Genetic and Non-Genetic Risk Factors Associated with Atrial Fibrillation

Abstract

Atrial fibrillation (AF) is the most common arrhythmic disorder and its prevalence in the United States is projected to increase to more than twelve million cases in 2030. AF increases the risk of other forms of cardiovascular disease, including stroke. As the incidence of atrial fibrillation increases dramatically with age, it is paramount to elucidate risk factors underlying AF pathogenesis. Here, we review tissue and cellular pathways underlying AF, as well as critical components that impact AF susceptibility including genetic and environmental risk factors. Finally, we provide the latest information on potential links between SARS-CoV-2 and human AF. Improved understanding of mechanistic pathways holds promise in preventative care and early diagnostics, and also introduces novel targeted forms of therapy that might attenuate AF progression and maintenance.

Graphical Abstract

1. Introduction

Cardiovascular disease (CVD) is the leading cause of mortality and morbidity in developed countries, accounting for one-third of global deaths.¹ Atrial fibrillation-related CVD has led to a total of 276,373 deaths between 2011 and 2018.² Deaths from conditions such as ischemic heart disease, hemorrhagic stroke, and rheumatic heart disease saw a decline in death rates due to epidemiologic changes. Notably, mortality rates due to atrial fibrillation (AF) increased, driven by changes in age- and sex- specific death rates.³ AF is the most common arrhythmic disorder that increases risk of heart failure (HF), myocardial infarction, chronic kidney disease, stroke, and death.^{4, 5} The growing incidence and prevalence of AF are likely due to a growing elderly population.

Medical management of patients with AF remains complex and may involve invasive procedures in up to 20% of cases.⁶ Although genetics may impact the development of AF, the cause of onset is difficult to pinpoint due to environmental and epigenetic risk factors. The first genetic association of AF was discovered in 1943 by Wolff⁷ and led to further investigation on potential genetic mechanisms that might contribute to AF inheritance. Polygenic factors make up a major segment of susceptibility to inherited AF⁸, making it likely that many different pathways are involved in the development of AF and the onset is based upon the individual. In addition to polygenic backgrounds, the variability and temporal occurrence of other environmental risk factors further complicate the mechanism of disease progression. AF may develop secondary to other variables such as hypertension, coronary artery disease, alcohol consumption, obesity, and diabetes, among others.⁹ Here, we review variants identified in ion channel genes and their contribution to the pathophysiology of AF. We also include an overview of recently identified variants in genes encoding cardiac structure, signaling, and their association with AF. This information might provide guidance in AF management through the development of novel diagnostics, design of targeted therapies and prediction of AF recurrence and outcomes.

2. Clinical Classification of AF

AF is a progressive disease that is self-perpetuating due to electrical dysregulation that affects structural remodeling and is diagnosed by electrocardiography. AF is classified by its temporal pattern and occurrence in patients, including paroxysmal, persistent, long-standing persistent, and permanent. Paroxysmal AF is classified by periods of AF that terminate spontaneously or with intervention and usually last less than seven days.¹⁰ Persistent AF is a sustained AF episode, typically lasts more than seven days, and requires medical interventions to return to sinus rhythm by pharmacologic or electrical cardioversion. Both paroxysmal and persistent AF may recur in patients. Long-standing persistent AF refers to AF lasting more than one year and when a rhythm control strategy is determined. Permanent AF refers to persistent AF longer than one year that is unable to be resolved by cardioversion or when cardioversion is not attempted and, in this condition, AF is the accepted rhythm for the patient.^{11, 12} Lone AF is a diagnosis of exclusion where AF cannot be attributed to any other concomitant CVD or comorbidity with a structurally normal heart on echocardiogram.¹⁰ Finally, non-valvular AF refers to AF in the absence of mitral stenosis, a mechanical or prosthetic valve or mitral valve repair.¹³

3. Pathogenesis of Atrial Fibrillation

AF is characterized by an increase in excitation rate and irregular activation of the atrium that results in dysfunction in atrial contraction.¹⁴ Normal cardiac rhythm is initiated in the sinoatrial node (SAN) leading to the contraction of the atria followed by ventricular excitation. Atrial structural and functional changes constitute atrial remodeling that promotes arrhythmia and is central to AF maintenance. There are fundamental determinants of how remodeling increases the likelihood of reentry. In a normal pathway, refractory tissue causes a unidirectional block of an ectopic beat.¹⁵ Atrial remodeling increases conduction time of the ectopic beat and shortens the refractory period in atrial tissue, slowing the entire circuit to allow the recovery of excitability outside the normal signaling pathway. Structural remodeling, particularly fibrotic infiltration, replaces dead cardiomyocytes thereby increasing conduction time and promoting ectopic activity. Fibrosis has been reported to cause AF progression, thus, fibrotic infiltration during structural remodeling is a potential therapeutic target for AF treatment.¹⁶

Electrical remodeling alters ion channel expression and function in a way that promotes AF.¹⁶ The refractory period of the atria is dependent on cardiac action potential duration (APD). Among other currents, APD is determined by inward L-type calcium (I_CaL) current, inward rectifier current (I_K1) and ultra-rapid delayed rectifier K⁺ current (I_Kur) balance.^{17, 18} During electrical remodeling, there is an I_CaL downregulation, reducing Ca²⁺ entry, and enhancement of I_K1 current. APD abbreviation is caused by changes in channel function and promotes multiple circuit reentry in the atria, leading to a possible mechanism for AF development.¹⁹ In addition, diastolic Ca²⁺ leak from the sarcoplasmic reticulum (SR) activates inward Na⁺ current through Na⁺/Ca²⁺ exchange initiating or maintaining AF.²⁰ The increase in cytosolic Ca²⁺ activates the Na⁺/Ca²⁺ exchanger causing depolarization, which produces ectopic action potentials.²¹ Abnormalities in ryanodine receptors (RyRs) - specialized SR Ca²⁺ channels that respond to transmembrane Ca²⁺ entry^{20, 21}- cause enhancement of SR Ca²⁺ leak underlie the phenomenon known as delayed afterdepolarization (DAD).²²

The pathogenicity of atrial remodeling via structural and electrical routes is further complicated with the introduction of inflammatory and oxidative processes. Oxidative stress has been correlated with Ca²⁺ handling abnormalities and elevated AF risk.²³ Notably, oxidative stress activates Ca²⁺/calmodulin-dependent protein kinase II (CAMKII) through oxidation at Met281/282 and oxidized CAMKII was increased in AF patients and has been implicated in AF inducibility.²⁴ Further, JNK2-dependent CAMKII activation has been implicated in conditions that promote AF.²⁵ Inflammation may be induced by several cardiac-related diseases (including but not limited to hypertension, congestive HF, and diabetes), where stress leads to endothelial dysfunction and arterial damage. Specifically, inflammatory modulators may lead to structural and electrical atrial changes by inducing atrial fibrosis, gap junction and intracellular calcium regulation. These changes ultimately increase ectopic activity and slow atrial conduction, which promotes re-entry.²⁶

Biomarkers for inflammation including C-reactive protein, tumor necrosis factor-α (TNF-α), and interleukin (IL)-6 are indicated as a risk factor for AF disease initiation and AF recurrence after electrical ablation.^{26, 27} These biomarkers have specific effects in altering inflammatory pathways, and they likely contribute to AF by affecting the electrical signaling of cardiomyocytes.²⁷ Regulation of these biomarkers has been linked to an inflammasome complex “NACHT, LRR, and PYD domain containing protein 3” (NLRP3). NLRP3 inflammasome is increased in the atrial samples of paroxysmal and persistent AF patients, promoting ectopic firing and AF-related substrates. Yao et al. proposed that NLRP3 acts by enhancing the expression of ryanodine receptor 2 (Ryr2), resulting in enhanced sarcoplasmic reticulum Ca²⁺ release that triggers DADs and ectopic firing.²⁸ The mechanisms underlying the effect of NLRP3 in the atria also include the increased function of I_Kur provoking action potential abbreviation.²⁷ An alternative pathway is that NLRP3 inflammasome leads to caspase-1 cleavage causing cytokine secretion to recruit leukocytes and induce fibrosis, which maintains AF-related substrates.²⁸ Atrial remodeling and fibrosis may be further exacerbated as AF leads to inflammation, creating a cyclical nature in which AF may progress. While the mechanism of inflammation and oxidative stress remain to be elucidated, their causative role in aggravating structural and electrical atrial modifications warrants special consideration of pathogenicity, comorbidities, and treatment.

