CD19-directed chimeric antigen receptor T-cell (CAR T) therapy has significantly improved outcomes for patients with relapsed or refractory B-cell non-Hodgkin lymphoma (R/R B-NHL) but is frequently complicated by immune effector cell-associated neurotoxicity syndrome (ICANS), a major cause of morbidity and mortality. Preclinical and early-phase clinical studies suggest that interleukin-1 blockade with anakinra may mitigate ICANS without impairing CAR T efficacy. However, real-world data evaluating the efficacy and safety of prophylactic anakinra remain limited. We performed a retrospective analysis comparing CAR T toxicities and clinical outcomes among patients treated with prophylactic anakinra versus controls using inverse probability of treatment weighting (IPTW), hypothesizing that anakinra prophylaxis would be associated with decreased severe ICANS. In 2023, our institution implemented a policy for anakinra prophylaxis for high-risk patients based on promising early-phase data. We conducted a single-center retrospective cohort study of 176 adult patients with R/R B-NHL who received CD19 CAR T therapy between 2018 and 2025. The analysis was restricted to patients meeting institutional criteria for anakinra prophylaxis (age ≥65 yr or receipt of a CD28 costimulatory domain CAR T product), excluding those with baseline ICE score <8. Patients treated with anakinra in the postpolicy era were compared with patients treated prior to the implementation of the policy. IPTW was used to balance baseline clinical and disease-related covariates between groups. After IPTW, patients receiving prophylactic anakinra had a higher incidence of any-grade ICANS compared with those who did not (40.0% versus 23.0%, P = .03), while rates of grade ≥3 ICANS were similar between groups (P = .62). In multivariate regression, anakinra prophylaxis was associated with increased odds of any-grade ICANS (aOR 3.22; 95% CI 1.39 to 7.46) but not grade ≥3 ICANS (aOR 1.84; 95% CI 0.61 to 5.5). Rates of all-grade cytokine release syndrome (CRS) were comparable (P = .74); however, in multivariate regression, anakinra prophylaxis was associated with increased odds of grade ≥3 CRS (aOR 17.83; 95% CI 1.30 to 245.21), though the estimate was imprecise due to sparse events. Grade ≥3 infections were more common in the anakinra cohort (21.5% versus 7.3%; P = .01) by day 30 and by day 90 (27.1% versus 10.4%; P = .01). Grade ≥3 infections remained associated with use of anakinra in multivariate regression (day 30: aOR 7.08; 95% CI 1.90 to 26.30, P = .004) (day 90: aOR 5.59; 95% CI 1.83 to 17.04; P = .003). No differences in response rates, event-free or overall survival were observed between groups. In this single-center historical comparison of high-risk patients receiving CD19 CAR T cells, prophylactic anakinra did not reduce severe ICANS and was associated with increased immune-mediated toxicities and infectious complications. Causal inference is limited by policy-driven treatment assignment, residual temporal confounding, and limited practical overlap. These findings suggest the need for caution in routine use of anakinra prophylaxis outside of clinical trials and underscore the importance of prospective studies to better define optimal toxicity mitigation strategies.