Abstract Body (Do not enter title and authors here): Background: The associations of specific blood amino acid (AA) metabolite levels with frailty and heart failure have been reported. However, the utility of comprehensive AA profiling in cardiovascular disease (CVD) patients remains unknown.
Aims: This study aimed to determine whether plasma AA profiling through hierarchical clustering analysis (HCA) can identify and clinically define a high-risk population prone to future adverse events.
Methods: We enrolled 267 patients (mean age 74 years, 60% male) with CVD (118 with ischemic heart disease, 85 with valvular heart disease, 33 with cardiomyopathy, and 31 with other CVD). In fasting plasma samples, 39 AA metabolite levels were measured using liquid chromatography-mass spectrometry. Hierarchical clustering was used for data analysis. Major adverse cardiovascular events (MACE) were defined as a composite of all-cause mortality, nonfatal myocardial infarction, ischemic stroke, and rehospitalization for worsening heart failure. During a one-year follow-up, MACE incidence was compared across HCA-based patient clusters using Kaplan-Meier (KM) curve analysis. The cluster with the worst prognosis was characterized clinically compared to other clusters.
Results: HCA enabled patients to be categorized into four clusters (Figure 1A). KM curves showed that Cluster 2 had a significantly worse prognosis than other clusters (log-rank test; p=0.0018) (Figure 1B). AA metabolism in Cluster 2 was characterized by the lowest values of total AA, non-essential AA (NEAA), essential AA (EAA), branched-chain AA (BCAA), BCAA/total AA ratio, and NEAA/EAA ratio. Compared to non-Cluster 2 patients, those in Cluster 2 were older (p=0.004), had lower body weight (p=0.02), a higher prevalence of chronic kidney disease (p<0.0001), higher levels of brain natriuretic peptide (p<0.0001) despite similar left ventricular ejection fraction, lower albumin levels (p=0.001), and more frailty indicators, such as muscle weakness (p=0.002), slower walking speed (p=0.0003), and lower scores on the Short Physical Performance Battery (p=0.002).
Conclusion: Comprehensive blood AA profiling can effectively identify high-risk CVD patient groups and provide relevant clinical characterization for prognosis.