Overall, AF can develop from substrates formed from reentry, structural/electrical remodeling, and ectopic firing. Many complex mechanisms including reentry, remodeling and ectopic firing cause changes in the activation of the typical signaling pathway to drive and sustain AF.²⁹ Maintenance of AF is further complicated by the cyclical nature of inflammatory and oxidative stress in AF initiation and recurrence. Although there have been major strides in identifying pathophysiological mechanisms that underlie AF, there are many aspects of these mechanisms that remain unclear. It is important to elucidate AF pathogenesis, as treatment is currently limited to preventing symptoms of AF rather than addressing causative mechanisms. Advances in the identification of AF drivers through optical mapping show promise in informing new therapeutic targets.

4. Genetic Diversity in Atrial Fibrillation

Variants in ion channel, structural, signaling, and transcription factor genes have all been linked to AF development and maintenance (summarized in Figure 1 and Table 1). Interestingly, non-ion channel genes have been linked to AF pathogenesis, where variants affect sarcomeric proteins, hinder cell-cell communication, or impact nuclear structure and transport.

Figure 1: Schematic of genetic variants identified in ion channel and non-ion channel genes.

Variants are added next to the proteins. Red text for key organelle locations. Black text for AF-linked proteins. Abbreviations include calsequestrin-2 (CASQ2), ryanodine receptor 2 (RyR2), potassium voltage-gated channel (K_V1.5), inward-rectifier potassium ion channel (Kir2.1), G-protein activated inwardly rectifying potassium channel (Kir3.1/3.4), voltage gated sodium channel (Na_V1.5), short stature homeobox 2 (SHOX2), Kruppel-like factor 15 (KLF15), Homeobox protein NKX2.5 (NKX2.5), hyperpolarization activated cyclic nucleotide gated potassium channel 4 (HCN4), atrial natriuretic peptide (ANP), potassium voltage-gated channel subfamily Q member 1/Potassium voltage-gated channel subfamily E regulatory subunit 1-5 (KCNQ1/KCNE1-5).

Table 1:

Genetic complexity of AF

Gene	Protein	Current	Examples of variants	Mechanism/suggested mechanism of pathogenesis
SCN5A	SCN5A, alpha subunit of Nav1.5	I Na	LOF - H558R31 N1986K32	LOF variants lead to a reduction in sodium current density and an increase in AF susceptibility.
			GOF - M1851V33 (R222Q, K1493R)33	GOF variants result an increased INa current window.
SCN1-4B	SCN1-4B, beta subunits of Nav1.5	I Na	SCN1B – (R85H, D153N)34	SCN1B: variants show reduced sodium current amplitudes.
			SCN1Bb - R214Q36	SCN1Bb: variant is implicated in Brugada syndrome but has also been associated with early onset lone-AF.
			SCN2B – (R28W, R28Q)34	SCN2B: variants in SCN2B (R28W and R28Q) display reduced peak sodium current amplitude.
			SCN3B – (R6K, L10P, M161T)37	SCN3B: variants in SCN3B (R6K, L10P, and M161T) affect the sodium current by altering steady-state inactivation, reducing peak current density, or both.
			SCN4B – (V162G, I166L)35	SCN4B: variants may affect the sodium current density and the voltage dependence of sodium channel activation or inactivation.
HCN4	HCN4	I f	LOF – (L573X, K530N)40	Loss of function of HCN4 channels.
			P883R41	Shifts the activation voltage of If to more positive potentials and increases the current density.
KNCJ2	Kir2.1	I K1	GOF - V93I42	The variant causes increase in IK1 channel activity.
KCNQ1/KCNE1-5	KCNQ1/KCNE1-5	I KS	KCNQ1- (S140G, G229D)44, R14C45, S209P46, (R231C, R231H)44	GOF variants in KCNQ1 have been reported to abbreviate the atrial APD and promote reentry.
			KCNE2- (M23L, I57T)48	GOF variants in KCNE2 and KCNE5 increase the IKS channel activity creating a substrate for AF susceptibility.
			KCNE4- E145D49
			KCNE5- L65F47
KCNJ5	Kir3.4	I KAch	(C171T, G810T, C834T)51	Variants may cause reduction in the APD and the effective refractory period in atrial myocytes, which may contribute to AF initiation and maintenance.
			G387R50	Loss of IKACh channel function.
KCNA5	Kv1.5	I kur	LOF – (T527M, A576V, E610K)53 E375X54	LOF effects on IKur prolong the effective refractory period and enhance early afterdepolarization predisposition.
			GOF – (E48G, A305T, D322H)55	Increase in IKur current and therefore shortening of the APD.
GJA1	Connexin-43		Frameshift - c.932delC56	Variant is mosaic and unequal distribution of the variant protein can result in difference in conduction velocities and promote electrical reentry.
GJA5	Connexin-40		(P88S, M163V, G38D, A96S)60	Variants cause abnormal gap-junction formation and weakened intercellular electrical coupling.
NPPA	ANP		Frameshift (c.456-457delAA)65	Causes extended half-life of mutant ANP.
			GOF– (A117V, S64R)66, 67	GOF effects in the IKS current, which leads to shortened APD.
LMNA and NUP155	Lamin A/C		LMNA-T488P71	No conclusive evidence.
	Nucleoporin 155		NUP155 72	Variants in NUP155 reduce nuclear envelope permeability by affecting the overall nuclear pore complex and lead to shortened APD
JPH2	Junctophilin-2		E169K73	Decreased binding of E169K-JPH2 to RyR2 and increase in SR Ca2+ leak.
SYNPO2L	Synaptopodin 2 like		LOF74	Lack of a protein product or formation of a dysfunctional protein.
TTN	Titin		TTNtv76	Increase in fibrosis.
MYH7	MyHC motor protein, β		Variation in MYH7 S1 domain78	Structural and hemodynamic changes are suggested.
MYL4	ALC-1		Frameshift-c.234delC79	Loss of function of MYL4.
NKX2.5	NKX2.5		(N19D, F186S, E21Q, T180A)81, 82	Reduction in the transcriptional activity of NKX2.5.
			F145S83	Decrease in the transcriptional activity of NKX2.5.
GATA4/5/6	GATA4/5/6		GATA4 – (M247T, A411V)85	May impact the transcriptional activity and the downstream genes.
			GATA5 - (Y138F, C210G)87	Remain to be addressed.
			GATA6 – (G469V, Q206P, Y265X)86	Reduction in the transcriptional activity.
GREM2	GREM2		Q76E91	Variant increases the inhibitory activity of GREM2 to BMP signaling.
PITX2	PITX2		Mechanism of pathogenesis 89	PITX2 has multiple gene targets.
SHOX2	SHOX2		G81E, H28399	The H283Q Variant represses BMP4 and ISL1.
KLF15	KLF15		LOF-K229*102	Variant creates a mutant KLF15 protein that delays the repolarization and prolongs the effective refractory period.
ZFHX3	ZFHX3		Variant rs7193343-T104	Association with AF.
TBX5	TBX5		LOF-P132S94	Reduction in the transcriptional activity of TBX5.

AF-associated genes including protein products, affected currents, identified AF variants and mechanisms of pathogenesis. LOF: Loss of function, GOF: Gain of function.

