The Atrium in Atrial Fibrillation - A Clinical Review on How to Manage Atrial Fibrotic Substrates.

Atrial fibrillation (AF) is the most common sustained arrhythmia in the population and is associated with a significant clinical and economic burden. Rigorous assessment of the presence and degree of an atrial arrhythmic substrate is essential for determining treatment options, predicting long-term success after catheter ablation, and as a substrate critical in the pathophysiology of atrial thrombogenesis. Catheter ablation of AF has developed into an essential rhythm-control strategy. Nowadays is one of the most common cardiac ablation procedures performed worldwide, with its success inversely related to the extent of atrial structural disease. Although atrial substrate evaluation remains complex, several diagnostic resources allow for a more comprehensive assessment and quantification of the extent of left atrial structural remodeling and the presence of atrial fibrosis. In this review, we summarize the current knowledge on the pathophysiology, etiology, and electrophysiological aspects of atrial substrates promoting the development of AF. We also describe the risk factors for its development and how to diagnose its presence using imaging, electrocardiograms, and electroanatomic voltage mapping. Finally, we discuss recent data regarding fibrosis biomarkers that could help diagnose atrial fibrotic substrates.

Introduction

Atrial fibrillation (AF) is the most common sustained arrhythmia and is associated with a substantial economic burden and significant morbidity and mortality (1, 2). AF can be asymptomatic or lead to symptoms such as palpitations, dyspnoea, and dizziness. The condition is associated with an increased risk of serious complications, including stroke (3), dementia (4), ventricular dysfunction, and death (3, 5). With a rising prevalence, it is estimated to affect nearly 17 million people in Europe by 2030, primarily driven by the aging of the population and increased survival with chronic cardiovascular diseases (3, 6–9).

In the past two decades, the knowledge of AF pathophysiology has led to significant developments in the treatment options, particularly regarding catheter ablation (10–13). Paroxysmal forms of AF are thought to primarily depend on triggers, primarily from the pulmonary veins (PV), while persistent forms involve a more significant modification of the atrial substrate (14, 15), promoting multiple re-entrant waves that maintain the arrhythmia.

Since a significant percentage of AF patients may have an indication for catheter ablation, analysis of the potential arrhythmogenic substrate is an essential part of the clinical evaluation of AF patients. Moreover, identifying the various risk factors promoting the development of a fibrotic substrate will enable a comprehensive approach to correct these factors, thus preventing the future progression of the arrhythmic substrate and increasing long-term therapeutic success.

Fibrotic atrial cardiomyopathy (FAC), a clinical entity proposed by Kottkamp (16), and one of the EHRAS atrial cardiomyopathy consensus classes (17), is a primary form of atrial pathology, characterized by extensive fibrosis as the substrate underlying atrial arrhythmias. This concept has been evolving ever since, and some authors (18) have used it more broadly to define significant atrial fibrosis due to several insults from different aetiologies. Understanding the multiple factors and mechanisms contributing to the complex development of atrial fibrosis and the management of AF based on arrhythmogenic substrates represents a challenge for interventional electrophysiology. At the same time, it may contribute to a more personalized approach, as the presence of atrial fibrosis and its characterization may guide the operator to modify the atrial substrate beyond PV isolation and estimate prognosis based on fibrosis characteristics.

Although atrial fibrosis has different clinical manifestations (like cardiac conduction disease and atrial thrombus formation), in this manuscript, we will review the role of the atrial fibrotic substrate in the context of AF. We will discuss the electrophysiology of the atria, the pathophysiology of atrial fibrillation, the molecular and genetic aspects, and the risk factors for fibrosis development. We will also review the principal elements of diagnosing the presence of atrial fibrosis using the 12-lead electrocardiogram, imaging, and electroanatomic voltage mapping and will discuss the most clinically relevant fibrosis biomarkers.

Pathophysiology of Atrial Fibrillation in the Fibrotic Atrial Substrate

Conceptual Framework for Atrial Fibrillation Pathophysiology

Atrial fibrillation has a multi-factorial nature and complex pathogenesis. Underlying mechanisms involve structural and electrical remodeling, autonomic nervous system dysfunction (19), and calcium dysregulation (20–23). The pathophysiological triangle for AF comprises triggers (for the arrhythmia initiation), a structural (typically fibrotic) substrate (for the maintenance of AF), and different modulators (that promote the propensity to AF through multiple potential mechanisms) (9, 16, 24, 25) (Figure 1). Re-entry is considered the primary mechanism for AF maintenance. Generally, it requires a vulnerable substrate characterized by slow conduction and short effective refractory periods, combined with a trigger to initiate the unidirectional block.

FIGURE 1

Pathophysiological dynamics in atrial fibrillation. Adapted from Kottkamp and Schreiber (24). JACC Clin Electrophysiol. AF, atrial fibrillation; PV, pulmonary veins.

Cardiac structural remodeling is characterized by atrial enlargement, a vital determinant of the persistence of AF-maintaining re-entry, and tissue fibrosis, characterized by the excessive accumulation of collagenous material in the extracellular space (12, 20, 26). Fibrotic atrial cardiomyopathy is a progressive disease with heterogeneous expressions, from mild to severe, and wide clinical variations, from asymptomatic to multiple arrhythmic manifestations (12, 16, 25, 27). Fibrosis is promoted by various risk factors (discussed below). It is involved in nearly all types of heart disease, including different ischemic and non-ischemic aetiologies (28). In many patients, AF can be understood as a manifestation of pre-existing atrial fibrosis, integrated into a gradual remodeling process (20, 27), albeit with a highly variable rate of progression determined by the dynamics of the fibrosis-promoting risk factors (29). In addition, AF itself promotes atrial fibrosis, which will contribute to AF progression and the development of therapeutic resistance in patients with long-standing arrhythmia (20, 30). Atrial fibrosis can interfere directly with impulse propagation by forming barriers to electrical conduction and separating the well-connected syncytium (31, 32). The increase in the extracellular matrix will disturb the continuity of the fibers bundle, causing local conduction disturbances (33). Additionally, direct electrical fibroblast-cardiomyocyte interactions may cause changes in cardiomyocyte electrophysiology (20, 34). Cardiac fibroblasts express multiple ion channels (35). Even though fibroblasts do not generate action potentials, they may influence cardiac electrophysiology by electrical coupling via gap junctions with cardiomyocytes (36). Finally, perivascular fibrosis around intracoronary vessels may impair oxygen and nutrient availability, promoting myocyte ischemia (37).

Prevalence and Mechanisms of Atrial Fibrosis

Cardiac fibrosis is pathological extracellular matrix (ECM) remodeling resulting in abnormal matrix composition (38). The cardiac ECM serves as a mechanical scaffold and is involved in the transmission of contractile force (39). The ECM consists of several proteins (40) like type I collagen (the most abundant protein), type III collagen, and a wide range of glycoproteins, glycosaminoglycans, and proteoglycans, and is a reservoir of stored latent growth factors and proteases, that can be rapidly activated following injury (41). Tissue remodeling results from an imbalance in the equilibrium of the normal synthesis process and degradation of ECM components (42). Extracellular matrix deposition is a physiologic and protective process essential for wound healing (Figure 2), but excessive or prolonged deposition can impair tissue function (43).

FIGURE 2

Fibroblast and Fibrin Activity in tissue healing. Visual representation of the pathophysiological process of reparative fibrosis after an injury to the cell.

