Mechanisms of atrial fibrillation

INTRODUCTION — There are many mechanisms responsible for and contributory to development of atrial fibrillation (AF). The chance of developing AF is tied closely to age; AF is rare before age 50 [1]. In addition to age, there are many types of cardiac and medical conditions that are also closely linked to AF. These include hypertension, coronary artery disease, heart failure, valvular heart disease, obesity [2], and sleep-apnea syndrome. It is well established that high levels of alcohol [3] can increase the probability of developing AF, and that hyperthyroidism can cause AF. Evidence for caffeine and energy drinks, while suspected to be contributory, is questionable [4]. Furthermore, while exercise can be protective against atrial fibrillation, endurance athletics may be a cause for atrial fibrillation [5]. It is also well established that AF is more common in individuals who have a first-degree relative who developed AF at a young age. There is also a variety of acute conditions that can initiate AF such as cardiac surgery, pulmonary embolus, and inflammatory lung conditions. (See "Epidemiology, risk factors, and prevention of atrial fibrillation".)

The precise mechanisms by which age and the other conditions listed above increase the propensity for development of AF are understood poorly (figure 1). However, these conditions may impact the triggers for AF, which commonly arise in the pulmonary veins or the substrate for maintenance of AF, which broadly relates to atrial size and the extent of fibrosis. Some of the factors that may play a role in the mechanisms of AF include autonomic tone, inflammation, atrial pressure and wall stress, and genetics. (See "Epidemiology, risk factors, and prevention of atrial fibrillation", section on 'Other factors'.)

This topic will provide a broad overview of the current understanding of the mechanisms of AF. This discussion will provide a relatively simplistic approach to a complex topic. The reader will be referred to a rapidly growing literature on this topic, including some comprehensive reviews [6].

DEFINITIONS — The following terms are defined to help the reader’s understanding of the material below:

●Trigger – Rapid firing of the left atrium from sites arising in the pulmonary veins often initiates atrial fibrillation (AF).

●Triggered activity – One of three mechanisms of cardiac arrhythmias (including automaticity and reentry). Triggered activity refers to additional depolarizations, which occur during or immediately following a cardiac tissue electrical depolarization and may initiate a sustained cardiac arrhythmia.

●Substrate – Mechanical and anatomic structure of the atria in which AF can occur and perpetuate.

●Substrate remodeling – Changes in the mechanical and anatomic macro, micro, and ultrastructure of the atrial substrate that result from the development of AF and increase the propensity for the development and maintenance of AF over time.

●Electrical remodeling – Changes in the atrial electrical properties (refractoriness and conduction) that result from the development of AF and increase the propensity for the development and maintenance of AF over time.

●Dispersion of refractoriness – A range of differences in the refractory period properties throughout the atrial tissue.

●Spatial heterogeneity of refractoriness – Dispersion of refractoriness manifest as variability in refractoriness throughout the atrial anatomy.

●Complex fractionated electrograms – Local measurement of electrical activity recorded as bipolar electrograms obtained from areas of the atrium that are low amplitude with multiple components.

●Reentry/reentrant mechanism – One of three mechanisms of cardiac arrhythmias (including automaticity and triggered activity). Reentry is the most common mechanism of cardiac arrhythmias and refers to the presence of one or more electrical circuit(s) in which electrical activation proceeds in a circular fashion to complete a self-sustaining circuit.

●Atrial anisotropy – Altered conduction properties related to directionality of conduction through atrial tissue.

BASIC ATRIAL ELECTROPHYSIOLOGY — The electrophysiologic properties of normal and fibrillating atria have been studied extensively [7]. A basic understanding of these properties is necessary to understand the pathologic processes that play a role in initiating and perpetuating atrial fibrillation (AF). In the aggregate, these electrophysiologic properties permit the development of very complex patterns of conduction and an extremely rapid atrial rate as seen in AF.

The atrial myocardium consists of so-called "fast-response" tissues that depend on the rapidly activating sodium current for phase 0 of the action potential. (See "Cardiac excitability, mechanisms of arrhythmia, and action of antiarrhythmic drugs".)

Normal atrial myocardium has the following properties [7-9]:

●A short action-potential duration.

●Cellular reactivation can occur rapidly due to the short refractory period (in contrast to Purkinje fibers and ventricular muscle).

●Very rapid electrical conduction can occur.

●The refractory period shortens with increasing rate.

In the aggregate, these electrophysiologic properties permit the development of very complex patterns of conduction and an extremely rapid atrial rate as seen in AF.

CLINICAL FACTORS ASSOCIATED WITH AF — The following are common clinical conditions associated with atrial fibrillation (AF) in developed countries, and the percent of AF cases in which they are found (see "Epidemiology, risk factors, and prevention of atrial fibrillation", section on 'Chronic disease associations') [6]:

●Hypertension (60 to 80 percent).

●Cardiovascular disease, including cardiomyopathy, valvular and coronary artery disease (25 to 30 percent).

●New York Heart Association class II to IV heart failure (30 percent).

●Diabetes (20 percent).

●Age.

Each of the first three is often associated with left atrial dilatation, which is important in the development of a substrate for AF and also may increase the probability of electrical firing from the pulmonary veins. (See 'Mechanisms of atrial fibrillation: triggers and substrates' below.)

The following section will discuss the link between these conditions and AF.

MECHANISMS OF ATRIAL FIBRILLATION: TRIGGERS AND SUBSTRATES — Atrial fibrillation (AF) may present as a paroxysmal (self-terminating AF within seven days), a persistent (one that lasts greater than seven days), or a long-standing persistent AF (continuous AF for 12 months or greater). The term “permanent AF” should be used when both the patient and physicians agree to not pursue strategies to restore or maintain sinus rhythm and AF continues unabated. (See "Atrial fibrillation: Overview and management of new-onset atrial fibrillation", section on 'Classification and terminology'.)

