Background: Although atrial fibrillation (AF) is the most common supraventricular arrhythmia worldwide, little is still known about its multifactorial pathogenesis.  Interstitial fibrosis with low-voltage areas (LVAs) is a well-recognized factor involved in left atrial (LA) remodeling and is supposed to sustain AF and to predict poor outcome after catheter ablation. However, the detection of an underlying atrial cardiomyopathy (ACM) as well as the relationship between LA electrical and anatomical remodeling and LA mechanical function are still a matter of debate. We sought to determine clinical and echocardiographic parameters for the diagnosis of ACM as determined by high-density endocardial contact mapping in a population of patients with AF. Moreover, we aimed to assess cut-off values for ACM diagnosis to improve patients selection criteria and identify optimal targets for ablation in a preprocedural setting. 

Methods: We prospectively enrolled 37 consecutive, ablation-naive patients (age 65 ± 15 years, 75% males) with AF (paroxysmal 46%, persistent 54%), between June 2020 and June 2021. Besides clinical parameters, 2D echocardiographic assessment of LA size (LA diameter, LAD; LA end-systolic volume, LAESV; LA end-diastolic volume, LAEDV) and function (LA ejection fraction, LAEF) as well as 3D speckle-tracking (3D-STE) echocardiographic assessment of global longitudinal strain (GLSr) as index of LA dyssynchrony before ablation were also considered.  Contact high-density (minimal map density of 500 points) LA bipolar voltage maps were constructed before ablation using 3D CARTO® 3 System. LVAs were determined with a local bipolar electrogram amplitude of 0.5mV. 

Results: Patients with persistent or paroxysmal AF did not significantly differ for clinical characteristics and cardiovascular comorbidities. However, patients with persistent AF showed significantly larger LA (46±1,9 mm vs 40±1,2; p=0,05). LAEF (28±3 vs 46±4; p=0,001) as well as GLSr (14±9 vs 24±13; p=0,023) significantly decreased in persistent than in paroxysmal AF patients. LVAs were significantly larger (50±11 vs 33±15%; p=0,001) in patients with persistent than those with paroxysmal AF. According to that we defined a relevant ACM as LVAs extending > 33% of the entire LA surface at 0.5 mV threshold on endocardial contact mapping. We observed a significant correlation between ACM extent quantified by endocardial voltage mapping and LAEF using standard 2D echocardiography (r = 0.41, p = 0.012). Similar results were obtained for speckle tracking parameters GLSr (r = 0.46, p = 0.002). Correlation with LAD was low (p=0,06). A LAEF of >42% predicted healthy LA without ACM despite the presence of AF with an area under the curve (AUC) of 0.819 (sensitivity 88.9%, specificity 25%), similar to a GLSr > 16% (AUC 0.849, sensitivity 89%, specificity 32%). 

Conclusion: Echocardiographic parameters as LAEF and GLSr enable a non-invasive method to estimate the presence of ACM prior ablation. No significant correlation was found between ACM and any clinical variable and cardiovascular comorbidities typically known as risk factors for atrial fibrosis, thus supporting the hypothesis of an independent arrhythmogenic process.