Comparative clinical frailty scale and hospital frailty risk score in identifying frailty and predicting mid-term outcomes in older patients with acute coronary syndrome: a multicenter cohort study in Vietnam

Background We aimed to compare the agreement between two common frailty assessment tools, Clinical Frailty Scale (CFS) and Hospital Frailty Risk Score (HFRS), and their ability to predict mid-term adverse outcomes in older patients admitted with acute coronary syndrome (ACS). Methods We conducted a prospective analysis of patients aged >= 60 admitted with ACS at multiple centers in Vietnam between July 2022 and June 2023. A cross-tabulation method was used to describe the correlation between CFS and HFRS. To test the predictive accuracy of HFRS for identifying patients with frailty according to CFS, we evaluated the area under the curves of receiver operating characteristic (ROC) analysis. Youden J index was used to identify a new optimal probability threshold for HFRS. We employed Cox regression models to investigate the association between frailty assessed by CFS, HFRS (using both old and new cut-offs), and 9-month mortality. Results We included 504 older patients admitted with ACS (median age 72.7 years; male: 59.9%). The correlation between CFS and HFRS was fair (AUC = 0.787, p < 0.010). HFRS had a sensitivity of 39.7% and a specificity of 79.2% to detect frailty based on CFS classification. The new optimal probability threshold of HFRS (>= 1.15 points) improved the instrument's performance with a significantly higher sensitivity of 90.2%. While frailty categorized by HFRS with the original cut-off did not impact mid-term all-cause and cardiovascular mortality, frailty according to CFS and HFRS with the new threshold was shown to be a predictor of mid-term all-cause and cardiovascular mortality (HR = 4.48, p < 0.001 vs. HR = 2.29, p = 0.001; HR = 5.19, p < 0.001 vs. HR = 1.99, p = 0.020). Conclusions Although a fair correlation existed between the CFS and the HFRS in older patients with ACS, HFRS demonstrated limited predictive validity for mid-term mortality. We advocate for a revised cutoff (HFRS >= 1.15 points) to enhance its sensitivity and predictive accuracy. Future research should prioritize the integration of additional clinical biomarkers and conducting longitudinal studies to assess the efficacy of targeted interventions informed by frailty scores, ultimately striving to improve outcomes in this vulnerable population.