This wide range of clinical presentations is likely due to an interaction between a trigger and the substrate (figure 1). AF is initiated by rapid firing (or triggers) from the pulmonary veins (PV). Early in the course of AF the atrium is relatively healthy and as a result sinus rhythm is spontaneously restored. As the substrate remodels further over time, AF no longer terminates spontaneously and becomes persistent. With more extensive remodeling of the atrium, it becomes increasingly difficult to maintain sinus rhythm and the patient and physician may agree no longer to attempt to maintain sinus rhythm, with the AF thereby becoming permanent [10].

Triggers of AF — It has been known for many years that a single focus firing rapidly in the atria can be a trigger for fibrillatory conduction throughout the atria [11]. The most common site of the rapid atrial firing that triggers AF is the PVs. Catheter ablation of AF depends in large part on the electrical isolation of the PVs from the remainder of the atrium. Electrophysiologic evaluation of the PVs has identified myocardial tissue that can lead to repetitive firing or even the presence of episodic reentrant activation in the veins [6].

Additionally, stretch can increase the propensity for rapid firing from the PVs as a result of stretch sensitive ion channels. [12]. It has been speculated that the mechanism of atrial stretch may help explain the association between AF and mitral regurgitation as well as various types of heart failure that increase left atrial pressure.

Role of premature atrial complex and other arrhythmia triggers — AF is initiated (triggered) predominantly by rapid firing from PVs. Much less commonly, AF can be triggered by non-PV sites of rapid firing (such as tissue near the PV including the Vein of Marshall, the superior vena cava, or coronary sinus) or by other types of supraventricular arrhythmias including atrioventricular nodal reentrant tachycardia (AVNRT), orthodromic AV reciprocating tachycardia, and atrial flutter [6,13-23]. In some patients, successful elimination of AF with catheter ablation requires both isolation of the PVs, as well as elimination of these non-PV triggers. (See "Atrial fibrillation: Catheter ablation" and "Atrial fibrillation: Surgical ablation".)

Role of atrial flutter and supraventricular tachycardias — Atrial tachycardia, atrial flutter, and other supraventricular tachycardias can initiate AF in predisposed patients. The interaction between these arrhythmias and AF is not well understood, but atrial flutter and AF commonly coexist.

In some instances, elimination of atrial flutter will diminish and/or eliminate episodes of AF. Nevertheless, elimination of the right atrial reentry circuit responsible for typical flutter frequently does not eliminate the predisposition to AF that is predominately a left-atrial problem in a large number of patients. Many studies have demonstrated that patients who undergo catheter ablation of typical atrial flutter have a very high probability of developing AF over the ensuing five years. This is true regardless of whether AF had been observed prior to development of typical atrial flutter. This has clinical implications when it comes to ablation, but also has implications for anticoagulation strategies and patient follow-up. Nevertheless, for most patients, it makes sense to try to eliminate the organized supraventricular tachycardia, especially if right-sided by ablation before considering PV isolation and/or other more extensive ablation procedures to eliminate AF, as the AF may be reduced or eliminated by eliminating the other tachycardia first.

Role of the autonomic nervous system — The autonomic nervous system plays an important role in the development and maintenance of AF [24-26]. Clinical studies using heart rate variability analysis in patients with AF suggest that fluctuation in autonomic tone may be a major determinant of AF in patients with focal ectopy originating from the PVs [27]. Studies have also demonstrated a change in heart rate variability after PV ablation [28], further suggesting that PV triggers may be at least partially modulated by autonomic activity. Another study showed that the occurrence of paroxysmal AF greatly depends on variations of the autonomic tone, with a primary increase in adrenergic tone followed by an abrupt shift toward vagal predominance [29].

Anatomic studies of the autonomic innervation of the atria also indicate that the PVs and posterior left atrium (PLA) have a unique autonomic profile with a rich innervation from sympathetic and parasympathetic nerves [30-35]. The autonomic nervous system may also be playing a role in the genesis of AF in diseased hearts [30,36,37]. Studies suggest that the parasympathetic and sympathetic nervous system may also be playing a role in creation of AF substrate in the setting of heart failure [36,37].

Both the sympathetic and parasympathetic nervous systems have been implicated in the genesis [30,38,39] and maintenance of AF:

●Sympathetic effects

•Early studies suggested that exercise-induced AF may be sympathetically driven [30,40].

•PV ectopic foci appear to be at least partially modulated by autonomic signaling, with sympathetic stimulation with isoproterenol frequently utilized to elicit these triggers in patients undergoing ablation for AF [41].

●Parasympathetic (vagal) effects

•The parasympathetic nervous system may contribute to AF in young patients with no structural heart disease [42].

•Animal studies show that vagal stimulation contributes to the genesis of AF by nonuniform shortening of atrial effective refractory periods, thereby setting up substrate for reentry. Vagal stimulation can also lead to the emergence of focal triggers in the atrium [43-45].

•Bezold-Jarisch-like "vagal" reflexes can be elicited during radiofrequency ablation and occur in and around the PVs. It has been suggested that elimination of these vagal reflexes during ablation may improve efficacy of AF ablation procedures [46].

•Vagal responsiveness also appears to decrease following ablation in the left atrium [47]. In some series, adding ganglionated plexi (GP) ablation to PV isolation appears to increase ablation success for AF [12]. Data suggest that areas in the atrium demonstrating complex fractionated atrial electrograms (CFAE) may represent a suitable target site for ablation; although several studies have reported that ablation at these sites may increase the efficacy of PV isolation procedures [48,49], enthusiasm for this approach has fallen over time. One possible explanation for the improvement in ablation success reported in these trials is that several CFAE sites anatomically overlie fat pads containing GPs [18,50]. As indicated above, autonomic denervation performed by GP ablation is thought to improve efficacy of AF ablation. (See "Atrial fibrillation: Catheter ablation" and "Catheter ablation for the treatment of atrial fibrillation: Technical considerations for non-electrophysiologists", section on 'Ablation techniques and targets'.)

