Key concepts: trigger and substrate

AF is characterised and defined by very rapid and uncoordinated atrial activity. Conceptually, the initiation and maintenance of AF can be linked to the interaction between a trigger and the substrate. A ‘trigger’ is a rapidly firing focus that can act as an initiator for the arrhythmia, the maintenance of which generally requires a ‘substrate’, that is, electrophysiological, mechanical and anatomical characteristics of the atria that sustain AF. Development of this substrate usually includes both electrical and structural elements of atrial remodelling. Electrical remodelling encompasses changes in the properties of ion channels affecting atrial myocardial activation and conduction, while structural remodelling refers to alterations in the tissue architecture, both microscopic (eg, fibrosis) and macroscopic (eg, atrial dilatation). This conceptual framework for the key concepts underlying the induction and maintenance of AF is summarised in figure 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Key concepts underlying the induction and maintenance of atrial fibrillation (AF). AF can be maintained by either re-entrant or rapid and sustained ectopic activity. Development of re-entry depends on the action of a trigger (usually from an ectopic beat) acting on vulnerable substrate. In normal hearts, atrial electrical properties are less likely to support the maintenance of AF. Atrial remodelling creates a substrate for re-entrant AF, by altering ion channel function and/or inducing tissue fibrosis. Remodelling can also make ectopic activity more likely by producing changes in Ca²⁺ handling that promote both triggered activity and re-entry. DAD, delayed afterdepolarisation; EAD, early afterdepolarisation. Reproduced with permission from Dobrev D, Nattel S. New antiarrhythmic drugs for treatment of atrial fibrillation. The Lancet 2010;375:1212–23.10.1016/S0140-6736(10)60096-7.⁶⁹

It is thought that there is a progression over time from a trigger-driven disease, through development of a functional atrial substrate, followed by predominant structural atrial remodelling.8 This would correspond to the clinical observation that AF is often initially paroxysmal, before progressing to a persistent and ultimately permanent form of arrhythmia (figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Progression in atrial fibrillation (AF) mechanisms over time. (A) Local ectopic firing. (B) Single-circuit re-entry. (C) Multiple-circuit re-entry. (D) Mechanisms underlying clinical forms of AF. Paroxysmal AF is mostly underpinned by local triggers/drivers, particularly from pulmonary veins (PV). As AF becomes more persistent and eventually permanent, re-entry substrates (initially functional and then structural) predominate. IVC, inferior vena cava; LA, left atrium; RA, right atrium; SVC, superior vena cava. Reproduced with permission from Iwasaki Y-ki, Nishida K, Kato T, et al. Atrial fibrillation pathophysiology: implications for management. Circulation 2011;124:2264–74.10.1161/CIRCULATIONAHA.111.019893.8

Basic atrial electrophysiology

In health, atrial cell depolarisation is mediated by a large and rapidly activating and deactivating inward Na⁺ current, and the slower L-type Ca²⁺ current. Repolarisation is also rapid due to activation of a series of voltage-gated K⁺ channels. Action potential duration and refractory period are shorter in the atria (particularly in the left atrium) compared with the ventricular myocardium, although there is significant regional heterogeneity within and between the atria, reflecting systematic differences in intra-atrial ion channel density.9 Overall, the atrial myocardium is more prone to the development of very rapid rates with complex patterns of conduction than the ventricular myocardium, even before considering the proarrhythmic effects of atrial remodelling.

Triggers for AF

Seminal studies by Haïssaguerre et al identified the muscular sleeves within the pulmonary vein (PV) ostia as the source of the ectopic beats triggering AF in many patients with paroxysmal AF.10 The myocardial sleeves within PVs appear to demonstrate key differences from the remaining atrial myocardium in terms of cellular electrophysiology,11 gross anatomy and fibre geometry; these changes appear to predispose the PV muscle sleeves to rapid focal firing or re-entrant activation.12 It follows that electrical isolation of the PVs from the rest of the atrium (termed ‘PV isolation’) is the cornerstone of catheter ablation of AF.

