Atrial Fibrillation: Epidemiology, Pathophysiology, and Clinical Outcomes (original) (raw)

. Author manuscript; available in PMC: 2018 Apr 28.

Abstract

The last three decades have been characterized by an exponential growth in knowledge and advances in the clinical treatment of atrial fibrillation (AF). It is now known that AF genesis requires a vulnerable atrial substrate and that the formation and composition of this substrate may vary depending on comorbid conditions, genetics, sex, and other factors. Population-based studies have identified numerous factors that modify the atrial substrate and increase AF susceptibility. To date, genetic studies have reported seventeen independent signals for AF at fourteen genomic regions. Studies have established that advanced age, male sex, and European ancestry are prominent AF risk factors. Other modifiable risk factors, including sedentary lifestyle, smoking, obesity, diabetes mellitus, obstructive sleep apnea, and elevated blood pressure predispose to AF and each factor has been shown to induce structural and electrical remodeling of the atria. Both heart failure and myocardial infarction increase risk of AF and vice versa creating a feed forward loop that increases mortality. Other cardiovascular outcomes attributed to AF, including stroke and thromboembolism, are well established and epidemiology studies have championed therapeutics that mitigate these adverse outcomes. However, the role of anticoagulation for preventing dementia attributed to AF is less established. Our review is a comprehensive examination of the epidemiological data associating unmodifiable and modifiable risk factors for AF and of the pathophysiological evidence supporting the mechanistic link between each risk factor and AF genesis. Our review also critically examines the epidemiological data on clinical outcomes attributed to AF and summarizes current evidence linking each outcome with AF.

Keywords: atrial fibrillation, epidemiology, risk factors, prognosis, pathophysiology

Introduction

The association of an irregular pulse and mitral stenosis was first described by Robert Adams in 1827, but it was not until the turn of the 20th century when William Einthoven invented electrocardiography that atrial fibrillation (AF) was first recorded on the electrocardiogram.1 Its pathogenesis and clinical importance gained enhanced appreciation in the 1990’s when early community-based studies including the Framingham Heart Study (FHS)2–4 provided critical epidemiological data on associated risk factors (RF) and clinical outcomes. These associations empowered scientists and clinicians by focusing their attention on specific disease models. Over the last 3 decades an explosion of research has yielded progress in the clinical treatment of AF at a time when AF is reaching epidemic proportions.

Our review provides an overview of the pathogenesis of non-valvular AF and a comprehensive examination of the epidemiological data associating various RFs with AF. For each RF we highlight key population studies supporting its association and critically review data on how the RF may lead to the development of the AF substrate and AF genesis. Lastly, we review clinical outcomes associated with AF and discuss possible mechanisms linking these associations. Our review focuses on the epidemiology and pathophysiology of AF rather than its clinical treatment.

1. Pathophysiology and Natural History AF

AF is characterized by high frequency excitation of the atrium that results in both dyssynchronous atrial contraction and irregularity of ventricular excitation. Whereas AF may occur in the absence of known structural or electrophysiological abnormalities, epidemiological association studies are increasingly identifying comorbid conditions, many of which have been shown to cause structural and histopathological changes that form a unique AF substrate or atrial cardiomyopathy.5

1.1 AF Initiation: Ectopic Firing

The prevailing hypothesis of AF genesis is that rapid triggering initiates propagating reentrant waves in a vulnerable atrial substrate. The relative importance of the initiating trigger may decrease as the AF substrate progresses and AF becomes more stabilized. Haissaguerre and colleagues first identified focal ectopic firing arising from myocyte sleeves within the pulmonary veins (PV) in patients with paroxysmal AF; ablation of these ectopic foci reduced AF burden, demonstrating their role in AF genesis.6 (Figure 1A) It is now known that the PVs have unique electrical properties and a complex fiber architecture that promote reentry and ectopic activity to initiate AF.7 Autopsy studies have identified pacemaker cells, transitional cells, and Purkinje cells within the PVs.8 The molecular basis for PV triggers has been primarily attributed to abnormal calcium Ca2+ handling. A diastolic leak of Ca2+ from the sarcoplasmic reticulum (SR) activates an inward Na+ current via Na+Ca2+ exchanger resulting in spontaneous myocyte depolarization (early or delayed afterdepolarization). Hyperphosphorylation of key regulatory proteins and enzymes including protein kinase A, calmodulin kinase II, phospholamban, and the ryanodine receptor type 2 are important contributors to SR Ca2+ overload and diastolic membrane instability.9, 10 A reentrant mechanism for PV triggers also has been described. Decremental conduction and repolarization heterogeneity within the PV enable localized reentry and may foster a focal driver for AF.11

Figure 1.

Figure 1

Rendering of left and right atrium showing various mechanisms of AF. A. Focal trigger arising from muscle sleeve of PV propagating into left atrium and initiating AF in the vulnerable substrate. B. Fixed or moving spiral rotor, a result of functional reentry, acts as a driver for AF. C. Circus movement around anatomic structures or scar generating micro and macro reentrant circuits. D. Perpetual propagation of multiple simultaneous wavelets mediated by both functional and structural reentry. E. Point source with fibrillatory conduction acting as driver for persistence of AF. F. Electrical dissociation between myocardial layers enabling reentry in three-dimensional construct. Abbreviations: SVC, superior vena cava; IVC, inferior vena cava; LAA, left atrial appendage; LSPV, left superior; RIPV, right inferior pulmonary vein; RSPV, right superior pulmonary vein; LIPV, left inferior pulmonary vein; CS, coronary sinus; PG, parasympathetic ganglia (yellow).

1.2 AF Perpetuation: Reentry

Whereas triggers are required for AF initiation, a vulnerable atrial substrate is equally important. Structural, architectural, and electrophysiological atrial abnormalities promote the perpetuation of AF by stabilizing reentry. The mechanism of reentry in AF remains controversial with two dominant hypotheses including reentrant rotors12, 13 or multiple independent wavelets.14 (Figure 1B–E) Advances in electroanatomic mapping and ablation technologies have yielded increasing evidence supporting the former mechanism.15, 16 Recent data supporting a third hypothesis, the double layer hypothesis, suggests that electrical dissociation of epicardial and endocardial layers also may facilitate reentry.17, 18 (Figure 1D)

For perpetuation of functional reentry, the propagating wavefront must complete one circus movement in a time period long enough for atrial tissue within that circuit to recover excitability (refractory period, ERP). Thus, slow conduction velocity and a short ERP promote reentry. Both reduce wavelength size increasing the likelihood of multiple simultaneous reentrant circuits and AF perpetuation.