In the normal heart, thin layers of perimysium and endomysium surround myocardial bundles and individual myocytes, respectively. The walls of the blood vessels also contain adventitial fibroblasts that contribute to the endomysial collagen network (44). In histologic analyses, two predominant types of myocardial fibrosis can be identified: interstitial fibrosis and replacement fibrosis (40). In a typical example of necrosis - myocardial infarction - necrotic cardiomyocytes are replaced by collagen-based scar, causing ‘replacement fibrosis.’ ‘Interstitial fibrosis’ (also called “reactive”) (45) describes the expansion of the endomysial and perimysial space caused by the net accumulation of ECM proteins in the absence of significant cardiomyocyte loss. The term ‘perivascular fibrosis’ is used to describe the expansion of the microvascular adventitia (46).

In the heart, ECM deposition is primarily mediated by the activation of fibroblasts in response to injury, transforming them into ECM-secreting myofibroblasts (47). Fibroblast-mediated fibrosis can affect every tissue and is a frequent pathological feature of chronic inflammatory diseases (48, 49). Similarly, expansion of the cardiac interstitium and deposition of ECM proteins are consistently noted in experimental models of heart failure (HF) and human patients with cardiomyopathic conditions, regardless of etiology (50).

Fibroblasts are the primary regulator of cardiac ECM. In response to disease stimuli, cardiac fibroblasts undergo cell state transitions to a myofibroblast phenotype (51). This transition is a dynamic state that underlies the fibrotic response (52). Most activated myofibroblasts in the infarcted and pressure-overloaded hearts derive from resident fibroblast populations (47). Myofibroblasts are fibroblast-smooth muscle cell hybrid that more effectively secretes and remodels the ECM positioned between all myocytes (53). Myofibroblasts have typically been defined by critical phenotypic features, including the de novo expression of markers including α-smooth muscle actin and periostin, increased production of ECM, and the ability to contract (54). Myofibroblasts are intimately associated with hypertrophic fibrotic scars in various injury models, and differentiation from fibroblast to myofibroblast is promoted by transforming growth factor-β (TGF-β), cytokines, the ECM, and other growth factors (55).

Several cytokines, chemokines, and growth factors are induced in the injured heart. In conjunction with elevated wall tension, specific signaling pathways and downstream effectors are mobilized to initiate myofibroblast differentiation (53). While the signaling mechanisms governing fibroblast to myofibroblast conversion are not fully elucidated, much has been discovered. Transforming growth factor β1 (TGFβ1) is considered a master regulator (56). The TGFβ-Smad signaling pathway has long been known to be involved in this process and is arguably one of the most potent inductive mechanisms (57). TGFβ drives fibroblast activation via the activation of phosphorylation of Smad2 and/or Smad3, which complex with Smad4, translocate to the nucleus and form a transcriptional complex that can directly bind to and transactivate essential ECM genes such as those encoding type I collagen (58). TGFβ may also work via a parallel non-canonical signaling pathway involving the activation of protein kinases such as p42/p44 MAPK (59). Several potential critical drivers of fibroblast activation post-MI include IL-1α/β, TGF-β1, collagen, fibronectin, osteopontin, thrombospondin-1, and secreted protein acidic and rich in cysteine (SPARC) as well as mechanical signals (e.g., scleraxis, TRPC, and MRTF/SRF) (60).

This is an area of intense and prolific investigation, leading to a rapid evolution of knowledge. Recent data on transcriptome maturation suggest that muscle blind-like 1 (MBNL1) is a post-transcriptional switch, controlling fibroblast state plasticity during cardiac wound healing (52). In this study, in healthy mice, cardiac fibroblast-specific overexpression of MBNL1 transitioned the fibroblast transcriptome to that of a myofibroblast and, after injury promoted myocyte remodeling and scar maturation.

Nonetheless, there are important existing knowledge gaps, the complete list of factors that involve fibroblast activation still needs to be identified, and the importance of individual factors ranked (51).

The gold standard for determining atrial structural remodeling is histology, which is challenging to apply in the clinical setting. Nevertheless, a few studies have included histological analyses, mainly in the surgical context. In these studies, hypertrophy of myocytes and areas of fibrosis, particularly in the left atrium, constitute the basis for AF in hypertensive patients (61). In addition, in valvular heart disease, severe fibrosis and hypertrophy with degenerative changes in atrial cardiomyocytes are the most prominent histologic findings in AF patients (62). Patients with long-standing persistent (‘chronic’) AF undergoing mitral valve surgery displayed abundant collagen fibers, inflammatory infiltrates, and sympathetic nerve twigs surrounding individual atrial cardiomyocytes, thus breaking up their clusters typically seen in sinus rhythm patients (63).

In a study investigating whether patients who develop postoperative AF show pre-existent alterations in right-atrial histopathology (64), the investigators analyzed samples from the right atrial appendage (immediately collected after opening the pericardium) from seventy patients undergoing elective coronary revascularization. The histologic abnormalities associated with the development of postoperative AF in 22 (31%) patients were interstitial fibrosis, vacuolization, and nuclear derangement of myocytes. In multivariate analysis, myocyte vacuolization and nuclear derangement represented independent predictors of postoperative AF.

Molecular Aspects of Atrial Fibrogenesis

The development of fibrosis is a highly complex, multifactorial, and patient-specific process (Figure 3). Despite the growing interest in the subject over the past few years, the precise molecular mechanisms and signaling pathways involved in developing the human AF substrate are not entirely understood. Nevertheless, three interrelated signaling pathways appear to play a central role: the renin-angiotensin system (RAAS), the transforming growth factor-β1 (TGF-β1), and the oxidative stress pathways (65–67).

FIGURE 3

Etiology of atrial fibrosis. Different pathological insults, risk factors, and certain genetic diseases induce atrial fibrosis. Atrial fibrosis is characterized by myofibroblast growth and extracellular matrix (ECM) remodeling.

The RAAS plays a crucial role in cardiac structural remodeling and the development of myocardial fibrosis in several diseases states, including cardiomyopathy (65). Activation of the RAAS induces oxidative stress, which contributes to cardiovascular inflammation, fibrosis, and dysfunction (68). Angiotensin-converting enzyme (ACE) overexpression results in atrial fibrosis in several animal models (69–71), whereas the use of ACE inhibitors delays atrial fibrosis and reduces AF vulnerability and AF progression (65, 72, 73). In the right atrial tissue of patients undergoing open-heart surgery (74), the increase of atrial collagen deposition observed in atrial samples from AF patients undergoing open-heart surgery was also attenuated in those previously under ACE-inhibitor therapy, and the atrial micro-capillary density in these patients was similar to patients in normal sinus rhythm. In agreement, other studies have shown that ACE-inhibitor therapy is associated with a significant reduction in recurrent AF (75–77).

TGF-β1 is implicated in tissue repair and development of fibrosis, including atrial myocardial fibrosis, by enhancing collagen synthesis (18).

Inflammation has been implicated in various AF-related pathological processes, including oxidative stress, fibrosis, and thrombogenesis (78). Inflammation and oxidative stress may promote AF, as suggested by increased C-reactive protein (CRP) and evidence of oxidative injury seen during AF (79–81). AF induces substantial oxidative stress in fibrillating atrial tissue (14).