Non-PV triggers have also been described,13 14 including (among others) the superior vena cava, coronary sinus, left atrial appendage, ligament of Marshall, crista terminalis and left atrial posterior free wall, potentially due to the presence of myocardial sleeves or regional atrial fibrosis at these sites. Non-PV triggers are more common in advanced subtypes of AF and in patients who have already undergone a catheter ablation procedure. Ablation targeting non-PV triggers may be a useful addition to therapeutic approaches in selected individuals.13 15 Similarly, ganglionated plexi, which are conglomerations of autonomic ganglia on the epicardial surface of the heart, may play a role in the initiation and maintenance of AF.16 Ablation of these plexi in addition to PV isolation led to improved freedom from atrial tachyarrhythmia compared with PV isolation alone in one small clinical trial.17 Finally, AF may occasionally also be triggered by other forms of supraventricular arrhythmia, such as atrioventricular nodal re-entrant tachycardia, atrioventricular re-entrant tachycardia and typical counterclockwise right atrial flutter.

These triggers for AF are themselves often initiated or maintained by ‘upstream’ processes including atrial stretch, ischaemia and autonomic imbalance. This could explain, at least in part, the clinical observation that AF is more common in conjunction with comorbidities predisposing to these processes, such as mitral regurgitation, myocardial infarction and vagal stimulation, respectively.

Arrhythmic mechanisms that sustain AF

In the early days of investigation of the pathophysiology of AF, macro re-entrant circuits were suspected to be the predominant electrophysiological mechanism by key clinical scientists such as Sir Thomas Lewis,18 but in vivo evidence was lacking. In 1959, Moe and Abildskov extended the idea of re-entry to that of ‘multiple wavelets’—that is, the presence of multiple simultaneous re-entrant circuits within the atria.19 Subsequent work provided mapping evidence of multiple re-entrant wavelets in animal and human atria,20 and Cox surgical maze procedure was designed to prevent sustained re-entry by compartmentalising the atrium into small and electrically isolated units. Unlike classic re-entrant circuits that rely on a central anatomic barrier or scar, so-called ‘leading circle’ re-entry in AF is thought to be functional, due to constant centripetal activation of the centre of the circuit resulting in continuous local refractoriness.

There remains a degree of controversy and uncertainty regarding the precise mechanisms that initiate and sustain AF. Some investigators have described ‘rotors’ or spiral waves as a special form of functional re-entry.21 In a rotor, the wavefront has a curved or spiral form, with the velocity of any specific portion of the wavefront depending on its degree of curvature. The area of wavefront with the highest curvature has the slowest conduction velocity; this results in functional block at the centre of the rotor due to the propagating wavefront being unable to invade this core of tissue. Critically, this means that rotors can meander through space as there is no area of truly refractory myocardium, in contrast to leading circle re-entry, which must remain fixed around the unexcitable centre. Mapping studies have shown that stable rotors can also anchor at certain sites (often around the PVs and in areas of heterogeneous atrial tissue)—wavefronts spreading away from the centre of the rotor then fragment, inducing chaotic and fibrillatory activity within the rest of the atrium.22 The current hypotheses for AF maintenance are summarised in figure 3, which illustrates how rotors may be compatible with the ectopic foci and multiple wavelets. This illustration of circuits that can involve the epicardial and mid-myocardial layers also highlights the significant challenges to invasive mapping and identification of rotors, since conventional electroanatomic mapping only directly interrogates the endocardial layer.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Current hypotheses for atrial fibrillation (AF) maintenance. (A) Diagram of AF maintenance near a pulmonary vein that has been hypothesised to be driven by ectopic focus (left), rotor (middle) or multiple wavelets (right). Different wavefronts are represented in purple. (B) Representation of the compatibility of rotor maintenance with other mechanisms. Rotors can be initiated by wavebreaks near an ectopic focus (left) and underlie endocardial or epicardial breakthroughs (middle). A drifting rotor, whose trajectory is depicted in blue, can be the driver of multiple and apparently disorganised atrial wavelets (right). Reproduced with permission from Guillem MS, Climent AM, Rodrigo M, et al. Presence and stability of rotors in atrial fibrillation: evidence and therapeutic implications. Cardiovasc Res 2016;109:480–92.10.1093/cvr/cvw011.21