•In a study of 40 patients with paroxysmal AF scheduled to undergo catheter ablation, individuals were randomly assigned to noninvasive transcutaneous low-level stimulation of the tragus (the anterior protuberance of the ear where the auricular branch of the vagus nerve is accessible) or to sham stimulation for one hour. Compared with control, low-level stimulation suppressed AF as measured by the decreased duration of atrial pacing-induced AF and an increased AF cycle length [51].

Maintenance of atrial fibrillation — In patients with persistent AF, the prevailing understanding of the mechanism is that, once triggered, the arrhythmia is maintained (sustained) by one or more abnormalities in the atrial tissue. This process may explain why the failure rate of PV isolation is as high as 40 to 60 percent at one year: The trigger(s) may have been treated but not the abnormalities that sustain AF once triggered (initiated).

The role of localized sources (electrical rotors and focal impulses) in the initiation and maintenance of AF was explored in the CONFIRM trial of 92 patients undergoing ablation procedures for paroxysmal or persistent (72 percent) AF [52]. Consecutive patients were prospectively treated (not randomly assigned) in a 1:2 case-cohort design with either conventional ablation at sources identified within the atria followed by conventional ablation or conventional ablation alone. Localized sources were identified in 97 percent of cases (70 percent rotors and 30 percent focal impulses) with sustained AF, each with an average of 2.1 sources. During a median of 273 days, patients treated with treatment of both sources and conventional ablation had a significantly higher freedom from AF (82.4 versus 44.9 percent).

Similar information was reported, indicating that “driver domains,” located in specific areas of the atria, act as unstable re-entry circuits that perpetuate atrial fibrillation in patients who have persistent AF [53,54].

Murine cell cultures show a differential ion channel gene expression associated with atrial tissue remodeling (ie, decreased SCN5A, CACN1C, KCND3, and GJA1; and increased KCNJ2) [55]. Fibrillatory complexity, increased in late compared with early stage cultures, was associated with a decrease in rotor tip meandering and increase in wavefront curvature.

Rotors are not the only explanation. In a study using high-density, simultaneous, biatrial, epicardial mapping of persistent and longstanding persistent AF in patients undergoing open heart surgery, several non-reentrant drivers were present in both atria in 11 or 12 patients with two to four foci per patient; foci were seen in both atria but generally in the lateral left atrial free wall, and likely acted as drivers. Reentry was not found to be the mechanism [56].

Likely, the substrate to maintain AF is a combination of reentrant activity and focal triggers. In a study of biatrial epicardial mapping of AF in sheep, wave propagation patterns were passing wave (69 percent occurrence, 68.6 percent of total time), point source (20.4, 13.1 percent), wave collision (4, 2.8 percent), reentrant wave (0.7, 6.3 percent), half-rotation (2.9, 4.4 percent), wave splitting (2.7, 4.3 percent), conduction block (0.05, 0.03 percent) and figure of eight reentry (0.05, 0.05 percent) [57]. Periods of repetitive activity were detected in the left and right atria.

The following sections describe factors that might contribute to the maintenance of AF.

Atrial remodeling — Atrial remodeling involves the concept that there are structural changes, such as fibrosis, or electrical changes, such as refractory-period dispersion or conduction delay, in the atria that can predispose to the development and maintenance of AF. In some instances, structural and electrical changes occur simultaneously. These processes can facilitate or create electrical reentrant circuits or triggers that can lead to AF [13,58]. It is also well established that the presence of AF results in remodeling of the atrium over time [7]. This explains the well-established concept that AF begets AF (figure 2). Thus, the longer a patient has been in continuous AF, the less likely it is to terminate spontaneously, and harder it is to restore and maintain sinus rhythm [59].

Electrical remodeling — Paroxysmal AF commonly precedes persistent AF. It has been suggested even after only a few minutes, AF induces transient changes in atrial electrophysiology that promote its perpetuation [14]. This might occur through a tachycardia-induced cardiomyopathy or through "electrical remodeling" of the atria by AF, leading to a progressive decrease in atrial refractoriness [14,15]. Electrical remodeling results from the high rate of electrical activation, which stimulates the AF-induced changes in refractoriness [60]. Tachycardia-induced changes in refractoriness are spatially nonuniform and there is increased variability both within and among various atrial regions [61]. It is possible that the change in atrial refractory period observed after an episode of AF predisposes to the spontaneous recurrence of AF in the days following cardioversion.

In addition to the shortening of the refractory period, chronic, rapid, atrial pacing-induced AF results in other changes within the atria, including an increase in the expression and distribution of connexin 43 and heterogeneity in the distribution of connexin 40, both of which are intercellular gap junction proteins ("gap junctional remodeling") [16,17]; cellular remodeling is due to apoptotic death of myocytes with myolysis, which may not be entirely reversible [18]; the induction of sinus node dysfunction, demonstrated by prolonged corrected sinus node recovery time, reduced maximal heart rate in response to isoproterenol, and lower intrinsic heart rate after administration of atropine and propranolol [19]; and an increase in P wave duration and intraatrial conduction time.

A clinical study evaluated the hypothesis of electrical remodeling by the use of atrial pacing-induced AF in patients with a history of supraventricular tachycardia [20]. AF significantly shortened the right-atrial effective refractory period after only a few minutes, and temporal recovery of the refractory period occurred over about eight minutes. Upon termination of AF, there was an increased propensity for the induction of another episode of AF that decreased with increasing time after the initial AF reversion. The second also tends to last longer than the first.

The time to recurrence was also evaluated in a review of 61 patients who had daily electrocardiogram (ECG) recordings using transtelephonic monitoring: 57 percent had recurrent AF during the first month after cardioversion, with a peak incidence during the first five days [21]. Among patients with recurrence, there was a positive correlation between the duration of the shortest coupling interval of premature atrial complex (PAC; also referred to a premature atrial beat, premature supraventricular complex, or premature supraventricular beat) after cardioversion, which correlates with the refractory period and the timing of recurrence (figure 3). (See "Atrial fibrillation: Cardioversion".)

