Atrial Cardiomyopathy and Atrial Fibrillation: Insight From Histopathological Analysis of Atrial Biopsies

Author's summary

Key gaps in atrial fibrillation management include limited knowledge of atrial cardiomyopathy and the absence of reliable markers reflecting its histopathology. Recent biopsy studies have uncovered changes beyond fibrosis, including intercellular space expansion, myofibrillar loss, reduced nuclear density, and atrial amyloidosis, offering deeper insight into atrial remodeling. Biopsy has also highlighted sex-related differences: lower atrial voltage in women is attributable to smaller myocardial mass rather than greater histopathological severity. Together, these findings enhance understanding of atrial cardiomyopathy and provide a stronger foundation for preventive, diagnostic, and therapeutic advances.

Abstract

Atrial cardiomyopathy is a progressive condition that promotes atrial fibrillation (AF) persistence and adverse outcomes. Although fibrosis has long been considered its hallmark, the histopathological basis in non-valvular AF remains incompletely defined due to limited tissue availability outside surgery or autopsy. To overcome this, we established an integrated approach combining intracardiac echocardiography-guided right atrial biopsy with high-density electroanatomic mapping in AF ablation patients. Our studies revealed that atrial voltage reduction reflects not only fibrosis but also intercellular space expansion, myofibrillar loss, reduced nuclear density, and compensatory cardiomyocyte hypertrophy. Notably, atrial biopsy detected atrial amyloidosis in ~7% of patients, often at an early stage, highlighting its potential for earlier diagnosis and intervention. Sex-based analysis revealed that women consistently had lower atrial voltage, attributable to smaller myocardial mass rather than more severe remodeling. These findings demonstrate that atrial biopsy enables patient-specific histopathological assessment, deepening mechanistic understanding of atrial cardiomyopathy and informing future strategies for AF management.

INTRODUCTION

Atrial fibrillation (AF) is the most common clinically encountered arrhythmia and is associated with elevated risks of stroke, heart failure, and mortality.¹⁾ Its progression parallels the development of atrial cardiomyopathy, characterized by pathological alterations in atrial tissue. The European Heart Rhythm Association (EHRA), the Heart Rhythm Society (HRS), the Asian Pacific Heart Rhythm Society (APHRS), and the Sociedad Latino Americana de Estimulación Cardíaca y Electrofisiología (SOLAECE) define atrial cardiomyopathy as “any complex of structural, architectural, contractile, or electrophysiological changes affecting the atria with the potential to produce clinically relevant manifestations”,²⁾,³⁾ encompassing structural, electrical, and functional remodeling (Figure 1).⁴⁾

Figure 1. Schematic illustration of atrial cardiomyopathy. Atrial cardiomyopathy encompasses structural, electrical, and functional remodeling. The substrates for AF are included within both structural and electrical remodeling components. Reproduced from Yamaguchi.⁴⁾ .

AF = atrial fibrillation; ERP = effective refractory period.

Structural remodeling involves histological changes in cardiomyocytes and the interstitium, clinically reflected as atrial enlargement.⁵⁾,⁶⁾ Electrical remodeling is defined by a shortened refractory period that stabilizes AF.⁶⁾ Its principal ionic mechanism is a reduction in L-type Ca²⁺ current, which shortens and impairs refractory period adaptation.⁶⁾ A concurrent decrease in inward Na⁺ current slows conduction velocity,⁷⁾ further facilitating reentry by shortening wavelength. Functional remodeling refers to AF-induced contractile dysfunction, often observed after cardioversion or spontaneous termination.⁸⁾,⁹⁾,¹⁰⁾,¹¹⁾,¹²⁾,¹³⁾,¹⁴⁾ This phenomenon, known as post-fibrillatory contractile dysfunction, arises mainly from suppression of L-type Ca²⁺ currents.⁶⁾,¹⁵⁾,¹⁶⁾

An essential concept in understanding AF mechanisms and treatment, including catheter ablation, is the structural substrate, often termed the “AF substrate.” Traditionally, this refers to structural abnormalities that promote AF maintenance. While AF is usually initiated by rapid triggers from thoracic veins such as the pulmonary veins or superior vena cava, its persistence depends on the presence of an AF substrate. These substrates can be broadly divided into electrophysiological substrates, marked by shortened refractory periods, and structural substrates, most often associated with fibrosis (Figure 1).⁴⁾ Structural substrates have typically been assessed using surrogate markers such as atrial bipolar voltage or abnormal local electrograms, though their histopathological basis has remained uncertain.

To address this gap, we developed an atrial biopsy technique performed during catheter ablation, allowing direct assessment of atrial cardiomyopathy and AF substrates.¹⁷⁾ By integrating histopathology with high-density electroanatomic mapping, this approach has provided novel insights into the pathology of atrial remodeling and AF substrates.¹⁷⁾,¹⁸⁾,¹⁹⁾,²⁰⁾,²¹⁾ This review highlights these findings, aiming to refine the understanding of AF pathophysiology and to guide future therapeutic strategies.

Ethical statement

This study was approved by the Ethics Committee of Saga University Hospital (approval reference numbers: 20200101 for HEAL-AF, 20200901 for HEAL-AF2, 202200401 for FUTURE-AF-S, 20220401for HEAL-AF3, 20240804 for HEAL-AF4, 20230203 for HISTORY, and 20230501 for LEARNMORE studies).

PROCESSES OF ATRIAL STRUCTURAL REMODELING: FOCAL VERSUS REGIONAL

Using histopathological changes from atrial biopsy at a single site to represent the entire atrium relies on the assumption that structural remodeling is a diffuse process. In this section, we outline the rationale for performing endomyocardial biopsy of the right atrial (RA) septum to estimate histopathological alterations in both atria of patients with non-valvular AF. We then review evidence on the nature of atrial remodeling—whether diffuse, focal, or showing left–right differences—based on findings from autopsy studies, surgical biopsy specimens obtained during open-heart procedures for valvular or other diseases, late gadolinium-enhanced cardiac magnetic resonance imaging (LGE-CMR), and electroanatomic mapping, particularly bipolar voltage mapping performed during catheter ablation.

Postmortem studies

A postmortem study by Platonov et al.²²⁾ demonstrated that, in AF patients without valvular disease, the degree of fibrosis, fatty infiltration, capillary density, and cardiomyocyte size was comparable across 5 atrial regions: the crista terminalis, Bachmann’s bundle, and the inferior, posterior, and superior pulmonary veins. In non-AF controls, no regional differences were observed. In patients with AF, fibrosis and fatty infiltration were consistently 2- to 3-fold higher at all sites and were associated with lymphomononuclear infiltration, with permanent AF showing greater fibrosis than paroxysmal AF. Similarly, in our supplementary postmortem analysis of 8 non-valvular cases, we found no significant differences between the RA septum and the anterior, roof, inferior, posterior, and lateral walls of the left atrium (LA).¹⁸⁾ Collectively, these findings indicate that histopathological remodeling in non-valvular AF is diffuse, affecting both atria.

Surgical biopsy studies

Corradi et al.²³⁾,²⁴⁾ reported greater fibrosis and myofibrillar loss in the LA posterior wall compared with the LA appendage in surgical biopsy samples from patients with chronic AF associated with mitral valve disease, suggesting heterogeneous remodeling. These differences may reflect their distinct embryologic origins: the smooth-walled LA body develops from the incorporated pulmonary veins, while the trabeculated LA appendage derives from the embryonic atrium.²⁵⁾ In a subsequent study, Corradi et al.²⁶⁾ examined atrial biopsies from the RA free wall and LA posterior wall in 19 patients with non-valvular persistent AF undergoing surgical ablation (16 with prior catheter ablation). Among 3 parameters—interstitial fibrosis, cardiomyocyte myocytolysis, and cardiomyocyte transverse diameter—only interstitial fibrosis was significantly greater in the LA posterior wall (4.6±3.9% vs. 2.3±2.2% in the RA), whereas the others exhibited no significant differences.²⁶⁾ Although prior ablation may have influenced results, these findings suggest modest histological variation between the RA and LA. In contrast, no such differences were observed between the RA free wall and LA posterior wall in 15 autopsy controls.²⁶⁾ More recently, Winters et al.²⁷⁾ analyzed atrial appendage samples obtained during cardiac surgery (95 left, 76 right) and found greater total extracellular matrix and endomysial fibrosis, along with smaller cardiomyocytes, in RA appendages. In surgical cases, the underlying type of valvular disease may contribute to regional variation within the LA and to histological differences between the RA and LA.