Notwithstanding this evidence of complex re-entrant mechanisms involving large areas of atrial myocardium, AF may in some cases also be driven by a rapid localised source of triggered discharge or micro re-entry. In this situation, the remainder of the atrial myocardium may be a bystander as suggested by a study demonstrating that ablation of these so-called driver domains terminated persistent AF in many cases, particularly in patients where AF had been persistent for less than 6 months.23

Overall, although AF is defined by the presence of chaotic atrial electrical activity, it is now recognised that the proarrhythmic mechanisms are extremely heterogeneous. The relative contributions of potential mechanisms appear to be widely variable between different individuals, and may also change over time within a single individual.

Development of a substrate for AF

The maintenance of AF is thought to reflect development of a vulnerable substrate as a result of electrical and structural remodelling, particularly within the left atrium. Aspects of atrial electrical remodelling include shortening of the refractory period due to downregulation of the Ca²⁺ current,24 accelerated repolarisation and hyperpolarisation of atrial cells due to increases in outward K⁺ currents,25 and conduction abnormalities due to altered expression and localisation of connexins that connect atrial myocytes.26 These changes all promote re-entry and chaotic patterns of atrial activation, and are closely related to autonomic nervous activity.27

The most prominent aspects of structural remodelling include progressive atrial dilatation, readily detected by transthoracic echocardiography.28 Atrial dilatation may support re-entry directly, but is also strongly correlated to the presence of fibrosis.29 Fibrosis appears to be of critical mechanistic importance to the development and maintenance of AF, by causing heterogeneity of electrical conduction and predisposing to re-entry. Atrial fibrosis results from activation of fibroblasts, and has classically been ascribed to ageing, comorbidities and risk factors, although experimental evidence for these assertions is somewhat lacking.29 In contrast, both animal studies30 and postmortem histological studies in humans29 support an association between AF and progressive atrial fibrosis—leading to the notion that ‘AF begets AF’, that is, that AF directly induces atrial remodelling that supports the further induction and maintenance of AF. This concept appears to tie in with the clinical observation that AF often progresses from infrequent paroxysms to more frequent and long-lasting episodes, and then persistent AF, although an alternative possibility is continuous evolution of the atrial phenotype over time, largely independent of the presence or absence of AF. Meanwhile, progressive structural atrial abnormalities have also been described in the absence of AF, suggesting an alternative paradigm where fibrotic atrial cardiomyopathy is, at least in some patients, an independent disease process that occurs first and predisposes to the subsequent development of arrhythmia.31

Cardiac magnetic resonance now offers the possibility of accurate, non-invasive and serial atrial imaging,32 including assessment of atrial tissue characteristics with atrial late gadolinium enhancement imaging (figure 4).33 Atrial late gadolinium enhancement is thought to reflect the presence of atrial fibrosis, although this is largely based on correlation with invasive electroanatomic data34 and studies of postablation atrial injury in animals,35 rather than histological validation in patients. Nevertheless, these advanced imaging approaches should in time increase our understanding of the predictors, natural history and clinical significance of atrial remodelling. Already, such studies have shown that patients with stroke of undetermined cause have more atrial late gadolinium enhancement than patients with an identified non-AF-related specific cause of stroke, suggesting a possible aetiological role for an underlying atrial cardiomyopathy without clinical evidence of AF.36 Meanwhile, the multicentre DECAAF (Delayed-Enhancement MRI (DE-MRI) Determinant of Successful Radiofrequency Catheter Ablation of Atrial Fibrillation) trial demonstrated that atrial late gadolinium enhancement was independently associated with likelihood of recurrent arrhythmia after catheter ablation of AF.37 Thus, increasing evidence suggests that atrial fibrosis may be an important marker of disease severity and predictor of clinical outcomes. However, it remains unclear whether atrial fibrosis is a potentially modifiable risk factor—this hypothesis is being tested in the ongoing DECAAF II trial (NCT02529319), which will test the efficacy of fibrosis-guided ablation in addition to conventional PV isolation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Left atrial tissue characterisation using late gadolinium enhancement (LGE)-MRI. Following acquisition of high-resolution LGE-MRI scans (step 1), the left atrial wall is identified and isolated by manually tracing the blood pool in each slice of the LGE-MRI volume (step 2). The mitral valve and extension of the left ventricle are manually excluded. Quantification of fibrosis is based on the relative signal intensity of LGE (step 3). A three-dimensional model of the left atrium (LA) is rendered with the maximum enhancement intensities being projected on the model surface; healthy tissue is depicted as blue, whereas any tissue with LGE is depicted as green and yellow (step 4). Reproduced with permission from Siebermair J, Kholmovski EG, Marrouche N. Assessment of left atrial fibrosis by late gadolinium enhancement magnetic resonance imaging: methodology and clinical implications. JACC Clin Electrophysiol 2017;3:791–802.33