In contrast to the normal situation in which the atrial refractory period shortens with an increase in rate (as in AF) and prolongs when the rate decreases, the refractory period fails to lengthen appropriately at slow rates (eg, with return to sinus rhythm) in patients with acute or persistent AF. The duration of AF has no significant impact upon the extent of these electrophysiologic changes [22].

Atrial electrical remodeling is reversed gradually after the restoration of sinus rhythm [23,62]. This may be one of the explanations for the early or immediate return of AF after cardioversion. In one study of 25 patients, the atrial refractory period increased and the adaptation of atrial refractoriness to rate was normal by four weeks after cardioversion [23]. In another report of 38 patients, the atrial refractory period increased by one week, with some variation in different regions of the atrium [62]. This observation has important clinical implications.

The mechanism for electrical remodeling and shortening of the atrial refractory period is not entirely clear; a possible explanation is ion-channel remodeling, with a decrease in the protein content of the L-type calcium channel [63]. Support for this comes from an animal study in which verapamil, an L-type calcium antagonist, prevented electric remodeling of short-duration AF (one day or less) and hastened complete recovery, without affecting inducibility of AF [64]. Similar findings have been noted in humans as verapamil, but not procainamide, prevented remodeling when given prior to the electrophysiologic induction of AF [65]. Oral diltiazem is also effective in some patients [66], while beta blockers had no effect on electrical remodeling in an animal model [60].

In comparison, cytosolic calcium overload, induced by hypercalcemia or digoxin, which increases the intracellular concentration of calcium by activating the sodium-calcium exchanger, enhances electrical remodeling [64,67,68]. The effect of digoxin, which is not due to its vagotonic activity, is associated with an increase in the inducibility and duration of AF [68].

Calcium leak from the sarcoplasmic reticulum may trigger and maintain AF. It is known that protein kinase A (PKA) hyperphosphorylation of the cardiac ryanodine receptor (RyR2), resulting in dissociation of the channel-stabilizing subunit calstabin2, causes sarcoplasmic reticulum (SR) calcium leak in failing hearts. This phenomenon seems to be involved in triggering ventricular arrhythmias.

Using similar logic, these proteins were investigated in atrial tissues from both dogs and humans with AF [69]. Atrial tissue in those with AF showed a significant increase in PKA phosphorylation of RyR2 and a decrease in calstabin2 binding to the channel. Channels isolated from dogs with AF had an increased open probability under conditions simulating diastole compared with channels from control hearts, suggesting that these AF channels could predispose to a diastolic SR calcium leak. The conclusion was that SR calcium leak due to RyR2 PKA hyperphosphorylation may play a role in the initiation and/or maintenance of AF. Other studies also suggest that RyR2 receptor-mediated calcium leak drives progressive development of an atrial fibrillation substrate in a transgenic mouse model [70,71].

The effects of calcium overload are quite complex. It is likely that triggers and substrates initiate short episodes of AF that then lead to calcium overload and over a period of minutes there is activation of the I _CaL current that increases I _K1, decreases I _Na, increases I_KACh, and decreases I_TO. This can affect the action-potential duration and allow for more reentry to occur. As reentry occurs, the substrate changes and there is remodeling through calcium handling abnormalities as well as mRNA transcription [59], and ultimately perhaps with protein decrease, changes in connexons, including, Cx40, that can affect conduction. The calcium-handling abnormalities can also lead to hypocontractility and atrial dilatation, thereby affecting even more the possibility of developing AF [59].

Both animal and human studies suggest that angiotensin II is involved in electrical and atrial myocardial remodeling [72,73] (see "Pathophysiology of heart failure: Neurohumoral adaptations", section on 'Renin-angiotensin system'). In an animal model, inhibition of angiotensin II with captopril or candesartan prevented shortening of the atrial effective refractory period and atrial electrical remodeling during rapid atrial pacing [72], while atrial tissue obtained during open heart surgery from patients with AF revealed downregulation of AT1 receptor proteins and upregulation of AT2 receptor [73]. The potential clinical importance of these changes is illustrated by the observations that angiotensin converting enzyme (ACE) inhibitors reduce the incidence of AF in patients with left ventricular dysfunction after myocardial infarction [74] and in patients with chronic left ventricular dysfunction due to ischemic heart disease [75]. (See "Angiotensin converting enzyme inhibitors and receptor blockers in acute myocardial infarction: Clinical trials".)

Another possible contributor to electrical remodeling and shortening of the atrial refractory period is atrial ischemia, which activates the sodium/hydrogen exchanger. The intravenous administration of HOE 642, a selective inhibitor of this sodium proton pump, to dogs undergoing rapid atrial pacing resulted in the lengthening of atrial refractoriness after one hour, while control dogs showed effective refractory-period shortening greater than 10 percent [76].

Role of fibrosis — The development of AF invokes atrial remodeling processes that involve electrophysiological and structural alterations that serve to maintain, promote, and propagate AF. In addition to electrophysiological alterations, such as shortening of the atrial action potential, increased dispersion of refractoriness, and conduction velocity shortening, morphological changes consist of fibrosis, hypertrophy, necrotic and apoptotic cell loss, and dilation [77].

Of these, fibrosis is considered especially important in the creation of AF substrate, especially in the setting of chronic atrial dilatation caused by heart failure. A canine model of heart failure has demonstrated a progressive increase in AF inducibility with increasing fibrosis [78]. An increase in conduction heterogeneity noted in this model is thought to play a major role in the creation of reentrant circuits in the dilated atria. Patients with AF also display increased atrial fibrous tissue content, along with increased expression of collagen I and III [79], as well as up-regulation of MMP-2 protein, and down-regulation of the tissue inhibitor of metalloproteinase, TIMP-1 [79]. Expression of the active form of MMP-9 and of monocyte chemoattractant protein-1, an inflammatory mediator, is increased in AF patients [80]. The left atrial free wall around the PV area presents particularly strong interstitial fibrotic changes [81-83].

