = 6). Next, we used cause-specific Cox regression models to identify factors associated with the efficacy of anakinra to treat refractory ICANS. In univariate analyses, follicular lymphoma (FL; reference: LBCL, HR = 4.46, 95% CI, 1.61-12.3, p = 0.004) was associated with shorter time to ICANS resolution from anakinra initiation. We could not confirm associations between time from ICANS onset to anakinra initiation, daily/total MP or anakinra dose, or clinical scenarios (initiation after dex failure, after dex and MP failure, simultaneous initiation with MP after dex failure) and time to ICANS resolution from anakinra initiation. Higher daily dex dose was associated with shorter time to ICANS resolution from CAR T-cell infusion (time-dependent td HR = 1.19, 95% CI, 1.04-1.36, p = 0.013). In a multivariable model including daily dex dose, disease type, and CAR costimulatory domain, daily dex dose (tdHR = 1.25, 95% CI, 1.05-1.48, p = 0.014) and FL (HR = 3.88, 95% CI, 1.08-13.9, p = 0.038) remained independently associated with shorter time to ICANS resolution from CAR T-cell infusion. CONCLUSION To our knowledge, this is the largest study of CAR T-cell therapy recipients treated with anakinra for steroid-refractory CRS/ICANS. Anakinra was primarily administered for steroid-refractory ICANS with improved ICE scores by 48 hours and a median time to resolution of 10 days. Higher dex doses remained associated with shorter time to ICANS resolution, highlighting that corticosteroids remain the cornerstone of treating refractory ICANS. More effective strategies for anakinra-refractory ICANS are critically needed.