Association of Fried’s frailty phenotype, but not the clinical frailty scale, with overall survival in older adults with acute myeloid leukemia.

e23222 Background: Frailty is a geriatric syndrome marked by increased vulnerability to stressors. “Fitness,” is utilized to guide treatment in acute myeloid leukemia (AML) where the intensity of initial induction may be high (anthracycline and cytarabine-based), intermediate (venetoclax combinations), or low (single hypomethylating agent or targeting therapy). Fried’s Frailty Phenotype (FP) combines subjective (self-reported exhaustion, weight loss > 5%, activity level) and objective (4-meter walk speed and grip strength) measures to categorize patients into fit, pre-frail, and frail. In prior work, we demonstrated an association of FP with survival for newly diagnosed older adults with AML. Clinical Frailty Scale (CFS) is a purely subjective score that utilizes chart data to stratify the degree of fitness. The incorporation of objective measures into clinical practice is limited due to time constraints, we sought to determine if CFS could similarly identify frailty and risk-stratify older adults with AML. Methods: Eighty-three patients 60 years or older with a new diagnosis of AML were enrolled on an IRB approved study from September, 2018 to May, 2023 at the Hospital of the University of Pennsylvania. FP was performed prior to initiation of therapy. A CFS certified trainee blinded to FP scores retrospectively reviewed electronic medical records to assign a CFS score. CFS was collapsed into variations defined as CFS fit (1-3), pre-frail (4), and frail (5 or greater); CFS Collapse 1 fit (1-2), pre-frail (3-4), and frail (5 or greater), and CFS Collapse 2 fit (1-4) and mild to moderately frail (5-6). Variations were based on the scale’s descriptions and literature. Results: Median patient age was 71 (range 60-91) years. By European LeukemiaNet 2017, 64% of patients had adverse-risk AML. 39% of patients received IC, 52% received HMA/ven, 6% received HMA alone or targeted therapy, and 2% received supportive care. By FP, 18% were fit, 38% were pre-frail, and 45% were frail. FP was significantly associated with OS (p = 0.047): 2-year OS was 57%, 42%, and 21% in fit, pre-frail, and frail. By CFS, 43% were fit, 32% pre-frail, and 24% frail. CSF was not significantly associated with OS (p = 0.495); 2-year OS was 24%, 35%, and 19% for fit, pre-frail, and frail patients, respectively. CSF collapse 1 (p = 0.249) and 2 (p = 0.070) were not significant. Conclusions: FP identified more patients as frail and was associated with a doubling of early mortality and inferior 2-year OS. CFS identified more patients as fit and was not associated with 2-yr OS. The discrepancy emphasizes the importance of objective measurements to capture frailty and the difficulty in determining frailty based on chart reviews in an acutely ill population. Future work to streamline objective measurements in frailty assessments to increase utilization in clinics given its likely importance for accurate risk-stratification.