Effect of estrogen and testosterone on cardiac electrophysiology and atrial fibrillation

Abstract

The prevalence and treatment outcomes of atrial fibrillation (AF) differ between sexes, yet the mechanisms underlying these differences remain poorly understood. Given that estrogen and testosterone are the major sex hormones, they may be involved in altering cardiac electrophysiology and risk of AF.

Preclinical studies demonstrate that estrogen and testosterone alter ion currents and cardiomyocyte calcium (Ca²⁺) handling via receptor-mediated mechanisms, but with opposite effects. Estrogen prolongs action potential duration (APD), whereas testosterone shortens it. A longer APD has antiarrhythmic effects in the atria, whereas a shorter APD is proarrhythmic. Testosterone also accelerates repolarization and enhances contractility, further increasing AF risk. However, most laboratory studies were done in animal ventricular models, limiting direct translation to human atrial physiology owing to species-specific electrophysiological differences.

Human population studies suggest that decreased exposure to endogenous estrogen increases AF risk in women. In males, there is no similar consensus for the relationship between endogenous testosterone exposure and AF risk in men. Evidence from hormone replacement, testosterone replacement, and androgen deprivation therapies is similarly inconclusive, reflecting confounding factors, variation in hormone formulations, and reliance on single hormone measurements.

This review provides novel insights into how estrogen and testosterone differentially influence cardiac electrophysiology and risk of AF in men and women. These findings underscore the need for personalized sex- and hormone-specific treatment strategies in AF management, addressing a critical gap in current clinical practices.

Key findings.

• ▪Receptor heterogeneity drives sex-specific cardiac electrophysiology changes: Chamber-, sex-, and development-dependent variability in estrogen (ERα, ERβ, GPER) and androgen (AR, AR45) receptor subtypes likely modulate ionic current profiles and underlie sex-specific susceptibility to atrial fibrillation (AF).
• ▪Estrogen and testosterone have opposite electrophysiological effects: Estrogen prolongs action potential duration (APD) by suppressing repolarizing potassium currents (IKr, IKs) and reducing calcium transients, while testosterone shortens APD by enhancing repolarizing currents (IKs, IK1, Ito) and increasing calcium transient amplitude and SR spark frequency, producing faster repolarization and greater contractility.
• ▪Estrogen is anti-arrhythmic while testosterone is pro-arrhythmic as per laboratory studies: In vitro and animal studies consistently demonstrate that testosterone promotes atrial re-entry and delayed afterdepolarizations, whereas estrogen mitigates these electrophysiological triggers, theoretically conferring protection against AF initiation.
• ▪Population studies show inconsistent associations between sex hormones and AF incidence: Population studies report conflicting results on association between estrogen or testosterone levels and AF risk in either sex, likely due to inadequate control of confounding factors and reliance on single, static hormone measurements.
• ▪Human induced pluripotent stem cell–derived atrial cardiomyocytes (hiPSC-aCMs) offer a novel direction for further laboratory research: hiPSC-aCMs offer a promising platform to study sex-specific electrophysiological effects with greater translational relevance than animal ventricular models, although limitations in maturity and hormone receptor expression remain.

Introduction

Atrial fibrillation (AF) is the most common arrhythmia worldwide, affecting more than 30 million people.^1,2 Treatment primarily focuses on rate and rhythm control and stroke prevention; however, antiarrhythmic drugs are often ineffective.³ Catheter ablation, although more effective at maintaining sinus rhythm, is still associated with significant AF recurrence rates.³ Furthermore, AF increases the risk of stroke, heart failure, and mortality, thus imposing a vast burden on patients and health care systems.^3,4

Despite increasing AF incidence, significant gaps remain in understanding its pathophysiology, particularly regarding the influence of sex hormones.⁴ Although age-related AF prevalence is lower in women than men, women frequently experience more severe symptoms and higher rates of complications, particularly stroke.^5,6 Treatment outcomes also differ between sexes; females are more susceptible to drug-induced arrhythmias such as torsade de pointes after antiarrhythmic drug therapy. Some studies suggest higher recurrence rates in women after catheter ablation.^4,6, 7, 8, 9, 10 Therefore, examining the influence of sex hormones on these outcomes is essential.

