Nucleus, CAR T

An Electronic Frailty Index Is Associated With Outcomes in Older Adults Receiving CAR T Therapy

Authors from Massachusetts General Hospital and Brigham and Women’s Hospital have found that a frailty score computed automatically from codes already sitting in the electronic health record separates older chimeric antigen receptor T-cell (CAR-T) recipients who do well from those who do not, with no added clinic assessment. Reported in Blood Advances, the study identifies very frail patients as a distinct high-risk group within an already selected older population.

Multiple myeloma and B-cell non-Hodgkin lymphoma (NHL) are diseases of older adults, yet registration trials enrolled younger patients and largely excluded those with comorbidity, and fewer than 20% of eligible older adults receive CAR-T. Chronologic age predicts neither efficacy nor toxicity reliably, and existing frailty assessments are heterogeneous and difficult to apply in real time. The investigators retrospectively reviewed 260 patients aged 65 years or older treated with a commercial CAR-T product at a single academic health system over an eight-year period, scoring each with the electronic frailty index (eFI), a validated claims-based measure that accumulates 31 age-related deficits from routine encounter data and sorts patients into robust, prefrail, frail, and very frail categories. Survival, toxicity, and healthcare utilization endpoints were modeled against frailty category with adjustment for age, performance status, product, and disease factors.

Median age was 73 years (range 65 to 91), and 70% had B-cell NHL. Over a median follow-up of 28.3 months, median overall survival (OS) was 44.2 months in robust/prefrail patients, 34.6 months in frail, and 19.5 months in very frail patients (p = 0.02), with median event-free survival (EFS) of 20.2, 12.7, and 5.9 months (p < 0.001). On multivariable analysis, very frail status predicted worse OS (hazard ratio 1.64; 95% CI, 1.01 to 2.67; p < 0.05) and EFS (HR 2.02; 95% CI, 1.34 to 3.04; p = 0.001), while frail status did not (p = 0.45 and p = 0.14). Odds of grade 3 or higher cytokine release syndrome or immune effector cell-associated neurotoxicity syndrome rose in frail (odds ratio 3.28; p = 0.048) and very frail patients (OR 4.84; p = 0.006), both of whom had fewer days alive and out of hospital (β = −2.37, p = 0.026; β = −3.79, p < 0.001). The eFI correlated with the Hematopoietic Cell Transplantation-specific Comorbidity Index (HCT-CI; r = 0.31, p < 0.001), but only the eFI predicted OS when both were modeled together. Frailty, not age, marks the older patients who fare poorly after CAR-T, and because the eFI is calculated passively from existing records it could be embedded in the electronic health record to flag that risk before treatment rather than after.

Reference

Enns JP, Korst MR, Horick N, et al. An electronic frailty index is associated with outcomes in older adults receiving CAR-T therapy. Blood Adv. 2026. https:doi.org/10.1182/bloodadvances.2026021361