Comparison of frailty instruments for predicting mortality and prolon ged hospitalization in acute coronary syndrome patients

OBJECTIVE: To evaluate the relative strengths of 3 frailty assessment instruments for predicting mortality and prolonged hospitalization in acute coronary syndrome patients. DESIGN: Prospective cohort study. SETTING: Acute cardiac care units in New Zealand. PARTICIPANTS: 1174 patients >70 years of age hospitalized with an acute coronary syndrome. INTERVENTIONS: The Clinical Frailty Scale (CFS), Edmonton Frail Scale (EFS) and Fried Criteria (Fried), were completed during hospital admission following an acute coronary syndrome when the patient was clinically stable. PRIMARY AND SECONDARY OUTCOME MEASURES: All-cause mortality over the next ~5 years and hospitalization for >10 days in the next year determined from national administrative data. RESULTS: During median follow-up of 5.1 years there were 353 deaths. Harrell's C-statistic for mortality for EFS was 0.663, Fried 0.648 and CFS 0.640 (p<0.001 for all). C-statistics for hospitalization >10 days (n = 267, 22%) were EFS 0.649, Fried 0.628, and CFS 0.584 (p<0.001 for all). Associations between increase in frailty scores and mortality were graded including in patients not classified as frail. The hazard ratio (HR) for mortality, adjusted for age and sex, for patients with an EFS score ≥9 (n = 197) compared to ≤2 (n = 331) was 5.0 (95% CI: 3.4-7.4). In models which included the Euroscore II or GRACE risk scores the EFS improved risk discrimination for both mortality and prolonged hospitalization more than the CFS and Fried. CONCLUSION: In older patients assessed following an acute coronary syndrome the EFS discriminated the risk of all cause mortality and prolonged hospitalization better than the CFS and Fried tests, and improved risk discrimination when added to clinical risk scores.