Atrial substrates that promote reentry are characterized by abnormalities of the atrial cardiomyocyte, fibrotic changes, and alterations in the interstitial matrix with primarily non-collagen deposits.5 These molecular and histologic changes impair normal anisotropic conduction (fibrosis and/or reduced cell coupling) and may shorten atrial ERP. For example, in familial AF, congenital abnormalities that lead to a gain in K+ channel function shorten ERP of atrial cardiomyocytes; whereas in heart failure (HF) a combination of atrial fibrosis and alterations in cardiomyocyte function result in both a slowing of conduction velocity and shortening of ERP. Thus, the development of and characterization of the vulnerable atrial substrate is specific to the predisposing AF RF.

1.3 Natural History

For decades the prevailing notion was that AF began with paroxysmal episodes that increased in frequency and duration causing progression to more persistent AF subtypes. This so-called “AF begets AF” postulate was based on early experimental data in goats showing that tachycardia induces electrophysiologic atrial remodeling resulting in persistence of AF.19 Regional heterogeneity and shortening of the atrial ERP20, 21 occurs within 30 minutes of tachycardia onset and is a result of adaptation to intracellular Ca2+ overload.22 In the FHS, only 10% of participants remained free of AF 2-years after incident AF and recurrent (26%) or sustained AF (34%) was common.23 But other studies suggest that this abiding notion is not ubiquitous. In the Canadian Registry of Atrial Fibrillation (CARAF) progression of paroxysmal AF to more persistent (chronic) AF subtype was 8.6% at 1-year and 24.7% by 5 years.24 The Euro Heart Survey followed 5,333 patients with AF for one year and found that 80% of patients with paroxysmal AF remained paroxysmal while 30% of patients with persistent AF progressed to permanent.25 Studies in patients with pacemakers allow for more robust assessment of AF burden and have shown that the majority of patients (54–76%) with paroxysmal AF remain paroxysmal.26, 27 One study showed that only 24% of patients with paroxysmal AF progressed to persistent AF in one year and that there was a progressive pattern of increasing arrhythmia burden in these patients except in the days prior to the development of persistent AF supporting the mechanism of tachy-mediated atrial remodeling.27

The most remarkable observation is that persistent AF may spontaneously switch to paroxysmal subtype;26 highlighting the complex natural history of AF, the limitations of experimental data, and the existing uncertainty in the mechanisms and factors that govern the clinical course of AF. In addition, the natural history of AF may change over time as the RFs contributing to AF onset shift in prevalence and severity (e.g. less smoking and lower blood pressures, higher prevalence of obesity), and primary (e.g., better hypertension control) and secondary (anticoagulation) prevention treatments evolve.28 Finally, the means for quantifying and assessing AF burden over time in various population studies is inconsistent leading to ascertainment bias and challenges in predicting the natural history of AF subtypes.

2 Risk Factors for Developing AF

2.1 Unmodifiable Risk Factors

2.1.1 Genetics

Epidemiology

Near the start of the millennium rare familial forms of AF were identified and loci were mapped to 10q22-24,29 6q14-16,30 and 11p15-5.31 Subsequent population-based studies showed that family history of AF is associated with a 40% increased risk of first-degree relatives developing AF.32, 33 The recognition of the heritability of AF in the general population has propelled the search for associated genetic loci.34

Classic Mendelian genetics and candidate gene approaches have been used to define the familial basis of AF. To date, at least fifteen AF-causing mutations in K+ channel genes or accessory subunit have been identified,34 including mutations in ABCC9 (IKATP), HCN4 (If), KCNA5 (IKur), KCND3 (IKs), KCNE1 (IKs), KCNE2 (IKs), KCNE3 (IKs), KCNE4 (IKs), KCNE5 (IKs), KCNH2 (IKr), KCNJ2 (IK1), KCNJ5 (IKAch), KCNJ8 (IKATP), KCNN3 (IAHP), and KCNQ1 (IKs). Gain-of-function mutations increase repolarizing K+ current shortening the action potential duration (APD) and atrial refractoriness. Loss-of function mutations delay repolarization and promote Ca2+ mediated after-depolarization that trigger AF.35, 36 Six variations in Na+ channel genes have been identified and these include SCN1B, SCN2B, SCN3B, SCN4B, SCN5A, and SCN10A.34 Gain-of-function mutations may increase AF vulnerability by increasing cellular hyperexcitability,37 whereas loss-of function mutations shorten ERP and slow conduction.38 Other important AF-causing genetic variants include mutations in the gap junctional protein-coding gene GJA5 that diminishes cell-cell coupling and promotes reentry by slowing conduction velocity and shortening the wavelength.39

Instead of isolating a specific AF-causing gene, genome-wide association studies (GWAS) seek to scan the entire genome for disease-related genetic variants in single-nucleotide polymorphisms (SNPs). The first GWAS for AF identified SNP rs2200733 located on chromosome 4q25 upstream of PITX2 in an Icelandic population, which was strongly replicated in samples from Sweden, United States, and Hong Kong.40 A meta-analysis of three AF susceptibility loci (4q25, 1q21, 16q22) showed that the chromosome 4q25 SNP rs2200733 was associated with 30% increased risk of recurrent atrial tachycardia following AF ablation.41 In a separate meta-analysis this locus was associated with 38% increased risk of cardioembolic stroke.42

In experimental mouse models, the Pitx2 gene, encodes a transcription factor important for the embryonic organogenesis of the asymmetric organs placed in the left side including the heart.43 Moreover, Pitx2c is involved in the formation of the PV muscle sleeves,44 the most common site of triggered activity in AF. Loss of function of Pitx2c gene plays a role in the development of AF, and the differentiation, proliferation, and expansion of pulmonary myocardial cells.45, 46

In separate GWAS, the SNP rs2106261 near gene ZFHX3 identified on locus 16q22 was associated with increased AF risk. A knockdown of ZFHX3 dysregulates Ca2+, shortens the APD, and promotes arrhythmia susceptibility.47

AFGen consortium was formed in 2008 to increase statistical power for identifying loci associated with AF.48 To date 17 independent susceptibility signals for AF at 14 genomic regions have been identified. These include KCNN3, PRRX1, CAV1, SYNE2, C9orf3, HCN4, and MYOZ1.49, 50 The identification of genes related to AF is still in an early stage but in the future may allow for the assessment of an individuals’ AF risk and the discovery of novel therapeutic targets.

2.1.2 Age

Epidemiology

Advancing age is the most prominent RF for AF.28 Understanding the influence of age on AF risk is important for assessing how changes in life expectancy will affect the prevalence of AF.