In many diseases, tissue inflammation is a significant trigger for fibrosis development (82). The inflammatory response is mediated by inflammasomes, which are intracellular multiprotein complexes that can trigger the host-defense response (83). The inflammasomes comprise a family of cytosolic pattern-recognition receptors called nucleotide-binding oligomerization domain (NOD)–like receptors (NLRs) that are involved in innate immune recognition of pathogen-associated molecular patterns as well as intracellular and extracellular damage-associated molecular patterns (84). Functionally, inflammasomes are sensors and receptors of the innate immune system that can induce inflammation in response to pathogens and molecules derived from host proteins. In response to these “cellular danger signals,” the inflammasomes activate caspase-1 and release both IL-1β and IL-18 via pores formed by the N-termini of gasdermin-D, which are cleaved by activated caspase-1 (85). Activation of the NLR family pyrin domain containing 3 (NLRP3) inflammasome is increased in patients with paroxysmal and long-standing-persistent AF (86), patients that go on to develop post-operative AF (87), and patients with risk factors for AF such as diabetes and obesity (88, 89) via both priming (increased expression of components of the NLRP3 inflammasome) and triggering (assembly of the NLRP3 complex) mechanisms. Activating NLRP3 selectively in atrial cardiomyocytes is sufficient to promote atrial structural remodeling (atrial hypertrophy), spontaneous premature atrial contractions, and inducible AF (86). The previously cited and fascinating clinical study (87) analyzed tissue from patients with postoperative AF, preexisting Ca2+-handling abnormalities, and activation of NLRP3-inflammasome/CaMKII signaling were evident in atrial cardiomyocytes.

Inflammasome signaling and downstream cytokine responses mediated by the inflammasome have been found to play an important role not only in wound healing but also in fibrosis.

Inflammasome activation induces the differentiation of quiescent fibroblasts to myofibroblasts (84). In addition, it is hypothesized that chronic dysregulation of the inflammasome promotes the differentiation of myofibroblasts, leading to excessive extracellular matrix accumulation and subsequent failure of the affected organ (90). The inflammasome regulates the secretion of IL-1β and IL-18 cytokines (91), and both are critical for repairing damaged tissue and play a role in fibrosis. Inflammasome-mediated activation of IL-18 in the myocardium is a crucial trigger for the cytokine cascade and macrophage infiltration in the heart, leading to adverse cardiac remodeling (92). However, what dictates the delicate balance between routine wound healing versus fibrosis is yet to be fully elucidated (90).

Several studies have linked fibrosis to perturbations in cardiac (myo)fibroblast calcium (Ca2+) handling and electrophysiology, providing a basis for future investigation of molecular targets for the prevention of fibrosis progression (93, 94). For example, transient receptor potential (TRP) channel remodeling has been implicated in profibrotic atrial remodeling in large animal models and human samples. TRP melastatin-related 7 (TRPM7) is a Ca2+-permeable channel upregulated in atrial fibroblasts from AF patients, likely in a TGF-β1-dependent manner (95). TRPM7 downregulation reduced basal AF fibroblast differentiation as well as TGF-β1 induced fibroblast differentiation in culture (95). Similarly, TRP canonical 3 (TRPC3) expression is upregulated in atria from AF patients, goats with electrically maintained AF, and dogs with tachypacing-induced HF, whereas TRPC3 knockdown decreased canine atrial fibroblast proliferation (96). Moreover, in vivo administration of the TRPC3 blocker pyrazole-3 suppressed AF in dogs while decreasing fibroblast proliferation and extracellular matrix gene expression (96). Various molecules have been associated with disturbances in atrial Ca2+ handling in AF. Patients with AF have elevated atrial endothelin-1 levels, associated with increased atrial preexcitation (97), inadequate Ca2+ leak, and increased intracellular overload. Additionally, mice with cardiac-specific knockout of liver kinase B1 (LKB1), a protein highly expressed in the heart and responsible for regulating myofilament response to Ca2+, developed early-onset atrial cardiomyopathy (98). Fibrosis progression has also been associated with atrial ion channel remodeling (36). Wiedmann et al. (99) showed that the TASK-1 [two-pore-domain potassium (K+) channel that contributes to the regulation of atrial action potential duration] is decreased in AF-prone transgenic mice, leading to both FAC and AF progression. For some K+ channels expressed in atrial fibroblasts, their profibrotic effects have been attributed to increasing the driving force for fibroblast Ca2+ entry, e.g., in the case of HF-related upregulation of KCNJ2, underlying the inward-rectifier K+ current (100). Similarly, mutations in the voltage-gated sodium channel have been associated with LA dilatation (101).

Genetic Basis of Atrial Fibrillation

Atrial fibrillation has precise genetic determinants, including common and rare gene variants with variable penetrance (17, 102–104). Over the last decades, multiple studies have observed familial aggregation of individuals with lone AF (105). A family history of AF in a first-degree relative independently increases AF risk twofold (7), with the most substantial risks associated with young age at AF onset and multiple affected relatives (106). Genome-wide association studies have identified genetic variants associated with increased susceptibility to atrial fibrillation, with the strongest hits clustering on chromosome 4q25, close to the gene for the homeobox transcription factor PITX2 and single nucleotide polymorphisms in T-box (TBX)5 (107–110). However, in most individuals, atrial fibrillation is a complex trait reflecting the combined effects of aging, genetic predisposition, comorbidities, and environmental factors (111). Both standard and rare genetic variants increase susceptibility to AF in the presence of specific risk factors (104).

Inherited arrhythmia syndromes are commonly known as ‘channelopathies,’ highlighting that mutations in genes encoding cardiac ion channels are the predominant cause of these conditions (112). There is considerable overlap in ion channel genes responsible for causing arrhythmia syndromes between atria and ventricles, with genetic defects recognized to cause episodic arrhythmias in either chamber (113). Similarly, in a significant percentage of patients atrial dilated cardiomyopathy with the fibrotic structural substrate may represent the counterpart of idiopathic ventricular dilated cardiomyopathy, which is often of genetic origin (114). Still, very little information is about the genetic causes of specific atrial cardiomyopathy. Nevertheless, some studies have identified variants in non-ion channel genes as a cause of primary arrhythmogenic atrial cardiomyopathy in the last years. Hodgson-Zingman et al. (115) reported a genetic mutation in the atrial natriuretic peptide gene – Natriuretic Peptide Precursor A (NPPA) – in a large family with AF. They demonstrated the novel observation of the effects of this neuro-hormone on the action potential of the atrial myocardium. Subsequent work has implicated this gene mutation in inherited atrial cardiomyopathy (114). They investigated the evolving arrhythmic substrate in 5 patients with isolated arrhythmogenic atrial cardiomyopathy, caused by NPPA gene mutation, with repeated electroanatomic mapping and tomographic evaluations and reported that the evolution of the arrhythmic patterns to sinus node disease with atrial standstill was associated with giant atria with extensive areas of low voltage and atrial scarring. They concluded that the evolution of the amount and distribution of atrial scarring/fibrosis constitutes the structural substrate for the different types of atrial arrhythmias in a pure genetic model of arrhythmogenic atrial cardiomyopathy.

Peng et al. (116) identified a family with heritable atrial cardiomyopathy manifesting as progressive atrial−selective electromechanical dysfunction, tachyarrhythmias, and bradyarrhythmias requiring pacemaker implantation. Myosin light−chain 4 (MYL4), encoding the atrial−selective essential myosin light chain, was identified as a candidate gene. Genetically modified rat models knocking out the MYL4 gene or knocking in the human MYL4 p.E11K mutation showed early atrial fibrosis associated with enhanced proapoptotic and profibrotic signaling associated with atrial cardiomyopathy featuring atrial arrhythmia, atrial contractile failure, and atrial enlargement. The C allele and CC genotype of rs4968309 in MYL4 were also associated with AF onset and recurrence in patients after catheter ablation (117).

