Abstract
This study investigates the role of gut microbiota in cardiovascular diseases, with an additional focus on pro-atherogenic metabolites. We use advanced network analysis and machine learning techniques to identify key microbial features linked to coronary artery disease (CAD) and heart failure with reduced ejection fraction (HFrEF). This cross-sectional study included 189 participants divided into three groups: coronary artery disease (n = 93), heart failure with reduced ejection fraction (n = 43), and controls (n = 53). Assessments included physical exams, echocardiography, dietary surveys, blood analysis, and fecal analysis. Gut microbiota composition was analyzed using next-generation sequencing (NGS) and quantitative polymerase chain reaction (qPCR). Statistical analysis methods for testing hypotheses and correlations, alpha and beta-diversity analyses, co-occurrence networks, and machine learning were conducted using Python libraries or R packages with multiple comparisons corrected using the Benjamini-Hochberg procedure. Significant gut microbiota alterations were observed, with higher Bacillota/Bacteroidota ratios in CAD and HFrEF groups compared to controls (p < 0.001). Significant differences were observed in α-diversity indices (Pielou, Chao1, Faith) between disease groups and controls (p < 0.001). β-diversity analyses also revealed distinct microbial profiles (p = 0.0015). Interestingly, trimethylamine N-oxide (TMAO) levels were lower in CAD and HFrEF groups compared to controls (p < 0.05), while indoxyl sulfate (IS) levels were comparable between the study groups. Co-occurrence network analysis and machine learning identified key microbial features linked to these conditions, highlighting complex interactions within the gut microbiota associated with cardiovascular disease.