4.1. Ion Channels

4.1.1. Sodium Voltage-Gated Channel subunits

Variants in sodium voltage-gated channel α subunit (SCN5A) have been previously linked to several arrhythmogenic cardiac diseases, including dilated cardiomyopathy, long QT syndrome, Brugada syndrome, ventricular fibrillation and AF (summarized in Figure 1 and 2). In a prevalence study of variants conferring risk to AF, common and rare SCN5A variants made up approximately 6% of the cohort, supporting SCN5A as a candidate gene for AF development and maintenance.³⁰ Broadly, the loss-of-function (LOF) variants likely lead to a shortened APD, which can confer AF susceptibility. H558R is a LOF variant located in the intracytoplasmic linker, leading to reduced sodium current density that leads to decreased conduction velocity, promoting a re-entrant circuit.³¹ N1986K is located in the cytoplasmic domain and demonstrated a role in channel regulation by modulating steady-state inactivation.³² However, the gain-of-function (GOF) variants (such as R222Q, K1493R, M1851V) (Figure 2) pose a more interesting question, with researchers postulating that the prolongation of the atrial APD caused by SCN5A variants may disrupt repolarization and provide a substrate for pathogenesis.³³

Figure 2: Genetic variants in SCN5A, HCN4 and PITX2 transcriptional network.

A. Variants linked to Na_V1.5 structural domains. Highlighting key loss of function (red star) and gain of function variants (green circle). Gain of function (red star) variants identified in HCN4. B. Pitx2 regulates the effector genes involved in Ca²⁺ signaling, Na⁺ currents, K⁺ currents and cell-cell communication. Abbreviations include: Na_V1.5: Voltage gated sodium channel, HCN4: Hyperpolarization activated cyclic nucleotide gated potassium channel 4, Tbx5: T-box 5, Pitx2: Paired-like homeodomain transcription factor 2, SHOX2: Short-Stature Homeobox-2, BMP4: Bone Morphogenetic Protein 4, ISL1: Insulin gene enhancer protein ISL1 and AF: atrial fibrillation.

All five sodium voltage-gated channel β subunits (encoded by SCN1B to SCN4B including SCN1Bb) are expressed in the atria and have been implicated in susceptibility to AF.^34,35 In a genetic screening of patients with lone AF, two LOF variants in SCN1B and two LOF variants in SCN2B were identified independent of variants in SCN5A in highly conserved extracellular domains.³⁴ Variants in the SCN1B (R85H and D153N) showed reduced sodium current amplitudes, while variants in SCN2B (R28W and R28Q) only displayed reduced peak sodium current amplitude.³⁴ Variants in SCN1Bb are typically implicated in Brugada syndrome but have also been identified in patients with early-onset lone AF (R214Q).³⁶ LOF variants in SCN3B (R6K, L10P, and M161T) were reported to affect the sodium current by altering steady-state inactivation, reducing peak current density, or both.³⁵ Variants in SCN4B (V162G and I166L) were reported to affect the sodium current density and the voltage dependence of sodium channel activation or inactivation.³⁷ Overall, the reduced function of the sodium channel affects the initiation and the duration of the action potential, shortening the refractory period and slowing conduction, which creates a substrate for re-entry and provides a possible mechanism for AF.³⁴

4.1.2. Hyperpolarization Activated Cyclic Nucleotide Gated Potassium Channel 4 (HCN4)

Variants in HCN4 have been linked to sinus node dysfunction, and the link between AF and sinus node disease makes HCN4 an important target for genetic studies of AF. All variants in HCN4 found in genetic screening of patients are heterozygous, and the mechanism of most HCN4 mutations is by LOF that decreases electrical activity by suppressing the frequency of action potentials.³⁸ Another potential mechanism in humans was proposed by Macri et al. in the variant P257S, suggesting that the mutation leads to a loss of channel function.³⁹ HCN4 channels with a L573X variant had insensitivity to cAMP, and a K530N variant acted by shifting the activation curve to more negative voltages⁴⁰ (Figure 2). Finally, the HCN4 (P883R) variant was presented in patients with a common KCNE1 variant, which could contribute to the onset of AF (Figure 2).⁴¹

4.1.3. Potassium Channels

Potassium inwardly rectifying channel subfamily J member 2 (KCNJ2) is considered a causative gene of Andersen syndrome, however, it has been implicated in short QT syndrome and familial AF. A GOF variant (V93I) in a highly conserved region across species in Kir2.1 was identified in 30 Chinese AF kindreds. The substitution was found in all affected members and is implicated in initiation and maintenance of AF by increasing I_K1 channel activity.⁴² I_K1 overexpression has been shown to hyperpolarize the resting membrane potential which has been implicated in AF pathogenesis.⁴³

The potassium voltage-gated channel subfamily Q member 1/potassium voltage-gated channel subfamily E regulatory subunit 1-5 (KCNQ1/KCNE1-5) complex constitutes the I_Ks current. LOF variants in the I_KS channel subunits have been linked with cardiac conditions such as long QT syndrome while GOF variants in the I_KS channel have mainly been linked to AF. GOF variants in KCNQ1 have been reported to abbreviate the atrial APD and promote reentry. A missense variant S140G was identified in the S1 transmembrane segment of KCNQ1 near the extracellular surface of the plasma membrane. Patch clamping data revealed that S140G alone did not produce a substantial current, however, when coexpressed with KCNE1/2, there was a significant increase in the current density.⁴⁴ Instantaneous activation of the I_KS channel was also reported in a novel KCNQ1 missense variant (G229D) in AF patients less than 50 years old.⁴⁴ Additionally, KCNQ1 variants R14C, S209P, R231C, R231H occurring in the N-terminus, S3 and S4, respectively, have been linked to AF.^44-46 GOF variants in KCNE2 (M23L and I57T), KNCE4 (E145D), and KCNE5 (L65F) increase the I_KS channel activity creating a substrate for AF susceptibility.^47-49

KCNJ5 encodes the G-protein-coupled inward rectifier potassium channel subtype 4 protein (Kir3.4) which is necessary for the formation of the acetylcholine/adenosine-induced potassium current, I_KACh.⁵⁰ Variants in KCNJ5 have been identified in patients with long QT syndrome type (LQTS) 13 and have been linked to the incidence of AF.⁵⁰ Zhang et al. assessed KCNJ5 SNPs and their association with lone AF in a Chinese Han population, identifying Kir3.4 SNPs C171T, G810T, and C834T as potential targets.⁵¹ These variants have been linked to a reduction in the APD and the effective refractory period in atrial myocytes, which may contribute to AF initiation and maintenance.⁵¹ A heterozygous mutation in Kir3.4, G387R, was discovered in a large Chinese family with LQTS13. This variant resulted in the loss of I_KACh channel function by disrupting the trafficking of Kir3.4 subunits. Moreover, the presence of AF was more common in patients with LQTS than in the general population, although, the molecular mechanism is unclear. Physiologically, the variant causes a decrease in I_KACh channel current which could promote the initiation and maintenance of AF by APD prolongation, slowed conduction and early afterdepolarizations.⁵⁰

Potassium voltage-gated channel subfamily A member 5 (KCNA5) encodes for the poreforming α-subunit (Kv1.5) of the ultra-rapid delayed rectifier potassium channel (I_Kur). I_Kur is mainly localized in the atrial cells where it plays a vital role in the repolarization of the action potential, implicating the importance of mechanistic understanding of the variants in KCNA5 for AF treatment.⁵² Notably, both LOF and GOF variants in KCNA5 have been found to increase AF susceptibility. In 2009, Yang et al. reported three LOF variants in KCNA5 (T527M, A576V and E610K) that significantly prolonged the effective refractory period and enhanced early afterdepolarization predisposition.⁵³ Variant E375X has also been shown to disrupt I_Kur function, lengthening the APD.⁵⁴ Christophersen et al. identified the first GOF AF-associated KCNA5 variants (E48G, A305T and D322H) that increased I_Kur current and therefore shortened the APD.⁵⁵ The discovery of GOF variants in KCNA5 suggest that further understanding of I_Kur and its function in the atria needs to be explored.

