A machine-learning pipeline applied to full-length 16S rRNA sequencing of stool samples from 33 STEMI patients identified five acute-phase gut taxa — Clostridium aldenense, Enterocloster bolteae, bacterium NLAE-zl-G101, Alistipes shahii, and Roseburia sp. — reproducibly associated with primary ventricular tachycardia/fibrillation (VT/VF). A reduced-feature support vector machine model achieved an AUROC of 0.846, with 85.7% sensitivity and 76.9% specificity. Functional enrichment analysis additionally flagged fucose degradation and purine catabolism pathways, with elevated predicted xanthine dehydrogenase — an enzyme whose activity drives uric acid production and oxidative stress.
The gut-heart axis has gained serious traction over the past decade, with TMAO and short-chain fatty acids already linked to cardiovascular risk. This study extends that framework into acute arrhythmia territory, suggesting that peri-infarct microbial dysbiosis — not merely chronic exposure — may shape electrophysiological vulnerability. The purine catabolism finding is mechanistically plausible: xanthine dehydrogenase-derived reactive oxygen species can directly impair cardiac ion channel function. However, the 33-patient cohort, with only 7 VT/VF cases, makes overfitting a serious concern despite FDR correction, and the microbiome signatures are correlational rather than causal. Reverse causation — where hemodynamic collapse reshapes gut flora — cannot be excluded. As a preprint posted to medRxiv and not yet peer-reviewed, these results must be treated as hypothesis-generating. Validation in larger, multicenter cohorts is essential before any clinical translation.