Though the prevalence of AF varies among different ethnic populations, epidemiological studies have consistently found a step-wise increase in the prevalence of AF with advancing age.51–53 For example, a population-based multicenter cohort study reported age-specific rates in individuals aged 65–74 and 75–84 years of 3.4 (95% confidence interval (CI), 1.4–7.0) and 8.6 (95% CI, 4.6–14.9) for Chinese, 4.9 (95% CI, 3.1–7.3) and 10.6 (95% CI, 7.2–15.1) for non-Hispanic blacks, 7.3 (95% CI, 4.7–10.7) and 9.4 (95% CI, 5.9–14.4) for Hispanics, and 13.4 (95% CI, 10.6–16.7) and 19.6 (95% CI, 15.6–24.3) for non-Hispanic whites, respectively.51

The incidence of AF also increases with advancing age. In a Scottish study the incidence rates per 1000 persons was 0.5 for age 45–54 years, 1.1 for age 55–64, 3.2 for age 65–74, 6.2 for age 75–84, and 7.7 for age ≥85 years.53

The FHS examined temporal trends in AF RFs. During the last 5 decades, age was observed to be the greatest RF for AF as compared with other RFs including male sex, body mass index (BMI), diabetes, smoking, alcohol consumption, systolic blood pressure, hypertension treatment, left ventricular hypertrophy, heart murmur, HF, and myocardial infarction.28 In the most recent time-period studied (1998–2007), participant age of 60–69, 70–79, and 80–89 years was associated with 4.98, 7.35, and 9.33 fold risk of AF, respectively, compared with individuals aged 50–59 years.28 Accordingly, a step-wise increment in age has been incorporated in AF risk prediction scores.54, 55 Among AF patients, those with age <65 years also have been found to be healthier, have a different AF RF-profile, and less in-hospital deaths compared with those age ≥65 years.56

2.1.2 Sex Differences

Epidemiology

There is now greater recognition that epidemiology of AF differs between men and women.57 The age-adjusted incidence of AF is higher in men compared to women in North American and European populations. In the FHS, the AF incidence (per 1,000 person-years) was 3.8 in men and 1.6 in women.28 The Olmsted County Minnesota Study58 and the Rotterdam Study59 reported the AF incidence (per 1,000 person-years) in men to be 4.7 and 11.5, respectively, compared to 2.7 and 8.9 in women. A higher incidence of AF in men also is observed in Asian populations; although less data are available.60, 61 Similarly, the age-adjusted prevalence of AF is higher in men than in women in North American and European populations. Among the Medicare beneficiaries of adults aged ≥65 years, the prevalence was 10.3% in men and 7.4% in women.62 The higher prevalence of AF in men is also observed globally in both high-income and low and middle-income countries.60 However, there are less consistent data in Asian countries with some studies showing higher prevalence in men;61, 63, 64 whereas others showed no sex difference.52, 65, 66

In North American and European populations, the lifetime risk for AF is similar between men and women despite higher AF incidence in men owing to their shorter life expectancy.57 The FHS reported the lifetime risks to develop AF at age 40 years were 26% for men and 23% for women.67 In the Rotterdam study, the lifetime risks at age 55 years were 23.8% for men and 22.2% for women.59 On the contrary, among Chinese adults, the lifetime risk was consistently lower in men compared to women across all age groups.52

Underlying RFs of AF and taller stature in men largely explain higher AF incidence in men.57 The CHARGE-AF Consortium reported that after adjusting for AF-related RFs, male sex was no longer an independent RF for AF.54 The population-attributable risks of the RFs for AF differ by sex.57 The population-attributable risk for AF of coronary disease is higher in men68 whereas the population-attributable risks of elevated systolic blood pressure57 and valvular disease68 are higher in women.

2.1.3 Racial Differences

Many of the early population-based studies in AF were limited by racial diversity. However, in the last decade there has been a determined initiative to better understand the racial and ethnic differences in AF prevalence, pathophysiology, and outcomes.

Numerous studies demonstrated that AF is less prevalent in individuals of African as compared with European ancestry.51, 69–73 These data appear counterintuitive given the higher prevalence of traditional AF RFs in African Americans (AAs) than whites.69,74 The Candidate-Gene Association Resource Study69 found that among AAs the risk of AF was independently associated with increasing percentage of European ancestry (hazard ratio (HR), 1.13; 95% CI, 1.03–1.23). Adjusted associations showed that with every 10% increase in European ancestry there was a 16%–20% increased risk of AF. The data are consistent with the hypothesis that either African ancestry has protective effect against AF or European ancestry enhances AF susceptibility.

Candidate SNP analysis performed in two cohorts, showed that the rs10824026 SNP located on the 10q22 genetic locus accounted for 11.4%–31.7% increased AF risk in white individuals as compared with AAs.75 Furthermore, the minor allele G of the SNP confers low AF risk49 and is more common in AAs (37.7–37.8%) as compared with white individuals (15.5–16.4%).75 In a separate study that combined 3 cohorts with European and AA descent, an intronic SNP rs4845625 of the IL6R gene was associated with AF in white individuals (relative risk, 0.9; 95% CI, 0.85–0.95) and in AAs (relative risk, 0.86; 95% CI, 0.72–1.03 with borderline significance p = 0.09).76 Finally, the SNP rs4611994 on chromosome 4 near PITX2 was associated with AF risk in AAs (HR, 1.4; 95% CI, 1.16–1.69) and this chromosomal locus is associated with AF in white individuals.40 Although traditional AF RFs are well recognized, there is increasing data that race, ethnicity, and attributed genetic ancestral variants may play a significant role in modulating AF susceptibility.

More recently the Multi-Ethnic Study of Atherosclerosis (MESA) reported the prevalence of AF in Hispanic and Asians residing in the United States. The age- and sex-adjusted incidence rates per 1000 person-years of AF were 6.1 in Hispanics and 3.9 in Asians as compared with 11.2 in white and 5.8 in AA individuals.51 These data are comparable to The Healthcare Cost and Utilization Project that reported that Hispanics and Asians had multivariable-adjusted lower risk of AF (HR, 0.78; 95% CI, 0.77–0.79) as compared with whites.74 Both studies showed that a larger proportion of AF in non-whites was attributed to the greater presence of traditional RFs particularly hypertension as compared with whites.

2.2 Modifiable Risk Factors

2.2.1 Physical Activity and Sedentary Lifestyle

Epidemiology

The relationship between level of physical activity and risk of AF has been described as nonlinear.77–79 Sedentary lifestyle is associated with higher risk of AF,80 but paradoxically extreme levels of physical activity also are associated with increased AF risk.81–83

A large retrospective cohort study of 64,561 patients showed a graded and inverse relationship between cardiorespiratory fitness (CRF) as objectively assessed with treadmill testing. The incidence rate of AF in patients with lowest CRF was 18.8% as compared to 3.7% in those with highest CRF. Every 1-MET increase in CRF was associated with a dose-dependent 7% reduction in AF risk.84 A meta-analysis of pooled data from seven studies showed that sedentary lifestyle was associated with increased risk of AF (odds ratio (OR), 2.47; 95%CI, 1.25–3.7) compared to moderate or intense physical activity.83