4.1.4. Gap Junctions

Thibodeau et al. identified a frameshift variant within Gap junction protein alpha 1, GJA1 (c.932delC) in 10 unrelated healthy and young patients diagnosed with lone AF.⁵⁶ GJA1 has been identified as a possible locus of an AF susceptibility variant at SNP rs13216675.⁵⁷ The variant is atrial-tissue specific and presents as a mosaic variant in lymphocyte DNA, leading to unequal distribution of the variant protein that could result in difference in conduction velocities and promote electrical reentry.⁵⁶ Wang et al.⁵⁸ completed an association study of this GJA1 SNP in a Chinese Han GeneID population and found significant association of the SNP with AF, even after adjustment for age, gender, and comorbidities (diabetes, hypertension, and coronary artery disease). Within the Chinese Han population, it was identified that the GJA1 SNP explained 1.8% of AF heritability. This study suggests a common susceptibility of the GJA1 SNP in association with AF across this population.⁵⁸

Gap Junction Protein Alpha 5 (Gja5) knockout mice demonstrated increased susceptibility to atrial arrhythmias, leading to further studies of the implications of GJA5 variants in human AF.⁵⁹ Gollob et al. identified four heterozygous variants (P88S, M163V, G38D, and A96S) in highly conserved regions of the connexin-40 (Cx40) protein from human patient data. These variants resulted in abnormal gap-junction formation due to intracellular retention of Cx40 and weakened intercellular electrical coupling.⁶⁰ Regulatory elements of Gja5 (Shox2, Nkx2.5) have been identified as other potential targets of AF. Ablation of the enhancer activity reduces expression of endogenous Gja5.⁶¹ Early studies showed that Cx40 deficient mouse models led to prolonged P-wave duration, which could be due to delayed cellular depolarization.⁵⁹ A mouse model based on clinical data of the A96S mutation demonstrated shortened atrial refractory periods, decreased conduction velocity between gap junctions, and higher susceptibility to inducible AF⁶², suggesting that the conduction abnormalities propagated by the A96S mutation contribute to AF development.

4.2. Structural and Signaling

The natriuretic peptide A (NPPA) gene encodes the atrial natriuretic peptide (ANP), which acts as a cardiac hormone to modulate the cardiovascular system. Specifically, ANP infusion was demonstrated to reduce atrial conduction time and therefore delay refractory period, which could provide an electrical basis on the effect of ANP on the conduction system of the heart.⁶³ Natriuretic peptides (NP) perform their effects by binding to NP receptors, including NPR-1, NPR-B and NPR-C. NPR-C is significantly expressed in atria and plays a role in the regulation of atrial conduction and the susceptibility to AF.⁶⁴ Hodgson-Zingman et al.⁶⁵ identified the first familial AF-associated variant in NPPA, a heterozygous frameshift mutation (c.456-457delAA) in exon 3 that extends the reading frame and leads to a fusion protein of the normal peptide and a carboxyl terminus. The suggested mechanism of disease development is by an extended half-life of ANP, as longer carboxyl termini are related to increased resistance to degradation.⁶⁵ Other NPPA variants (A117V and S64R) lead to GOF effects on the I_KS current, which leads to shortened APD.^{66, 67} Studies in murine models detail the same results seen in clinical studies, where overexpression of an AF-linked Nppa mutation made mice more susceptible to AF, mediated by cardiac sodium, calcium, and potassium channels.⁶⁸ Cheng et al. developed a model of a missense I137T mutation that showed spontaneous AF and pacing-induced AF. The mutation led to the activation of cardiac inflammatory pathways by tumor necrosis factor-α (TNF-α), nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) and fibrosis.⁶⁹ Overall, GOF variants in NPPA are related to AF by electrical or inflammatory phenotypes that lead to familial or lone AF.

Nuclear integrity and transport have been implicated in AF pathogenesis through disease-associated variants in_Lamin A/C (LMNA) and Nucleoporin 155 (NUP155), which encode lamin A/C and nucleoporin 155, respectively. Lamin A/C is important in the structure of the inner nuclear membrane, and variants identified affect interactions with nucleoporin 155.⁷⁰ Brauch et al. performed a genetic screen to identify variants in LMNA that underlie AF without associated dilated cardiomyopathy (DCM), and implicated a T488P substitution as a possible pathogenic cause.⁷¹ Variants in NUP155 reduce nuclear envelope permeability by affecting the overall nuclear pore complex, which can lead to shortened APD.⁷² Overall, studies of possible variants in LMNA conclude the presence of some pathogenic role, but further studies are needed, and it is likely that AF develops alongside another disease like DCM. Factors downstream of NUP155 must be further studied to better holistically understand the mechanism by which variants may have a causal drive in AF pathogenesis.

Junctophilin 2 (JPH2) plays an important role in the structure of junctional membrane complexes that contain RyR2 and are responsible for SR Ca²⁺ handling. Knockdown of Jph2 in mouse models demonstrated reduced Ca²⁺ induced Ca²⁺ release and acute heart failure. Beavers et al. identified an E169K variant in a patient with juvenile-onset lone AF. E196K knock-in mice demonstrated decreased binding of Jph2 to Ryr2 and increased frequency of spontaneous Ca²⁺ release events. Reduced JPH2 levels therefore may destabilize RyR2 and promote abnormal SR Ca²⁺ release events, leading to AF development.⁷³

Using Finnish biobanks of genotype data, Clausen et al. predicted a rare Synaptopodin 2 Like (SYNP02L) LOF variant in AF development. SYNP02L encodes the cytoskeletal heart-enriched actin-associated protein (CHAP), where adult and embryonic cardiac tissues are mostly comprised of the isoform CHAPa and CHAPb respectively. Knockdown of these 2 Chap isoforms in zebrafish leads to disorganized sarcomeres as well as decreased cardiac contractility, suggesting that SNYP02L is linked to structural function of sarcomeres. Interestingly, both Myozenin 1 (MYOZ1) SYNP02L are located on chromosome 10q22, suggesting that either may affect AF pathogenesis, although MYOZ1 was previously identified as the causative gene.⁷⁴ Typically, MYOZ1 is involved in cardiac muscle calcineurin signaling to stabilize the sarcomere.⁷⁵ Overall, the 10q22 locus modulates sarcomere function that, when disrupted, causes higher incidence of AF.

TTN encodes Titin, the giant sarcomere protein that is expressed throughout the heart. TTN has been previously linked to ventricular cardiomyocyte structure and cardiomyopathies. Whole-exome sequencing of families affected by AF has identified a truncating TTN variant (TTNtv). Recapitulation of this variant in zebrafish models leads to a sarcomere defect that increased fibrosis and predisposed to arrhythmia.⁷⁶ Other rare variants associated with AF have been identified in genes affecting the cardiac structure including, Myosin Heavy Chain 7 (MYH7), Myosin Binding Protein C3 (MYBPC3), and Myosin Light Chain 4 (MYL4). MYH7 encodes a slower MyHC motor protein-β that is found in heart and in type I skeletal muscle fibers. MYH7 is increased in atrial tissues of patients with chronic AF as previously reported.^{77, 78} Atrial Light Chain-1 (ALC-1) is a protein encoded by MYL4 and is expressed in fetal and adult cardiac atrial tissue. A frameshift variant in MYL4 has been reported in association with a recessive form of AF.⁷⁹

4.3. Transcription Factors

Homeobox protein NKX2.5 (NKX2.5) is a cardiac-specific homeodomain transcription factor important in atrioventricular embryonic development. NKX2.5 continues to be expressed into adulthood. NK family members interact with GATA factors to activate target gene expression during development. NKX2.5 is essential for developing a boundary between the atria and SAN, as NKX2.5 acts to downregulate SAN function. Loss of Nkx2.5 leads to more pacemaker-like myocardial tissue, therefore increasing expression of Hcn4 but reducing levels of Cx40.⁸⁰ Genetic screening has revealed variants in NKX2.5 (N19D, E21Q, T180A, and F186S) that occur in the amino-terminal domain and the homeodomain, likely affecting the transcriptional ability of NKX2.5 by direct inhibition.^{81, 82} The E21Q variant was first identified in patients with congenital atrial septal defect, and later associated with AF.⁸² Decreased transcriptional activity of NKX2.5 was reported in an F145S variant occurring as autosomal dominant familial AF, likely leading to AF pathogenesis.⁸³ Ablation of NKX2.5 function may affect AF pathogenesis by disrupting the boundary between atrial and SAN function, affecting downstream genetic targets important to cardiogenesis.

