Histological validation of clinical classification models of atrial cardiomyopathy

Abstract Background Atrial cardiomyopathy (AtCM) associates with atrial fibrillation (AF) and higher stroke risk. Early, non-invasive identification of AtCM would aid risk-stratification and prevention. A new AtCM staging system using ECG, echocardiography and circulating biomarkers was recently suggested (1). There is an urgent need for histological validation of the suggested AtCM classifiers. Purpose Histological evaluation of non-invasive AtCM qualifiers using human atrial tissue samples. Methods Left (LA) and right (RA) atrial tissue samples were obtained in patients undergoing cardiac surgery (bypass graft/valvular surgery, n=136) in the RACE V Tissue Bank study. Pre-operative rhythm history, digital ECG, transthoracic echocardiography (TTE) and biomarker levels were used to group patients in AtCM stages: PersAF (history of persistent/permanent AF), Severe AtCM (LAVi >50 ml/m2 or LAEF<35%), Moderate AtCM (LAVi 34-50 ml/m2, or LAEF 35-50% AND NT-proBNP >250 pg/ml, or history of paroxysmal AF), Mild AtCM (no AF, LAVi≤34 ml/m2 AND P-terminal Force V1> 5 mV x ms OR P wave duration > 120 ms), no AtCM (no AF AND LAVi ≤ 34 ml/m2 AND LAEF≥50% AND NT-proBNP < 250 pg/ml AND P-terminal Force V1 ≤ 5 mV x ms AND P wave duration ≤ 120 ms). Tissue samples were stained (WGA/CD31/Vimentin) to quantify fibrosis, fibroblast density, myocyte diameter, capillary size and density. Histological analyses were adjusted for age and sex. Post-operative rhythm was monitored continuously (2.5 years) using implantable loop recorders. Results Patients with severe AtCM had more endomysial fibrosis (LA:+0.40 µm, p=0.009 ; RA: 0.27 µm, p=0.08) and larger cardiomyocytes (LA:+0.90 µm, p=0.01, RA: +0.28 µm , p=0.32) than patients with no or subclinical (Mild) AtCM (Fig.1). Moderate AtCM patients also showed LA myocyte hypertrophy (LA: +1.02 µm, p<0.01) compared to No/Mild AtCM patients, but had similar levels of LA endomysial fibrosis (p=0.81). Severe AtCM patients had a higher risk of late post-operative AF (OR: 2.29, p=0.05), with a trend towards higher burden (+8.24%, p=0.14). Tissue samples clustered (k-means) in 4 histologically distinct clusters, reflecting fibrosis and/or myocyte hypertrophy (Fig.2). Patients with fibrotic remodeling more often had a history of AF, while patients with myocyte hypertrophy and patients with both fibrotic and hypertrophic remodeling more often had a history of heart failure with reduced ejection fraction and/or AF. Conclusion Non-invasive AtCM markers reflect histological remodeling and a higher risk of late post-operative AF. Clusters of fibrotic, hypertrophic and combined atrial remodeling were distinguished.