Postmortem studies on transthyretin amyloidosis type and atrial natriuretic peptide type atrial amyloidosis

In a forensic autopsy series of 1,698 Japanese cases, cardiac transthyretin amyloidosis (ATTR) amyloidosis was identified in 44 patients (2.6% overall; 8.8% in those ≥80 years; mean age 85.4±5.7 years; 22 men; 5 with AF).²⁸⁾ In this study, the atrial septum was subdivided into right and left sides, and amyloid deposition was separately assessed in the endocardium and myocardial interstitium. While deposition within the myocardial interstitium showed no left–right differences in frequency or severity, more extensive involvement was observed in the RA septal endocardium.

The same group subsequently analyzed 325 autopsy cases, identifying 20 with sporadic ATTR cardiac amyloidosis, all of which exhibited LA myocardial deposition.²⁹⁾ In 14 of these cases, semiquantitative analysis demonstrated more severe ATTR deposition in the LA than in the RA. This represented the first autopsy-based evidence of left–right asymmetry, showing a clear LA predominance.²⁹⁾ Notably, the frequency of atrial natriuretic peptide (ANP)-derived amyloid did not differ significantly between atria.²⁹⁾

By contrast, Kawamura et al.³⁰⁾ examined 100 autopsy hearts (65–89 years) and found atrial amyloid in 91% of cases, almost exclusively ANP type, with ATTR type in only 4. ANP-derived deposits showed marked LA predominance across the septum, appendages, and free walls, typically located in the subendocardial interstitium and adjacent myocardium. Similarly, Steiner et al.³¹⁾ reported that ANP amyloidosis preferentially affects the LA, with deposition unevenly distributed and most pronounced in the anterior wall compared with the posterior wall or appendage. Collectively, these autopsy data indicate that atrial amyloid distribution varies by amyloid type: ATTR amyloidosis tends to favor the LA myocardium while also showing RA septal endocardial involvement, supporting the utility of RA septal biopsy for ATTR detection.¹⁹⁾,²⁰⁾ In contrast, ANP type amyloidosis shows consistent LA predominance, raising the possibility that RA septal biopsy may underestimate its burden.

Late gadolinium-enhanced cardiac magnetic resonance imaging studies

LGE-CMR is a key tool for diagnosing and assessing prognosis in ventricular cardiomyopathies. LGE reflects either marked interstitial expansion or delayed gadolinium clearance.³²⁾ It serves as a surrogate for myocardial fibrosis in chronic infarction and various non-ischemic cardiomyopathies, but can also indicate interstitial expansion from edema or inflammatory infiltration, as seen in acute myocarditis or cardiac sarcoidosis.³²⁾,³³⁾ In addition, interstitial amyloid deposition may produce LGE.³²⁾ More recently, efforts have focused on visualizing and quantifying LA fibrosis with LGE-CMR.³⁴⁾,³⁵⁾ Typically, LGE-magnetic resonance imaging (MRI) demonstrates visually identifiable focal regions on 3-dimensional (3D) reconstructions; however, histopathological validation of these findings is limited. Several challenges remain in detecting atrial fibrosis with LGE-CMR.³⁶⁾ These include the absence of standardized imaging protocols and analysis methods, as well as limited spatial resolution due to the thin atrial wall (2.0–2.5 mm).³⁾,³⁷⁾,³⁸⁾ Importantly, LGE-CMR is not well suited for diffuse fibrosis because “normal” myocardium is nulled to highlight focal lesions, thereby obscuring background interstitial expansion.³⁹⁾ This is particularly relevant given autopsy evidence that atrial fibrosis is largely diffuse.²²⁾ Consequently, atrial LGE may not always correspond to fibrosis but may reflect partial volume effects from thin atrial walls, adjacent extracardiac tissue,³⁶⁾ or, in advanced AF, high signal from slow-flowing blood in areas such as the LA appendage. Further histopathological validation is needed, ideally by assessing multiple biopsy sites within the same patient—both in focal LGE regions and in apparently normal regions. Such comparative studies could clarify whether focal LGE represents discrete lesions or simply more advanced stages of diffuse remodeling, in which fibrosis is already widespread.

Electroanatomic mapping studies

With modern electroanatomic mapping during AF ablation, several thousand bipolar voltage recordings can be acquired and projected onto a 3D reconstruction of the atrial endocardium. Low-voltage areas (LVAs), commonly defined by bipolar voltage <0.5 mV, are generally considered surrogate markers of local fibrosis.¹⁷⁾,⁴⁰⁾ Clinical studies have shown that both the presence and burden of LVAs strongly predict atrial tachyarrhythmia recurrence after AF ablation.⁴¹⁾,⁴²⁾,⁴³⁾,⁴⁴⁾,⁴⁵⁾,⁴⁶⁾,⁴⁷⁾,⁴⁸⁾,⁴⁹⁾,⁵⁰⁾ These findings suggest a central role of LVAs in AF maintenance and reinforce the assumption that LVAs correspond to fibrosis, though histological confirmation remains limited. Our recent work demonstrated that: (1) bipolar voltage reduction in the LA of AF patients is diffuse, with severity ranging from minimal to severe; and (2) LVA extent correlates strongly with global LA voltage (V_GLA) reduction, regardless of the threshold used (0.1–1.5 mV).¹⁷⁾ Thus, LVAs appear to represent diffuse atrial remodeling rather than focal degeneration, with their prevalence depending on thresholding and extent. Notably, LVAs are most frequently observed in the anterior wall and septum, areas inherently characterized by lower voltages compared with inferior and lateral walls.¹⁷⁾ Their distribution therefore reflects V_GLA reduction rather than localized pathology. Furthermore, RA septal voltage correlated linearly with V_GLA,¹⁷⁾,¹⁸⁾,²¹⁾ supporting the concept of diffuse bi-atrial remodeling. Voltage variability across atrial regions likely reflects intrinsic anatomical differences such as wall thickness and morphology.³⁷⁾,³⁸⁾ Importantly, discrepancies between LVAs and LGE-MRI findings in identifying LA substrates have been consistently reported, independent of LGE methods.⁵¹⁾,⁵²⁾ Considering that LVAs in non-valvular AF represent local manifestations of diffuse remodeling,¹⁷⁾,¹⁸⁾,²¹⁾ such inconsistencies are likely an expected phenomenon.