Although the underlying molecular mechanisms that lead to the development of atrial fibrosis are complex, work suggests that the TGF-beta pathway may be an important contributor to the development of fibrosis (especially in the setting of increasing atrial stretch/dilatation resulting from congestive heart failure) [84-86].

Role of inflammation and oxidative stress — Emerging evidence suggests a significant role of inflammation in the pathogenesis of AF [87]. Evidence includes elevated serum levels of inflammatory biomarkers in patients with AF, the expression of inflammatory markers in atrial tissue from AF patients, and beneficial effects of antiinflammatory drugs in the setting of experimental AF [88]. Inflammation is suggested to be linked to various pathological processes, such as oxidative stress, apoptosis, and fibrosis that promote the creation and perpetuation of AF substrate. Several of the downstream effects of inflammation in the heart are thought to be mediated by oxidative stress [89].

Indeed, studies in patients with AF demonstrate increased generation of reactive oxygen species (ROS) in the fibrillating atrium compared with normal atria [90,91]. Several major enzymatic sources of ROS have been implicated in AF. Of these, NAPDH oxidase (specifically its NOX2 isoform) has been shown to be elevated in humans with AF in a variety of studies [92,93]. Other sources of ROS implicated in AF include uncoupled nitric oxide synthase [94] and xanthine oxidase [95]. In addition to the increase in ROS noted in tissue from patients with AF, experimental evidence suggests that ROS may be implicated not only in promoting AF but also in maintaining atrial arrhythmia. The administration of antioxidants such as vitamin C or statins (which are known to have pleiotropic antioxidant effects) decreased AF inducibility in canine models of tachypacing-induced AF [96,97]. Antioxidants such as vitamin C and n-acetylcysteine have been administered to patients undergoing cardiac surgery and have been shown to decrease postoperative AF [98,99]. These early results are encouraging and warrant further investigation of inflammation and oxidative stress as viable therapeutic targets in patients with AF.

Recent data indicate an association of gut microbiome dysbiosis and AF [100,101]. In a recent meta-analysis of 14 studies on 2479 patients, biomarkers of gut dysbiosis (eg, short-chain fatty acids, trimethylamine N-oxide, lipopolysaccharides, and bile acids) were shown to be related to AF. Whether inflammation plays a role in this association is uncertain. The microbiome may also be related to postoperative AF [102]; proposed mediators of this association include increased postoperative inflammation and electrolyte abnormalities.

Reentrant mechanism — Maps of AF in animals and humans suggest that this arrhythmia is caused by multiple wandering wavelets (figure 4), and these may be due to heterogeneity of atrial refractoriness and conduction. In addition, the response of atrial activity to adenosine infusion suggests a reentrant rather than a focal mechanism [103]. Adenosine increases the inward potassium rectifier current, which shortens refractory periods and would accelerate reentrant circuits. In contrast, this effect would slow an automatic or triggered focus. In a series of 33 patients with AF undergoing electrophysiology study, adenosine increased the dominant frequencies, supporting reentrant rather than focal sources for the perpetuation of AF.

It has been suggested that at least four to six independent wavelets are required to maintain AF [104]. These wavelets rarely reenter themselves but can re-excite portions of the myocardium recently activated by another wavefront, a process called random reentry [7,105-107]. As a result, there are multiple wavefronts of activation that may collide with each other, extinguishing themselves or creating new wavelets and wavefronts, thereby perpetuating the arrhythmia (figure 5).

The reentrant circuits are therefore unstable; some disappear, while others reform. These circuits have variable but short cycle lengths, resulting in multiple circuits to which atrial tissue cannot respond in a 1:1 fashion. As a result, functional block, slow conduction, and multiple wave fronts develop [107].

Patients with AF may have increased dispersion of refractoriness. This correlates with enhanced inducibility of AF and spontaneous episodes [108] likely related to unstable reentry circuits. Some patients have site-specific dispersion of atrial refractoriness and intraatrial conduction delays resulting from nonuniform atrial anisotropy [109]. This appears to be a common property of normal atrial tissue, but there are further conduction delays to and within area surrounding the AV node in patients with induced AF, suggesting an important role for the low right atrium in the genesis of AF.

Abnormalities in restitution as well as the spatial distribution of such abnormalities can be related to the persistence of AF. In one study, monophasic action potential recordings were evaluated in patients with AF [110]. The action potential duration was plotted as a function of the preceding diastolic interval, and the slope of the action potential duration versus the diastolic interval (the restitution curve) was determined. If the slope was greater than one, oscillations occurred that may cause localized conduction delay or block resulting in a wave break giving rise to atrial fibrillation.

These different patterns of conduction are reflected in the morphology of electrograms recorded with mapping during induced AF. Single potentials were indicative of rapid uniform conduction, short double potentials indicated collision, long double potentials were indicative of conduction block, while fragmented potentials were markers for pivoting points or slow conduction (figure 6) [111,112].

Sites of fragmented potentials or complex fractionated atrial electrograms are potential targets for radiofrequency ablation to terminate AF as they may represent critical areas from which AF originates and perpetuates. (See "Atrial fibrillation: Catheter ablation".)

This phenomenon has been termed “microreentry” to distinguish it from classic reentry in which the same reentrant pathway is repetitively traversed. The impulse may circulate around a central line of functional block, so-called leading circle reentry; this type of reentry tends not to be stable but rather to drift through the atria until it is extinguished. The perpetuation of AF may also depend importantly upon macroreentry around natural orifices and structures in the atrium, which provides a rationale and anatomic landmarks for ablative treatment. The collision of wavefronts cancels many atrial depolarizations that might otherwise reach the AV node, resulting in a slower heart rate than might otherwise have occurred (figure 7A-B).

Although multiple wandering wavelets probably account for the majority of AF, one study reported nine patients in whom a single, rapidly firing focus was identified with electrophysiologic mapping [113]. Organized and rapid atrial activity with a centrifugal and consistent pattern of atrial activation resulted from this focus, but it fired irregularly with striking and abrupt changes in atrial cycle lengths. In most of the patients, the focus was near the ostia of great vessels and was amenable to radiofrequency ablation (figure 8 and figure 9).