Interestingly, in silico and functional studies suggest that atrial fibrillation-associated genetic variants generate an arrhythmogenic atrial cardiomyopathic substrate (111). A better understanding of AF heritability will improve AF prediction models and be the next step toward more efficient personalized treatment strategies (118). Nevertheless, most patients have significant acquired risk factors predisposing to this fibrotic response, which we will describe in the next section.

Risk Factors for the Development of Atrial Fibrosis

Numerous risk factors have been identified as contributors to fibrosis and AF’s development and dynamic progression. They include but are not limited to advanced age, HF, obesity, hypertension, sleep-disordered breathing, and diabetes (119).

Aging

Incidence and prevalence of AF are age-dependent (7, 120), with increasing fibrosis being a characteristic of the aging heart (49, 121). Age-associated changes of the atria include global and regional reductions in atrial voltage with an increased heterogeneity, conduction slowing (with alterations of the wavefront propagation), prolongation of atrial refractoriness, fractionated electrograms, and double potentials (122, 123).

Heart Failure

Atrial fibrillation and congestive HF are commonly encountered together, and each condition predisposes to the other. In the Framingham cohort (124), HF was significantly associated with AF risk in both sexes (OR, 4.5 for men and 5.9 for women). The prevalence of AF in patients with HF increases from <10% in those with New York Heart Association (NYHA) functional class I HF to approximately 50% in those with NYHA functional class IV HF (125). Moreover, diastolic dysfunction appears to be a potent precursor of AF, with an independent, graded relationship between the severity of diastolic dysfunction and the development of AF (126). Patients with HF who have concurrent AF have worse outcomes (127).

The HF and AF share common mechanisms, including myocardial fibrosis and dysregulation of intracellular calcium and neuroendocrine function (128). In animal models of HF induced by rapid ventricular pacing, there was a more significant atrial interstitial fibrosis than in AF induced by rapid atrial pacing (129). This study’s histological analysis displayed extensive interstitial fibrosis accompanied by cell loss, degenerative changes, and hypertrophy. The connective tissue was composed of increased numbers of fibroblasts, large amounts of collagen, ground substance, and occasionally fat cells. These changes were more extensive in LA. Subsequent work (130) revealed that apoptosis, leukocyte infiltration, and an increased cell death rate occur before arrhythmogenic atrial structural remodeling associated with experimental HF. These authors suggested that apoptosis is more likely associated with the pathophysiological mechanisms leading to the AF substrate rather than a result of AF per se.

Although both experimental paradigms promote AF, the atrial cellular electrophysiological substrate produced by HF is different from that seen with atrial tachycardia-induced remodeling. Similarly, HF and cAF produce distinct electrical and calcium-handling remodeling in human atrial samples, with repolarization shortening in cAF but not HF (131). By contrast, protein levels of ECM components are significantly increased in HF patients (131), suggesting a significant role of re-entry-promoting structural remodeling in AF development. Thus, HF promotes the presence of AF (by producing an altered substrate), and, in turn, the presence of AF worsens the prognosis of the patient with HF. It should be highlighted that CHF has different dynamic components with distinct time courses, which can further modulate the interaction between AF and HF (29).

Obesity

There is a strong correlation between obesity and AF (132–134). In a meta-analysis of 16 studies (135), obesity increased the risk of developing AF by 49% in the general population. Additionally, obesity is usually accompanied by several other risk factors predisposing to developing AF (136). Epicardial adipose tissue is metabolically active (137), with its cardiometabolic risk being comparable to other visceral fat stores. Specifically, it can directly affect the atrial myocardium by releasing adipokines, which promote inflammation and fibrosis (138). Epicardial adipose tissue accumulation is closely associated with atrial and ventricular arrhythmias and electrocardiographic signs associated with arrhythmogenesis (139). Patients with AF have higher levels of epicardial adipose tissue than controls, and those with chronic AF are more likely to have a higher volume of epicardial adipose tissue than those with paroxysmal AF (140, 141). Volume or thickness of epicardial adipose tissue, measured on cardiac computed tomography (CT) and cardiac magnetic resonance (CMR), are predictors of the presence, severity, and recurrence of AF (142).

Hypertension

There is a well-established association between hypertension and AF (143–146). Although, from an epidemiological perspective, it is still unclear whether the risk of AF rises linearly with blood pressure (BP) or whether there is a BP threshold above which the risk increases (147). In the Framingham study, hypertension added an excess risk for AF of 50% in men and 40% in women (124). In animal models of experimental hypertension, the high BP rapidly induced LA hypertrophy, fibrosis, and inflammation (14). In humans chronic systemic hypertension with left ventricular hypertrophy is accompanied by atrial remodeling characterized by slowing of global and regional conduction, increase in low voltage areas, and easier inducibility of sustained AF (148). LA enlargement and associated P-wave changes predict AF occurrence in hypertensive patients (149, 150). Hypertension is also a risk factor for arrhythmia recurrence after AF ablation, but it is unclear whether this is independent of other factors such as atrial size (17).

Sleep-Disordered Breathing

Sleep-disordered breathing (SDB) has been linked to long-term adverse outcomes and is proposed as an additional and independent risk factor for cardiovascular diseases (151). SDB is highly prevalent among AF patients (from 21 to 74%), promotes arrhythmogenesis, and impairs treatment efficacy (152, 153). There is evidence that SDB – may – promote atrial fibrosis in animal models. Previous studies showed that SDB induces conduction slowing decreases matrix metalloproteinase-2, changes atrial connexin-43 expression and distribution, and significantly increases atrial fibrous tissue content (154, 155). In the clinical setting, the atria of SDB patients have extensive areas of low voltage and conduction abnormalities (156). A meta-analysis of observational studies concluded that SDB was associated with AF recurrence after catheter ablation (157). Patients with SDB had a 31% greater risk of AF recurrence after successful catheter ablation than patients without SDB. Importantly, in the same study, the efficacy of catheter ablation for AF was similar between patients without SDB and those with SDB undergoing continuous positive airway pressure treatment.

Diabetes Mellitus

Diabetes mellitus (DM) is an independent risk factor for the development and progression of AF (158). Patients with DM have a 40% higher risk of developing AF than patients without DM (159). They have increased levels of angiotensin II, TGF-β signaling, adipose tissue, systemic inflammation Campo (160), larger atria, lower atrial voltage, and higher recurrence of AF after ablation (161). Evidence of widespread fibrotic deposits in the atria was also found in DM animal models (162).

Sex

There are well-established sex differences in AF regarding its epidemiology, with a lower age-specific prevalence in women (women presenting at a later age) and its clinical presentation, with women more likely to be symptomatic (163). Globally the number of men and women with AF are similar since, on average, women live longer than men (164), and after 75 years of age, about 60% of the people with AF are women. Women with AF have an increased risk of stroke and death compared to men (165), which, besides differences in treatment, might in part be explained by the interesting observation that women with AF have a more significant atrial fibrosis burden, which may predispose them to more AF-associated complications. In a study with CMR in 939 patients (166), advancing age and female sex were associated with a higher burden of atrial fibrosis in AF patients. Women with a prior history of stroke also had more fibrosis than women and men without a history of stroke. In another study, female sex and AF persistence were independently associated with the presence of fibrosis on delayed enhancement CMR (167). The delayed enhancement was variably distributed in this population but more frequently detected in the posterior wall. Thus, females may have a higher probability of the presence of atrial fibrosis and atrial myopathy. Despite these observations, the mechanisms underlying differences between sexes are mainly unknown.

