Abstract Background The pathophysiology of atrial fibrillation (AF) involves a continuum of structural, hemodynamic, and electrophysical remodeling that coexist in a complex interplay, often involving increased diffuse myocardial fibrosis. Purpose To examine whether low dose spironolactone can reduce the extent of diffuse fibrosis in the left atrium (LA) and left ventricle (LV), evaluated using post-contrast T1 myocardial relaxation time (T1 time) and myocardial strain derived from cardiac magnetic resonance imaging (CMR), in patients with paroxysmal and persistent AF. Methods CMR sub-study population of 100 individuals from INSPIRE-AF, an investigator-initiated, randomized, controlled, double-blind study (NCT02764619). Study participants with paroxysmal (n=49) and persistent (n=51) non-valvular AF and LV ejection fraction ≥45% were randomized to spironolactone 25 mg once daily (SPL-group, n=49) or placebo (PL-group, n=51) in addition to standard treatment. CMR scans were conducted at inclusion and after 12 months of treatment. LA and LV measures, including volumes, ejection fraction, and peak systolic longitudinal strain were assessed using cine CMR feature tracking. T1 times were estimated using a previously validated modified look-locker inversion-recovery sequence. Commercially available software (Circle, Calgary, Canada) was used. All participants were followed for recurrences of AF, and all recurrences were documented with a standard 12-lead electrocardiogram. The associations between changes in mean volume, strain, and T1 times were examined using multivariable linear regression, adjusted for baseline age, sex, and AF type. Results LV post-contrast T1 time increased significantly in the SPL group (mean increase in SPL=1,97 ms [95% CI: 0,169; 3,323] vs. PL=0,21 ms [95% CI: −0,67; 1,11], p=0,027 for difference). However, the mean change of post-contrast LA T1 time did not differ between groups. LV peak longitudinal strain was also significantly improved in the SPL group (mean decrease in SPL: 19,82% [95% CI: −0,85; 40,48] vs. PL: 14,15% [95% CI: −36,21; 7,92], p=0,047). Similarly, LA stroke volume corrected for body surface area increased in the SPL-group (mean SPL: 34,33 ml/m2 [95% CI: −2,17; 42,62] vs. PL: 8,46 ml/m2 [95% CI: −20,13; 5,94], p=0,013). All changes remained significant after adjusting for age, sex and systolic blood pressure. No difference was found in arrhythmia burden between groups (P=0,89). Conclusion Low dose spironolactone may reduce the burden of diffuse myocardial fibrosis in AF. However, there was no reduction in the burden of arrhythmia. Acknowledgement/Funding Region of Southern Denmark, University of Southern Denmark, Danish Heart Association