Work

The gut microbiome and vascular disease — A systematic review of mechanistic insights and implications for vascular surgery

Year: 2026

Type: review

Abstract: Objective Observational studies repeatedly link gut microbial features and metabolites to vascular disease, but it is unclear which associations are reproducible, which survive confounder adjustment, and how they relate to experimental evidence. This review asks which taxa and metabolites are consistently associated across independent human studies; whether population, phenotype, method and sample type explain divergent findings; and where the human data converge with, diverge from, or leave untested the animal and interventional literature. Methods PubMed/MEDLINE, Embase, Web of Science and the Cochrane Library were searched from inception to 31 December 2025 for human studies of gut microbiome composition or gut-derived metabolites reporting atherosclerotic, aneurysmal, peripheral arterial, calcific or arterial-stiffness outcomes. Screening and duplicate extraction followed PRISMA 2020. Design was classified explicitly, the covariates in each primary analysis were extracted, and risk of bias was appraised with the Newcastle-Ottawa Scale or the Joanna Briggs Institute checklist. Findings were then contrasted with animal and human interventional literature identified by targeted searching. Results All 29 included reports were observational (19 cross-sectional, 3 case-control, 7 cohort) and none was interventional. Reproducible signals were enrichment of Proteobacteria-affiliated genera (5 reports) and of oral-origin organisms (3 reports), and depletion of butyrate producers in carotid disease. The carnitine-pathway intermediates gamma-butyrobetaine and trimethyllysine were independently associated with cardiovascular death and with major adverse limb events (HR 1.93, 95% CI 1.35-2.76) in two independent cohorts. TMAO was outcome-specific: associated with 5-year mortality in peripheral arterial disease (adjusted HR 2.06, 95% CI 1.36-3.11) and with ankle-brachial-index-defined disease in diabetes (OR 1.051), but lower in large-artery stroke than in controls and unrelated to carotid or limb outcomes once renal function was modelled. Sixteen of 29 reports fitted a multivariable model and only 3 entered renal function; compositional findings were almost exclusively unadjusted. Risk of bias was good in 11 reports, moderate in 16, low in 2. Animal models causally implicate TMAO in atherogenesis and aneurysm formation and show butyrate and Roseburia intestinalis to be protective, but germ-free experiments and the only human faecal transplantation trial are discordant. Conclusions A small number of gut microbial signals are reproducible, but most are unadjusted for the confounders — renal function in particular — that plausibly generate them. The carnitine-pathway metabolites carrying the strongest prognostic signal in surgical populations have never been tested in an animal vascular model. Microbiome profiling has no current role in vascular surgical practice.

Source: JVS Vascular Science

DOI: 10.1016/j.jvssci.2026.100454