Summary of Section 2

In non-valvular AF, findings from autopsy studies and voltage mapping indicate that structural remodeling occurs diffusely across the atria. This supports the rationale that local pathological assessment via atrial biopsy can yield insights into global histopathological changes, providing a theoretical foundation for its use in evaluating atrial degeneration. Regarding amyloid deposition, sampling errors may arise in early stage disease; however, in advanced atrial amyloidosis, deposition typically involves the entire atrium, reducing the likelihood of such errors.⁵³⁾

DEVELOPMENT OF ATRIAL BIOPSY

Histopathological analysis of atrial tissue is critical for understanding atrial structural remodeling. Early human studies relied primarily on surgical biopsy specimens obtained from the free wall and/or appendages of the atria during open-heart surgery, most often in patients with mitral valve disease.²³⁾,²⁴⁾,⁵⁴⁾,⁵⁵⁾,⁵⁶⁾ Later work expanded to surgical myocardial biopsies in patients with lone or non-valvular AF undergoing surgical ablation.²⁶⁾,⁵⁷⁾ However, because of the technical challenges posed by the thin atrial wall, histological assessment in patients not undergoing open-heart surgery has remained limited. In 1997, Frustaci et al.⁵⁸⁾ introduced fluoroscopy-guided endomyocardial biopsy of the RA septum in paroxysmal lone AF, reporting frequent lymphomononuclear infiltrates with adjacent cardiomyocyte necrosis—findings consistent with acute or chronic myocarditis—alongside other established histological abnormalities. Subsequent case reports also described atrial biopsy for diagnosing atrial myocarditis.⁵⁹⁾,⁶⁰⁾ More recently, Sepehri Shamloo et al.⁶¹⁾ reported a series in which atrial biopsy was performed under combined transesophageal echocardiography (TEE) and fluoroscopic guidance to evaluate the histopathological substrate of AF.

In 2023, we developed an intracardiac echocardiography (ICE)- and fluoroscopy-guided endomyocardial atrial biopsy technique, initially applied in 22 patients,¹⁷⁾ followed by a larger cohort of 243 patients undergoing catheter ablation for non-valvular AF.¹⁸⁾ We subsequently expanded our experience and recently reported outcomes in 1,000 patients, including 37 procedures performed under TEE guidance (Figures 2 and 3).²⁰⁾

Figure 2. ICE- or TEE-guided atrial biopsy. (A) Fluoroscopy was used for approximate positioning of the bioptome during catheter ablation. (B) Representative ICE image showing the limbus of the FO. (C) Representative biopsy specimens, with white arrows indicating a sample containing only endocardial tissue, while the other samples include myocardial tissue. (D) Fluoroscopic view of the bioptome and TEE probe during LA appendage occlusion from the LAO projection. (E, F) Biplane TEE images during atrial biopsy demonstrating the bioptome tenting the limbus of the FO. Reproduced from Otsubo et al.²⁰⁾ .

Ao = aorta; CS = coronary sinus; FO = fossa ovalis; ICE = intracardiac echocardiography; LA = left atrium (or left atrial); LAO = left anterior oblique; RA = right atrium (or right atrial); SVC = superior vena cava; TEE = transesophageal echocardiography.

Figure 3. Feasibility and safety of echocardiography-guided atrial biopsy. This study evaluated the feasibility and safety of echocardiography-guided atrial biopsy in a consecutive cohort of 1,000 patients undergoing catheter ablation for atrial tachyarrhythmias or percutaneous LA appendage occlusion. Biopsies were obtained at the limbus of the FO using either a 5.5-Fr (n=233) or 7.0-Fr (n=767) bioptome under intracardiac (n=963) or transesophageal (n=37) echocardiography guidance, in conjunction with fluoroscopy. For histological analysis, 5 tissue samples were collected from the same site. Biopsy was successfully completed in 996 patients (99.6%). Based on histological depth, patients were classified into Group A (samples including myocardium beyond the endocardium; n=885) and Group B (endocardial-only samples; n=111). Multivariate logistic regression identified larger LA volume, use of a 5.5-Fr bioptome, and amyloid deposition as independent predictors of endocardial-only biopsy (Group B; p=0.009, p<0.001, and p=0.001, respectively). No biopsy-related complications were observed. Amyloid deposition was detected in atrial samples from 80 of 996 patients (8%). Reproduced from Otsubo et al.²⁰⁾ .

AF = atrial fibrillation; FO = fossa ovalis; LA = left atrium (or left atrial); LAAO = left atrial appendage occlusion; RA = right atrium (or right atrial).

The RA septum, the preferred biopsy target, has a relatively thick endocardium (mean 0.85 mm),¹⁸⁾ which can occasionally impede penetration and result in retrieval of endocardium-only samples, precluding myocardial analysis.²⁰⁾ In our series, biopsy was successful in 996 of 1,000 patients (99.6%), with myocardial tissue obtained in 885 cases and endocardium-only samples in 111.²⁰⁾ Predictors of inadequate myocardial sampling included larger LA volume, use of a 5.5-Fr rather than a 7.0-Fr bioptome, and the presence of amyloid deposition (Figure 3). Notably, no major complications occurred, and no significant late events were detected during follow-up. Based on these findings, use of a 7.0-Fr bioptome (7.0 Fr, 104 cm, 5.20 mm³ tip volume; Cordis, Miami Lakes, FL, USA) is preferable to improve sampling success. In patients with atrial amyloidosis, ATTR type deposits are consistently found in the RA septal endocardium,¹⁹⁾ allowing diagnostic confirmation even with endocardium-only samples, though such findings may themselves serve as predictors of amyloidosis. Thus, atrial biopsy retains diagnostic value even in these cases.

Nevertheless, obtaining myocardial tissue is crucial for assessing structural remodeling, and operators should aim for myocardial samples whenever feasible. Practical strategies include: (1) targeting the limbus of the fossa ovalis (FO), where bioptome positioning can be consistently visualized using ICE or TEE (Figure 2); (2) preferring a larger bioptome (7.0 Fr); (3) performing multiple biopsies at the same site (3–4 attempts may be required, akin to “digging a hole”); and (4) stopping further attempts if specimens float in solution, as this may indicate penetration into the fat layer and risk of perforation.²⁰⁾ These steps enhance the likelihood of successful sampling while minimizing complications. Moreover, the technique could be applied in other TEE-guided interventions such as transcatheter aortic valve implantation, LA appendage closure, and transcatheter mitral repair. Our cumulative experience suggests that atrial biopsy performed during catheter ablation is safe; however, further multicenter studies are needed to validate safety and clarify its clinical utility.

HISTOPATHOLOGICAL EVALUATION OF ATRIAL BIOPSY SAMPLES

In 1977, Thiedemann and Ferrans⁵⁴⁾ described both light microscopic and ultrastructural abnormalities in atrial specimens from patients with mitral valve disease, including cardiomyocyte hypertrophy, marked cellular degeneration, and extensive interstitial fibrosis. They noted that atrial myocardium displayed more pronounced and widespread degenerative changes compared with ventricular myocardium, remarking that “the occurrence of such extensive alterations in ventricular myocardium would be incompatible with survival of the patient.” Similar structural abnormalities were subsequently reported in surgical specimens from patients with mitral valve disease and other cardiac conditions.²³⁾,²⁴⁾,⁵⁵⁾,⁵⁶⁾ These alterations were consistently observed not only in patients with underlying heart disease but also in those with lone or non-valvular AF,²⁶⁾,⁵⁷⁾ indicating that atrial structural remodeling is not exclusive to AF. In 1997, Frustaci et al.⁵⁸⁾ introduced fluoroscopy-guided endomyocardial biopsy of the RA septum in patients with paroxysmal lone AF, reporting a high prevalence of lymphomononuclear infiltrates accompanied by necrosis of adjacent cardiomyocytes—findings suggestive of acute or chronic myocarditis—alongside the common histological changes described above. Building on this foundation, we quantitatively assessed several histopathological parameters: fibrosis extent (%fibrosis), intercellular space extent (%intercellular space), severity of myofibrillar loss (%myofibrillar loss), adipocyte content (%adipocytes), myocyte size, myocardial nuclear density (a surrogate for cardiomyocyte number), and myocyte disarray.¹⁸⁾ Furthermore, we evaluated the presence of lymphomononuclear infiltrates as potential indicators of acute or chronic myocarditis. The following section discusses each of these histopathological features in greater detail, particularly in relation to electroanatomic mapping findings and their prognostic significance.