Interesting and seemingly contradictory, male endurance athletes are at increased AF risk. In a prospective case-control study, individuals who had performed <2000 hours of lifetime-accumulated high-intensity exercise had an attenuated risk of lone AF (OR, 0.38; 95%CI, 0.12–0.98) as compared with sedentary individuals. But the AF risk in those with ≥2000 hours of high-intensity exercise was increased (OR, 3.88; 95%CI, 1.55–9.73).77 The type of exercise modulates AF risk with endurance sports, such as marathon running77 or long distance cross-country skiing,81 conferring the highest AF risk. A meta-analysis showed that there was over a 5-fold increased risk of AF in athletes than non-athlete referents.85

Sex-differences in the association of physical activity with AF have been identified. In men a U-shaped association of AF risk and physical activity was observed where moderate physical activity was found to confer lowest risk (OR, 0.81; 95% CI, 0.26–1.00) and intense activity conferred the highest risk (OR, 3.30; 95% CI, 1.97–4.63). In women, however, the association was inverse and linear. Increasing physical activity was associated with progressive and decreasing AF risk with ORs of 0.91 (95% CI, 0.77–0.97) and 0.72 (95% CI, 0.57–0.88) for moderate and intense activity, respectively.83

Pathophysiology

Sedentary lifestyle is known to increase risk of AF RFs including hypertension,86 obesity,86 and diabetes87. Obstructive sleep apnea (OSA) is common in obese individuals and has been associated with sedentary lifestyle.88 These conditions have been shown to independently induce structural and electrical remodeling of the atrium. Physical inactivity also increases systemic inflammation,89 which may induce atrial remodeling and has been associated with AF.90–94 Finally, sedentary lifestyle is associated with autonomic dysfunction and elevated sympathetic tone, which enhances afterdepolarization triggering and AF susceptibility.

In endurance athletes, the pathogenesis of AF has been attributed to two primary mechanisms. First, increased vagal tone in these individuals95, 96 may shorten and increase the dispersion of atrial ERP promoting PV firing and localized reentry. Second, long-term endurance training causes progressive cardiac remodeling including left atrial enlargement, which may promote AF.97–99 Atrial fibrosis and increased AF susceptibility has been observed in a rat model of prolonged, intensive exercise.100

2.2.2 Smoking

Epidemiology

Smoking is associated with incident AF.54, 101 The FHS showed that within the last 50 years, the frequency of smoking among participants with new-onset AF has decreased. Between 1998–2007, only 12.7% of AF-affected participants were smokers as compared to 15.6% in the prior decade.28

The Rotterdam Study found that both former and current smoking were equally associated with increased AF risk.102 In the Atherosclerosis Risk in Communities Study (ARIC) study, the multivariable-adjusted incidence of AF was 1.58 times higher in ever smokers (former and current) and two-fold higher (HR 2.05) in current smokers as compared with non-smokers.101 A dose-response association was observed with increasing cigarette-years.101 In the CHARGE-AF consortium, incident AF was 1.44 times higher in current smokers as compared with nonsmokers.54 Smoking is a RF for AF across various races and ethnicities.54, 101, 103

Finally, secondhand tobacco exposure also has been associated with risk of AF.104 Exposure during gestational development or early childhood is associated with approximately 40% increased risk of AF.105

Pathophysiology

Smoking is thought to increase AF susceptibility through indirect and direct mechanisms. Smoking may increase myocardial ischemia by increasing systemic catecholamine and myocardial work, reducing oxygen carrying capacity, and promoting coronary vasoconstriction.106 In addition, smoking accelerates atherosclerosis through effects on lipids, endothelial function, oxidative stress, inflammation, and thrombosis.107 These effects may indirectly increase AF susceptibility by predisposing to atrial ischemia, myocardial infarction, and HF. Reduced lung function and chronic obstructive pulmonary disease also increase vulnerability to AF.108

Smoking and nicotine also have been shown to directly contribute to the AF substrate. In a case-control study of patients undergoing coronary artery bypass, the volume of atrial fibrosis in smokers was shown to be dose-dependent, and in non-smokers nicotine was shown to induce a pattern of collagen type III expression in atrial tissue culture that was similar to that observed in smokers.109 In a dog model, nicotine induced interstitial fibrosis and increased AF susceptibility.110 The profibrotic effect of nicotine was attributed to down regulation of atrial micro RNAs, miR-133, and miR-590, which in turn increased transforming growth factors (TGF-b1 and TGFb2) and connective tissue growth factor. Finally, nicotine prolongs the APD by blocking the inward rectifier potassium channel (Ik1 and Kirs)111 and reduces the transient outward potassium current (Ito).112 Prolongation in the APD may increase arrhythmia susceptibility, but the pro-arrhythmic effect of nicotine has not been confirmed.

2.2.3 Obesity

Epidemiology

Both obesity and elevated BMI predispose to established AF RFs including hypertension,113 diabetes mellitus,114 myocardial infarction,115 left ventricular hypertrophy,116 left atrial enlargement,117 left ventricular diastolic dysfunction,118 HF,119 and OSA.2 However, when accounting for these concomitant conditions, population studies show that obesity and elevated BMI independently increase risk for AF.

Numerous population-based studies have shown an association between elevated BMI and increased risk of AF.120–122 A meta-analyses of 5 studies found that obesity confers a 49% increased risk of developing AF.123 A dose-response relationship was observed with each 1 unit increase in BMI associated with a 3%–4.7% increase in AF risk.120, 124, 125 Other measures of obesity including abdominal circumference and total fat mass have been associated with 13%–16% increase in AF risk per 1 SD over 10-year follow-up.126 Importantly, obesity is a powerful predictor of incident AF even when regression analyses have been adjusted for OSA, which commonly co-exists in such patients.125

Pathophysiology

Excess AF risk associated with obesity has been attributed to left atrial enlargement,120 increased left ventricular mass,116 and diastolic dysfunction.118, 127, 128 In a sheep model of obesity, increasing weight correlated with increased left atrial volume and pressure, ventricular mass, and pericardial fat. Histologic analysis revealed that myocardial lipidosis, fibrosis, and inflammatory infiltrates increased progressively with increasing weight. These pathologic changes were associated with decreased conduction velocity and increased AF.129 Whereas no change in atrial ERP was observed in the sheep model, a clinical study of 63 patients undergoing PV isolation reported that elevated BMI was associated with short atrial ERP and slower atrial conduction velocity,130 properties that promote reentry.