The GATA Binding Protein (GATA) family encompasses six transcription factors, where GATA-4, −5, and −6 are expressed in the heart. Human variants in GATA4 are associated with congenital heart abnormalities, linking GATA expression to a regulatory role in cardiac development. It has been postulated that GATA4 and GATA5 co-regulate to act as direct regulators of SCN5A, which encodes the sodium channel Nav1.5, whose dysfunction has been implicated in cardiac arrhythmias, including AF.⁸⁴ Genetic screening of patients presenting with lone AF have resulted in the identification of several GATA4 variants including M247T, and A411V.⁸⁵ These variants were heterozygous, autosomal dominant, and linked to decreased transcriptional activity, having possible roles in decreasing levels of NKX2.5 and other target proteins, which could have further downstream effects on cardiac development and function or electrical activity. Variants in GATA5 and GATA6 are also implicated in AF, and lead to possible structural defects or changes in cardiac gene regulation that impact AF susceptibility.^{86, 87}

Paired-like homeodomain transcription factor 2 (PITX2) (Figure 2) belongs to the homeobox transcription factor family PITX, which is crucial for left-right asymmetry development of internal organs. PITX2c is the major cardiac isoform⁸⁸, and it has been shown in mouse models that Pitx2 is expressed in the left atrium and pulmonary vein.⁸⁹ It is postulated that Pitx2 inhibits the pacemaker gene program in the left atrium, as demonstrated by an increase in Shox2, Nppa, Kcnq1 and Hcn4 in response to Pitx2 deficiency.⁸⁹ A mouse model with a heterozygous Pitx2c mutation displayed shortened action potential duration, suggesting a potential electrical pathway for Pitx2 to affect the development of AF.⁸⁸ However, PITX2 overexpression was also shown to increase the activity of the KCNQ1 and KCNE1 potassium channels, strengthening the possibility that PITX2 variants could lead to AF by electrical remodeling through action potential alteration.⁹⁰ Gremlin 2 (GREM2) modulates bone morphogenetic protein (BMP) signaling as an upstream antagonist of PITX2. Muller et al. identified a GREM2 (Q76E) variant in a family presenting with AF.⁹¹ The Q76E variant showed an overactivity of GREM2 that caused reduction of cardiac contraction rate and interfered with contraction propagation in zebrafish atrial cardiomyocytes, contributing to a possible pathogenesis mechanism for AF.⁹¹

Several gene-regulatory networks have been linked to AF and specifically epigenetic and transcriptional networks have been reviewed in relation to AF.^{23, 92} Here, we provide an example of crosstalk between PITX2 and T-box 5 (TBX5). PITX2 and TBX5 are transcription factors that are correlated to AF susceptibility.^{23, 92} TBX5 belongs to the T-box family of transcription factors and variants in TBX-5 have been correlated with changes in PR interval, QRS duration, QT interval and AF.⁹³ A heterozygous missense variant P132S was identified in a patient presenting with lone AF, where the variant led to decreased transcriptional activity.⁹⁴ Tbx5 deficient mice showed APD prolongation, delayed afterdepolarizations, spontaneous diastolic depolarizations and other findings of atrial conduction dysfunction and arrhythmogenic triggers. Interestingly, their phenotype was reversed or rescued by reduced Pitx2c.⁹⁵ Howevith Ca²⁺ handling dysfunction ⁹⁶. Further, ETV1, which is a transcription factor that negatively regulates TBX5 is increased in AF patients.⁹⁷ Taken together, low levels of TBX5 or PITX2c can predispose to AF.²³

Short-Stature Homeobox-2 (SHOX2) is a homeodomain transcription factor important for cardiac development by aiding in SAN development, with possible downstream targets of Nkx2.5 and Hcn4.⁹⁸ Genetic screening of AF patients with prolonged PR intervals identified variants in SHOX2, G81E and H283Q (Figure 2). The H283Q variant represses Bone Morphogenetic Protein 4 (BMP4), which plays a role in cardiogenesis and Insulin gene enhancer protein ISL1 (ISL1) and initiates SAN development, therefore impairing function of SHOX2.⁹⁹ Homozygous deletions of Shox2 in mice are embryonically lethal due to severe cardiovascular defects and abnormal expression of Cx40 and Cx43, in addition to Nkx2.5, which aids in regulation of the connexins. Li et al. discovered a heterozygous nonsense mutation (R194X) that led to half of the homeobox being deleted, which could disrupt DNA sequence recognition and binding. The truncated protein failed to activate BMP4 and ISL1 promoters, which could be a molecular mechanism in which AF was induced by impairing SAN development and function.¹⁰⁰

Kruppel-like Factor 15 (KLF15) belongs to a family of Kruppel-like factors (KLF) that act as transcriptional regulators of DNA-binding. KLF15 is specifically expressed in myocytes and fibroblasts and plays a role in repressing hypertrophy and fibrosis. KLF15 modulates downstream targets such as ANP and BMP to inhibit hypertrophic activity.¹⁰¹ Li et al.¹⁰² identified a KLF15 variant (K229*) that creates a truncated KLF15 protein from a study of a large Chinese family affected with AF. This variant has complete penetrance and is transmitted in an autosomal dominant pattern. The affected individuals also presented with premature ventricular contraction, and a few presented with ventricular tachycardia and hypertrophic cardiomyopathy. It is hypothesized that the KLF15 loss-of-function variant contributes to AF by delaying the repolarization and prolonging the effective refractory period.¹⁰² The truncated variant likely delays the repolarization and prolongs the effective refractory period due to ablated transcriptional activity on the KChIP2 promoter and decreased the inhibitory effect of KLF15 on the connective tissue growth factor (CTGF).¹⁰² Typically, KChIP2 forms a transient outward potassium current (I_to) channel with K_v4.3. A deficiency of KChIP2 could cause an absence of I_to that would lead to prolonged action potential, therefore increasing susceptibility to arrhythmia.¹⁰² CTGF is a pro-fibrotic signaling molecule, and Klf15-null mice had increased CTGF levels that increased cardiac collagen deposition¹⁰³, making it likely that the KLF15 LOF variants lead to CTGF-induced atrial fibrosis and therefore AF.

Zinc Finger Homeobox 3 (ZFHX3) is a transcription factor on chromosome 16q22 containing multiple homeodomains and zinc finger sites first implicated in AF susceptibility by a GWAS.¹⁰⁴ ZFHX3 is involved in the regulation of many processes including tumor suppression and cell proliferation. Importantly, ZFXH3 is crucial in cardiac development by mediating myogenic differentiation. ZFHX3 inhibits STAT3, which regulates inflammatory processes, and increased expression of STAT3 has been linked to pacing-induced AF.¹⁰⁵ In cells derived from HL-1 myocytes, Zfhx3 knockdown increased Ca²⁺ leak, which has previously been linked to lone AF. The knockdown model demonstrated a shortened action potential duration that can facilitate the maintenance of re-entrant circuits in AF, which in combination with Ca²⁺leak could lead to the pathogenesis of AF.¹⁰⁶

5. Additional Risk Factors

There are several environmental factors predicted to increase AF susceptibility, with novel factors continuing to be identified in recent years. The use of the term ‘environmental’ is more inclusive than solely the extracorporeal environment, but encompassing any factor that is non-genetic in origin. Of these, there are several factors that have been well categorized: age, body weight, and a litany of cardiovascular comorbidities (such as hypertension, myocardial infarction, and structural cardiac abnormalities), but there has been an increasing amount of research into other factors, including moderate alcohol usage, temperature, frequency of smoke inhalation, and airborne particulate matter.