Identifying Fibrotic Atrial Substrates by Electrocardiogram

The electrocardiogram (ECG) during normal sinus rhythm could be a tool to characterize the fibrotic substrate and predict AF risk (168). Interestingly in a study with 285,933 individuals (169), compared with the reference group (P wave duration of 100–105 ms), individuals with very short (≤89 ms; hazard ratio [HR] 1.60, 95% confidence interval [CI] 1.41–1.81), and very long P-wave duration (≥130 ms; HR 2.06, 95% CI 1.89–2.23) had an increased risk of incident AF.

The rationale behind using ECG as a prediction tool is that atrial remodeling is associated with an increased risk of AF and can be detected by a shift in the P-wave axis (170). The terminal force of the P wave during sinus rhythm in lead V1 (PTFV1) correlates with LA anomalies (171). A PTFV1 > 0.06 mm/s is associated with an increased risk for the development of AF (hazard ratio 4.02, 95% confidence interval 1.25–17.8; P = 0.018) (172, 173), and PTVF1 is independently related to cryptogenic, cardioembolic and ischemic strokes (174, 175). Suppose this evidence of inter-atrial conduction block is present in the absence of chamber enlargement or ischemia (especially in elderly patients with P-wave duration >140 ms). In that case, it can be a marker of short-term development of AF, an association called Bayés syndrome (176). Furthermore, PTFV1 ≥0.04 mm/s, along with P-wave duration ≥125 ms and P-wave dispersion ≥40 ms, are predictors of AF recurrence post-PVI (P wave duration >125 ms had 60% sensitivity, 90% specificity, positive predictive value of 72% and negative predictive value of 83.7%) (177).

Recently, a deep convolutional neural network trained on >1 million 12-lead resting ECGs predicted new-onset AF within 1 year (178). This model classified 62% of all patients who experienced an AF-related stroke within 3 years of the index ECG as being at high risk for new-onset AF. Thus, electrocardiographic analysis during sinus rhythm could be an additional tool to detect the presence of a vulnerable atrial substrate. However, the exact pathophysiological features and mechanisms detected by such approaches remain incompletely understood. A better insight into the underlying mechanisms may help improve early tailored treatment to prevent substrate progression and the occurrence of adverse outcomes. Even though the ECG is one of the oldest ancillary exams in cardiology, the widespread availability of standardized digital ECGs provides an opportunity for deep learning to make a significant clinical impact in cardiac electrophysiology, including characterization of the atrial cardiomyopathy (179).

Characterizing Atrial Structural Remodeling by Non-Invasive Imaging

Non-invasive imaging is a powerful tool for identifying patients with atrial fibrosis. Echocardiography, CT, and CMR are useful for assessing the LA structure.

Echocardiography

Echocardiography is the modality of choice for screening patients with cardiac pathologies, including those involving the LA (180). Given the non-uniform nature of remodeling (17), real-time 3D echocardiography (3DE) technology compared with CMR reference is more accurate than conventional 2D-based analyses, resulting in fewer patients with undetected atrial enlargement [3DE-derived LA Volume values showed higher correlation with CMR than 2DE measurements (r = 0.93 vs. r = 0.74 for maximal LAV; r = 0.88 vs. r = 0.82 for minimal LAV)] (181).

Increased LA size on echocardiography is associated with a higher recurrence rate of AF treated with ablation (182) and an increased risk of stroke in patients with non-valvular AF (183, 184). Assessment of LA function can be performed by pulsed-wave Doppler measurements. Certain features, like LA active relaxation and contraction, are altered in AF patients compared to subjects with sinus rhythm, regardless of LA size and age (185). Total atrial conduction time during sinus rhythm can be estimated as the interval from the beginning of the P-wave on the body-surface ECG to the peak A’-wave on the tissue-doppler imaging (TDI) tracing of the LA lateral wall on echocardiography. This echocardiography-derived PA-TDI duration reflects electrical and structural changes to the atria (186). PA-TDI is prolonged in AF patients, including those without overt cardiovascular disease (idiopathic AF) (187), and is associated with AF recurrence after ablation in paroxysmal AF patients (188). Two-dimensional speckle-tracking echocardiography, a method to quantify atrial deformation, has also been used as a sensitive marker to detect early functional remodeling before anatomical alterations occur (189). Reduced atrial strain, as calculated using speckle-tracking, has been correlated with reduced atrial compliance and increased fibrosis. In a study by Rivner et al. (168), it was reported that global compliance tended to be an independent determinant of the presence of low-voltage zones (odds ratio 1.347, P = 0.046) and is a predictor of the development of AF and AF recurrence after ablation.

Cardiac Computed Tomography

Cardiac CT is a method with excellent spatial resolution compared to CMR (190), enabling accurate assessment of atrial volume and LA wall thickness. On the other hand, it has a low contrast-to-noise ratio, which reduces its ability to distinguish between normal myocardium and scar (191). Before catheter ablation, LA volume and LA asymmetry (asymmetry over 60% predicted AF recurrence with 74% sensitivity and 73% specificity), predict the likelihood of maintaining sinus rhythm post-AF ablation (192, 193). CT-based local wall deformations correlate better with extended low-voltage areas than other remodeling surrogates (194). The progression of the shape of the LA roof determined by CT correlates with the development of non-PV arrhythmic substrate in patients undergoing AF ablation (195, 196).

Cardiac Magnetic Resonance

Cardiac magnetic resonance has become the gold standard in volumetric LA structure and function assessments.

Contrast-enhanced CMR with gadolinium is additionally used to detect atrial fibrosis (197, 198) and can non-invasively identify atrial scar, which has been shown to spatially correlate with low-voltage areas (199). Other studies have reported the feasibility of delayed-enhancement CMR to quantify fibrosis in the LA and show that a high degree of delayed enhancement in the LA is associated with a more complex and extensive ablation with AF termination as the endpoint (200). Spragg et al. (201) reported a sensitivity and specificity of LGE for discrimination of low-voltage areas of 0.84 and 0.68, respectively. Delayed-enhancement CMR correlates with surgical biopsy results and is strongly associated with AF recurrence after catheter ablation (202, 203).

Marrouche’s group introduced the Utah scoring system (204). This score classified patients by the extent of enhanced LA area into four groups: 1 (<5%), 2 (5–20%), 3 (20–35%), and 4 (>35%). In this study, procedural outcomes were predicted by the baseline LA scar burden. During follow-up, all patients in group 1 were free of AF, but in group 4, only 4% of patients remained AF free. Other studies also found that LA fibrosis detected by CMR is associated with appendage thrombus and spontaneous contrast (205) and is an independent risk factor for stroke in patients with AF (206, 207).

Technological developments have expanded CMR use, and atrial 4-dimensional flow CMR recently emerged as a novel non-invasive approach that characterizes Campo’s atrial flow dynamics. It allows measurement of 3D blood flow and the derivation of stasis maps, providing visualization and quantification of potentially thrombogenic stasis in the LA and left atrial appendage (208–211). Similarly, non-invasive digital atrial twins based on patient-specific CMR imaging can integrate anatomical, structural, and functional determinants of atrial electrophysiology and arrhythmogenesis (29). Proof-of-concept studies have shown the promise of this approach for guiding AF ablation, and initial randomized trials comparing simulation-guided versus standard clinical therapy are ongoing (119). With these recent advances, CMR imaging can provide comprehensive images of the heart in patients with various cardiac diseases, adding prognostic value (212).