The impact of sex hormones extends beyond AF, extending to other cardiac conditions, such as stroke and takotsubo cardiomyopathy.11, 12, 13, 14 Physiologically, sex hormones may explain such differences, particularly in conditions such as AF, which is an electrical rhythm disorder. Differences in cardiac electrophysiology between men and women are partly responsible for these variations, and these disparities become more pronounced after puberty as sex hormone production increases.15, 16, 17

This review summarizes current evidence on the effects of sex hormones on cardiac electrophysiology and their contribution to AF, aiming to help formulate an explanation for sex-specific treatment outcome differences in AF and inform more tailored management options in men and women. We limit our review to 2 major sex hormones—estrogen and testosterone. We first examine laboratory studies with predominantly animal models exploring the hormonal effects on cardiac electrophysiology. We then look at human population studies assessing hormone levels and AF incidence in men and women.

Effect of estrogen and testosterone on cardiac electrophysiology

Estrogen and testosterone receptor variability

Estrogen (ERs) and androgen receptors (ARs) are strongly expressed on cardiomyocytes (CMs), highlighting the importance of the 2 sex hormones in cardiac function.¹⁶ Sex hormone receptors are differentially expressed within myocardial tissue, with notable variability across sexes, cardiac chambers, and developmental stages. Differences in receptor densities and isoforms contribute to heterogeneity in cardiac electrophysiology by modulating key ionic currents and repolarization dynamics.

ARs are expressed at relatively low levels in the myocardium; however, 1 AR splice variant, AR45, is more abundant and has been shown to influence repolarization reserve in CMs. AR45 upregulates the human ether-à-go-go–related gene (HERG), which encodes the rapidly activating delayed-rectifier potassium current (I_Kr), a major determinant of cardiac repolarization.¹⁷ Increased AR45 stimulation enhances HERG expression and the resulting repolarizing potassium (K⁺) current, thereby stabilizing repolarization and therefore exerting antiarrhythmogenic effects.¹⁷

Given that sex receptor expression and localization are influenced by hormone availability, it is expected that AR messenger RNA (mRNA) expression is more pronounced in male myocardium than in female myocardium.¹⁸ However, studies investigating this were derived from diseased cardiac samples (cardiomyopathy and aortic stenosis), reducing generalizability to healthy tissue.¹⁸ Furthermore, different isoforms of AR can have opposing effects on arrhythmia risk: α-AR isoforms can seem cardioprotective, reducing arrhythmias, whereas β-AR may accelerate myocardial injury when overexposed to testosterone.¹⁹ However, the effects of AR isoforms on human myocardial cells under physiological testosterone levels on cardiac arrhythmia remain to be further investigated.

In contrast, ERs within the heart are more extensively studied. The most abundant isoforms expressed in CMs include ERα, ERβ, and the G protein–coupled estrogen receptor (GPER). Through both genomic and nongenomic mechanisms, these receptors modulate Ca²⁺ handling and L-type Ca²⁺ currents, thereby influencing myocardial excitability and repolarization reserve.²⁰

Women express higher overall levels of estrogen receptors owing to greater circulating estrogen, yet isoform distribution varies. ERα expression was approximately 4 times greater in female than in male rat myocardium, whereas ERβ levels are comparably low in both sexes.²¹ Interestingly, although GPER levels were comparable between male and female left rat ventricles at baseline, there was a significant increase in GPER expression only in female rat hearts upon exposure to pressure overload.²² The relative contributions of each atrial receptor to cardiac electrophysiology and arrhythmia susceptibility remain unclear.

Receptor distribution also varies across cardiac chambers and developmental stages. In murine models, ERα mRNA expression is greatest in the right atrium and left ventricle during early postnatal development. However, there is a more pronounced increase in atrial tissue (6–18-fold) than in ventricular tissue (2–5-fold) as the heart matures, the result being that ERα mRNA expression was higher in adult rat atria than ventricles.²³ In contrast, ERβ expression in the left atrium is initially the greatest but decreases dramatically by approximately 20-fold with maturation, such that, in the end, ERβ levels become relatively uniform across chambers except the right ventricle where they were 2-fold higher.²³ Such temporal changes in estrogen receptor expression may have important implications for arrhythmogenesis, particularly because AF predominantly occurs in the adult heart.

