CIBMTR real-world data: about one-third of multiple myeloma patients receiving cilta-cel are frail, with 12-month PFS of 62.7% and OS of 72.8% and doubled ICANS risk

Among real-world recipients of cilta-cel for relapsed/refractory multiple myeloma, about one-third (183/541, 33.8%) met the adapted simplified frailty score definition of frailty, and their 12-month progression-free survival of 62.7% and overall survival of 72.8% were lower than the 75.9% and 90.4% seen in nonfrail patients. The authors note that no prior analysis had specifically examined outcomes of cilta-cel among frail patients, and that most existing real-world data came from ide-cel or from cilta-cel cohorts of fewer than 20 patients; this study addresses that gap with 541 registry patients. Based on prospectively collected multicenter CIBMTR registry data, 541 patients, median follow-up 12 months, analyzed with Kaplan-Meier methods and log-rank tests with 95% confidence intervals reported.

Frailty did not reduce the likelihood of an initial response: overall response rate was 82.8% in frail versus 88.5% in nonfrail patients, and frailty was not independently associated with ORR (OR 1.03) or complete response (OR 1.17) on multivariable analysis. This positions frailty as affecting durability of response rather than the ability to induce an initial response, distinguishing it from assumptions that frailty mainly limits efficacy. Response was assessed per IMWG criteria, competing-risk analysis (Gray's test) accounted for progression, death, or subsequent autologous transplant, and multivariable logistic regression supported the findings.

Frailty was independently associated with higher neurotoxicity risk: any-grade ICANS occurred in 32.2% of frail versus 17.6% of nonfrail patients, with a multivariable odds ratio of 2.01; grade ≥2 ICANS occurred in 11.5% versus 4.8%. The authors note this is consistent with prior real-world analyses showing increased ICANS with CAR-T in frail older adults, but this study is the first to quantify the association in a cilta-cel population using multivariable modeling. ICANS was graded by American Society for Transplantation and Cellular Therapy consensus criteria; multivariable logistic regression adjusted for disease status, marrow burden, prior BCMA therapy, and other covariates, with consistent results in a sensitivity analysis using 20 imputed data sets.

Prolonged cytopenia beyond day 30 was more common in frail patients (30.6% vs. 21.2%), 12-month treatment-related mortality was 6.8% versus 3.6% (p = .12, not statistically significant), while clinically significant infection (49.2% vs. 45.8%) and second primary malignancy (6.6% vs. 3.1%) rates did not differ significantly between groups. These safety data add to the toxicity profile of cilta-cel in frail patients beyond ICANS and suggest that prolonged cytopenia did not translate directly into higher infection rates. Safety endpoints came from systematic registry collection; prolonged cytopenia had explicit laboratory thresholds (ANC <500 × 10⁶/L and/or platelets <20 × 10⁹/L beyond day 30), though the treatment-related mortality comparison did not reach statistical significance.

The results apply to adults with relapsed/refractory multiple myeloma receiving commercial cilta-cel in the real world, particularly those who would not have met CARTITUDE-1 eligibility criteria because of ECOG status, prior BCMA therapy, thrombocytopenia, or neutropenia. They support assessing frailty before infusion and using that assessment to design risk-stratified bridging strategies, cytopenia-mitigation bundles, infection-prevention algorithms, and standardized neurologic monitoring. The authors note that prospective studies are still needed to validate frailty-adapted care pathways for cilta-cel.

As an observational registry study, residual confounding remains possible, for example from clinical decisions around referral, bridging therapy, and lymphodepletion intensity. The frailty definition used an adapted simplified score with HCT-CI in place of the Charlson Comorbidity Index and does not capture all geriatric domains such as cognition or gait speed, which could misclassify risk. Event adjudication for delayed neurotoxicity and infection can vary by center, and median follow-up of 12 months remains modest for late events such as second primary malignancies. In addition, the study included only patients who received infused product meeting quality specifications, which may have selected a relatively fitter population. The authors state that prospective studies are needed to validate frailty-adapted care pathways.