Abstract Background Recent advances support the concept that pre-persistent Atrial Fibrillation (AF) and paradoxical embolism do not explain the wholeness of embolic strokes, thus suggesting a broad Atrial Cardiomyopathy (AC). Pathogenic mechanisms underlying AC are still largely unknown. Folate cycle disorders are a yet underrated dysmetabolism. MTHFR-inherited defects could hinder both endothelial and circulating endothelial progenitor cell (EPCs) functioning, therefore providing one-shot explanation to both atrial stasis (increasing atrial fibrosis and generating atrial hypocontractility even in sinus rhythm) and endothelial dysfunction (ED). Purpose This study aims to enquire for the hypothesis that: 1) atrial fibrosis (AFib – intended as a relative percentage of low voltage area in the context of left atrial endocavitary voltage mapping) would relate to folate cycle disorders (intended as both: a)MTHFR C677T inherited mutations and b)bone-marrow function disorders, here referring to erythropoiesis diversions) and 2) AF patients would show dysfunctional EPCs. Methods We studied 59 consecutive patients admitted to the cardiology Unit of the General Hospital “F.Miulli”, with preserved EF, subjected to AF ablation. AFib was quantified by bipolar peak-to-peak voltage areas (<0,5 mV) with respect to the wholeness of the picked voltage points. Blood count cell was evaluated by a commercial laboratory test. EPCs isolation and characterization were performed by Ficoll-Hypaque gradient and following flow cytometry analysis for cell surface antigens: CD45, CD34, CD133, VEGFR2 and KDR. EPCs functional wound healing assay was performed to determine the number of EPCs migrated to the “wound”, measuring the percentage of relative wound closure compared with matched-controls. Results Baseline characteristics did not differ between Sample and Control groups. % of Afib significantly differs between C677T MTHFR homozigosis patients (n=15) with respect to non-C&//T MTHFR homozygosis patients (n=44 p < 0,02). Once univariate analysis was performed, subsequent multivariate analysis highlights highest fit once merged RBC, RDW-SD and folates values were inputed: Goodness of fit was proper, modelling good (Low figure - R2=0,39; p=0,0001). Either RBC, RDW-SD and folates coefficient reached significance (p < 0,0001; p < 0,01; p < 0,05 respectively). Number of EPCs significantly differs between AF patients and matched controls (p < 0,001). Conclusions Our findings support the hypothesis that genetically determined folates dysmetabolism (MTHFR dysfunction) promotes AFib via a complex cardiac-bone marrow networking involving circulating EPCs and unraveled by erythropoiesis diversions. Such results suggest a pathogenic role of folate cycle disorders in the AC development