Temporal dynamics in sex hormone receptor activation are crucial for understanding hormonal effects across populations. During puberty, surges in circulating estrogen and testosterone drive receptor upregulation and sensitization, contributing to the development of sex-specific electrophysiological profiles. In contrast, the effects of aging on receptor expression are less consistent. Age-related declines in hormone levels may alter receptor expression or signaling efficiency, but evidence remains conflicting. For example, 1 study reported reduced ERα transcription in rat myocytes alongside increased GPER expression, affecting Ca²⁺ handling and repolarization,²⁴ whereas other studies observed increased ERα expression with age in both sexes, independent of estrogen therapy.²⁵ These findings indicate that the cardiac response to a given hormone concentration can vary with age, reflecting changes in receptor expression and sensitivity over the lifespan. Therefore, incorporating temporal receptor dynamics is essential to understanding sex- and age-specific patterns of arrhythmia susceptibility.

Most available data on receptor distribution and subtype expression are derived from in vitro or animal models, and human studies are often limited to diseased myocardium. A more detailed understanding of receptor variability in healthy human tissue across sexes, cardiac regions, and developmental stages will be essential to clarify the mechanisms by which sex hormones influence cardiac electrophysiology and arrhythmia risk.

Sex hormones

Estrogen and testosterone affect cardiac electrophysiology through both genomic (modifying ion channel transcription) and nongenomic pathways.^26,27 They alter action potential duration (APD) by modulating repolarizing currents and cardiac contractility by regulating CM Ca²⁺ handling.

Estrogen

Estrogen suppresses repolarizing currents, prolonging APD (Figure 1). A study of guinea pig ventricular CMs showed that estrogen suppressed the rapidly activating component of the delayed-rectifier K⁺ channel, I_Kr, with no apparent effect on the slowly activating delayed-rectifier K⁺ current (I_Ks), the slowly activating component of the delayed-rectifier K⁺ channel.²⁸ However, a study of rabbit ventricles suggested that high estrogen levels reduce 1 of the transcripts of I_Ks mRNA, lowering I_Ks currents and increasing rates of sudden cardiac death.²⁹ The heterogeneous effect estrogen seems to have on I_Ks may be caused by contrasting methodologies. The first study used a voltage-clamp functional assay,²⁸ whereas the second study relied on mRNA quantification, which provides only an indication of the genetic message and does not reflect translational or post-translational modifications.²⁹ In addition, electrophysiological characteristics vary across animal species. Although guinea pigs have a fixed 16-day estrous cycle, rabbits have irregular cycles varying between 3 and 12 days, thus serving as poorer models for studying changes in estrogen.30, 31, 32

Figure 1.

The effects of estrogen and testosterone on cardiac electrophysiology. A: The effects of estrogen and testosterone on repolarizing K⁺ currents and cardiac action potential duration (APD). B: The myriad of effects of estrogen deficiency on cardiomyocyte Ca²⁺ handling proteins. The figure was created using BioRender.com. Ca²⁺ = calcium; I_K1 = inwardly rectifying potassium current; I_Kr = rapidly activating delayed-rectifier potassium current; I_Ks = slowly activating delayed-rectifier potassium current; I_to = transient outward potassium current; K⁺ = potassium; LTCC = L-type calcium channel; NCX = sodium-calcium exchanger; RyR = ryanodine receptor; SERCA = sarcoplasmic reticulum calcium ATPase.

Estrogen also exerts complex effects on Ca²⁺ handling (Figure 1). Estrogen deficiency induced by ovariectomy in mice models led to greater sarcoplasmic reticulum (SR) Ca²⁺ content, greater Ca²⁺ transient amplitude, faster Ca²⁺ transient rise and decay, and more frequent Ca²⁺ sparks. These translate to increased excitation–contraction coupling (ECC) and enhanced contractility.33, 34, 35 Similarly, chronic deficiency of ovarian hormones increases L-type inward Ca²⁺ current (I_CaL) in guinea pig CMs, which in turn activates ryanodine receptor 2–mediated Ca²⁺ release, also known as Ca²⁺-induced Ca²⁺ release.³⁶ This has a positive inotropic effect and is negated by timely estrogen replacement,^33,37 suggesting that estrogen deficiency leads to dysregulation of Ca²⁺ handling, which promotes a proarrhythmic state (Figure 1).

Several important limitations must be considered. First, early-life ovariectomy prevents myocardial tissue from being exposed to normal pubertal estrogen levels.^34,35 This early estrogen withdrawal contrasts with other more commonly used models in which tissue exposure to estrogen is present into adulthood and may more closely simulate the effect of estrogen on Ca²⁺ homeostasis in the myocardium.³⁵ Second, the elimination of estrogen synthesis from the ovaries owing to ovariectomy results in upregulation of estrogen synthesis from other sources.^34,35 Ovariectomy-induced increased adiposity and aromatase activity can modify Ca²⁺ handling, which is not appreciated in animal models.^34,35 Third, ovaries also produce significant levels of testosterone even after menopause.³⁴ Therefore, testosterone deficiency is a confounding factor in ovariectomy models.

