153 Agreement among frailty assessment tools and their association with quality of life in older patients with heart failure: a comparison of four instruments

Introduction Frailty is highly prevalent in older patients with heart failure (HF) and is associated with poor clinical outcomes. Many tools are available for assessing frailty, however, very few studies have simultaneously evaluated different frailty tools and their relation to quality of life (QoL) within the same cohort of patients. We studied the agreement between 4 commonly used frailty tools and their association with QoL and functional dependence amongst older patients with HF. Methods We recruited older adults (aged ≥ 65) with HF from a hospital-based HF clinic and assessed their frailty using 4 tools: Clinical Frailty Scale (CFS), Fried Frailty Phenotype (FP), Short Physical Performance Battery (SPPB) and Edmonton Frail Scale (EFS). Patients were classified as frail and non-frail according to cut-offs specified by each instrument. Functional dependence and QoL were assessed using Barthel Index (BI) and Kansas City Cardiomyopathy Questionnaire (KCCQ-12), respectively. Kappa statistics was used to assess the level of agreement between the frailty tools. Results We studied 150 participants [median age: 80 (range 65–94) years, 65% male]. Thirty-eight percent had HF with reduced ejection fraction; One fifth had NYHA III/IV symptoms; median NT-proBNP 2137 ng/L (IQR = 1120–5028). The SPPB model identified the highest proportion of patients as frail (69%), followed by CFS (67%), FP (57%) and EFS (47%) (table 1). Forty-five patients (30%) were classified as frail by all 4 tools (figure 1). There was moderate agreement between SPPB vs CFS (ĸ = 0.49) and SPPB vs FP (ĸ = 0.43) and fair agreement between FP vs CFS (ĸ = 0.29), FP vs EFS (ĸ = 0.25), CFS vs EFS (ĸ = 0.27) and SPPB vs EFS (ĸ = 0.220), all p < 0.001 (table 2). Of the frailty tools, CFS had the strongest correlation with BI (r = -0.703, p < 0.001), whilst EFS had the strongest correlation with KCCQ-12 (r = -0.627, p < 0.001). Worse frailty status determined by any frailty tool was significantly associated with poorer QoL and functional dependence after adjusting for age and NYHA class. Conclusion SPPB had moderate agreement with FP and CFS while EFS had low agreement with other frailty tools. Frailty was significantly associated with functional dependence and QoL in older patients with HF. Our observations have implications for the application of clinical frailty tools in clinical practice and in research. Conflict of Interest None