Pericardial fat also has been implicated in the pathogenesis of the obesity-AF relation. Cross-sectional studies have shown that pericardial fat is associated with prevalence, severity, and recurrence of AF.131, 132 A recent meta-analysis of 23 studies correlated epicardial adipose tissue (EAT) with AF after adjusting for traditional RFs (OR, 1.47 per standard deviation EAT increase; 95% CI, 1.17–1.84).133 A local paracrine effect of EAT mediated by inflammatory cytokines,134 growth and remodeling factors,135 angiogenic factors, and adipocytokines may lead to the development of the AF substrate.135, 136 EAT location on CT imaging correlates with high dominant excitation frequency during electoanatomic mapping in patients undergoing AF ablation.137

2.2.4 Diabetes Mellitus

Epidemiology

The FHS showed that men and women with diabetes had a 40% and 60% increased risk of AF, respectively.2 Level of blood glucose may be more predictive than actual diagnosis of diabetes in older adults.138 A meta-analysis of cohort and case-control studies found that patients with diabetes or impaired glucose homeostasis had a 34% greater risk of AF than individuals without diabetes.139 A causal association is supported by evidence that worse glycemic control and longer duration of diabetes are associated with increased AF risk.140 The estimated risk of AF increases by 3% per additional year of diabetes. The risk of AF in patients with diabetes for >10 years was 64% but only 7% in those with diabetes ≤5 years.

Pathophysiology

Glucose intolerance and insulin resistance appear to mediate the development of the AF substrate.141 The molecular mechanism by which insulin resistance alters cardiac structure is complex and involves impaired mitochondrial function and oxidative stress, which alter the transcription and translation processes essential for cardiac adaptation.142,143 In a rat model of diabetes, prolonged intra-atrial conduction time and diffuse interstitial fibrosis was observed, predisposing to increased arrhythmogenicity.144 In patients undergoing AF ablation, abnormal glucose metabolism was associated with prolonged atrial activation time and reduced bipolar voltages with electroanatomic mapping, a finding consistent with atrial fibrosis or scar.145 Finally, autonomic dysfunction also has been implicated.146

2.2.5 Obstructive Sleep Apnea

Epidemiology

OSA is highly prevalent147 and has been associated with other AF RFs including hypertension, diabetes, coronary heart disease, myocardial infarction, and HF.148 The Sleep Heart Health Study found a 4-fold increase in the prevalence of AF with OSA and one third of participants had arrhythmia during sleep.149 The Olmsted County Study similarly found that OSA and its severity strongly predicted 5-year incidence of AF (HR, 2.18; 95% CI, 1.34–3.54). In older individuals only the magnitude of nocturnal oxygen desaturation was predictive of AF.150 Similarly, a meta-analysis of five of prospective studies reported that OSA was associated with about a two-fold increased odds of post-operative AF.151 Patients with OSA have a higher recurrence of AF after cardioversion152 and catheter ablation (relative risk, 1.25; 95% CI, 1.08–1.45).153

The impact of OSA on AF outcomes was studied in the ORBIT-AF registry.154 Patients with OSA had more severe symptoms and were at higher risk of hospitalization (HR, 1.12; 95% CI, 1.03–1.22) than those without OSA, but had similar mortality, risk of stroke, or myocardial infarction. Patients with OSA who were treated with CPAP were less likely to progress to permanent AF subtype than those who were untreated (HR, 0.66; 95% CI, 0.46–0.94).

Pathophysiology

Electroanatomic mapping in patients undergoing AF ablation has been used to characterize the AF substrate associated with OSA.155 Observed structural changes included increased atrial size and expansive areas of low voltage or electrical silence, which indicate fibrosis, loss of atrial myocardium, or electrical uncoupling. Prolonged and regional disparities in atrial conduction also were seen. AF associated with OSA tends to be refractory to cardioversion and catheter ablation particularly in patients with untreated OSA, highlighting the expansive atrial remodeling associated with OSA.152 In a rat model of AF, OSA was associated with atrial conduction slowing attributed to connexin-43 down-regulation and increased atrial fibrosis. Such atrial remodeling promoted the persistence of AF.156

Several mechanisms may account for the development of AF and the AF substrate in patients with OSA. First, surges of sympathetic activity induced by hypoxia and the chemoreflex near the end of an apneic episode result in transient blood pressure rises.157 Second, vigorous inspiratory efforts during apnea accentuate the fluctuation of intrathoracic pressure increasing left atrial volume (stretch) and pressure.158 Third, an increase in oxidative stress signaling159 and systemic inflammatory mediators160 may promote atrial remodeling. Fourth, hypercapnia acutely prolongs ERP and slows conduction velocity, but with return of eucapnia delayed recovery of conduction has been associated with increased AF vulnerability.161 Fifth, negative tracheal pressure shortens atrial ERP and atrial monophasic action potential via vagal stimulation, which enhances AF inducibility.162

2.2.6 High Blood Pressure

Epidemiology

In the FHS the RF-adjusted OR for AF was 1.5 and 1.4 in men and women with hypertension, respectively.163 Later studies including the FHS found a limited association with mean arterial pressure, but found that pulse pressure was highly predictive of AF risk.164 The CHARGE-AF consortium observed that both systolic and diastolic blood pressure were predictive of AF risk.54 In addition, systolic blood pressure that approaches the upper limit of normal is associated with increased AF risk in healthy, middle aged men165 and women.166 The Women’s Health Study also showed that when an individual’s blood pressure remained elevated at follow-up visits, the risk of AF was higher as compared to those whose subsequent blood pressure recordings were lower suggesting a role for secondary prevention.166 The recent 50-year analysis of the FHS showed that while the rate of treated hypertension increased and severe hypertension became less prevalent, the population-attributed risk of AF was unaffected suggesting that anti-hypertensive therapy does not completely eliminate the elevated AF risk associated with hypertension.28

Pathophysiology

Increased left atrial size is a well-established, independent predictor of AF,163, 167 but other pathologic features of chronic hypertension including left ventricular hypertrophy163, 167, 168 and impaired diastolic dysfunction169 are also associated with AF. Common to all is an elevated left ventricular end-diastolic pressure, which increases left atrial pressure and volume. Atrial remodeling is associated with slower and more heterogeneous atrial conduction and increased PV firing. In addition, increased left atrial mass supports multiple reentry circuits.

Studies of animal models of hypertension have reported that early left atrial remodeling is characterized by atrial dilation with hypertrophy, reduced atrial ejection fraction, increased refractoriness, and prominent inflammatory infiltrates.170 Chronically, interstitial fibrosis and conduction slowing and heterogeneity are observed.170, 171 Moreover, increased atrial apoptosis has been observed.171 Electrophysiology studies in patients with chronically treated hypertension but without AF, have shown global conduction slowing, regionally delayed conduction in the crista teminalis and increased AF inducibility.172

Animal studies have suggested that the renin-angiotensin-aldosterone system (RAAS), which stimulates myocyte hypertrophy and intracellular fibrosis, also may contribute to atrial remodeling.173,174, 175 While upstream RAAS blockage was effective in animal models for reducing AF remodeling, two separate meta-analyses have reported that the benefit of RAAS blockade was limited to patients with HF or left ventricular hypertrophy.176, 177