In summary, estrogen prolongs APD by dampening repolarizing K⁺ currents and reduces contractility by reducing Ca²⁺ transients and I_CaL. However, most evidence arises from studies on animal models, which have significant heterogeneity in cardiac electrophysiology and hormonal cycles. For instance, APD effects are mainly caused by I_Kr and I_Ks changes in guinea pigs and I_to changes in rats.38, 39, 40 This reduces comparability with human CM electrophysiology owing to the variation in estrogen exposure. In addition, Ca²⁺ release plays a bigger role in ECC for mice than guinea pigs.³⁴ Finally, most studies were done on ventricular CMs, which have different inherent electrophysiological properties from atrial CMs (aCMs).²¹

Testosterone

Testosterone seems to exert effects opposite to estrogen by shortening the APD. In castrated male rabbits supplemented with testosterone, APD was significantly shortened compared with placebo, suggesting an inverse relationship between testosterone and APD.⁴¹ Similarly, men with hypogonadism have an increased prevalence of prolonged QTc interval.⁴² The shorter repolarization period in males can be attributed to increased repolarizing currents and faster Ca²⁺ decay (Figure 1).⁴³

Animal studies demonstrate that testosterone increases repolarizing currents. A guinea pig study using isolated Langendorff-perfused hearts suggested that testosterone accelerated repolarization by shortening APD through enhancing I_Ks. In addition to chromanol 293B, an I_Ks inhibitor, APD remained the same.⁴⁴ A canine study further found that testosterone affected myocardium repolarization via the inward rectifying K⁺ current channel, inwardly rectifying K⁺ current (I_K1), and the transient outward K⁺ current (I_to) channel, I_to proteins (Kir_2.1, Kv_4.3).⁴⁵ Female and male dogs were castrated and given testosterone and estrogen, respectively. Testosterone-treated castrated females and normal males had greater Kir_2.1 and Kv_4.3 expression than estrogen-treated castrated males and normal female dogs, potentially offering antiarrhythmogenic properties.⁴⁵

A shorter APD can also be attributed to shorter repolarization arising from faster decay of intracellular Ca²⁺ at the end of systole.^46,47 However, the mechanism behind this is unclear. Given that most of the postsystole Ca²⁺ dispersion is performed by SR Ca²⁺ transport ATPase (SERCA), one would expect SERCA to be increased or more active in males.⁴⁸ Nevertheless, there are opposing opinions as to whether sex hormones influence SERCA. Through its regulatory mechanism, phospholamban (PLB) inhibits SERCA-mediated SR Ca²⁺ uptake,⁴⁹ so low levels of PLB could explain faster Ca²⁺ decay. However, similar PLB levels and mRNA levels were found in both male and female mice hearts.¹⁵ Given that phosphorylated PLB cannot inhibit SERCA Ca²⁺ uptake, decreased PLB phosphorylation by enzymes such as protein kinase A or calmodulin-dependent protein kinase II could explain the faster Ca²⁺ decay in males instead and can be explored in future studies.¹⁵ Although the sodium-Ca²⁺ exchanger (NCX) is also involved in Ca2+ dispersion, evidence regarding sex-related differences in NCX remains inconsistent, limiting conclusions about the exact mechanisms driving faster Ca²⁺ decay in males.⁵⁰

Testosterone also enhances cardiac contractility by increasing the systolic transient Ca²⁺ rise.^15,47,51 First, males have larger and more frequent Ca²⁺ sparks.^52,53 Second, male CMs have a greater ECC gain, with more Ca²⁺ released from SR for every unit of I_CaL.^47,54 However, the mechanism behind the greater ECC gain is less well understood. Although it was previously hypothesized that SR Ca²⁺ content is greater in males, leading to more Ca²⁺ released from the SR during contraction, studies show similar or even reduced SR Ca²⁺ content in male CMs.^15,54 Therefore, the ECC gain disparity across sexes is likely caused by other testosterone-mediated changes in Ca²⁺ handling.

