Abstract
Heart Failure (HF) impacts approximately 64 million people globally. While overall incidence of HF is relatively stable across countries, the overall number of HF patients is increasing due to aging populations. Many articles examine the microbiome in HF, however, studies from humans have not been analyzed systematically. The aim of this meta-analysis is to bridge this gap by analyzing previously published data on human HF patients with untargeted metabolomics to understand whether microbially-mediated metabolites are consistently important for HF status. A systematic survey of the literature identified 708 articles discussing HF, the microbiome, and metabolomics. Of these, 82 were primary studies of HF patients, 61 studied human adults, 23 included an untargeted metabolomics measure, and 3 studies had data that was usable and publicly accessible. These studies include a GCMS study from stool, NMR of saliva and exhaled breath condensate, and LCMS from left ventricle of HF patients undergoing transplantation and unused donor hearts. Significant differences were observed from PCA between HF and controls for stool and left ventricle, but not saliva or EBC samples. OPLS-DA was conducted for stool and ventricle samples, and further revealed significant group differences. Univariate testing with FDR correction revealed 8 significant microbially-relevant metabolites (p < 0.005 after correction), most notably asparagine from left ventricle and 2-methylbutyryl carnitine from stool. Though there is much discussion of the microbiome in health outcomes in HF, there is limited research from human populations. Some microbial co-metabolites from both stool and heart were significantly associated with HF.