3 Clinical Outcomes

3.1 Stroke

Epidemiology

AF is associated with increased risk of stroke and transient ischemic attack;178, 179 furthermore, AF-related strokes increase the risk of long-term disability or death.3 In the FHS the attributable risk of AF for stroke was 1.5% among 50–59 years olds whereas among 80–89 years olds it was 23.5%.179 Often AF is asymptomatic and consequently subclinical; however, in patients with implanted cardiac devices including pacemaker and defibrillators the burden of AF including subclinical AF may be accurately assessed. Atrial tachyarrhythmia (atrial rate >190 beats per minute) for longer than 6 minutes has been associated with an increased risk of clinical AF (HR, 5.56; 95% CI, 3.78–8.17) and ischemic stroke (HR, 2.50; 95% CI, 1.28–4.89).180

The risk of stroke with AF is variable and is modulated by other RFs including age ≥65, hypertension, diabetes, prior stroke/transient ischemic attack/thromboembolism history, vascular disease, HF, and female sex.181–183 Stroke risk models incorporating these RFs have been validated.184

Strokes in AF patients are associated with increased morbidity and mortality. In the Copenhagen Stroke Study, compared with individuals without AF, patients with AF had higher rates of in-hospital death (OR, 1.7; 95% CI, 1.2–2.5), longer hospital stay (50 days verse 40 days, p <0.001), and lower rates of discharge home (versus care facility, OR, 0.60; 95% CI, 0.44–0.85).185 Moreover, the infarct was larger and more commonly involved the cerebral cortex in patients with AF. The same study also showed that the odds for silent infarcts were similar for patients with AF and non-AF (OR, 0.99; 95% CI, 0.65–1.5).185 Correspondingly, the FHS showed increased 30-day mortality in AF-associated strokes than non-AF strokes (OR, 1.84; 95%CI, 1.04–3.27). Individuals with AF had worse 1-year survival following stroke and increased risk of stroke recurrences compared to those with non-AF strokes.3

Pathophysiology

Thrombogenesis in AF is not fully elucidated. A confluence of factors including blood stasis, endothelial dysfunction, and prothrombotic state has been implicated.

Attention has focused on the left atrial appendage (LAA). Animal models of AF demonstrate atrial contractile dysfunction from reduced myofibrillar sensitivity to Ca2+ 186 and intracellular Ca2+ transients.187 Clinically, reduced LAA emptying velocity on transesophageal echocardiography is associated with presence of spontaneous echo contrast, increased LAA thrombus, and stroke.188 In vitro studies have attributed spontaneous echo contrast to erythrocytes and fibrinogen interaction under low flow and shear stress.189 The role of LAA in AF-related stroke is further supported by efficacy of LAA closure devices for reducing strokes in patients with AF.190

Observational studies have revealed abnormalities in coagulation in patients with AF related strokes. Increases in prothrombin fragment and thrombin-antithrombin complexes have been observed in patients with AF-related strokes191 and in those with transesophageal spontaneous echo contrast.192 Other hemostatic factors have been implicated in contributing to the hypercoagulable state including fibrinogen, D-dimer, factor VIII, and von Willebrand factor.191, 193–196 However, in an adjusted model, the FHS showed that such abnormalities are ascribed to AF RFs and presence of cardiovascular disease rather than AF alone.197 Finally, inflammation may mediate endothelial dysfunction and hypercoagulability.

3.2 Extracranial Systemic Thromboembolism

Epidemiology

Compared to AF-related stroke, relatively less is known about epidemiology of extracranial systemic embolism events (SEEs). Data from the Danish Atrial Fibrillation Cohort showed an association between hospital diagnosis of AF and increased relative risk of SEEs in men (relative risk, 4.0; 95% CI, 3.5–4.6) and women (relative risk, 5.7; 95% CI, 5.1–6.3). Nearly half of SEEs occurred in patients between the ages of 70 and 79 years. The highest risk period for SEE was during the first year of incident AF, which is consistent with stroke data.198

More contemporary data are derived from a pooled analysis of four AF antiplatelet and anticoagulation trials with 37,973 patients from >40 countries.199 During the mean follow-up of 2.4 years, there were 221 SEEs accounting for 11.5% of clinically apparent embolic events. The incidence rates per 100 person-years for SEE and stroke were 0.24 and 1.92, respectively. Anatomically, SEEs were more likely to involve the lower extremities (58%) and mesenteric circulation (22%) while involvement of splenic and renal circulation was less common. Finally, increased morbidity and mortality was associated with SEE as 64% of patients required an interventional procedure or amputation and 24% died within 30 days.

Pathophysiology

The mechanisms underlying thromboembolism with SEE are similar to that of AF-related strokes.

3.3 Dementia

Epidemiology

Whereas dementia and AF share similar RFs including advancing age, obesity, diabetes, and hypertension, AF is associated with an adjusted increased risk of cognitive impairment,200 201 dementia,200, 202–205 Alzheimer’s dementia,206 and vascular dementia205 in patients with and without a history of stroke. In patients with normal baseline cognitive function and no history of stroke, meta-analysis of eight studies found a significantly increased risk of incident dementia in those with AF (HR, 1.42; 95% CI, 1.17–1.72).203 In patients with history of stroke, two meta-analyses have shown that AF is associated with ~ 2.5-fold adjusted risk of cognitive impairment and dementia.200, 202 Twenty-year follow-up of the Rotterdam Study reported AF patients < 67 years of age had greatest risk of dementia. The dementia risk increased with AF duration (exposure); whereas there was no increased risk associated with AF duration in those ≥ 67 years of age.204

Pathophysiology

Multiple mechanisms may explain the association of AF and dementia. Nearly one third of patients with AF have been observed to have silent brain infarcts on brain magnetic resonance imaging207 and micro thromboembolisms with covert infarction have been implicated as one possible mechanism. The FHS showed that over the last three decades, the incidence of dementia including dementia associated with AF has decreased.205 During this time period there has been improved use of anticoagulation in individuals with AF supporting the hypothesis that anticoagulation may reduce AF-associated dementia. This supposition is supported by a retrospective study of patients receiving long-term warfarin showing that incident dementia was 2.4 times higher in individuals with AF versus those without AF and that the dementia risk in those with AF and non-AF was significantly mitigated by increasing time in therapeutic range.208

A second possible mechanism may involve cerebral hypoperfusion associated with AF.209, 210 Interestingly one study indicated that the effect of AF on cerebral perfusion was most pronounced in younger patients (<50 years) and that no difference was observed in those >65 year of age209 consistent with epidemiological data showing age-dependent dementia risk in AF patients.204

3.4 Heart Failure

Epidemiology

HF is both a RF and an adverse clinical cardiovascular outcome associated with AF. The association was first recognized in the 1940s211 and it is now established that HF and AF often co-exist,2 predispose to the other,212 and share common RFs including hypertension, diabetes, coronary disease, and valvular disease.