In summary, testosterone shortens APD by enhancing repolarizing currents and accelerates Ca²⁺ decay through an unknown mechanism, leading to faster repolarization. Testosterone also boosts cardiac contractility by increasing Ca²⁺ transients. However, rats and guinea pigs have different action potential morphology, as discussed earlier, whereas dogs have a stronger I_K1 and a greater repolarization reserve than humans.⁵⁵ Therefore, it is necessary to investigate the electrophysiological effect of testosterone in human CMs.

Electrophysiological effect of estrogen and testosterone on AF from in vitro and animal studies

From an electrophysiological standpoint, testosterone generally promotes AF risk, whereas estrogen seems protective. Testosterone shortens APD, boosts transient Ca²⁺ rise in systole, and increases SR Ca²⁺ spark frequency, whereas estrogen exerts opposite effects. First, a shorter APD increases the likelihood of signal re-entry and thus atrial arrhythmia.⁵⁶ Although a longer APD increases susceptibility to early afterdepolarizations (EADs), aCMs are resistant to EADs and EADs have been shown to trigger AF only when they occur in pulmonary vein myocytes.57, 58, 59, 60 Second, a greater systolic transient Ca²⁺ rise increases myocardial contractility, which increases AF risk.⁵⁶ Finally, more frequent SR Ca²⁺ sparks increase intracellular Ca²⁺, activating NCX, which results in delayed afterdepolarizations and promotes atrial arrhythmia.⁵⁰

Although the effects of estrogen and testosterone on cardiac electrophysiology have been well demonstrated, this alone cannot be used to predict their overall effects on AF owing to their widespread nonelectrophysiological effects. For instance, estrogen and testosterone also affect sympathetic nervous system activity through changes in sensitivity to catecholamines.^61,62 The resultant effect includes cardiac structural remodeling and changes in the secretion of atrial natriuretic peptide, which further alters the risk of developing AF.^63,64

Discussing the numerous nonelectrophysiological effects of both hormones on AF risk is outside the scope of this review. To get an idea of the overarching effect of sex hormones on AF risk, we review population studies that investigate serum hormone levels and assess AF incidence in females and males over the years.

Effect of estrogen and testosterone on AF in female population studies

Estrogen in pregnancy

Endogenous estrogen levels increase markedly in pregnancy.⁶⁵ Thus, studies assessing AF risk in pregnancy offer a different perspective on estrogen’s role in AF.

A retrospective population study found a relatively lower AF prevalence among pregnant patients.⁶⁶ Interestingly, a significant increase in AF cases was observed in the third trimester, coinciding with a peak in estrogen levels, as confirmed in a subsequent meta-analysis.⁶⁷ However, the increased AF incidence in the final trimester of pregnancy is more likely associated with other physiological adaptations such as plasma volume expansion, red cell mass increase, and cardiac structural remodeling, rather than elevated estrogen levels.⁶⁸ In addition, other hormones, particularly progesterone, experience drastic changes in this period, further casting doubt on the true extent of the role of estrogen in AF.⁶⁵ A Mendelian randomization study provided evidence that greater live births were associated with greater AF risk, possibly owing to cumulative lifetime exposure to estrogen levels across multiple pregnancies.⁶⁹

Estrogen in menopause

During menopause, estrogen levels decrease significantly and the primary estrogen changes from estradiol to the weaker estrogen estrone. Cohort studies of menopausal women generally suggest a mild protective role for estrogen in AF.

A nationwide Korean study found that women with fewer reproductive years (and hence decreased lifetime exposure to endogenous estrogen) had higher AF risk than women with later menopause.⁷⁰ Similarly, a United Kingdom cohort study reported increased AF incidence in women with premature (<40 years) menopause.⁷¹

In contrast, a smaller European prospective cohort study of primarily European women found no significant association between menopausal age and AF incidence.⁷² However, this study had limited power (n < 2000) and failed to control for hormone replacement therapy (HRT) use.

Overall, these findings support that prolonged endogenous estrogen exposure may protect against AF or, conversely, that its absence may increase AF risk. However, we must also consider that menopause is associated with not only a decrease in estrogen levels but also increased body mass index, metabolic syndrome, and blood pressure, all of which contribute to AF risk.⁷³ Whether the postmenopausal predisposition results directly from reduced estrogen levels or indirectly through established AF risk factors is unclear.⁷⁴

Estrogen replacement therapy

HRT is commonly prescribed to alleviate menopausal symptoms, leading to supraphysiological levels of exogenous hormones. Nationwide cohort studies of menopausal women in Taiwan and Korea have assessed the effect of exogenous estrogen on AF risk.^75,76

The Taiwanese study found that estradiol-only HRT significantly reduced AF risk, whereas the Korean study found no such benefit. Interestingly, the Korean study found that those currently on estradiol plus progesterone HRT had increased AF risk, whereas past users showed decreased risk. Both studies also found conjugated equine estrogen HRT significantly increased AF incidence, which was further corroborated by a randomized controlled trial.⁷⁷

This conflicting evidence suggests that estradiol may reduce AF risk, whereas conjugated equine estrogen may increase it, highlighting the importance of formulation and administration route in evaluating AF risk.