Identifying Atrial Fibrosis by Electroanatomic Voltage Mapping

Electroanatomic voltage mapping (EAVM) plays an essential role in diagnostic and therapeutic mapping and ablation in AF patients (17), providing information regarding local voltage abnormalities that may be used as a surrogate marker of myocardial health (213). Animal studies have demonstrated the histological correlation between low voltage areas (LVA) and atrial scars (214). However, the methodology for defining LVA has not been standardized, and a clear voltage threshold for abnormality has never been histologically validated (215). Several studies have compared the voltage maps with CMR findings, correlating the bipolar voltage with late gadolinium enhancement (203, 214, 216). In one of these studies, the overlap of late gadolinium enhancement areas with LVA (defined as <0.5 mV) had a sensitivity of 84% but a specificity of only 68% (201). One potentially confounding factor is the atrial rhythm during mapping, an important determinant of voltage (217). The voltage of bipolar signals during AF is significantly reduced compared to sinus rhythm (218). It has been suggested that the correlation between LVA and posterior LA delayed enhancement on CMR (219) is significantly improved when acquired during AF compared to sinus rhythm because the fixed-rate and wavefront characteristics present during sinus rhythm may not accurately reflect underlying functional vulnerabilities responsible for AF maintenance. This is an important area of investigation to improve the correlation between voltage mapping and atrial fibrosis. New technological developments like omnipole mapping and dynamic voltage attenuation may further enhance the detection of the abnormal atrial substrate (220).

During the last decade, technological developments of EAVM have helped identify different arrhythmia patterns and locations, generating new insights into the pathophysiological mechanisms of AF. In addition, progress in body-surface mapping and computer processing has allowed non-invasive mapping of atrial activation with increasing accuracy (221). For example, Metzner et al. (222) reported that non-invasive epicardial and endocardial electrophysiology systems produce comparable characterization of rotational sources with invasive mapping. A non-invasive evaluation of segmented images (223), used to construct personalized 3D models of the fibrotic atria with biophysically realistic atrial electrophysiology, demonstrated that AF in fibrotic substrates is perpetuated by re-entrant drivers (rotors). This and several other observations have led to the hypothesis that fibrillation mechanisms may exist along a continuous spectrum, with the specific electrophenotype determined by the degree of remodeling of the underlying myocardial substrate (224) particularly the extent of atrial fibrosis.

Finally, the presence of LVA could be considered in thromboembolic risk stratification, as the presence of LA LVA correlates with a higher incidence of previous stroke or the presence of pre-existing procedure-independent silent cerebral events on cerebral delayed enhancement MRI (225).

Biomarkers of Atrial Fibrosis

Several biological markers reflecting atrial stress, inflammation, endothelial dysfunction, kidney dysfunction, and atherosclerosis have been associated with future AF events, further supporting the correlation between inflammation (and fibrosis) and atrial dysfunction in a population at risk for AF (226). The utility of these markers is the possible identification of the presence of atrial myopathy during incipient stages of the disease and the identification of ‘high-risk’ patients for thromboembolic complications and stroke. For example, inflammation and fibrosis biomarkers (CXCL16, FABP3, PIGF, and MMP-9) were higher in subjects with worse LA reservoir function (227) in a population at risk of AF. Furthermore, blood-derived biomarkers (such as markers of inflammation, coagulation activity, cardiovascular stress, myocardial injury, and cardiac and renal dysfunction) can contribute to refining risk assessment for stroke outcomes and mortality in the presence of AF (228), since currently used clinical scores (e.g., CHA2DS2-VASc) only provide modest discrimination of stroke risk. Recent studies of biomarkers in AF have shown that they significantly improve risk stratification (229, 230).

Troponin

Elevated troponin levels have been associated with an increased incidence of AF (231–233). However, the optimal cut-off to determine the risk of AF is unclear. There are no AF primary prevention studies using troponin screening, so it is unclear how detectable troponin levels in the absence of AF will change clinical management (234). Troponin levels increase immediately after AF ablation. More significant elevation of troponin levels is related to favorable outcomes after ablation and more significant reversal of structural remodeling. In multivariate analysis, the TnT level was the only independent predictor for responders (odds ratio 90.1; 95% confidence interval 14.95–543.3; P < 0.0001) (235). The reason for this paradoxical observation may be the presence of healthy myocardium, as more troponin T would be released by radiofrequency ablation in a healthy LA than in a ‘sick’ LA (in which the myocardium had already degenerated into fibrous tissue) and postprocedural troponin levels may therefore reflect preservation of healthy LA myocardium.

Natriuretic Peptides

Some studies have shown that natriuretic peptides are elevated in patients with paroxysmal AF compared with matched controls in sinus rhythm (236, 237). Natriuretic peptides levels fall rapidly after restoring sinus rhythm (238, 239). However, the usefulness of natriuretic peptide (NT-proBNP) levels to predict the maintenance of sinus rhythm after successful cardioversion remains controversial (240, 241). Nevertheless, the addition of NT-proBNP to the CHADS2 and CHA2DS2-VASc risk stratification models significantly improves discrimination performance (242). Hijazi et al. (228) reported an adjusted hazard ratio of 4.0 (95% confidence interval, 3.2 to 5.0; P < 0.001).

Collagen

There are two significant biomarkers of collagen metabolism, the procollagen type-III N-terminal propeptide (PIIINP) and collagen type-I carboxy-terminal telopeptide (ICTP). PIIINP reflects collagen synthesis and degradation, whereas ICTP reflects collagen degradation only (243). In a large cardiovascular disease-free, multi-ethnic, and middle-aged sample population, PIIINP and ICTP predicted new onset of AF during a median follow-up of 10 years (244). A combination of circulating biomarkers reflecting excessive myocardial ICTP is also associated with higher AF prevalence, incidence, and recurrence after ablation (245). In the later study, the adjusted hazard ratio for AF recurrence was 3.4 (p = 0.008).

ST2

Suppression of tumorigenicity 2 (ST2) is a member of the IL-1 receptor (IL-1R) family that plays a major role in immune and inflammatory responses (246). In recent years, knowledge about ST2’s role in the pathophysiology of cardiovascular diseases has expanded, with strong links to myocardial dysfunction, fibrosis, cardiovascular stress, and remodeling (247). Although ST2 concentrations do not improve risk discrimination in AF patients (231) its levels are of prognostic value in patients on anticoagulation (248). Concentrations of sST2 were also significantly associated with the risk of mortality, even after adjusting for the CHA2 DS2 -VASc score [HR: 1.007 (1.001–1.013); P = 0.014] (248). ST2 has been proposed as a screening tool to detect atrial fibrosis in AF patients to allow a more aggressive therapy (234). Likewise, ST2 may be an objective biomarker to predict the risk of arrhythmia recurrence after ablation, emergency admission, or HF events (249, 250). In a study by Kim et al. (251), mean ST2 was higher in AF, persistent AF, and symptomatic AF and decreased after ablation. Another study, in a population of patients scheduled for cryoballoon catheter ablation (252) analyzed the relationship between ST2 and recurrence of AF. ST2 was the only independent parameter predicting AF recurrence [sensitivity: 77.3%, specificity: 79.5%; area under the curve was 0.831 (p < 0.001)]. It might therefore be a useful marker for detecting patients with high-grade fibrosis who will benefit less from ablation.

Galectin-3

Galectin-3 is a marker of myocardial fibrosis, which may be involved in AF-promoting atrial remodeling (253). Galectin-3 levels correlate with LA volume and are increased in AF patients (254). In a study that compared galectin-3 levels in patients with myocardial infarction and with or without AF (255) the patients with AF had higher levels of C-reactive protein (p < 0.01) and galectin-3 (p < 0.05) than those without AF. Patients with high galectin-3 had 4.4 times greater odds of having AF. Galectin-3 levels were lower in patients without AF (p < 0.01) than in those with permanent/persistent AF.