HF as a Risk Factor for AF

In major HF trials, the prevalence of AF in patients with HF ranges from 13%–27%213–215 and the prevalence increases with increasing New York Heart Association Functional Class.216 In the FHS, HF was associated with 4.5-fold risk of AF in men and 5.9-fold risk in woman.2 Other epidemiological studies showed 2.67–3.37 fold risk of AF associated with HF.217, 218 Incremental reductions in systolic function are associated with increasing AF risk.167 HF with preserved ejection fraction also confers an increased risk of developing AF (HR adjusted for age and sex, 3.75; 95% CI, 2.19–6.40) with grade IV diastolic dysfunction as assessed with echocardiography conferring the highest risk.169

HF as an Outcome Associated with AF

The FHS uniquely reported the joint incidence of AF and HF and their temporal relationship. Among 931 participants with HF, 24% had prior or concurrent AF and 17% subsequently developed AF. One fifth of participants had AF and HF detected on the same day,212 demonstrating the closely interlinked pathophysiology. The incidence of first HF in FHS participants with AF is 33 per 1000-person years,212 which is comparable to that observed in the Danish nationwide cohort study.219 In a contemporary FHS cohort the incidence of HF was markedly higher in participants with AF than the incidence of AF in those with antecedent HF.220 In other words, AF begets HF more than HF begets AF.

The association of AF on HF subtype has also been reported. AF precedes HF with preserved ejection fraction (HFpEF) more commonly than HF with reduced ejection fraction (HFrEF). Among patients with prevalent AF, the adjusted HRs of incident HFpEF and HFrEF were 2.34 (95% CI, 1.48–3.70) and 1.32 (95% CI, 0.83–2.10), respectively.220 The finding that AF is more predictive of HFpEF is consistent with increased incident AF in patients with HFpEF versus HFrEF and suggests shared common mechanisms.169

In the FHS, the combination of AF and HF was associated with reduced survival.212 Among patients with prevalent AF, incident HF was associated with increased all-cause mortality compared to those without HF (HR, 1.25; 95% CI, 1.04–1.51).220 In one of the largest, worldwide studies, HF was found to be the leading cause of death one year after new-onset of AF accounting for nearly a third of all deaths.221 A meta-analysis showed a significantly higher all-cause mortality in AF patient with HFrEF than those with HFpEF (relative risk, 1.23; 95% CI, 1.12–1.36) despite similar risk of stroke and HF hospitalizations.222

Pathophysiology

The strong association of HF and AF has been attributed to shared mechanisms that lead to neurohormonal and proinflammatory activation, which induces myocardial inflammation and fibrosis. The atrial substrate with HF is characterized by atrial fibrosis and abnormalities in Ca2+ handling. It is distinct from the electrophysiologic changes associated with AF-induced atrial remodeling.223 Studies in dogs224 and humans225 indicate that HF induces an AF-susceptible atrial substrate without significantly altering atrial ERP (except at rapid rates) or ERP heterogeneity. These findings differ from the “AF begets AF” model with reduced atrial ERP. Histologic studies in HF dogs revealed structural changes including interstitial fibrosis with cellular hypertrophy, degeneration, and loss. These changes were associated with regions of delayed conduction and AF susceptibility. In human, extensive atrial fibrosis has been observed at autopsy in patients with dilated cardiomyopathy226 and areas of low voltage or electrical silence (scar) and fractionated or delayed potentials (slow conduction) have been identified during electrophysiology studies in patients with HF (but no AF).225 Neurohormonal activation is central to the generation of the AF substrate with HF and the milieu of profibrotic mediators are induced though angiotensin-dependent and independent pathways.227 Meta-analyses demonstrate benefit of upstream RAAS inhibition in AF patients with HF, but not in AF patients with other comorbidities; highlighting the unique pathophysiology of AF with HF.176, 177 Oxidized calmodulin-dependent protein kinase II has been shown to be a molecular signal that is increased by RAAS activation and is a common promoter of sinus node dysfunction, AF, and HF; thus reducing this kinase with RAAS block may explain the benefit of upstream therapy in patients with both AF and HF.228, 229

Increased trigger activity also is associated with HF and may increase risk of AF in the vulnerable substrate by generating a rapid burst of ectopic firing or maintaining a focal driver. In dogs, HF prolongs the APD and increases the phosphorylation of key regulatory kinases and phosphatases including calmodulin-dependent protein kinase II. The net effect of these HF-mediated changes is to enhance Ca2+ uptake into the SR promoting after-depolarization initiation.186 Similarly, HF has been shown to increase calcium sparks in cardiomyocytes isolated from the PVs of rabbits.230

Finally, tachycardia and shortening of diastolic filling time associated with irregular ventricular activation with AF further impair diastolic relaxation and promote clinical HF, which further induces atrial remodeling leading to the perpetuation of AF.

3.5 Myocardial Infarction

Epidemiology

As with HF there is a bidirectional relationship between AF and myocardial infarction (MI). Coronary heart disease is associated with an increased risk of AF,231 but AF also is associated with increased risk of MI.232 In the REGARDS cohort AF was associated with a 2-fold increased risk of MI,233 that was greater in women (HR, 2.16; 95% CI, 1.41–3.31) versus men (HR, 1.39; 95% CI, 0.91–2.10) and AAs (HR, 2.53; 95% CI, 1.67–3.86) versus whites (HR, 1.26; 95% CI, 0.83–1.93). The Olmsted study reported similar unadjusted risk of coronary ischemic event, but after adjusting for age the incidence was higher in men than woman.234 Among those with newly diagnosed AF, the risk of death after coronary ischemic events was higher in woman than men (HR, 2.99; 95% CI, 2.53–3.53 versus HR, 2.33; 95% CI, 1.94–2.81). Interestingly, as observed with strokes and SEEs, the event rate of coronary ischemic events was highest within the first year of incident AF (4.7%, 95% CI, 3.9–5.6) and subsequently declined to 2.5% per year.

Pathophysiology

The mechanisms linking AF to MI are not completely understood. First, both AF and MI have overlapping RFs that may lead to the development of AF and MI in parallel. For instance, both AF and coronary heart disease are associated with proinflammatory and prothombotic states. Second, MI in AF may be attributed to coronary artery thromboembolism. Third, myocardial ischemia may arise from supply-demand mismatch in the setting of tachycardia associated with AF. Fourth, MI may lead to left ventricular remodeling that may predispose to AF.