Small reentrant sources, called rotors, may drive or maintain AF in some cases. These rotors result in a hierarchical distribution of frequencies throughout the atria that may be identified with spectral analysis of intracardiac recordings. Ablation of such sites has terminated paroxysmal AF, suggesting that they may play an important role [114], but it is not clear that the rotors are responsible for AF or are fixed in most instances. AF may be chaotic and have wavelets and rotors that are secondary rather than the predominant cause of AF [115]. However, antral pulmonary venous reentrant and focal drivers may be responsible for AF [54]. The complexity of such drivers increase with prolonged AF. These sites are often localized near the PV orifices in patients with paroxysmal AF, and are more often localized to the left or right atria in patients with chronic AF [103].

The fibrillating atrium cannot be captured by pacing when the atrial electrograms are disorganized. This observation supports the presence of microreentry, since there is no excitable gap (or it is very small) to permit capture. However, when type I (figure 9) AF (which has organized atrial electrograms) is induced by rapid atrial pacing, the fibrillating atrium can be captured with rapid atrial pacing, suggesting the presence of an excitable gap [116].

ROLE OF THE ATRIOVENTRICULAR NODE — The atrioventricular (AV) node regulates the number of atrial impulses that reach the ventricle. The ventricular rate in atrial fibrillation (AF) is typically irregularly irregular, with a ventricular rate that may be slow, moderate, or rapid depending on the capacity of the AV node to conduct impulses. The rate of AV nodal conduction is dependent upon multiple factors, including electrical properties of the node and the influence of the autonomic nervous system [117]. In addition, the use drugs such as digoxin, calcium channel blockers, or beta blockers may influence AV nodal function. There also may be a circadian rhythm for both AV nodal refractoriness and concealed conduction, accounting for the circadian variation in ventricular response rate [118].

AV nodal tissue consists of so-called "slow response" fibers, which depend on a mixed calcium/sodium current. This current is often called the inward calcium current, since in a normal physiologic environment, the ions are almost exclusively calcium. The mixed current uses a kinetically slow channel and is responsible for phase 0 depolarization. (See "Cardiac excitability, mechanisms of arrhythmia, and action of antiarrhythmic drugs".)

These characteristics lead to properties that are quite different from "fast-response" tissue in the atria, which as noted above, depend on an inward sodium current that uses a kinetically fast channel for phase 0 depolarization [8,119]:

●Partial and complete reactivation returns only 100 ms or more after return to the diastolic potential (versus 10 to 50 ms in the atria).

●The refractory period changes little as a function of rate.

●Conduction velocity is relatively slow, ranging from 0.01 to 0.1 m/s.

●Unlike tissue generating a fast action potential that has an all-or-none response (ie, the velocity of impulse conduction is similar at all stimulation rates until block occurs), tissue that generates a slow action potential exhibits a graded or decremental response, in which the velocity of impulse conduction slows as the stimulation rate increases.

●As noted above, the ventricular rate usually ranges 90 and 170 beats/min. Ventricular rates below 60 beats/min are seen with AV nodal disease, drugs that affect conduction, and high vagal tone as can occur in a well-conditioned athlete. Ventricular rates above 200 beats/min suggest catecholamine excess, parasympathetic withdrawal, or the existence of an accessory bypass tract as occurs in the preexcitation syndrome. The QRS complexes are widened in the last setting and must be distinguished from a rate-related or underlying bundle branch block.

In the classical view, the AV node is bombarded by impulses from the fibrillating atria. Some impulses traverse the AV node and reach the specialized infranodal conduction system and then the ventricles. However, most atrial impulses penetrate the AV node from varying distances and then are extinguished when they encounter the refractoriness of an earlier wavefront; this phenomenon of concealed conduction in turn creates a refractory wave that affects succeeding impulses. The failure of the refractory period to shorten with increasing rate (as occurs in the atria) further decreases the likelihood of an impulse traversing the AV node.

Anatomically distinct AV nodal inputs, called the slow and fast pathways, are involved in the ventricular response to AF. The importance of these pathways has been demonstrated in radiofrequency ablation studies in which ablation reduced the number of beats that successfully reached the infranodal conduction system and the ventricles [120-123]. (See "Atrioventricular nodal reentrant tachycardia".)

In addition to its intrinsic properties, the AV node is richly supplied and affected by both components of the autonomic nervous system. AV conduction is enhanced and refractoriness reduced by the sympathetic fibers, and conduction reduced and refractoriness prolonged by the parasympathetic fibers.

The net effect of the electrophysiologic properties of the AV node is that the rate of conduction into the specialized infranodal conduction system is (fortunately) much slower than the rate of the fibrillating atria. In some cases, the high degree of refractoriness in the AV node with AF results in high-grade or third-degree block. In this setting, the pacemaker that controls the ventricles is below the AV node. (See "The electrocardiogram in atrial fibrillation".)

In patients with the preexcitation syndrome, the AV node is bypassed by "fast-response" tracts, which activate and reactivate much faster than the AV node and are therefore capable of rapid conduction. The development of AF in such a patient can result in very rapid transmission of atrial impulses to the ventricles [123] and can rarely cause ventricular fibrillation [15]. (See "The electrocardiogram in atrial fibrillation".)

It is also important to recognize that the presence of an accessory pathway can increase the propensity for development of AF, particularly if the pathway conducts only antegrade. The mechanism(s) for how accessory pathways affect the presence of AF is uncertain. In patients with AF who have Wolff-Parkinson-White (WPW) syndrome, catheter ablation of the accessory pathway is indicated to lower the sudden death risk but also to decrease the probability of recurrent AF.