An important point is that individuals receiving HRT have different characteristics from the general female population. First, to qualify as a candidate for estrogen-only therapy, a uterus must be absent, owing to the increased risk of endometrial cancer upon unopposed estrogen.^78,79 Second, many AF risk factors, such as obesity, diabetes, and smoking, are often contraindications for HRT therapy, but common in the general population, potentially creating healthy-user bias in the HRT cohort.⁷⁸ Differences in compliance with HRT regimes are yet another limitation and could be mitigated with regular monitoring of adherence. In addition, whether these results can be generalized to other races must be considered.⁸⁰

Another possible reason for the differences in results from HRT trials exploring estrogen’s effect on the risk of AF could be HRT timing. Earlier initiation of HRT after the commencement of menopause may reduce cardiovascular events and atherosclerotic progression, as per the controversial timing hypothesis.⁸¹ Given that atherosclerosis increases AF risk, it raises the question of whether AF risk is similarly dependent on the timing of HRT initiation, raising the question of another potential confounding factor.

Testosterone in women

3 prospective cohort studies have explored the relationship between testosterone and AF in women. 1 study found that higher testosterone levels were associated with a higher risk of AF, a pattern supported by a large retrospective cohort study of women with polycystic ovarian syndrome, characterized by abnormally high testosterone levels.^82,83 In postmenopausal females, there are links to increased cardiovascular disease generally,⁸⁴ although there is a lack of research in AF specifically. Conversely, other prospective studies found no effect.^85,86 Discrepancies may arise from women being included at different stages after menopause, given its progressive decline over time. In addition, variations in testosterone assays across studies may further explain inconsistencies, particularly among the already low testosterone levels present in females.

Effect of estrogen and testosterone on AF in male population studies

Estrogen in men

Studies investigating the relationship between estrogen and AF in men yield conflicting results. 1 study found that higher estrogen levels increased AF incidence, but another found no association.^85,87 The latter study involved older, predominantly Caucasian men, potentially contributing to the discrepancy. Moreover, estrogen levels fluctuate significantly throughout the year, and both studies relied on single estrogen measurements.⁸⁸ It is unlikely that a singular value is an accurate measurement of the dynamic estrogen level over time. Longitudinal studies with repeated estrogen measurements are required for a better understanding of this relationship.

Testosterone in men

Among 4 large cohort studies published in the last decade, 2 reported a positive association between testosterone and AF, whereas the other 2 found a negative association.^54,55,57,58 In addition, another study identified no association with total testosterone but a negative association for free dihydrotestosterone, a more active metabolite of testosterone.⁸⁹

Multiple explanations may explain this discrepancy. First, nonlinear effects may exist. Both low and high levels of testosterone may similarly affect AF risk. Second, observational studies are inherently biased and have residual confounding. For example, only 1 study adjusted for PR interval baseline and presence of significant murmur, and another adjusted for estimated glomerular filtration rate, all of which are recognized risk factors for AF.90, 91, 92, 93 Third, differences in ethnic composition varied across different studies, with some featuring predominantly Caucasian populations and others having more homogeneous populations. The potential confounding role of ethnicity when determining the relationship between testosterone and AF risk remains understudied.⁷⁰ In addition, sex-hormone-binding globulin levels increase with age, where older people have decreased bioavailable testosterone for the same measured level of total testosterone.⁹⁴ Across the studies, the average age of participants ranged from 49 to 76 years, so total testosterone levels may not reflect bioavailable testosterone levels accurately.⁹³ Finally, similar to the earlier estrogen studies, testosterone levels were also singularly measured for a baseline value and not a true reflection of the dynamic hormone levels over time.⁹⁵