3.6 Venous Thromboembolism

Epidemiology

Increased BMI, obesity, and smoking are associated with venous thromboembolism (VTE)235–238 as well as AF. Potential direct causal relationship between AF and VTE has been proposed, but needs to be studied further.239, 240

A few studies have reported increased risk of VTE in AF and vice versa. In a retrospective cohort study based on the national administrative database in Taiwan, the risks of VTE (adjusted HR, 1.74; 95% CI, 1.36–2.24) and pulmonary embolism (adjusted HR, 2.18; 95% CI, 1.51–3.15) were both higher in the AF group compared to non-AF referents.241 A Norwegian administrative database study reported that AF was associated with an increased risk of pulmonary embolism (adjusted HR, 1.83; 95% CI, 1.16–2.90) but not VTE (adjusted HR, 1.04; 95% CI, 0.64–1.68).239 In addition, the same study found that individuals with incident VTE were subsequently at a higher risk of developing incident AF (adjusted HR, 1.63; 95% CI, 1.22–2.17) compared to those without VTE.240 It should be noted that in both the Taiwanese and Norwegian cohorts, a significant proportion of individuals had pre-exiting RFs for VTE including lower extremity fracture, recent surgery, cancer, and immobility.239, 241 In comparison to individuals with AF or VTE alone, those with AF and VTE were older and had higher mean BMI. In the Norwegian study, the mean age and BMI were 64 years and 26.9 kg/m2 for AF alone, 57 years and 26.7 kg/m2 for VTE alone, and 68 years and 29.2 kg/m2 for AF and VTE combined.240

Pathophysiology

The mechanisms underlying the AF-VTE association are inexplicable. Several studies have shown that AF is associated with a hypercoagulable state attributed to elevated hemostatic factors including fibrinogen, D-dimer, prothombin fragment, factor VIII, and von Willebrand factor;191, 193–196 however, many of these studies were not adjusted for co-existing cardiovascular RF. In an adjusted model, the FHS showed no significant difference in levels of fibrinogen, von Willebrand factor, or tissue plasminogen activator suggesting that co-existing RFs rather than AF may explain elevated thrombotic risk.

3.7 Mortality

Epidemiology

The FHS was one of the first studies to report that AF had a multivariable-adjusted association with increased risk of death.4 In addition, the study observed a significant interaction, such that AF diminished the survival advantage generally enjoyed by women; the multivariable-adjusted OR for death in men and women was 1.5 and 1.9, respectively. At 10-year follow up, 61.5% of men with AF between 55 to 74 years of age had died compared to 30.0% of men in the same age group without AF. A similar trend was found in women with 57.6% of those with AF dying by 10-year follow up as compared to 20.9% in those without AF. The increased risk was consistent across all decades of age from 55–95 years. Even in individuals without clinical evidence of cardiovascular or valvular disease at baseline, AF was associated with a 2-fold increased risk of death.4 A retrospective study of Medicare beneficiaries 65 years or older showed that death was the most frequent AF outcome with an incidence of 19.5% at 1 year and 48.8% at 5 years after initial diagnosis.242 A recent systematic review and meta-analysis of 64 studies that included 1,009,501 patients with 149,746 (14.8%) having AF, found a pooled relative risk of death that was 1.6 (95% CI, 1.39–1.53); although there was marked heterogeneity.243 In fourteen studies cardiovascular mortality was assessed and the pooled relative risk associated with AF was 2.03 (95% CI, 1.79–2.3).

It is noteworthy that there is growing evidence that AF is associated with an increased risk of sudden cardiac death (SCD). Pooled-analysis of the ARIC and Cardiovascular Health Study cohorts showed that AF was associated with more than a doubling of the risk of SCD SCD compared to participants without AF (HR, 2.47; 95% CI, 1.95–3.13).244 A meta-analysis of 7 studies found the relative risk of SCD was 1.88 (95% CI, 1.36–2.6); although significant study heterogeneity was present.243 In the RE-LY trial, 37.4% of all deaths and 60.4% of cardiac deaths were attributed to either SCD or death due to progressive HF.245 For comparison only 9.8% of all deaths were attributed to stroke or hemorrhage.

A meta-analysis of anti-thrombotic studies has shown that oral anticoagulation reduced risk of all-cause mortality with an absolute risk reduction of 1.6% as compared with control or placebo.246 In anticoagulated patients with AF increased mortality is largely driven by cardiovascular causes rather than non-hemorrhagic stroke or systemic embolism.245, 247–249 Strokes comprised a small portion of deaths from AF. In the ROCKET-AF trial cardiovascular deaths occurred over two times more often than strokes.247 Predictors of higher all-cause mortality included HF (HR, 1.51; 95% CI, 1.33–1.70) and age ≥75 (HR, 1.69; 95% CI, 1.51–1.90). Thus further advances in anticoagulation strategies may have little effect on improving overall mortality in AF.

Conclusion

Over the last 50 years, the FHS and other epidemiological studies have yielded a breadth of data associating various RFs with risk of AF and providing insight into their mechanistic link to AF genesis. However, many questions remain. Will genetic studies improve AF risk assessment, identify novel therapeutic targets, and help guide treatment strategies for both primary and secondary prevention of AF? By what degree does RF modification alter the atrial substrate, AF burden, and clinical outcomes? What are the target goals for RF modification and how will genetics alter these targets? Ongoing and future epidemiological, translational, and clinical studies may provide insight into these unanswered questions and improve clinical outcomes in patients with AF.

Figure 2.

Figure 2

AF risk factors induce structural and histopathological changes to the atrium that are characterized by fibrosis, inflammation, and cellular and molecular changes. Such changes increase susceptibility to AF. Persistent AF further induces electrical and structural remodeling that promotes perpetuation of AF. AF also may lead to the development of additional AF risk factors that further alters the atrial substrate. Finally, AF is associated with several clinical outcomes. *There are limited data supporting the association.

Acknowledgments

Source of Funding:

This work was supported by Velux Foundation and Boston University School of Medicine and the National Heart, Lung, and Blood Institute’s Framingham Heart Study (contract: NIH/NHLBI N01-HC25195l; HHSN268201500001I; 2R01HL092577; 1R01HL128914; 1P50HL120163).

Abbreviations and acronyms

AA

African Americans

AF

Atrial fibrillation

APD

Action potential duration

ARIC

Atherosclerosis Risk in Communities Study

BMI

Body mass index

CI

Confidence interval

CRF

Cardiorespiratory fitness

EAT

Epicardial adipose tissue

ERP

Effective refractory period

FHS

Framingham Heart Study

GWAS

Genome-wide association study

HF

Heart failure

HFpEF

Heart failure with preserved ejection fraction

HFrEF

Heart failure with reduced ejection fraction

HR

Hazard ratio

LAA

Left atrial appendage

MESA

Multi-Ethnic Study of Atherosclerosis

MI

Myocardial infarction

OR

Odds ratio

OSA

Obstructive sleep apnea

PV

Pulmonary vein

RAAS

Renin-angiotensin-aldosterone system

RF

Risk factor

SCD

Sudden cardiac death

SEE

Systemic embolism events

SNP

Single-nucleotide polymorphism

SR

Sarcoplasmic reticulum

VTE

Venous thromboembolism

Footnotes

Disclosures:

Laila Staerk has received research funding from Boehringer Ingelheim.

Jason A. Sherer: none

Darae Ko: none

Emelia J. Benjamin: none

Robert H. Helm: none

References