Histological factors associated with electroanatomic characteristics

Fibrosis is widely regarded as the primary contributor to the AF substrate, reflected by low-voltage,⁴¹⁾,⁴²⁾,⁴³⁾,⁴⁴⁾,⁴⁵⁾,⁴⁶⁾,⁴⁷⁾,⁴⁸⁾,⁴⁹⁾,⁵⁰⁾ fractionated electrograms,⁶²⁾,⁶³⁾,⁶⁴⁾ and zones of slow conduction⁶⁵⁾,⁶⁶⁾; however, direct histopathological validation is still lacking. In our study, we demonstrated that LA voltage was diffusely and uniformly reduced, and that voltage measured at the RA septal biopsy site correlated linearly with both regional and V_GLA.¹⁷⁾,¹⁸⁾,²¹⁾ These results suggest that the histological properties of the RA septum reflect not only local but also LA tissue characteristics. Since obtaining LA biopsy specimens—particularly from targeted sites such as LVAs—is challenging outside of cardiac surgery, we propose that the integration of high-density bi-atrial electroanatomic mapping with RA biopsy represents the most practical current strategy for investigating histopathological–electrophysiological relationships (Figure 4).

Figure 4. Towards the elucidation of atrial cardiomyopathy. The development of atrial biopsy techniques has marked a critical advance in understanding atrial cardiomyopathy, though their primary limitation is the inability to sample beyond the RA septum. By combining this approach with high-density electroanatomic mapping—which has shown that atrial voltage reduction reflects diffuse remodeling across both atria—histopathological insights into atrial cardiomyopathy have expanded substantially. Incorporating multi-omics analyses is anticipated to address remaining questions regarding its pathophysiology. Progress in elucidating atrial cardiomyopathy will rely on integrative research encompassing electroanatomic mapping, sex-related differences, early detection of cardiac amyloidosis, atrial functional assessment, event risk stratification, ablation outcomes, molecular mechanisms, genetic factors, and the role of CHIP. Collectively, these approaches are expected to guide the development of novel therapies and drug discovery.

CHIP = clonal hematopoiesis of indeterminate potential; LA = left atrium (or left atrial); RA = right atrium (or right atrial).

In our series of 230 patients undergoing AF ablation,¹⁸⁾ high-density mapping during high RA pacing with a grid-mapping catheter (Advisor™ HD Grid; Abbott, St. Paul, MN, USA) enabled comprehensive assessment of bi-atrial bipolar voltage and identification of fractionated electrograms and local slow conduction zones. Among the histological parameters, not only fibrosis but also increased intercellular space, myofibrillar loss, and reduced myocardial nuclear density were significantly associated with voltage reduction at both the RA biopsy site and across the LA. Increased intercellular space correlated with a higher proportion of fractionated electrograms, while fibrosis, intercellular expansion, and myofibrillar loss were associated with the extent of LA slow conduction zones. These findings indicate that atrial remodeling extends beyond fibrosis alone, and that electroanatomic mapping does not fully capture the underlying histopathological complexity.¹⁸⁾

Overall, the reduction in viable myocardium appears to be the principal driver of voltage decline. Yet, the associations between individual histological parameters and voltage were relatively modest. This likely reflects 2 key points: (1) histological changes are interdependent, evolving together to reduce viable myocardium rather than progressing in isolation; and (2) atrial voltage and mass are also influenced by age,¹⁸⁾ sex,²¹⁾ LA volume,¹⁸⁾ and wall thickness,³⁷⁾,³⁸⁾ all of which vary regionally. The moderate correlation between RA biopsy voltage and global or regional LA voltage may therefore also reflect interindividual variation in atrial architecture and wall thickness, particularly at the limbus of the FO. These observations parallel findings by Glashan et al.⁶⁷⁾ in non-ischemic cardiomyopathy, where both wall thickness and viable myocardial content determined ventricular voltage, with fibrosis causing a linear decline by reducing viable myocardium.

Notably, female sex was associated with lower atrial voltage, although not with any specific histological parameter, suggesting that sex-related differences are more likely attributable to inherent myocardial mass.²¹⁾ This topic will be further addressed in the discussion of sex differences (Section 6).

Interstitial change: fibrosis and increase in intercellular space

Fibrosis is a hallmark feature of the AF substrate, promoting arrhythmogenesis by disrupting fiber bundle continuity, inducing localized conduction abnormalities, and enhancing non-uniform anisotropic conduction.²²⁾,⁵⁷⁾,⁶⁸⁾,⁶⁹⁾ Our findings revealed that interstitial alterations associated with electroanatomic properties involve not only fibrosis but also expansion of the intercellular space. Transmission electron microscopy demonstrated plasma components and immature collagen within these expanded spaces, suggesting that intercellular space enlargement may precede collagen fiber accumulation and polymerization, i.e., the development of fibrosis.¹⁸⁾ Similar to collagen deposition, intercellular space expansion may contribute to electrical uncoupling of side-to-side myofibrillar connections, leading to slow conduction or zigzag transverse propagation that facilitates reentry.⁶⁸⁾ Both fibrosis and intercellular space expansion were more prominent in long-standing persistent AF than in paroxysmal AF, and their extent correlated with atrial tachyarrhythmia recurrence after AF ablation.¹⁸⁾ These observations indicate that intercellular space expansion, alongside fibrosis, constitutes an important histological substrate for AF. The mechanisms underlying intercellular space expansion remain unclear. In a canine model of chronic AF with mitral regurgitation and rapid atrial pacing, reverse electrical remodeling was observed within 7–14 days following cardioversion; however, atrial enlargement, ultrastructural alterations, and atrial dysfunction persisted. While spontaneous AF decreased, susceptibility to induced AF remained, underscoring that structural remodeling—rather than electrical remodeling alone—is a key determinant of AF progression. Notably, this model also exhibited marked expansion of the intercellular space, reinforcing the concept that such expansion represents an early stage of interstitial remodeling that contributes to atrial enlargement and AF persistence.⁷⁰⁾ If intercellular space expansion indeed constitutes the initial step of interstitial remodeling, timely therapeutic intervention may prevent irreversible fibrotic progression. Figure 5 shows 2 cases in which AF recurred immediately as persistent AF after catheter ablation, ultimately leading to discontinuation of rhythm control. Both cases demonstrated advanced interstitial fibrosis (%fibrosis: A, 19.1%; B, 19.8%). These findings highlight the clinical importance of elucidating the mechanisms underlying intercellular space expansion and preventing irreversible collagen accumulation to improve outcomes in AF and atrial cardiomyopathy.

Figure 5. Histological findings from 2 representative AF cases in which rhythm control failed. Histological findings from 2 representative AF cases are presented, both of which were refractory to rhythm control using catheter ablation and antiarrhythmic drugs, ultimately necessitating a shift to rate control. (A) A 59-year-old man with persistent AF and (B) a 44-year-old woman with persistent AF. Both patients showed extensive advanced interstitial fibrosis (%fibrosis: A, 19.1%; B, 19.8%). In each case, persistent AF recurred within one month following ablation. Antiarrhythmic therapy remained ineffective even beyond the blanking period, and repeated electrical cardioversions were unsuccessful. As a result, rhythm control was discontinued, and a rate control strategy was implemented.

AF = atrial fibrillation.