A recent meta-analysis showed that higher galectin-3 levels might be associated with an increased risk of AF recurrence in catheter ablation patients (256).

Transforming Growth Factor-β1

Several stimuli that promote AF converge in increased expression levels of TGF-β1, which in turn provokes interstitial fibrosis (257). TGF-β1 is a profibrotic cytokine and a central growth factor involved in regulating atrial fibrosis (258). Upon binding to its receptor, TGF-β1 leads to the activation of intracellular signaling cascades, which ultimately alter the expression of genes involved in differentiation, chemotaxis, and proliferation (259).

High plasma levels of TGF-β1 have been correlated with increased LA volumes and reduced bipolar voltage on electroanatomic voltage mapping (260), but results regarding its effect on the incidence of AF have been contradictory (261–263).

Therefore, researchers have been actively looking for additional biomarkers to predict therapeutic failure or morbidity mortality in patients with AF in the last decades. The measurement of several of the above-described biomarkers, such as troponin and natriuretic peptides, has consistently been demonstrated to improve risk prediction in addition to the clinical risk stratification models. However, there are still several significant uncertainties concerning its real value in detecting underlying atrial disease, partly due to these biomarkers’ unspecific nature.

Treatment

Determining the degree of the atrial fibrotic substrate is important for deciding the therapeutic strategy. This determination may help identify patients with the highest probability of success if treated with ablation therapy or, on the contrary, the population of patients with evidence of advanced disease and should be managed with medical treatment only.

Medical Treatment

Preclinical and Investigational Pharmacological Agents That May Directly Modulate the Fibrotic Substrate

As previously stated, diffuse excessive production and deposition of ECM is the primary manifestation of fibrosis. Despite extensive research in this area, there is a lack of efficacious therapies for inhibiting or reversing cardiac fibrosis, mainly due to the complexity of the cell types and signaling pathways involved (264). As an example, a clinical study that sought to determine whether matrix metalloproteinase (MMP) inhibitor, PG-116800, reduced left ventricular (LV) remodeling after myocardial infarction (MI) failed to show this objective and improve clinical outcomes (265).

The TGFβ is one of the regulators in the heart remodeling after injury. By this, the targeting of the TGFβ signaling pathway (266) has been explored as a potential therapy to inhibit fibrosis. In this context, Inhibitors of TGFβ Receptors I and II (receptors that activate the TGFβ signaling) have been tested. These inhibitors have demonstrated that they reduce myocardial fibrosis in animal models of Chagas disease and myocardial infarction (267, 268). Despite these beneficial effects, increased mortality and inflammation were observed (269, 270), and in long-term inhibition, cardiac toxicity (271). Nonetheless, novel TGFβ receptor inhibitors in experimental studies to treat cardiac fibrosis revealed an improved pharmacokinetic profile (272) and minimal toxic effects (273).

In the clinical setting, Pirfenidone (a drug that is inhibitory of TGFβ signaling) is approved as an oral anti-fibrotic drug for the treatment of idiopathic pulmonary fibrosis (274), and its use in treating cardiac fibrosis is actively studied (275). The recently published PIROUETTE trial (276) showed that pirfenidone appeared to be beneficial at reducing myocardial fibrosis among patients with heart failure with preserved ejection fraction. This medication was associated with a modest reduction in myocardial fibrosis, as assessed by cardiac MRI, compared with placebo. Despite this observation, presently, its clinical significance is still undetermined.

Tranilast is a drug used to treat allergic disorders (277–279) and dermatological diseases (280) has been demonstrated to inhibit collagen deposition by inhibiting fibroblast proliferation (281) and limit TGF-β-induced collagen synthesis by keloid-derived fibroblasts (282). Suppression of TGF- β, and more specifically, its action against tumor cells, is a proven action of tranilast that could have practical therapeutic applications (283, 284). Laboratory studies conducted on various populations of fibroblasts showed that tranilast inhibited its proliferation (285, 286). Notwithstanding clinical and experimental evidence supporting the anti-fibrotic effects of tranilast, its prolonged use can have hepatic toxicity (287), limiting its clinical application.

After the observation that in Galectin-3 knockout mice, left ventricular hypertrophy was prevented and left ventricular function was ameliorated (288), the possible application of Galectin-3 inhibition as a therapeutic target capable of slowing the progression of cardiac fibrosis (289) has been considered. Up to date, no clinical studies have been published.

As it is thought that endothelin can play a role in the pathophysiology of fibrosis, the endothelin receptor blockade has been considered another potential therapeutic target, but clinical studies have shown disappointing results (290, 291).

Clinical Use of Pharmacological Agents That May Modulate the Fibrotic Substrate

The guidelines from the principal Cardiac Societies (292) recommend established HF therapies that target neurohumoral pathways and may reduce mortality in HF patients, at least in part through inhibition of progressive structural remodeling (Table 1). Some of these therapies have also been studied in the population of patients with AF, and data on the potential usefulness in modifying the substrate using this so-called ‘upstream therapy’ have been published. These pharmacologic agents are ACE inhibitors (as explained previously), AT1 receptor blockers, mineralocorticoid antagonists (293, 294), and β-adrenoceptor blockers. However, the results have not been consistent (295–297). Other therapies like statins have also been studied because these agents seem to exert antifibrotic effects, modulate metalloproteinases, and interact with endothelial nitric oxide synthase that protects atrial myocardium during ischemia. However, results have been either neutral or inconsistent (298–301).

TABLE 1

Potential antifibrotic mechanisms of currently used pharmacologic agents.

Pharmacologic agent	Target	Mechanism	Description
B eta- blockers	β−adrenergic−receptor	Blocking the effects of beta-agonism, therefore reducing sympathetic overactivity decreasing inflammation (302). Anti-inflammatory and pro-angiogenic properties (increase cytokine IL-10; decrease cytokine IL-1β) (303).	β−adrenergic−receptor agonists, including epinephrine and norepinephrine, could produce cardiac hypertrophy and fibrosis in vivo (304). Cardiac fibroblasts have adrenergic receptors, and stimulation of the β2−adrenergic receptor leads to increased proliferation of human and rodent cardiac fibroblasts (305, 306).
Mineralocorticoid receptors antagonists	mineralocorticoid receptors	Reduce the proinflammatory and profibrotic effects of aldosterone (307–309)	Aldosterone stimulates the expression of profibrotic molecules, such as transforming growth factor-β1 (TGF-β1), plasminogen activator inhibitor 1 (PAI-1), endothelin 1 (ET-1), placental growth factor (PGF), connective tissue growth factor (CTGF), osteopontin, and galectin-3 (310).
ACE inhibitors/Angiotensin receptor blockers	Prevent the hydrolysis of Angiotensin I to Angiotensin II	Angiotensin-converting enzyme (ACE) promotes inflammation in the heart, kidney, and vasculature through Angiotensin II as the effector (311, 312). Angiotensin II induces fibrosis via the stimulation of TGF-β (313).	ACEIs reduce inflammation and fibrosis through the reduction of IL-6 and TNF-α (312). ACEIs have an effect on reducing TGF-β1, TGF-β2, and Th2 cytokines (314). Induce the apoptosis of cardiac fibroblasts (315). Reduction of sST2 (316).
Statins	Pleiotropic effects, e.g., anti-inflammatory, antifibrotic, and immune-modulatory (317).	Statins may have a beneficial effect on various factors that promote fibrosis, such as endothelial dysfunction, VEGF, IL-6, and TNFα (318). They have been found to improve endothelial function, exert an anti-inflammatory effect and lower the expression of VEGF (319).	Reduce the differentiation of MRC5 fibroblasts into myofibroblasts (320). Induce fibroblast apoptosis (321). Suppress epithelial-mesenchymal transition (EMT) by attenuating TGF-β signaling (322, 323). Reduce the expression of transforming growth factor (TGF)-β1, connective tissue growth factor (CTGF), RhoA and cyclin D1 (324, 325). Inhibition of geranylgeranylated Rho protein (326).