Unexpected ventricular rates — The ventricular response to AF characteristically is irregularly irregular although it may appear regular in the presence of complete AV block. The usual ventricular rate in AF is between 90 and 170 beats per minute in the absence of AV node disease, drugs that affect conduction, or enhanced vagal inputs. Ventricular rates that are clearly outside this range suggest some concurrent problem:

●A ventricular rate below 60 beats per minute, in the absence of AV nodal blocking agents, suggests AV nodal disease that may be associated with the sinus node dysfunction. (See "Sinus node dysfunction: Epidemiology, etiology, and natural history".)

●A ventricular rate above 170 beats per minute suggests thyrotoxicosis, catecholamine excess, parasympathetic withdrawal, or the existence of an accessory bypass tract in the preexcitation syndrome. (See "Epidemiology, risk factors, and prevention of atrial fibrillation".)

SPECIFIC CLINICAL SITUATIONS

Late recurrent AF after catheter ablation — The etiology of late recurrent atrial fibrillation (AF) following pulmonary vein isolation (PVI) has been debated. In some cases, triggering foci outside of the PVs may initiate AF [124-127]. Alternatively, persistence of the substrate for maintaining AF (abnormal electrical properties of the atria themselves) may be more important than the triggering foci, especially in chronic AF.

However, there is increasing evidence that when AF does recur late after PVI, it often represents incomplete electrical isolation of the PVs, either due to resumption of conduction across the ablation scar or to residual conduction in PVs that were not successfully ablated. Most [128-131], but not all [132], studies of the former mechanism support the hypothesis that resumption of PV-left atrial (LA) conduction is associated with an increased risk of recurrent AF. However, recurrent conduction across ablated lesions is more common than clinically evident recurrent AF [130,133].

Pre-existing LA scarring may predispose patients to late recurrence. In a series of 700 consecutive patients undergoing first-time PVI, scarring was detected in 6 percent [134]. These patients had a much higher rate of recurrence than those without scarring (57 versus 19 percent).

Possible causes of scarring include atrial remodeling and inflammation. The patients with scarring had significant elevations in serum C-reactive protein (CRP) compared to those without scarring (5.9 versus 0.31 mg/L). This is consistent with other studies showing a relationship between serum CRP and AF [135]. (See "Epidemiology, risk factors, and prevention of atrial fibrillation", section on 'Inflammation and infection'.)

After cardiac surgery — AF occurs frequently (approximately one of four patients) after cardiac surgery. Nonuniform atrial conduction is greatest on days two and three in this setting, and the longest atrial conduction time is greatest on day three after open heart surgery; these abnormalities coincide with the time of greatest risk for AF [136]. The degree of atrial inflammation after surgery in dogs was associated with a proportional increase in the inhomogeneity of atrial conduction and in the duration of AF; antiinflammatory therapy decreased the inhomogeneity [137]. Nevertheless, the mechanism of AF in the postoperative period is likely multifactorial. It is important to note that in most of the patients, especially those without a prior history of AF, that the AF is self-limited, and antiarrhythmic drug therapy can usually be stopped two to three months following surgery when the inflammation has subsided. (See "Atrial fibrillation and flutter after cardiac surgery".)

Hyperthyroidism — It is well established that hyperthyroidism can increase susceptibility to development of AF. As a consequence, all patients with new onset AF should have some measure of thyroid function tested. Successful treatment of the hyperthyroidism often results in elimination of the AF.

Obesity — Obesity has been associated with AF and it is possible that both are related mechanistically [138]. In a sheep model, weight gain was associated with increased left atrial volume, fibrosis, inflammatory infiltrates, and lipidosis. There was reduced conduction velocity in atrial tissue and increased inducible and spontaneous AF with obesity. Atrial endothelin-A and -B receptors, endothelin-1, atrial interstitial and cytoplasmic transforming growth factor beta1, and platelet-derived growth factor were higher with obesity. In a clinical study of 110 patients undergoing AF ablation versus 20 reference patients without AF, pericardial fat volumes were associated with AF, its chronicity, and its symptom burden. Pericardial fat predicted AF recurrence post-ablation [139]. Associations persisted after adjusting for body weight but body mass index was not associated with these outcomes in multivariate-adjusted models. In another report [140], weight management with subsequent weight loss was associated with improved AF symptom burden scores, symptom severity scores, number of episodes, and cumulative duration of AF. This preliminary information does not yet prove that obesity causes AF by any specific mechanism. In a study of atrial sheep myocytes, acute, short-term incubation in free fatty acids resulted in no differences in passive or active properties of isolated left atrial myocytes but stearic acid reduced membrane capacitance and abbreviated the action potential duration, likely due to a reduction of the L-type calcium and of the transient outward potassium currents [141].

Sleep apnea — Sleep apnea has been associated with AF [142]. It is uncertain if this association is mediated by other shared comorbidities (eg, obesity, alcohol). Observational data suggest that the treatment of sleep apnea is associated with reduction in AF. Potential mechanisms underlying the association between sleep apnea and AF include atrial distention, increased systemic inflammation and oxidative stress, autonomic perturbations, hypoxia, hypercapnia, effects on connexin dysregulation and increased fibrosis, metabolic dysregulation, and the possible role of epicardial fat secretome [142].

Alcohol — Alcohol is a known provocateur of AF. The mechanism may be a direct toxic effect. Alcohol can reduce L-type calcium and sodium current density in rats. Furthermore, there can be an increase in I _KACH activity, shortening the action potential. Additionally, alcohol can shorten pulmonary vein action potential duration by increasing I_TO and cause an upregulation in protein expression of the potassium channel Kir 3.1. Alcohol can affect sympathetic and parasympathetic modulation. Alcohol can damage cellular gap junctions, cause inflammation with oxidative stress, autonomic dysregulation, and affect atrial electrical channels, thereby leading to AF [3].