Myofibrillar loss

The most prominent parenchymal alteration in atrial cardiomyocytes is the progressive loss of sarcomeres, which begins in the perinuclear region and extends eccentrically into the sarcoplasm—a process known as myocytolysis.⁷¹⁾ Transmission electron microscopy has confirmed this finding as myofilament lysis.¹⁸⁾ In some cases, the perinuclear region instead contains abundant glycogen granules, suggesting intra-cardiomyocyte glycogen accumulation caused either by excessive glucose availability or impaired glycogen catabolism, with variable degrees of deposition.⁷¹⁾,⁷²⁾ The extent of myofibrillar loss was independently correlated with voltage reduction, but showed no association with fractionated electrograms, AF type, or atrial tachyarrhythmia recurrence following catheter ablation.¹⁸⁾ Thus, although parenchymal changes may not directly contribute to the AF substrate, they remain a hallmark of parenchymal degeneration. Notably, macrophage infiltration was more pronounced in patients with greater myofibrillar loss,¹⁸⁾ likely reflecting phagocytic removal of degenerated myofilaments. The etiology of myofibrillar loss remains uncertain. However, reports of reversibility in the human left ventricle⁷³⁾ raise the possibility that atrial myofibrillar loss may also be reversible. Future investigations should clarify the mechanisms driving myofibrillar loss and assess its prognostic implications and impact on atrial function.

Myocardial nuclear density

Nuclear density, a surrogate marker of myocardial cell number, emerged as a significant histological correlate of voltage reduction.¹⁸⁾ However, this relationship lost significance after adjustment for age, reflecting the documented decline in nuclear density with aging.¹⁸⁾,¹⁹⁾ Although cardiomyocytes are traditionally regarded as non-proliferative, limited proliferative capacity has been reported.⁷⁴⁾,⁷⁵⁾,⁷⁶⁾ Autopsy studies have shown age-related reductions in cardiomyocyte number in the ventricles,⁷⁷⁾ and our atrial biopsy data similarly demonstrate a progressive decline in myocardial nuclear density with age.¹⁸⁾,¹⁹⁾ In parallel, cardiomyocyte hypertrophy was commonly observed, consistent with earlier atrial biopsy studies.²³⁾,²⁴⁾,²⁶⁾,²⁷⁾,⁵⁴⁾,⁵⁵⁾,⁵⁶⁾,⁵⁷⁾

An inverse association between nuclear density and cardiomyocyte size was evident.¹⁸⁾,²¹⁾ If hypertrophy were solely an adaptive response to pressure overload, increases in both myocardial mass and bipolar voltage would be expected. Instead, decreased nuclear density combined with hypertrophy was independently associated with voltage reduction,¹⁸⁾ suggesting that atrial hypertrophy reflects compensation for cardiomyocyte loss rather than pure adaptation to loading conditions.

Recently, we conducted a more detailed analysis of the relationship between nuclear density and mean cardiomyocyte size, based on the following assumptions, conditions, and formula (Figure 6)²¹⁾:

Figure 6. Hypothetical relationship between myocardial nuclear density and cardiomyocyte size. A detailed analysis was conducted to examine the relationship between atrial cardiomyocyte size and nuclear density. From this, we derived an equation based on the following assumptions and conditions, showing that cardiomyocyte size (2 × R) is proportional to the inverse square root of nuclear density under these circumstances.

Assumption: A decrease in myocardial volume caused by reduced nuclear density (i.e., cardiomyocyte loss) is compensated by hypertrophy of the remaining cardiomyocytes.

Conditions: Interstitial remodeling is constant; atrial cardiomyocytes are mononucleated; and no new cardiomyocyte proliferation occurs.

• Assumption: A reduction in myocardial volume due to decreased nuclear density, i.e., cardiomyocyte loss, is compensated for by cardiomyocyte hypertrophy.

• Conditions: Interstitial remodeling remains constant, atrial cardiomyocytes are mononucleated, and no cardiomyocyte proliferation occurs.

Formula: A = πR² × χ × B

In this equation, A represents the myocardial volume (mm³), R is the cardiomyocyte radius, χ denotes nuclear density (per mm²), and B is a constant.

Rearranging the equation yields the following:

$$R = \sqrt{A\mathrm{/(}B\pi {\rm X})}$$

$$= a \frac{1}{\sqrt{{\rm X}}}, \left(where a is a constant\right)$$

This equation demonstrates that, under the stated assumption and conditions, cardiomyocyte size is proportional to the inverse square root of the nuclear density. Accordingly, the inverse square root of the nuclear density was used as the x-axis when plotting the relationship between nuclear density and cardiomyocyte size (Figure 7).

Figure 7. Relationship between myocardial nuclear density and cardiomyocyte size. Relationship between myocardial nuclear density and cardiomyocyte size is shown for men (A), women (B), and the overall cohort (C), illustrating a non-linear association. Assuming that a reduction in myocardial volume due to decreased nuclear density (χ) is offset by cardiomyocyte hypertrophy, the cardiomyocyte radius is proportional to the inverse square root of nuclear density (χ). When the inverse square root of χ (= X) was plotted on the x-axis, a linear correlation with cardiomyocyte size (y-axis) was observed for both men (D) and women (E). The regression lines intersected, indicating that among patients with preserved nuclear density, women had smaller cardiomyocyte size (F). Reproduced from Nakashima et al.²¹⁾ .

The relationship was assessed in 308 patients, including 282 with AF and 26 without AF, with analyses performed separately for men and women.²¹⁾ Nuclear density was inversely associated with mean cardiomyocyte size in both men and women in non-linear manner (Figure 7). When the inverse square root of nuclear density was plotted on the x-axis, a strong linear correlation with cardiomyocyte size was observed in both women (r=0.773) and men (r=0.618), supporting the concept that cardiomyocyte hypertrophy may serve as a compensatory response to the loss of cardiomyocyte number.

DNA damage, a known driver of ventricular cardiomyocyte death and adverse prognosis in heart failure,⁷⁸⁾,⁷⁹⁾ may also contribute to atrial cell loss, dysfunction, and remodeling. Nuclear density decline was linked to aging but not to AF duration or arrhythmia recurrence,¹⁸⁾ suggesting that it is unlikely to serve as a direct AF substrate. Nevertheless, reduced nuclear density was associated with voltage reduction, which in turn correlated with inducibility of macroreentrant atrial tachycardia (AT) in LVAs (<0.5 mV).¹⁸⁾ These observations highlight decreased cardiomyocyte number as a fundamental feature of atrial cardiomyopathy progression. Loss of cardiomyocytes, together with compensatory hypertrophy, may also impair atrial contractile function, thereby predisposing to heart failure and thromboembolic events.

Myocarditis

In our initial analysis of 230 atrial biopsy specimens, no cases fulfilled the Dallas criteria for myocarditis—defined as a myocardial inflammatory infiltrate accompanied by necrosis and/or degeneration of adjacent non-ischemic cardiomyocytes.¹⁸⁾,⁸⁰⁾ This contrasts with the findings of Frustaci et al.,⁵⁸⁾ who performed RA septal biopsies (2–3 samples per patient; mean 2.8) in 12 patients (mean age, 32 years) with paroxysmal lone AF refractory to antiarrhythmic therapy. In that series, lymphomononuclear infiltrates with associated myocyte necrosis were identified in 8 patients (5 with fibrosis and 3 without), resulting in a myocarditis diagnosis in 66% of cases, including 25% with active myocarditis. One possible explanation for this discrepancy lies in patient age: the Frustaci cohort was considerably younger (mean age, 32 years; range, 19–43 years) compared with our population (mean age, 67±12 years; range, 21–91 years). This raises the possibility that a stronger inflammatory component may contribute to AF pathogenesis in younger individuals. Nevertheless, even after extending our analysis to 1,000 biopsy cases (mean age, 67±12 years; range, 20–97 years), we did not identify clear histological evidence of myocarditis.²⁰⁾ In the initial series of 230 patients, immunohistochemical analysis was performed in 60 cases to characterize inflammatory cell infiltration, including CD3-positive T lymphocytes, CD20-positive B lymphocytes, CD45-positive leukocytes, CD11c-positive M1 macrophages, and CD163-positive M2 macrophages.¹⁸⁾ As anticipated, mild infiltration of these immune cell subsets was consistently observed, supporting the hypothesis that inflammatory processes, while not meeting the Dallas criteria, may nonetheless play a contributory role in atrial structural remodeling.