5.1. Age

Age is the most well-known risk factor independently associated with AF. Incidence and prevalence of AF increase dramatically with age, with an approximate 40% lifetime risk and 50% of AF patients aged 75 and older.¹⁰⁷ As a larger proportion of the population ages, AF prevalence will continue to rapidly increase in the future. The risk of developing AF doubles with each decade gained, ultimately resulting in a risk above 20% by 80.¹⁰⁸ Aging leads to structural and functional cardiovascular changes that may impact elderly predisposition to AF development.¹⁰⁹ Age is also associated with increased risk of complications from AF, with older patients having a higher incidence of stroke. Comparatively, older patients have a higher prevalence of persistent AF and comorbidities like HF, hypertension, and valvular heart disease.¹⁰⁸

5.2. Obesity

Obesity is defined as a BMI greater than 30 kg/m². Obesity is linked with lone AF, persistent AF, and reduced ablation success in AF patients. The pathophysiology has been described as multifactorial, but not fully elucidated.¹¹⁰ In a murine obese model, sodium, potassium, and calcium channel remodeling occurred, resulting in induced AF.¹¹¹ Obesity is an established risk factor for AF and is correlated with reduced efficacy of rhythm control therapies.¹¹² It was recently shown that obesity enhances the NLRP3 inflammasome in the atria and promotes AF development. Additionally, NLRP3 was shown more active in the atrial tissues of obese patients and sheep.¹¹³ A high-fat-diet in mice induced obesity associated with enhanced activity of NLRP3 that caused a pro-arrhythmic substrate for AF. Further, selective inhibition of NLRP3 was successful in halting the development of reentrant substrate and the abnormal Ca²⁺ release in obese mice. Overall, this data supports a role of NLRP3 inflammasome in the development of obesity induced AF.¹¹³ Obesity is also associated with hemodynamic changes that increase susceptibility to AF, including, an increase in total blood volume and a sustained increase in cardiac output, which causes left ventricular enlargement and hypertrophy in addition to left atrial enlargement associated with the increased BMI.¹¹⁰ Visceral fat and its secretome has been reported to have electrophysiological effects in terms of prolongation of the action potential of the left atria and increase in late sodium current and L-type calcium current in addition to the autonomic dysregulation.¹¹⁰ Additionally, an important pathophysiological factor of obesity in AF is the addition of epicardial adipose tissue and fibrosis.¹¹⁰ This fatty and fibro-fatty tissue has a paracrine function due to inflammation and fibrotic-related remodeling.^{110, 114}

5.3. Comorbidities

AF risk is significantly associated with other cardiovascular diseases, as substrates for disease pathogenesis can form from structural or electrical defects of other cardiovascular conditions. In a study of subjects 65 years and older, valvular heart disease, increased left atrial size, coronary artery disease, and systolic blood pressure were indicators of risk for AF development.¹¹⁵ Importantly, hypertension, myocardial infarction, and structural abnormalities confer strong independently associated risk of AF. Specifically, hypertension leads to left ventricular hypertrophy and arterial stiffening that ultimately results in atrial remodeling that increases risk of AF development.¹¹⁶ Myocardial infarction increases the risk of AF due to atrial pressure that leads to left atrial dilatation, with a recent study implicating that 12.5% of myocardial infarction patients develop AF.¹¹⁷ Inherited cardiomyopathies, caused by variants in genes encoding structural proteins and including hypertrophic cardiomyopathy, arrhythmogenic cardiomyopathy, familial dilated cardiomyopathy and left ventricular non-compaction cardiomyopathy, are associated with atrial remodeling and histological modifications, which predispose to AF.¹¹⁸ Finally, AF was strongly associated with both types of HF, HF with preserved ejection fraction and HF with reduced ejection fraction in the Framingham Heart study and notably, the development of HF among AF patients was associated with a 2 to 3-fold increase in mortality.⁴

The cardiac system is also target for endocrine regulation and endocrine dysfunction is linked to several cardiovascular diseases including AF.¹¹⁹ Disorders in hormones of the hypothalamus-pituitary axis including growth hormone, thyroid-stimulating hormone and others have been linked with AF development.¹¹⁹ Importantly, AF is prevalent in diabetes mellitus, which is a metabolic disorder characterized by insulin deficiency in Type-1 and insulin resistance in peripheral tissues in Type-2.^{120, 121} Persistent hyperglycemia in diabetes causes the production of glycation end products that can infiltrate the myocardium resulting in hypertrophy and fibrosis and renders the heart more conductive to AF.¹¹⁰ Calcitonin, which is secreted by the thyroid gland and plays a role in bone metabolism, was shown recently to be produced by atrial myocytes. Atrial specific knockdown of this hormone causes atrial fibrosis and promotes spontaneous AF and interestingly patients with persistent AF have lower levels of myocardial calcitonin compared to individuals with normal sinus rhythm.¹²²

5.4. Postoperative Atrial Fibrillation

Postoperative AF (POAF) is a serious concern, as it affects between 20-40% of patients after cardiac surgery^{123, 124} and between 10-20% of patients after non-cardiothoracic procedures.¹²⁴ Atrial arrhythmias are the most common rhythm disturbances postoperatively.¹²⁵Post-cardiac surgery AF is typically lone, brief, and asymptomatic, with onset typically occurring 2-4 days after surgery. These periods of AF may recur in the first week postoperatively and return to normal sinus rhythm is often spontaneous independent of intervention. POAF leads to increased morbidity, mortality and stroke risk, and specifically increases the risk of subsequent AF by eightfold ¹²⁴ However, risk for AF recurrence is about 20% lower in patients experiencing post cardiac surgery related AF as compared to non-thoracic surgery related AF.¹²⁶ Several mechanisms for POAF have been postulated, and relate to the multifactorial nature of AF in general.¹²⁴ Overall, β-blockers administered prophylactically have led to a protective factor, and therefore adrenergic responses are thought to be involved in POAF. Additionally, several inflammatory factors including IL-2, IL-6, and C-reactive protein have been linked to POAF.¹²⁴

POAF is exacerbated by existing atrial defects and remodeling, as triggers such as ectopic firing and re-entry promotes AF occurrence. In line with knowledge of adrenergic roles in POAF, drugs that increase sympathetic tone increase risk of POAF.¹²⁴ Age is a major risk factor in developing POAF, and the number of elderly patients undergoing cardiac surgery is expected to increase and may lead to increased incidence of POAF. Additional risk factors include hypertension, diabetes, and pre-existing cardiac disease. Hemodynamic mechanisms likely affect atrial remodeling, and diabetes may contribute to increased inflammatory markers that increase the risk of POAF.¹²⁵ AHA clinical guidelines call for the use of β-blockers perioperatively unless contraindicated, and drugs such as amiodarone and nondihydropyridine calcium channel blockers may be used in patients with contraindications. In addition, colchicine has been implicated in reducing incidence of POAF and decreasing hospital stay length.^{13, 127}

5.5. Alcohol

An important risk factor for AF is the relative consumption of alcohol. Recent research dedicated to relative quantity of alcohol consumed and AF initiation can reveal molecular mechanisms that lead to pathophysiological changes. At low doses of alcohol consumption, there is inconsistent data on the role of alcohol in AF development. In a recent European cohort study of 107,845 individuals of which 5854 developed AF, a Cox regression analysis of alcohol consumption was non-linearly and positively associated with AF with significance at one drink (12g/day).¹²⁸ In contrast to this, previous longitudinal studies have shown the lack of a significant increased risk on AF pathogenesis at this quantity of alcohol consumption.¹²⁹ Additionally, in a study of patients with AF before undergoing ablation, it was concluded that mild drinkers did not have the same electrical abnormalities that moderate drinkers possessed. Despite this, when compared to the lifelong nondrinker cohort, the mild drinkers possessed a significant increase in global complex potentials as well as some regional low-voltage zones in the septum and lateral wall.¹³⁰ In an attempt to understand the molecular mechanism that defines incident atrial fibrillation, Yan et al dosed rabbits with alcohol, simulating a binge drinking behavior. In conjunction with human heart samples with a history of binge drinking, it was concluded that binge alcohol activates JNK2, which in turn phosphorylates CAMKII. CAMKII activation intensifies SR Ca²⁺ mishandling and underlies the pro-arrhythmogenic mechanism of alcohol-induced AF.¹³¹

5.6. Smoking and Vaping

It has been shown that there is an association between AF and smoking.¹³² In a survey, it was seen that there is an association between self-reported current smoking and AF.¹³² Despite this, the study states that there was no association between those that report former smoking and AF. A previous study by Chamberlain et al. further stratified current and former smokers into sections at 800 cigarette-years of smoking.¹³³ This study design allows us to recognize how different amounts of smoking associate with AF. Not only did this study show that there is an association between AF and smoking, but that there is an increased risk of developing AF with increased smoking habits.