Debatable results have also been reported for the effects of fish oils (327). These conflicting data may be related to diverse study populations, differences in AF history, and concomitant diseases, resulting in heterogeneous baseline remodeling, in combination with the limited reversibility of structural remodeling once it has been established (328).

Ablation

Although pulmonary vein isolation (PVI) is very effective in maintaining sinus rhythm in patients with paroxysmal AF (329), it is much less so in persistent AF, with a reported 5-year AF freedom rate of 20% after a single and 45% after multiple procedures (330, 331). Progression of paroxysmal to persistent forms of AF occurs in 4–15% of patients per year, depending on risk factors (332–334). In recent years, several studies have shown that the earlier the treatment of patients with AF, the better the results regarding arrhythmia recurrence, hospitalization, and repeat procedures (329, 335–337). Also, early catheter ablation was superior to antiarrhythmic drug therapy in patients with drug-refractory paroxysmal AF in delaying progression to persistent AF (338). This suggests that early intervention can slow the substrate development and the progression to established atrial fibrosis and highlights a potential role for (non-invasive) characterization of the atrial substrate in guiding therapeutic decisions.

Considerable research has shown that patients with persistent AF have a more advanced atrial disease than those with paroxysmal AF (339), and different studies have shown that the extent of fibrosis before ablation is independently associated with the likelihood of AF recurrence (340). The higher the burden of atrial fibrosis, the lower the probability of sinus rhythm maintenance. These observations led to the assessment of various strategies for modifying the arrhythmic substrate beyond PVI. Clinical and experimental studies suggest that re-entrant drivers (i.e., rotors) might maintain persistent AF (341, 342). A clinical study to evaluate the relationship between fibrosis imaged by delayed-enhancement CMR and atrial electrograms in persistent AF reported that 90 percent of complex fractionated atrial electrogram sites occur at non-delayed-enhancement and patchy delayed-enhancement LA sites (343). A fascinating animal study (344), which analyzed the histological characteristics of Complex Fractionated Electrograms (CFAE) with atrial myocardial thickness and fibrosis sites, found the presence of a thicker wall and a more significant amount of fibrosis. The atrial myocardium was significantly thicker at CFAE sites (1757.5 ± 560.5 μm) than at non-CFAE sites (1279.5 ± 337.2 μm) (p = 0.036). At CFAE sites, it was filled with substantially more considerable fibrotic tissue than at non-CFAE sites (22.8 ± 6.9% versus 7.2 ± 4.7%, p < 0.001).

A 3D, biophysically detailed computational modeling study in patient-derived atrial models with individualized fibrosis distributions – derived from late enhancement CMR - showed that AF is inducible by programmed electrical stimulation in models with a sufficient amount of fibrosis. The induced AF is perpetuated by re-entrant drivers that persist in spatially confined regions. The latter areas constitute boundary zones, between fibrotic and non-fibrotic tissue characterized by high fibrosis density and entropy values (223).

Invasive electrical mapping data from a 64-pole basket catheter have been employed in a clinical computational mapping approach, revealing sustained electrical rotors and repetitive focal beats during human AF (345). Based on these observations, these authors have pioneered the CONFIRM study (346), in which the ablation in patients with persistent AF, guided by the computational mapping approach when compared with the conventional approach, showed higher freedom from AF (82.4% versus 44.9%; p < 0.001) after a single procedure. Although this focal impulse and rotor modulation (FIRM) approach initially gained some popularity, a meta-analysis that evaluated the results of PVI versus PVI + FIRM ablation demonstrated no therapeutic benefit of the additional focal impulse and rotor modulation approach over PVI alone (347).

A personalized substrate modification has been tested in the last decade, essentially with ablation targeting LA LVA or guided based on CMR-derived fibrosis patterns. Four strategies have been evaluated: Isolation of fibrotic areas (Box Isolation of the Fibrotic Area, BIFA) (348), homogenization of the LVA (349, 350), selective ablation of atrial LVA (351) and different combinations of part of these strategies (352). A meta-analysis that included studies with linear ablation or ablation of complex fractionated electrograms found disappointing results (353). In this meta-analysis in comparison with PVI alone, the addition of complex fractionated atrial electrograms (CFAE) ablation [RR 0.86; 95% confidence intervals (CI) 0.64, 1.16; P = 0.32] or left atrial linear ablation (LALA) at the roof and mitral isthmus (RR 0.64; 95% CI 0.37, 1.09; P = 0.10) offered no significant improvement in arrhythmia-free survival. However, adjunctive CFAE ablation was associated with significant increases (P < 0.05) in procedure and fluoroscopy times.

Another meta-analysis that analyzed specifically studies with a voltage-guided substrate modification by targeting LVA in addition to PVI found that this approach was more effective, safer, and with a lower proarrhythmic potential than conventional approaches (354). A common finding in different studies is that the absence of LVA identifies patients who respond well to a PVI−based ablation strategy (355).

Nonetheless, a recently published randomized controlled trial (VOLCANO trial, (356)) demonstrated that LVA ablation, in addition to PVI, had no beneficial impact on rhythm outcomes in patients with paroxysmal AF undergoing AF ablation. Patients with LVAs showed lower AF−recurrence−free survival rates (88%) than those without LVA (57%, P < 0.0001; C, 53%, P < 0.0001), and so, the presence of LVA strongly predicted AF recurrence (356). Similarly, recently presented (357) results of the DECAAFII trial suggest that fibrosis-guided ablation was not superior to conventional PVI in reducing atrial arrhythmia recurrence but significantly increased adverse events.

This is an area of active research, and probably soon, we will have data that will allow us to perform ablation tailored to the substrate observed in the particular patient. Based on the information from the different studies cited above, the most consistent finding is that the greater the degree of the atrial fibrotic substrate, the less probable is sinus rhythm maintenance.

Conclusion

Catheter ablation for AF has developed as an important rhythm-control strategy and nowadays is one of the most common cardiac ablation procedures performed worldwide. A rigorous assessment of the presence of atrial fibrotic substrate is important for determining the treatment options and as a predictor of long-term success after catheter ablation.

In humans, the progression from paroxysmal AF to persistent forms is marked by structural alterations of the atrial tissue (358). Although the clinical phenotype (paroxysmal vs. persistent AF) typically determines therapeutic choices, it may not be the primary driver determining the success of catheter ablation treatment. Instead, the underlying factors, primarily the extent of atrial fibrosis, may be decisive.

Despite the complexity of atrial substrate evaluation, we currently have several diagnostic resources (imaging, ECG, biomarkers, etc.) that enable a comprehensive assessment and quantification of the extent of LA structural remodeling and the presence of fibrotic atrial substrates. Given the central role of fibrosis in AF pathophysiology and therapy, such a comprehensive understanding is expected to improve AF management and patient outcomes.

Author Contributions

PC reviewed the current literature for the present manuscript, wrote the outline, composed the manuscript, and provided critical editing of the manuscript. SL performed background research for the manuscript and provided the necessary editing of the manuscript. MO and JH reviewed the literature and offered critical editing of the manuscript. All authors contributed to the article and approved the submitted version.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.