Summary of Section 4

Figure 8 depicts the progression of atrial structural remodeling.⁴⁾ Histopathologically, this process involves not only fibrosis but also expansion of the intercellular space (which may precede fibrosis), myofibrillar loss, decreased cardiomyocyte nuclear density, and amyloid deposition, all of which correlate with reductions in V_GLA. Interstitial alterations, including increased intercellular space and fibrosis, serve as substrates for AF. As atrial structural remodeling advances, V_GLA declines, leading to the enlargement of LVAs, a higher prevalence of fractionated electrograms, expansion in both the number and width of slow conduction zones, greater inducibility of AT, and increased recurrence of atrial tachyarrhythmias following catheter ablation. These histological changes are not independent but evolve synergistically, collectively reducing the amount of viable myocardium.

Figure 8. Schematic representation of atrial structural remodeling progression. The upper panel presents examples of high-density bipolar voltage maps of the LA recorded during high RA pacing at 100 beats per minute. The lower panel shows light microscopy images of atrial biopsy samples from patients with non-valvular AF. The progression of atrial structural remodeling, reflected by reduced V_GLA, is associated not only with fibrosis but also with interstitial changes such as increased intercellular space preceding fibrosis, myofibrillar loss, decreased cardiomyocyte nuclear density, and amyloid deposition. Interstitial alterations, including intercellular space expansion and fibrosis, serve as AF substrates. As remodeling advances, V_GLA declines, leading to enlarged LVAs, a higher number of fractionated electrograms, expanded slow conduction zones, increased inducibility of AT, and greater recurrence of atrial tachyarrhythmias after catheter ablation. Reproduced from Yamaguchi.⁴⁾ .

AF = atrial fibrillation; AT = macroreentrant atrial tachycardia; LA = left atrium (or left atrial); LVA = low-voltage area; RA = right atrium (or right atrial); V_GLA = global left atrial voltage.

ATRIAL AMYLOIDOSIS

Cardiac amyloidosis is a progressive disorder associated with high rates of heart failure, arrhythmias, and poor prognosis.⁸¹⁾,⁸²⁾ AF is common in these patients, with a reported prevalence of 30–50%.⁸³⁾,⁸⁴⁾ Postmortem studies in elderly populations consistently demonstrate cardiac amyloid deposits, often limited to the atria. Over 50 years ago, Hodkinson and Pomerance reported a ~50% prevalence of senile cardiac amyloidosis in autopsies of individuals aged ≥60 years, with nearly half exhibiting deposits confined to the atria.⁵³⁾ With aging, atrial amyloidosis often progresses and may eventually involve the ventricles and extracardiac tissues.⁵³⁾ This early report also noted associations between cardiac amyloidosis, AF, and heart failure, which strengthened with more extensive amyloid deposition. Isolated atrial amyloidosis, in which ANP forms the fibrils, was subsequently recognized as a distinct entity and is now synonymous with ANP type amyloidosis.⁸⁵⁾,⁸⁶⁾,⁸⁷⁾

In a recent study, we assessed the prevalence and clinical characteristics of amyloid deposition in 578 patients with non-valvular AF undergoing AF ablation, including both atrial and ventricular biopsies.¹⁹⁾ This section reviews the prevalence of atrial amyloidosis, associated clinical features, correlations with atrial voltage, and prognostic implications. Definitions used in our study included: atrial amyloidosis—any amyloid deposition in the atrium; ventricular amyloidosis—any deposition in the ventricles; cardiac amyloidosis—deposition in either or both atria or ventricles; confined atrial amyloidosis—deposition in atria but not ventricles; and confined ventricular amyloidosis—deposition in ventricles but not atria.¹⁹⁾

Prevalence of atrial amyloidosis and clinical characteristics

Amyloid deposition was detected in atrial samples from 40 patients (7%, 95% confidence interval [CI], 5–9%) within the atrial biopsy cohort (n=578) and in right ventricular (RV) samples from 13 patients (3%, 95% CI, 2–6%) among 385 who underwent concomitant RV biopsy.¹⁹⁾ Of the 26 patients with atrial amyloidosis who also had RV biopsies, 13 exhibited no amyloid deposits in the RV, consistent with confined atrial amyloidosis. Compared with clinically diagnosed cardiac amyloidosis patients, those with biopsy detected atrial amyloidosis had lower troponin T levels and less left ventricular hypertrophy (mean interventricular septal thickness 11.2±2.4 mm; mean left ventricular posterior wall thickness [LVPW] 11.3 ± 2.2 mm), suggesting early-stage disease. In contrast, compared with patients without atrial amyloidosis, those with atrial deposits were older and showed higher troponin T and NT-proBNP levels, more pronounced left ventricular hypertrophy, and more advanced atrial remodeling, as indicated by reduced V_GLA, the presence of LVAs defined as <0.5 mV (LVA_0.5), and a higher prevalence of heart failure with preserved left ventricular ejection fraction (LVEF). The prevalence of atrial amyloidosis increased with age, reaching 30–40% in subgroups with LVA_0.5 and/or LVPW ≥12 mm. In patients ≥60 years, LVPW ≥12 mm and LVA_0.5 had sensitivities of 20% and 31% and specificities of 98% and 97%, respectively. Using more lenient thresholds (LVPW ≥11 mm and LVA defined as <1.0 mV), sensitivity increased to 43% and 69%, while specificity remained 92%. Logistic regression confirmed that age, LVPW, and the presence of LVA_0.5 were independent predictors of atrial amyloid deposition.

By AF type, the prevalence of atrial amyloidosis was 8% in paroxysmal AF, 8% in persistent AF, and 4% in long-standing persistent AF. By heart failure type, prevalence was 7% in heart failure with reduced LVEF, 18% in heart failure with midrange LVEF, and 18% in heart failure with preserved LVEF.

Recurrence-free survival for atrial tachyarrhythmia and persistent AF after catheter ablation was similar between patients with and without atrial amyloidosis. However, atrial amyloidosis was independently associated with the composite outcome of all-cause mortality, stroke, and heart failure hospitalization after ablation (adjusted hazard ratio, 2.71; 95% CI, 1.10–6.70, adjusted for age, paroxysmal AF, and V_GLA).

Amyloid typing and characteristics of each type

Amyloid phenotyping of atrial samples identified ATTR in 25 patients (63%), light-chain amyloidosis in 6 patients (15%), and ANP amyloidosis in 5 patients (13%). In 4 patients (10%), the amyloid type could not be determined due to minimal deposits. In RV samples, typing revealed ATTR in 11 patients (85%) and AL in 2 patients (15%). All ATTR cases involved at least the atrial endocardium. Of the 25 ATTR cases, 7 samples contained only endocardial tissue, likely reflecting limited sampling from the thickened endocardium,²⁰⁾ while 18 samples (72%) included myocardial layers and 6 (24%) showed deposits confined to the endocardium. Patients with mild atrial amyloid deposition were more likely to have absent or minimal RV deposits compared with those with moderate to severe atrial deposition. Importantly, no cases of confined ventricular amyloidosis were observed.

Atrial amyloidosis and atrial cardiomyopathy

As noted previously, atrial cardiomyopathy involves not only fibrosis but also other histological changes such as increased intercellular space, myofibrillar loss, and decreased myocardial nuclear density, all contributing to atrial voltage reduction.¹⁸⁾ The present study further highlights amyloid deposition as an additional and significant factor in atrial cardiomyopathy.¹⁹⁾ Histological quantification of atrial samples revealed that patients with atrial amyloidosis exhibited pronounced structural degeneration—including reduced myocardial nuclear density, expanded intercellular space, and severe myofibrillar loss—even in cases with only mild amyloid deposits. In contrast, the extent of fibrosis remained relatively limited, consistent with prior observations.⁸⁸⁾ Importantly, analysis showed that atrial degeneration was often advanced even when ventricular amyloidosis was still in its early stages. While the underlying mechanisms and causal links between amyloid deposition and structural degeneration remain unclear, understanding these processes will be essential for developing targeted therapies for atrial amyloidosis and its associated histopathological changes.