There is increasing evidence that vaping, or the use of electronic cigarettes, increases the risk of cardiac arrhythmias in addition to a myriad of other cardiac issues¹³⁴, predominantly through alterations in heart rate variability.¹³⁵ While there has been comparative research between traditional smoking and electronic cigarettes in the development of ventricular arrhythmias, there is less information concerning atrial arrhythmias.¹³⁶ Alarmingly, a recent case study identified a young, healthy male with an incident of lone AF development attributed to electronic cigarettes. ¹³⁷ According to a review by Benowitz and Fraiman, the use of electronic cigarettes provides a lesser risk of CVD than traditional cigarettes.¹³⁸ However, there has been more research into traditional smoking, while the toxic particulates ingested from electronic cigarettes have not been fully investigated. The negative effects of vaping and electronic cigarettes are apparent in their association with subsequent tobacco cigarette usage.¹³⁹

5.7. Particulate Matter and Air Quality

There has been increasing evidence in recent decades that air pollutants amplify risks for cardiovascular disease.¹⁴⁰ There are three types of particle pollutants: PM_0.1, PM_2.5 and PM₁₀. “PM_2.5” is the term used to define particulate matter 2.5 micrometers in diameter and smaller while “PM₁₀” refers to particulate matter that is 10 micrometers in diameter and smaller. The ultrafine fraction of PM (PM_0.1) is less than 0.1 μm in size. The World Health Organization (WHO) has listed environmental guidelines on the concentrations of PM_2.5 and PM₁₀ (25 μg/m³ 24-hour mean and 50 μg/m³ 24-hour mean, respectively).¹⁴¹ PM_2.5 exposure has been associated with AF onset in implantable cardioverter defibrillator (ICD) patients according to a recent meta-analysis of current literature.¹⁴⁰ In a continuation of these previous studies, Gallo et al. set out to determine if there is an association of longer-term PM_2.5 and PM₁₀ exposure and AF in the patients with implantable devices, concluding that there is an association between AF and PM_2.5 and PM₁₀ exposure when the exposure is above the WHO threshold for acceptable exposure of particulate matter in volume of air.¹⁴² This data suggests the need for environmental policy changes to lower the risk of AF.

5.8. Obstructive Sleep Apnea

Obstructive Sleep Apnea (OSA) is a disorder where a person stops breathing multiple times during sleep due to an obstructed airway from the tongue or the airway dilator muscles. OSA leads to increased atrial pressure, inflammation, hypoxia and hypercapnia which result in high blood pressure, cardiac dysfunction, oxidative stress, and cardiac electrical remodeling.¹⁴³ However, in a prospective study of 2912 people without a history of AF, Tung et al. reported no association between OSA and incident AF (defined by the obstructive apnea- hypopnea index), but rather there was only an association between central sleep apnea and incident AF.¹⁴⁴ Using a logistic regression model, the patients with central sleep apnea were 2- to 3-times more likely to develop AF. ¹⁴⁴ In a recent study, the epidemiology and the diagnostic accuracy of the commonly used screening tools for OSA were assessed in a hospital cohort of 107 patients with AF.¹⁴⁵ It was determined that OSA is frequently undiagnosed in the hospital cohort and that level 3 home sleep study device holds great promise in the diagnostic accuracy of patients with AF.¹⁴⁵ The continuous positive airway pressure ventilation (CPAP) is the current standard of care for OSA, However, Huang et al. proposed the need for novel therapies to improve the efficacy of controlling AF.¹⁴⁶ They postulated that neuromodulation could be a target for therapy, whether by beta-blockers, ablations of the carotid body or ganglionated plexi, or renal denervation.¹⁴⁶ Despite these studies, further research is still needed to clarify the association of OSA to AF as an independent risk factor.

5.9. COVID-19

The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) virus results in significant cardiovascular complications.¹⁴⁷ Coronavirus disease 2019 (COVID-19), as with many serious viral infections, causes myocardial inflammation, leading to both structural and electrophysiological remodeling that explain the resultant cardiac arrhythmias from the infection.¹⁴⁸ COVID-19 infection has been shown to result in AF in hospitalized patients, where severity of infection is implicated in AF susceptibility.¹⁴⁹ Despite this, studies have revealed that AF does not change the mortality rate in patient populations.¹⁵⁰ However, elevated troponin levels were found in COVID-19 patients with AF, suggesting cardiac damage.¹⁵⁰ A recent study found that AF occurred in 17.6% of patients where AF was an independent predictor of hospital mortality.¹⁵¹ Interestingly, it was also noted that there were geographical differences in the prevalence of AF in COVID-19 patients with lower prevalence in the Asia Pacific region, but this is consistent with previous analyses of epidemiology of AF.^{152, 153}

More investigations are needed to truly elucidate if there is a difference in mortality among COVID-19 patients with and without AF. Although there is a consensus that AF can result from COVID-19 infection, the pathophysiological complexity and multifactorial nature of the illness leaves the precise mechanism currently unknown. Several hypotheses were proposed to explain the pathogenesis of COVID-19 related AF¹⁵⁴, including a reduction in the angiotensin-converting enzyme 2 (ACE2) receptor availability, CD-147 or extracellular matrix metalloproteinase inducer and sialic acid-spike protein interaction. In addition, inflammatory cytokine storm, endothelial damage, acid-base imbalance in the acute illness phase and activation of the sympathetic nervous system by the infection were also considered¹⁵⁴ (Summarized in Figure 3). ACE-2 has been connected to heart function, hypertension and diabetes and notably also has been identified as a major entry point for the SARS virus.¹⁵⁵The viral coat of SARS-CoV-2 expresses a protein called SPIKE or S-protein that possesses a receptor-binding area that binds to the extracellular domain of ACE2.¹⁵⁶ Cleavage of the S-protein takes place by the transmembrane protease serine-2 (TMPRSS2) and the virus uses ACE2 to invade multiple cells of the host.¹⁵⁴ Cardiovascular tissues or cells that express ACE are at high risk for virus infection. ACE2 internalization causes a decrease of ACE2 at the cell surface and suppresses a key pathway for angiotensin II (AngII) degradation and generates cardioprotective Ang1-7. The increase in the overall ratio of AngII to Ang1-7 exacerbates myocardial hypertrophy, oxidative stress, fibrosis and dysfunction.^{156, 157} Loss of ACE2 was also shown to promote epicardial tissue inflammation¹⁵⁸, pericarditis ¹⁵⁹and pericardial effusion. Taken together, all these events may predispose to AF.¹⁵⁴

Figure 3: Diagrammatic illustration of COVID-19 related AF pathogenesis.

Infection initiated by SARS-CoV-2 causing endothelial damage, triggering ACE-2 signaling pathway, generation of a cytokine storm, electrolyte imbalance and activation of the sympathetic nervous system via the acute phase of the infection. Abbreviations include severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), angiotensin-converting enzyme 2 (ACE2), Interleukin-1 Beta (IL-1β), Interleukin 2 (IL-2), Interleukin 6 (IL-6), interferon gamma (IFN-γ), Interferon gamma-induced protein 10 (IP-10), Tumor necrosis factor (TNF-α), monocyte chemoattractant protein (MCP)-1, macrophage inflammatory protein (MIP-1α), Interleukin-18 (IL-18) and NACHT, LRR, and PYD domain containing protein 3 (NLRP3).