GENETICS OF AF — Over the last decade, a preponderance of evidence suggests a large genetic contribution to atrial fibrillation (AF) [143,144]. Having a family member with AF is associated with a 40 percent increased risk for the arrhythmia [145]. Initially, traditional genetic techniques such as linkage analysis led to the discovery of rare, monogenic causes of AF. The first such study identified a genetic locus for AF using a series of related families with early onset AF [146]. A later study identified the first gene for familial AF [147]. Using a large Chinese kindred with autosomal dominant AF, they found a gain-of-function mutation in KCNQ1 (the gene encoding the α subunit of the potassium channel current, I_Ks). Since then, several additional gain-of-function variants have been identified in KCNQ1 [148,149]. In addition to KCNQ1, mutations have been identified in other potassium channels genes, including KCNA5 [150], KCND3 [151], and KCNJ2 [152], and accessory subunits KCNE1 [153], KCNE2 [154], KCNE3 [155], and KCNE5 [156,157].

The majority of these functionally validated, AF-associated potassium channel variants have a gain-of-function channel, with an expected shortening of the atrial action potential duration and atrial refractory period. Variation in sodium channel subunits has also been identified as an important factor in the development of familial AF, with AF-causing variants observed in both the major cardiac sodium channel alpha subunit SCN5A [158] and its associated beta subunits [159,160]. Several variants have also been identified in genes that do not directly alter the atrial action potential, but instead would be expected to cause AF through alternative mechanisms, eg, somatic mutations in GJA5, which encodes the gap junctional protein; connexin 40, a frameshift mutation that resulted in early truncation of NPPA [161], which encodes for the precursor for atrial natriuretic peptide; and genetic variation in several developmentally related cardiac transcription factors, ie, NKX2.5, PITX2, GATA4, GATA5, and GATA6 [160,162,163].

Genome-wide association studies (GWAS) have been used to identify genetic loci associated with AF. GWAS rely on the unbiased comparison of common single-nucleotide polymorphisms (SNPs) throughout the genome, with SNPs occurring with different frequency in individuals with a disease versus controls being used to localize disease-related genetic loci. The first GWAS performed for AF identified a region on chromosome 4q25, which was associated with AF in those of European and Asian descent [164]. Subsequently, these findings were broadly replicated in individuals of European, Asian, and African descent [165,166]. Genetic variants on chromosome 4q25 that are most significantly associated with AF reside about 150 kilobases upstream of the nearest gene PITX2. PITX2 encodes the paired-like homeodomain transcription factor 2, which helps determine cardiac laterality, suppresses the default expression of a sinoatrial nodal gene programme in the left atrium, and encodes the pulmonary venous myocardium [167]. In addition, PITX2 is associated with formation of the pulmonary veins. These findings are particularly interesting in light of the fact that AF triggers frequently arise in the pulmonary veins. In addition to the role of PITX2 in development, studies demonstrate a role for the PitX2c transcript in expression of gene-encoding ion channels, calcium cycling proteins, and gap junctions; these direct electrophysiological influences likely lead to formation of substrate for triggered activity as well as reentry [168].

The AF-associated ZFHX3 locus was shown to cause atrial dysregulation in knockout mice. Specifically, ZFHX3 knockout lead to significant increase in incidence and burden of AF, with alterations in conduction velocity, atrial action potential duration, calcium handling and the development of atrial enlargement and thrombus, and dilated cardiomyopathy [169].

Related analysis identified the same genomic region as being associated with an increased risk of cardioembolic stroke [160,170] and a prolonged PR interval [171]. To date, GWAS have identified 14 genomic regions of susceptibility for AF, with 17 independent signals at these loci [172]. These include the ZFHX3 gene that encodes a zinc finger homeobox transcription factor [173], the KCNN3 gene that encodes the SK3 potassium channel [174], and the PRRX1 gene that encodes a member of the paired-related homeobox gene family [173]. Whole exome and genome sequencing has been increasingly used to identify rare variants associated with AF [160]. For example, Oleson et al reported a much higher prevalence of rare variants in genes associated with AF (KCNQ1, KCNH2, SCN5A, KCNA5, KCND3, KCNE1, 2, 5, KCNJ2, SCN1-3B, NPPA, and GJA5) in early onset, lone AF patients than in the background population [175]. This approach is beginning to identify rare candidate variants in genes not previously linked to other types of Mendelian disease and thus may offer new insights into AF pathogenesis and disease pathways that could ultimately provide novel therapeutic targets for this common condition.

A 2018 meta-analysis of genome-wide association studies (GWAS) for AF to date, consisting of more than 500,000 individuals, sought to identify AF-associated genes at the GWAS loci by performing RNA sequencing and expression quantitative trait locus analyses in 101 left atrial samples (which is the most relevant tissue for AF) [176]. A transcriptome-wide analysis was also performed; this analysis identified 57 AF-associated genes, 42 of which overlap with GWAS loci. The identified loci-implicated genes enriched within cardiac developmental, electrophysiological, contractile, and structural pathways.

There have been several attempts to create polygenic risk scores in order to incorporate the influence of multiple disease-associated genetic risk variants on the risk of development of AF [177].

SUMMARY

●Introduction – The precise mechanisms by which age and other risk factors such as hypertension, coronary artery or valvular heart disease, or heart failure increase the propensity for development of atrial fibrillation (AF) are poorly understood (figure 1). These conditions may affect the triggers of or the substrate for the maintenance of AF. (See 'Introduction' above.)

●Triggers and substrate – These mechanisms are complex and involve a dynamic interplay between the triggers and substrate abnormalities. It is likely that short-lived episodes are due to specific triggers, including autonomic perturbations, focal discharges, specific reentry circuits in the pulmonary veins (PVs), and effects of stretch, whereas inflammation, dilatation, fibrosis, repolarization abnormalities, and conduction disturbances allow for perpetuation of episodes of AF. (See 'Mechanisms of atrial fibrillation: triggers and substrates' above.)

●Triggers of AF – AF is most often initiated (triggered) by rapid firing from the PV. (See 'Triggers of AF' above.)

●Electrical remodeling – Paroxysmal AF commonly precedes chronic AF. This suggests that, in addition to other predisposing factors, AF may play a role in its own natural history. (See 'Electrical remodeling' above.)

●Autonomic nervous system – This likely influences the initiation and perpetuation of AF. (See 'Role of the autonomic nervous system' above.)