SEX DIFFERENCE IN ATRIAL CARDIOMYOPATHY

AF shows notable sex differences. Although AF develops later and is less prevalent in women, they experience higher morbidity and mortality compared with men.⁸⁹⁾,⁹⁰⁾ Electrophysiological and ablation studies consistently identify female sex as an independent risk factor for LA LVAs, a surrogate for advanced atrial structural remodeling,¹⁷⁾,⁹¹⁾,⁹²⁾,⁹³⁾ and women also demonstrate lower recurrence-free survival after ablation.⁹⁴⁾ The mechanisms underlying these sex-related differences remain poorly understood, particularly from a histopathological perspective. To investigate this, we recently assessed sex differences in atrial structural remodeling in patients with non-valvular AF using an integrated approach combining high-density bi-atrial voltage mapping with RA biopsy (Figure 4).²¹⁾ We tested 3 hypotheses: (i) women have inherently lower atrial voltage in both atria than men, even without AF; (ii) women exhibit more advanced histopathological remodeling; and (iii) cardiomyocyte size is inherently smaller in women prior to hypertrophic changes, potentially contributing to lower atrial voltage.

Sex differences in atrial voltage

Women exhibited significantly lower voltage at the biopsy site, as well as reduced global and regional LA voltage across all 6 regions (anterior, roof, posterior, septal, inferior, and lateral walls) in both the AF (n=282) and non-AF (n=41) groups, indicating that lower atrial voltage in women is independent of AF.²¹⁾ Within the AF group, LVA_0.5 was more frequently observed in the LA of women. For patients who underwent both LA voltage mapping and atrial biopsy, V_GLA was ranked in descending order and stratified into quartiles (Q1–Q4) separately for women and men (Figure 9). Notably, even in the highest quartile (Q4), women consistently demonstrated lower voltage than men, supporting the presence of an intrinsic sex difference in atrial voltage. LVA_0.5 was primarily observed in the lowest quartile (Q1) and occurred more frequently in women.

Figure 9. Sex differences in atrial remodeling. Atrial voltage, including V_GLA, is consistently lower in women than in men, both in patients with AF and in those without AF. However, there are no sex differences in the extent of fibrosis, intercellular space, severity of myofibrillar loss, myocardial nuclear density, or the prevalence of amyloid deposition. In the minimal histological change group, women are more likely to have smaller cardiomyocytes. V_GLA cut-off values for predicting recurrence after ablation differ by sex: 4.6 mV for men and 3.3 mV for women. Reproduced from Nakashima et al.²¹⁾ .

AF = atrial fibrillation; LA = left atrial (or left atrium); V_GLA = global left atrial voltage.

Sex differences in histopathology

Contrary to the initial hypothesis, comparisons of fibrosis extent, intercellular space, myofibrillar loss severity, myocardial nuclear density, and the prevalence of amyloid deposition between women and men—both in AF and non-AF patients—revealed no significant sex differences in any parameter (Figure 9). Subgroup analyses stratified by V_GLA quartiles likewise showed no sex-related differences in histological features. Additionally, dividing the AF group into 3 age categories (<60, 60–74, and ≥75 years) demonstrated no sex differences across any histological measure. Multivariate linear regression adjusting for age, body surface area, AF type, and LA volume further confirmed that female sex was not significantly associated with any histopathological parameter. These findings suggest that the lower atrial voltage observed in women is likely due to inherent differences in atrial myocardial mass rather than histological remodeling. Our results align with Platonov et al.,²²⁾ who found no sex differences in atrial fibrosis extent in 30 autopsy cases without valvular disease. By contrast, Winters et al.²⁷⁾ reported that female sex was associated with greater endomysial fibrosis in atrial appendage samples from patients undergoing cardiac surgery, indicating that female sex may contribute to atrial fibrosis. These discrepancies could be attributed to differences in biopsy site (atrial septum vs. atrial appendage) or variations in patient populations, particularly regarding valvular vs. non-valvular heart disease.

Sex differences in cardiomyocyte size

A semiquantitative histological score was calculated based on %fibrosis, %intercellular space, %myofibrillar loss, and nuclear density, with scores ranging from 0 to 3, reflecting minimal to severe changes.¹⁸⁾,²¹⁾ Minimal histological change was defined as a total score ≤4, and patients with individual scores ≥2 for any parameter were excluded.²¹⁾ All other patients were classified as having advanced histological changes. In this study, 12 AF patients and 22 non-AF patients met criteria for minimal histological change, forming the minimal change group, while the remaining 274 patients were categorized as having advanced histological changes.

Within the minimal histological change group, no significant difference was observed in mean cardiomyocyte size between women and men (11.8±2.6 vs. 13.3±2.4 µm, p=0.095). However, the proportion of small cardiomyocytes (<10 µm) was significantly higher in women (p=0.029) (Figure 9). In contrast, no sex differences in cardiomyocyte size were observed in the advanced histological change group.

Myocardial nuclear density decreases with atrial voltage reduction. As described above, a non-linear inverse relationship was observed between nuclear density and mean cardiomyocyte size (Figure 7), supporting the idea that cardiomyocyte hypertrophy may compensate for the loss of cardiomyocytes. Autopsy studies have similarly shown that ventricular cardiomyocyte numbers decline with age, accompanied by an increase in cell volume per nucleus.⁷⁷⁾ In the AF group, no significant sex differences were found in nuclear density, mean cardiomyocyte size, or in the relationship between these 2 variables.

By contrast, in the minimal histological change group, which mostly included patients without AF, women had a significantly higher proportion of small cardiomyocytes (<10 μm). This may partly explain women’s inherently lower atrial myocardial mass prior to structural remodeling. Men and women are born with roughly the same number of cardiomyocytes, which does not substantially increase after birth⁹⁵⁾,⁹⁶⁾; cardiac growth is instead primarily driven by increases in cell volume.⁹⁵⁾,⁹⁶⁾ In animal models, female rats have a higher proportion of small cardiomyocytes than males, although this difference diminishes with pathological hypertrophy,⁹⁷⁾,⁹⁸⁾,⁹⁹⁾ suggesting that female cardiomyocytes retain a greater capacity for hypertrophy.

Overall, these findings support the hypothesis that women inherently have smaller cardiomyocytes and lower atrial myocardial mass. However, compensatory hypertrophy in response to cardiomyocyte loss may attenuate these sex differences in cell size.²¹⁾

Sex differences in the inducibility of macroreentrant atrial tachycardia

We previously reported that the inducibility of LA macroreentrant tachycardia following pulmonary vein isolation was strongly associated with a reduction in V_GLA (V_GLA cut-off: 4.2 mV, area under the curve=0.958), which corresponds to the presence of LVA_0.5, independent of histopathological parameters.¹⁸⁾ Our findings also demonstrated that women inherently have lower atrial myocardial mass and voltage.²¹⁾ Assuming that structural remodeling progresses similarly in both sexes, women are expected to reach the threshold for LVA_0.5 development and AT inducibility earlier than men (Figure 10). This may explain the higher prevalence of LVAs and the greater inducibility of atrial tachycardia observed in women (Figure 9). Consequently, the increased susceptibility to AT in women may partly account for the higher post-ablation arrhythmia recurrence reported in previous studies.⁹⁴⁾ Supporting this, earlier reports have documented a greater vulnerability to AT in women; for instance, one study reported a significantly higher recurrence rate of atrial tachycardia in women after thoracoscopic AF ablation,¹⁰⁰⁾ whereas another found a higher proportion of women among patients with de novo LA macroreentrant tachycardia.¹⁰¹⁾

Figure 10. Schematic illustrating the conceptual shift from the previous to the new hypothesis. The earlier hypothesis suggested that women exhibit more advanced structural remodeling—such as greater histological fibrosis—implying a faster remodeling process than men, which was thought to result in lower atrial voltage. The new hypothesis proposes that women inherently possess smaller myocardial mass, leading to consistently lower atrial voltage independent of histological remodeling. This inherently lower voltage is associated with an earlier onset and higher prevalence of LVAs and increased inducibility of AT.