Testosterone replacement therapy

Testosterone replacement therapy (TRT) is used to restore physiological levels in men with hypogonadism or as gender-affirming therapy in transgender men. Exogenous testosterone administration in men with hypogonadism seems to increase the risk of AF even if total testosterone levels are not high. A pooled meta-analysis in 2005 found 5 cases of AF in 651 TRT recipients compared with just 1 AF case in 433 placebo recipients, although the difference was not statistically significant.⁸⁶ A randomized controlled trial in 2010 was terminated early owing to a higher rate of cardiovascular-related events in patients on TRT.⁸⁷ This was corroborated by a randomized controlled trial in 2023 looking at more than 5000 men that found a significantly increased risk of AF in men on TRT.⁹⁶ In contrast, a retrospective cohort study with more than 75,000 hypogonadal men found that veterans with normalized testosterone levels after TRT had a statistically significant decrease in AF incidence compared with both those with non-normalized levels who received TRT treatment and those who remained untreated.⁹⁷

Although the cohort study’s large sample size is a strength, its observational study design renders it prone to bias, as well as the duration of treatment and exclusion of participants with subsequent normal testosterone levels in the controls. Interestingly, the 2 randomized controlled trials that found significant adverse cardiovascular effects from TRT looked only at testosterone gel application, whereas both the meta-analysis and cohort study also included other forms of TRT such as patches and injections. Different TRT preparations have different pharmacokinetic properties with gels having a longer effective half-life than patches and injections, resulting in longer periods of supraphysiological testosterone levels than transdermal routes.^98,99 Further research is needed to stratify the results by testosterone formulation.

Research on gender-affirming therapy and AF is sparse; however, preliminary findings in a US population study suggest no increase in AF incidence.¹⁰⁰

Androgen deprivation therapy

Androgen deprivation therapy (ADT), used to treat prostate cancer, also affects AF risk. A Norwegian longitudinal cohort study found a statistically significant increased risk of AF in patients receiving ADT for longer than 18 months,¹⁰¹ whereas a South Korean cohort study found no significant association.¹⁰² These differences likely reflect varying population demographics and comorbidity profiles, with the Norwegian ADT group being older with significantly elevated baseline cardiovascular disease, potentially inflating AF risk despite statistical adjustments.

The 2022 European Society of Cardiology guidelines highlight AF risk associated with 2 ADT medications—degarelix and abiraterone.¹⁰³ Relugolix, a gonadotropin-releasing hormone antagonist in the same class as degarelix, was not associated with AF risk.¹⁰³ In addition, abiraterone, a CYP17 inhibitor, was associated with significantly more cases of AF than other ADT medications.¹⁰⁴ This may result from an increased level of steroids such as aldosterone, which has been linked to AF at high levels.^105,106 Increased rates of AF were also observed in safety endpoints of 2 phase 3 trials of abiraterone and enzalutamide, although overall rates were low, and the studies were not specifically powered for AF outcomes.^107,108 Therefore, it is likely that the arrhythmogenic effects of ADTs, if any, are related to mechanisms beyond testosterone suppression.

In summary, although experimental evidence supports hormonal modulation of cardiac electrophysiology, population-level associations among estrogen, testosterone, and AF remain inconsistent (Table 1). Findings are similarly inconsistent in the subgroup of patients receiving hormone therapy, varying by formulation, dosage, timing, and patient characteristics, with a myriad of confounding factors complicating interpretation (Table 1). More robust population studies with repeated hormone measurements, greater control of confounding factors, and stratification by hormone formulation and ethnicity are needed to clarify the relationship between sex hormones and AF risk.

Table 1.

Results of population studies investigating the effect of estrogen and testosterone on AF risk in women and men

Effects of hormones on AF risk in general populations
Study authors	Study population	Hormone studied	Relationship with AF risk
Yang et al.70 (2019)	Women with natural menopause	Estrogen
Honigberg et al.71 (2019)	Women with premature menopause	Estrogen
Magnani et al.72 (2012)	Women with natural menopause	Estrogen	No effect
Magnani et al.87 (2014)	Men	Estrogen
O’Neal et al.85 (2017)	Men	Estrogen	No effect
Zeller et al.82 (2018)	Women	Testosterone
O’Neal et al.85 (2017)	Women	Testosterone	No effect
Berger et al.86 (2019)	Women	Testosterone	No effect
O’Neal et al.85 (2017)	Men	Testosterone
Berger et al.86 (2019)	Men	Testosterone
Zeller et al.82 (2018)	Men	Testosterone
Magnani et al.87 (2014)	Men	Testosterone
Rosenberg et al.89 (2018)	Men	Testosterone	No effect
Effects of hormone treatments on AF risk
Study authors	Study population	Hormone studied	Relationship with AF risk
Tsai et al.76 (2016)	Menopausal women on HRT	Estradiol
		CEE
Lee et al.75 (2021)	Menopausal women on HRT	Estradiol
		CEE
Perez et al.77 (2012)	Menopausal women on HRT	CEE
Calof et al109 (2005)	Men on TRT	Testosterone	No effect
Basaria et al110 (2010)	Men on TRT	Testosterone	No effect
Sharma et al.97 (2017)	Hypogonadal men	Testosterone
Forster et al.101 (2022)	Men on ADT	Testosterone
Kim et al.102 (2021)	Men on ADT	Testosterone	No effect