SARS-CoV-2 infection results in systemic inflammation and over activation of the immune cells resulting in a “cytokine storm” consisting of elevated levels of cytokines including IL-1β, IL-2, IL-6, interferon gamma (IFN-γ), interferon gamma-induced protein 10 (IP-10), TNF-α, monocyte chemoattractant protein (MCP)-1, macrophage inflammatory protein (MIP-1A) and other mediators.¹⁶⁰ These cytokines may cause apoptosis or necrosis of myocardial cells resulting in conduction disturbances.¹⁵⁴ Elevated levels of IL-6 were reported in non-COVID-19 survivors compared to survivors¹⁶¹ suggesting that increased mortality can be attributed to hyperinflammation. Notably, IL-6 has known proatherogenic effects including vascular smooth cell proliferation, platelet and endothelial cell activation. In addition, coronary atherosclerotic plaques are prone to rupture in case of hyperinflammation resulting in cardiac injury as was previously reported.¹⁶² Notably, SARS-CoV-2 has also been suggested in the activation of NLRP3 inflammasome.¹⁶³ Previous research has also suggested the link between NLRP3 inflammasome and the development of AF in atrial myocytes and NLRP3 activity was shown increased in patients with paroxysmal and chronic AF.²⁸ The mechanisms underlying the pro-arrhythmic effects of NLPR3 in atria include, the abnormal diastolic RyR2-mediated SR Ca²⁺ leak with the generation of delayed afterdepolarizations¹⁹, the amplified function of I_Kur along with re-entry causing shortening of the action potential duration and finally atrial remodeling.^{27, 154}

Another mechanism for AF in COVID-19 is the activation of cardiac immune response in response to cardiac inflammation and in particular of the CD4+ T cells that play a significant role in cardiac remodeling.¹⁶⁴ These cells were also shown to affect the infection clearance and prognosis in COVID-19.¹⁶⁵ Atrial stiffness is an independent risk factor and also a known mechanism for AF initiation. COVID-19 infection increases the stiffness of atrial and ventricular tissues as a result of cardiac leakage and/or pulmonary hypertension.^{165, 166} Pulmonary hypertension develops as a COVID-19 complication as a result of “pulmonary intravascular coagulopathy” caused by increased lung inflammation¹⁶⁷. Of note, AF is common in patients with pulmonary hypertension and it has been previously reported that pulmonary hypertension combined with increased heart rate further intensifies the risk of AF.¹⁶⁸ Interestingly, patients diagnosed with COVID-19 typically show an increased heart rate.¹⁶⁵ Therefore, increased AF susceptibility in COVID-19 patients could be attributed to pulmonary vascular dysfunction and atrial tissue stiffness.

In a study performed on 138 patients from Wuhan, who were hospitalized with COVID-19, arrhythmia was reported in ~17% of the general cohort and ~44% of patients admitted to the intensive care unit.¹⁶⁹ The most common pathologic arrhythmias in patients with COVID-19 include AF, atrial flutter and monomorphic or polymorphic ventricular tachycardia. In an observational study analyzing 160 patients hospitalized due to COVID-19, new-onset AF in COVID-19 patients was shown to have a notable effect on prognosis. New-onset AF during hospital stay has been shown an independent risk factor of in-hospital embolic events. Additionally, new-onset AF has been associated with an increase in hospital stay length and worse clinical features during hospitalization.¹⁷⁰ Management of AF in patients with COVD-19 is very challenging as AF is a common cause of embolism and stroke if not treated with anticoagulation therapy when indicated. COVID-19 by itself is a predisposing factor to thrombosis due to hyper-inflammation, endothelial dysfunction and other factors.¹⁷¹ While comparing the data from a large international patient registry including more than five thousand patients, 11% of patients experienced AF and 1.6% experienced atrial flutter during hospitalization. AF and atrial flutter were more reported in patients with pre-existing heart failure. New onset AF or atrial flutter was associated with increased mortality, but the increased risk was limited to males aged 60 to 72 years.¹⁷² Finally, there is a current need for greater retrospective analyses of cardiac health and arrhythmias in patients that have recovered from COVID-19.

6. Conclusions and Future Implications

An important aspect of AF management is preventative care, as several identifiable risk factors allow for early screening to reduce potential disease development. Risk factors include both non-modifiable factors such as age and race and modifiable factors such as body mass index (BMI), smoking, and hypertension.¹⁷³ Standardizing medical management of AF is unlikely, as each method of care is unpredictable depending on the pathophysiology and the symptoms of the individual. Consequently, advances in early clinical intervention are necessary to identify patients at risk for AF and prevent or manage the disease. Several studies have evaluated the impact of lifestyle and modifiable risk factor management of AF.¹¹² However, further work is still needed to improve patient education, to establish risk management programs, and to implant newer techniques to monitor the progress of risk management. Importantly more randomized studies are critical to evaluate the effect of risk management in AF patients.¹¹²

With the advent of new technologies, the ability to treat AF in a patient-specific way is pivotal to transforming a medical approach that prioritizes the individual affected and specifically works to combat their variation of disease pathology. One exciting potential for novel clinical screening is the rapid development of artificial intelligence¹⁷³, as the predictive model provides a unique approach to identifying individuals at risk of developing AF. Genetic studies aid in the progression of individualized medicine by identifying novel therapeutic targets. In specific consideration of AF, rare and monogenic variants and their mechanisms are most easily elucidated. The polygenic nature of most AF cases, combined with important environmental factors, presents a unique challenge in describing disease pathogenesis.¹⁷³ Based on clinical variant data, novel animal models can be developed to track the protein pathway and interactions to identify possible routes of AF pathogenesis, helping to elucidate mechanisms and potential therapeutic targets. Current studies of the I_KACh channel implicate genetic therapies as a potential mechanism to rescue an AF disease phenotype. Yoo et al. targeted atrial myocytes with a NOX2 shRNA inhibitor induced by injections into canine models. The study found that I_KACh channel inhibition by NOX2 decreased oxidative injury and prolonged atrial APD, although it had no effect on fibrosis and inflammation.¹⁷⁴ Continuing to understand novel gene treatments of AF can provide ways to target systematic disease prevention and mitigation.

Noninvasive optical mapping provides an approach to detect patient-specific sources and substrates of AF, and can inform subsequent therapeutic targets and techniques. Identifying the processes that lead to AF in its early stages would increase the effectiveness of already existing therapies and lead to the development of novel treatments. Advances made in accurately identifying AF drivers through the utilization of the adenosine challenge and Multi-Electrode Mapping (MEM) revealed critical AF drivers in ex vivo human heart studies.¹⁷⁵ In human clinical application of the adenosine challenge, researchers reported that there was a stabilization of driver patterns that lead to a more accurate site for ablation.¹⁷⁵ In addition to the adenosine challenge, the usage of machine learning models on Fourier spectrum features has increased the efficiency of differentiating AF drivers and non-drivers compared to standard mapping techniques.¹⁷⁶ The integration of MEM and 3D fibrosis analysis has also been reported to aid in the detection of AF drivers and non-drivers which can eliminate the false positives and increase the success of targeted ablation.¹⁷⁷

In summary, AF is defined by complex mechanisms linked by genetic and non-genetic risk factors that propagate disease development and maintenance. Current clinical diagnosis and treatment is informed by incomplete pathways, complexity of the condition and changes in symptom over life-course. Advancement in AF treatment is restrained by limitations of experimental models in comparison to human patients, lack of models that manifest spontaneous AF and translation from animal models to human patients.¹⁷⁸ Therefore, the continued study of genetic markers and non-genetic risk factors is paramount to discovering novel diagnostic and therapeutic targets. Clinical management that focuses on treatment of symptoms lacks a systematic way to prevent and/or manage AF.