AT = macroreentrant atrial tachycardia; LVA = low-voltage area; V_GLA = global left atrial voltage.

Previous versus new hypothesis of sex differences in atrial structural remodeling

Previously, it was thought that women exhibited more advanced histopathological remodeling. However, using our integrated approach—combining detailed histological analysis of RA biopsy specimens with high-density bi-atrial voltage mapping—we now propose that there are no significant sex differences in histopathological parameters at any stage of structural remodeling in patients with non-valvular AF. Figure 10 illustrates the transition from the previous hypothesis to this new understanding. Notably, myocardial mass, as reflected by V_GLA, is inherently lower in women even before structural remodeling begins.

This intrinsic difference may explain why women more frequently exhibit LVAs, defined by thresholds such as <0.5 mV, and a higher inducibility of AT —both strongly linked to absolute voltage reduction.¹⁷⁾,¹⁸⁾,²¹⁾ Our findings also revealed a sex-specific difference in the V_GLA threshold predictive of AF recurrence, with a lower cut-off in women (3.3 mV) than in men (4.6 mV).²¹⁾ When translated into LVA measurements, these thresholds corresponded to LVA_0.5 >3.6 cm² for women and a more moderate LVA_1.0 >2.9 cm² for men.²¹⁾

Recent studies have explored not only the prognostic value of LVA-based assessment but also the efficacy of LVA ablation in combination with pulmonary vein isolation.⁴⁴⁾,⁴⁵⁾,⁴⁶⁾,⁴⁷⁾,⁴⁸⁾,⁴⁹⁾,⁵⁰⁾ Considering sex differences in these contexts may provide new insights if additional LVA ablation is performed under the assumption that LVAs represent the AF substrate.

The sex difference in atrial myocardium may be attributed to inherently smaller cardiomyocyte size in women. Additionally, women appear to have a greater hypertrophic reserve—a higher capacity for cardiomyocyte hypertrophy—which is an important consideration for future studies on sex differences in atrial function, heart failure, and clinical outcomes. If similar sex-specific characteristics exist in the ventricular myocardium, this could have particular relevance for investigations into sex differences in heart failure, especially heart failure with preserved LVEF.¹⁰²⁾,¹⁰³⁾

CHALLENGES IN STUDYING THE TEMPORAL EVOLUTION OF ATRIAL HISTOPATHOLOGY

Atrial cardiomyopathy is a progressive disorder marked by structural, electrical, and functional remodeling of the atrial myocardium (Figure 1).²⁾,³⁾ Its progression is driven by factors such as aging, metabolic stress, inflammation, and comorbidities, and can be further accelerated by conditions like AF and heart failure.¹⁰⁴⁾,¹⁰⁵⁾ Figure 8 illustrates this progressive remodeling, reflected as a reduction in atrial voltage with corresponding histopathological changes. As remodeling advances, LVAs emerge, the proportion of fractionated electrograms increases, slow conduction zones develop, AT becomes more readily inducible, and the risk of atrial tachyarrhythmia recurrence rises. A recent observational study of 1,488 patients undergoing AF ablation demonstrated that both the presence and extent of LVAs were associated with adverse long-term composite outcomes—including death, heart failure, and stroke—independent of AF recurrence or other confounding factors.¹⁰⁶⁾ Based on our previous research on atrial amyloidosis, it is plausible that some of these adverse outcomes may be partially attributable to underlying atrial amyloidosis.¹⁹⁾

To date, however, no studies have analyzed serial histological changes in atrial remodeling over time within the same individual. Most human studies are cross-sectional, limited to a single time point per patient, comparing histological features across patients with different AF durations or types, or between AF patients and controls, rather than performing repeated biopsies in the same patient.²⁾,³⁾,⁴⁾ As a result, current understanding of the temporal evolution of atrial histopathology in humans is largely extrapolated from cross-sectional studies and animal models rather than direct longitudinal human data.

We are currently conducting an ongoing study—Histological Evaluation of Atrial Fibrillation Substrate Based on Atrial Septum Biopsy (HEAL-AF 3) (Japanese UMIN Clinical Trial Registration UMIN000059261)—in which patients who underwent an initial biopsy at their first ablation receive a second biopsy during a repeat ablation for recurrent arrhythmia. This design allows for direct assessment of histological changes over time. Figure 11 shows 2 representative cases: in Figure 11A, the second biopsy demonstrated progression of interstitial expansion compared with the initial biopsy; in Figure 11B, interstitial expansion was already present initially, while the second biopsy revealed progression to interstitial fibrosis. This approach may provide novel insights into the true histopathological basis of atrial structural remodeling.

Figure 11. Two representative cases illustrating the histological progression of atrial cardiomyopathy. (A) A 67-year-old man with long-standing persistent AF (duration: 3 years) underwent an initial ablation, followed by a second session 12 months later due to recurrence. Compared with the initial biopsy, the second session specimens showed further expansion of intercellular space. (B) A 62-year-old woman with persistent AF (duration: 6 months) underwent an initial ablation, followed by a second session 2 years and 9 months later for recurrence. In her case, the initial biopsy already demonstrated intercellular space expansion, which had progressed to fibrosis by the second session.

AF = atrial fibrillation.

LIMITATION OF ATRIAL BIOPSY

RA septal biopsy was the sole source of histopathological material in our studies.¹⁷⁾,¹⁸⁾,¹⁹⁾,²⁰⁾,²¹⁾ Although integrated with bi-atrial electroanatomic mapping, RA septal findings may not fully reflect histopathological alterations in other atrial regions, particularly the LA. In addition, atrial structural remodeling can be regionally heterogeneous, as reported in patients with valvular heart disease and atrial amyloidosis. Future studies directly evaluating multiple atrial regions are therefore warranted.

FUTURE PERSPECTIVE

This review summarizes advances in understanding atrial cardiomyopathy through atrial biopsy studies. Given the diffuse and progressive nature of atrial structural remodeling, these underlying changes may help explain the limited effectiveness of additional substrate ablation in AF catheter ablation strategies.⁵⁰⁾,¹⁰⁷⁾,¹⁰⁸⁾ Pulmonary vein and superior vena cava ablation, targeting the major triggers of AF, remains well established and central to therapy. However, future interventions should shift from strategies that destroy myocardium toward approaches that prevent or mitigate myocardial degeneration. Lifestyle modification also represents a key component of therapy.¹⁰⁹⁾,¹¹⁰⁾

Clinically, future research should delineate how these findings can be translated into stratified management. For example, in patients with biopsy-detected atrial amyloidosis, earlier initiation of disease-modifying therapies may improve outcomes. Regarding sex-related differences, the adoption of sex-specific voltage cut-offs for risk prediction and outcome stratification should be further investigated. Moreover, optimization of ablation strategies—including adjustment of power settings according to atrial myocardial mass—may allow for safer and more individualized treatment approaches.

In this context, developing a histopathology-based classification of atrial cardiomyopathy and AF will be essential. Accurate surrogate markers that reflect the spectrum of histopathological myocardial changes are also needed. Finally, comprehensive genomic and multi-omics analyses,¹¹¹⁾,¹¹²⁾,¹¹³⁾ including studies of clonal hematopoiesis of indeterminate potential,¹¹⁴⁾ will be critical to further elucidate the pathophysiology of atrial cardiomyopathy.