There is conflicting evidence on the association between estrogen or testosterone and risk of AF in either sex. The direction of statistically significant relationships is described with arrows.

ADT = androgen deprivation therapy; AF = atrial fibrillation; CEE = conjugated equine estrogen; HRT = hormone replacement therapy; TRT = testosterone replacement therapy.

Future work

Animal studies show that estrogen and testosterone alter ionic currents and APD in CMs.¹¹¹ However, given the inherent limitations of animal models, more future studies should focus on human-induced pluripotent stem cell–derived CMs (hiPSC-CMs). This approach will allow the development and investigation of sex-specific differences in both male and female cardiac models.^112,113 These cells are better than standard CMs for studying sex differences because they can be maintained in culture for a long time, allowing long-term monitoring of hormonal effects.^112,114,115

Interestingly, a recent study on hiPSC-CMs found that estrogen upregulated I_CaL in female CMs, contrasting previous animal model findings.^112,116 However, these hiPSC-CMs were ventricular CMs similar to the animal studies discussed earlier.¹¹² Given that aCMs have significantly different electrophysiological properties from ventricular CMs, these findings are not fully applicable to the effect of estrogen in the atria.¹¹⁷ Recent studies have found a method to develop hiPSC-aCMs by manipulating retinoic acid signaling.¹¹⁴ Therefore, further studies could examine estrogen and testosterone effects using these atrial models.

Despite their promise, hiPSC-aCMs still have notable limitations. Owing to their relative immaturity, they differ substantially from adult aCMs in both electrophysiological and structural properties. Conventionally cultured hiPSC-aCMs exhibit reduced functional expression of key ion channels such as Nav1.5 and I_to, decreased I_K1 density, and increased acetylcholine-activated K⁺ current activity, resulting in a lower action potential upstroke velocity and a more depolarized resting membrane potential.^114,118 Structurally, these cells display shorter, disorganized sarcomeres and lack the T-tubule network that facilitates efficient ECC in mature CMs.¹¹⁹

However, recent advances have significantly improved their physiological relevance. The groundbreaking development of 3-dimensional culture systems combined with precise electrical and mechanical stimulation regimes has been shown to enhance myofibril alignment, promote sarcolemmal maturation, and even induce T-tubule formation in hiPSC-aCMs, bringing their ultrastructural properties much closer to the adult cardiac phenotype.120, 121, 122, 123, 124 Although current hiPSC-aCMs may still not be able to fully replicate the complexity of adult aCMs, they offer an underexplored modality for investigating sex differences in human aCMs.

Conclusion

We found that animal studies show that estrogen prolongs APD and reduces cardiac contractility, whereas testosterone does the opposite. This suggests that testosterone may elevate the risk of AF, whereas estrogen seems to confer protection from an electrophysiological perspective. Importantly, most laboratory studies used animal ventricular CMs, so findings have limited applicability to humans. Future studies should use hiPSC-aCMs.

From a clinical perspective, our results highlight the importance of incorporating hormonal factors into AF prevention and management. Understanding how estrogen deficiency and testosterone excess contribute to AF risk can guide the development of personalized therapeutic strategies. Clinicians may even need to consider hormone levels when evaluating AF risk and tailoring treatments, particularly in populations with hormonal imbalances such as postmenopausal women and men with hypogonadism.

Finally, the current conflicting evidence from population studies underscores the need for further research. Limitations include inadequate control of confounding factors and reliance on singular measurements of dynamic hormones. Future studies should address these limitations through randomized controlled trials with periodic hormone measurements. Addressing these gaps could lead to more targeted and effective therapies, ultimately improving patient outcomes in AF management and reducing the disparities in AF outcomes between men and women.

Disclosures

The authors have no conflicts to disclose.