Predictive Biochemical Model of Frailty and Mortality After Acute Myocardial Infarction

Abstract

Frailty, characterized by reduced resistance to stressors, is associated with adverse outcomes in patients with myocardial infarction. The Fried score is commonly used to assess frailty but has several limitations. This study aimed to evaluate the relation between frailty and blood biomarkers and their predictive value for long-term mortality using a biochemical model. A total of 2 cohorts of elderly patients (>65 years old) hospitalized for acute coronary syndrome were included. Geriatric assessments and several blood biomarkers were measured. The predictive models for frailty were developed using logistic regression. The survival models were also developed using Cox regression. Among 466 patients, 9 biomarkers were significantly associated with frailty. Between these biomarkers, white blood cells count, hemoglobin, and fibrinogen showed the highest predictive power. Model 1, without growth differentiation factor 15 (GDF-15), showed a better accuracy in predicting the mortality than the Fried score. Model 2, with GDF-15, had a stronger correlation with frailty but had a lower predictive power for survival. Frailty is associated with dysregulation in the physiological systems and several biomarkers were linked to this fact in our study. However, the inclusion of GDF-15 did not significantly improve the model's predictive ability. Frailty assessment using blood biomarkers can provide valuable prognostic information in elderly patients with acute coronary syndrome.