Alterations in myocardial redox state in atrial remodelling and AF

More recent efforts have focused on identifying the underlying cellular and molecular mechanisms that lead to atrial remodelling. Alterations in myocyte nitroso-redox state have been closely linked to the initiation, development and maintenance of AF.38 Redox signalling can affect downstream targets in various subcellular compartments via effects on transcription factors, direct protein transnitrosylation or targeting of other signalling molecules. For example, angiotensin-II-induced oxidation of Ca²⁺/calmodulin-dependent protein kinase II results in increased sarcoplasmic reticulum Ca²⁺ leak through the ryanodine receptor, and contributes directly to increased susceptibility to AF in mice.39

Alterations in redox signalling in AF are complex, and the atrial sources of reactive oxygen species have been shown to differ with the duration and substrate of AF.40 Recent work has identified that atrial-specific upregulation of a small non-coding RNA (miR-31) leads to depletion of neuronal nitric oxide synthase and repression of dystrophin (which binds neuronal nitric oxide synthase in the myocardium).41 The disruption in neuronal nitric oxide signalling leads to shorter action potential duration and loss of rate-dependent adaptation in action potential duration, creating a proarrhythmic substrate.41 This suggests that local inhibition of relevant miRs in the atrial myocardium could reverse atrial remodelling and potentially act as a novel adjunct to current therapeutic strategies, assuming that tissue-specific delivery strategies can be developed.

Genetics of AF

Individuals with a family member affected by AF have a 40% greater risk of incident AF than those without an affected family member, after adjusting for AF risk factors.42 Genome-wide association studies have progressively identified more risk variants and genes that underlie the observation of familial risk, enriched within cardiac developmental, electrophysiological, contractile and structural pathways.43 A recent such study of over 1 million individuals identified 142 independent risk variants at 111 loci, corresponding to 151 gene candidates likely to be involved in AF pathogenesis.44 Pathway and functional enrichment analyses have further highlighted fetal heart tissue and pathways related to cardiac development as being functionally relevant in AF pathogenesis, implying that such genes and pathways either act in the developing heart to influence the future risk of AF, or that they are activated in the adult heart as a response to stress.44

The genes and pathways identified from such approaches may allow new insights into AF pathophysiology, and potentially reveal new therapeutic targets. While genetic testing is not currently undertaken routinely in patients with AF, this could change rapidly if polygenic risk scores can be identified that, for example, contribute to the clinical classification of AF phenotype, aid stroke risk stratification or predict response to catheter ablation.

Beyond the atrium: is AF a systemic disease?

As detailed above, several decades of detailed investigation have yielded fundamental insights into the pathophysiology of AF and the associated alterations in the cellular, molecular, electrophysiological and structural architecture of the atria. More recently, it has become increasingly recognised that AF is more than just an atrial disease, with documented associations with systemic inflammation, endothelial dysfunction, cardiometabolic disturbance and wider abnormalities in myocardial structure and function.45

Longitudinal and multiparametric cardiac magnetic resonance studies show that even patients with apparently lone AF have significantly impaired ventricular myocardial energetics (figure 5), coronary microvascular dysfunction and subtle reduction in left ventricular performance, which fail to normalise following catheter ablation.32 46 This suggests that AF may actually be the consequence (rather than the cause) of an occult cardiomyopathy, which is unaffected by successful rhythm control, and that adjunctive therapies may be needed to target the ongoing drivers of the disease process.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Left ventricular energetics in patients with lone atrial fibrillation (AF). The ³¹P magnetic resonance spectra and derived PCr/ATP ratios are shown in a control subject (A, top panel) and a patient with lone AF before catheter ablation (A, bottom panel). Despite a significant reduction in AF burden at a median of 7 months after ablation (p<0.001) (B), there was no change in PCr/ATP ratio (p=0.57) (C, left panel), with myocardial energetics remaining significantly impaired compared with matched control subjects in sinus rhythm (p=0.001) (C, right panel). 2,3-DPG, 2,3-diphosphoglycerate; PCr, phosphocreatine; PDE, phosphodiester. Reproduced with permission from Delgado V, Di Biase L, Leung M, et al. Structure and function of the left atrium and left atrial appendage: AF and stroke implications. J Am Coll Cardiol 2017;70:3157–72.10.1016/j.jacc.2017.10.063⁷⁰ (originally adapted from Wijesurendra et al 32

These observations are also in keeping with the strong epidemiological associations between AF and other cardiac, metabolic and systemic comorbidity.47 An exemplar is obesity, which is the strongest modifiable risk factor for AF,48 with a Mendelian randomisation study indicating a direct causal relationship.49 50 Further evidence for the clinical relevance of obesity and other systemic diseases in AF comes from emerging randomised and cohort studies demonstrating dramatic improvements in AF burden and symptoms following weight loss and risk factor control.^W51–W53 The mechanism by which obesity predisposes to AF is unclear, but much interest has focused on the potential role of epicardial fat, which is closely associated with AF phenotype and recurrence. A body of work indicates that epicardial fat may influence the triggers and substrate for AF through a number of mechanisms, including fatty infiltration of the atrial myocardium, induction of atrial fibrosis and activation of inflammatory and oxidative stress pathways.^{W54 W55} Other potential mechanisms include left atrial enlargement, left ventricular hypertrophy^W56 and altered cardiac energetics.^W57

Finally, the intriguing possibility that AF progression is linked to vascular risk via hypercoagulability that influences atrial vascular remodelling and fibrosis is being assessed in the ongoing RACE-V cohort study (Reappraisal of Atrial Fibrillation: Interaction Between HyperCoagulability, Electrical Remodeling, and Vascular Destabilisation in the Progression of Atrial Fibrillation; NCT02726698).

Implications for current and future rhythm control strategies

Progressive advances in our knowledge of the mechanisms of AF have directly translated into current rhythm control strategies, both pharmacological and interventional. It is important to emphasise that rhythm control strategies have generally not demonstrated any prognostic benefit compared with rate control strategies in patients with AF.^{W58 W59} Similarly, the recently published CABANA (Catheter ABlation vs ANtiarrhythmic Drug Therapy in Atrial Fibrillation) trial failed to demonstrate an improvement in the composite outcome of mortality, stroke, bleeding and cardiac arrest in patients randomised to PV isolation via catheter ablation compared with those randomised to rate and rhythm control with medical therapy, despite a significant reduction in recurrent AF in the former group.^W60 Catheter ablation remains commonly indicated solely for the improvement of symptoms,^W61–W63 although evidence for improvement in symptoms and/or quality of life derives from open-label studies including CABANA,^W62 where the mean differences in quality of life and symptom scores between the groups at 12 months were of questionable clinical significance, particularly given the possibility of a larger placebo effect resulting from an interventional therapy. Definitive evidence of symptomatic benefit from ablation would require more rigorous blinded comparisons of ablation with a sham procedure, and such trials are currently lacking.

Meanwhile, the recent and relatively small CASTLE-AF (Catheter Ablation versus Standard Conventional Therapy in Patients with Left Ventricular Dysfunction and Atrial Fibrillation) randomised trial also demonstrated a significant improvement in the composite primary endpoint of mortality and the rate of hospitalisation for worsening heart failure with ablation compared with medical therapy, suggesting prognostic benefit of ablation in selected patients with AF and heart failure.^W64 These data are consistent with the results of prior smaller randomised trials comparing catheter ablation versus medical therapy in patients with AF and heart failure,^W65 and a trend towards benefit of catheter ablation compared with medical therapy in patients with a history of congestive heart failure or New York Heart Association II–IV symptoms in CABANA.^W60 Meanwhile, the soon-to-report EAST trial (Early Treatment of Atrial Fibrillation for Stroke Prevention Trial; NCT01288352) has been designed to test the hypothesis that early and structured rhythm control therapy (with antiarrhythmic drugs and catheter ablation) can prevent AF-related complications compared with usual care in a less selected group of patients with AF.

Pharmaceutical approaches with antiarrhythmic drugs are targeted to reverse the effects of atrial electrical remodelling, mainly by prolonging the atrial effective refractory period and lengthening atrial action potential duration, thereby reducing the propensity for induction and maintenance of AF. Although theoretically potentially proarrhythmic, these medications are relatively safe in clinical use, with an incidence of major ventricular arrhythmia of just 0.8% over the median 4-year follow-up in CABANA.^W60 Nevertheless, side effects remain a limiting factor to the long-term acceptability of such medications at therapeutic doses for many individuals.

Over the last two decades, catheter ablation has become a mainstream rhythm control strategy, established in clinical guidelines.^W61 PV isolation is the mainstay of catheter ablation, and can be used as a first-line treatment in patients with paroxysmal AF, in whom it is at least as effective as antiarrhythmic drug therapy.^W61 Persistent AF is much more challenging, as no approach to atrial substrate modification with ablation has proven effective. In the pivotal STAR-AF II trial (Substrate and Trigger Ablation for Reduction of Atrial Fibrillation Trial Part II), patients were randomised in a 1:4:4 ratio to ablation with PV isolation alone, PV isolation plus ablation of electrograms showing complex fractionated activity, or PV isolation plus additional linear ablation across the left atrial roof and mitral valve isthmus. The results were sobering: around 40% of patients experienced recurrent AF after ablation with no statistically significant differences between the groups, although there was a trend to more recurrent AF in both groups who received additional ablation compared with those treated with PV isolation alone.^W66 Further work is therefore still needed to understand if, and how, the atrial substrate for AF can be ameliorated by ablation. In this context, the results of the recent RACE 3 trial are also particularly relevant.^W67 Patients with early persistent AF and mild to moderate heart failure randomised to targeted therapy of underlying conditions (consisting of both pharmacological and lifestyle interventions) had improved maintenance of sinus rhythm at 1 year compared with those randomised to conventional therapy.^W67 Similarly, weight loss^W51 and improvement in cardiorespiratory fitness^W68 appear to be associated with a reduction in AF burden and symptom severity. This supports the paradigm that more holistic therapy, beyond catheter ablation alone, is likely to be required for the successful treatment of persistent AF.

Summary and conclusions

AF is a complex arrhythmia that is characterised and defined by rapid and uncoordinated atrial activity. The pattern of atrial electrical activity in AF is not completely understood, but can include complex re-entrant mechanisms as well as localised focal discharges and micro re-entry. The initiation and maintenance of AF is dependent on the presence of both trigger and substrate, including electrical and structural atrial remodelling. Paroxysmal AF often precedes persistent AF, consistent with experimental evidence showing that AF can itself induce atrial remodelling that contributes to the further maintenance of AF.

Atrial remodelling and AF often reflect the combined effects of several discrete and interacting pathophysiological processes, both inherited and acquired, although there is significant heterogeneity in the balance of the contributions of each of these mechanisms in any one individual. AF is closely associated with advanced age, the presence of comorbidities and systemic disease, cardiometabolic disturbance, and wider abnormalities in myocardial structure and function, consistent with a pathophysiological basis that goes beyond the atrial myocardium, although the precise mechanisms linking extra-atrial pathology to AF remain poorly defined.

Pharmacological and interventional therapeutic approaches to rhythm control in AF mainly address alterations in atrial electrophysiology and triggers for AF. The limitations of current approaches are particularly pronounced in patients with persistent AF and/or advanced structural atrial remodelling. Further mechanistic, translational and clinical studies are needed to improve understanding of AF mechanisms and pathophysiology, and direct development of novel therapeutic approaches.

The additional references can be found in online supplementary file 1.