Does Testosterone Therapy Increase Risk of Atrial Fibrillation? Appraisal of the Contemporary Scientific and Clinical Evidence

Abstract

Here we review the evidence whether “testosterone” (T) therapy in hypogonadal men contributes to onset and increases the incidence of atrial fibrillation (AF). A MEDLINE review of the literature was performed, and relevant studies were analyzed and discussed. The effect of T therapy (TTh) on the incidence of AF is a controversial clinical issue that remains hotly debated. Few studies have purported that TTh increases the incidence of AF. However, a closer look at these studied suggest that several confounding factors may have contributed to the authors’ conclusion that T increases the incidence of AF. On the contrary, a much larger number of studies have demonstrated that T deficiency contributes to incidence of AF and TTh does not increase the incidence of AF. Because this clinical issue is of critical importance in management of men with hypogonadism, it is critical that more clinical studies with AF as the primary outcome measure need to be considered. However, to date, none of the clinical trials have focused on AF as the primary outcome and in most of the reported studies AF was a secondary or tertiary outcome measure. Thus, lack of accurate clinical data on this issue will remain in the helm of debate, until new clinical trial data becomes available. We conclude that the findings of numerous studies suggest that low circulating levels (T deficiency) are associated with greater incidence of AF and higher endogenous T levels or TTh in hypogonadal men are associated with lower incidence of AF.

INTRODUCTION

Atrial fibrillation (AF) is a major public-health concern associated with increased mortality and morbidity including stroke, heart failure (HF), and cognitive decline. AF is a multifactorial and heterogeneous disease. Among the many factors contributing to development of AF are hypertension, HF, valve disease, coronary artery disease, obesity and congenital heart disease. AF may be the result of multiple structural, electrophysiologic, ionic, and molecular changes related both to aging itself and to aging-related comorbidities and cardiovascular risk factors, such as hyperlipidemia, hypertension, testosterone (T) deficiency, metabolic syndrome, diabetes, sarcopenia and frailty. The underlying molecular mechanisms of AF onset and persistence in hypogonadal men are poorly understood. We performed a MEDLINE review of the literature and relevant studies were analyzed and discussed.

T is a sex steroid hormone that plays an important role in maintaining cardiomyocytes, vascular smooth muscle growth and function. T regulates the expression and activity of endothelial nitric oxide synthase (eNOS) and downregulates the nuclear factor kappa-light-chain-enhancer (NF-κB) pathway, thus producing anti-inflammatory, antifibrotic, and vaso-protective effects further maintaining atrial structural and functional activity. T is thought to be important in maintaining the growth and function of the myocytes and its deficiency may contribute to vascular disease. T deficiency is thought to increase atrial fibrosis, impaired gap junctional communication and produces an imbalance in its autonomic activity and in turn contributing to onset of AF. Recent studies have produced conflicting results with some suggesting that T contributes to the incidence of AF and others argued that T is protective. In this review we summarize the contemporary finding of the clinical literature and discuss the effects of T deficiency and T therapy (TTh) on the onset of incidence of AF.

METHODS

We should point out from the outset that this is a narrative review of the literature and not a systematic review. A MEDLINE search was performed for the literature review, with key words including “testosterone,” “cardiovascular disease,” “atrial fibrillation,” “arrhythmia.” Additional review was performed of articles that postulated that higher T levels may increase the risk of AF incidence with TTh. Studies were selected based on the findings that T levels or T treatment may reduce or increase the incidence of AF and then categorized as those postulating that T increases AF incidence vs. those that propose that TTh reduces AF incidences.

1. Does TTh contribute to higher incidence of AF?

To shed light on this critical question, we have reviewed a host of studies in the contemporary clinical and scientific literature that have attempted to address this key question. The relationship between T levels and onset of AF appears to be complex and very controversial. Below we present summary of clinical studies suggesting higher T levels or TTh may be linked to increased incidence of AF. We present a summary of clinical studies that argue that T deficiency increases the incidence of AF and TTh may be protective. We also summarize some of the findings from the basic science studies in animal models that demonstrate T deprivation contributes to higher incidence of AF.

1) Studies suggesting that TTh or high circulating T levels increase the risk of AF

A recent study by Bonnet et al [1] using data from the TriNetX Research Network compared 117,908 cis men with hypogonadism treated with T with 1:1 propensity-score matched cis men not treated with T (Table 1). After 5 years of follow-up, cis men treated with TTh had a lower risk of myocardial infarction with no differences for stroke or mortality. However, these men exhibited higher risks of AF and acute pulmonary embolism/deep vein thrombosis (APE/DVT). The authors suggested that TTh increases the onset of AF.

Table 1. Clinical studies claiming higher T levels or TTh in hypogonadal men increase the risk of AF.

Study	Type of study	No. of subjects	Duration of follow-up (y)	Comments
O’Neal et al, 2017 [9]	The Multi-Ethnic Study of Atherosclerosis	4,883	10.9	Higher levels of endogenous bioavailable T contribute to AF development in men
Berger et al, 2019 [8]	The prospective Atherosclerosis Risk in Communities (ARIC) study	9,282	2	Higher plasma total T concentration was associated with a modestly greater incidence rate of AF in men
Lincoff et al, 2023 [4]	Clinical trial RCT	5,200	4	TTh is associated with increased incidence of AF
Tran et al, 2024 [7]	Post-hoc analysis participants in the ASPirin in Reducing Events in the Elderly (ASPREE)	4,570	4.4	Circulating T concentrations within the high-normal range are independently associated with an increased risk of incident AF amongst healthy older men
Bonnet et al, 2025 [1]	Retrospective study	117,908	5	T treatment in cis men with hypogonadism was associated with a higher risk of AF

T: testosterone, TTh: T therapy, AF: atrial fibrillation, RCT: randomized clinical trial.

Close examination of the findings of this study, however, raised several concerns and suggested that confounding factors may have contributed to their findings. For example, the study did not account for variables that may contribute to AF such as comorbidities including diabetes, hypertension, chronic obstructive pulmonary disease, obstructive sleep apnea, periodontal disease, congestive HF, peripheral vascular disease, coronary artery disease, hyperthyroidism, rheumatic valve disease, nonrheumatic valve disease, structural heart disease, and cardiomyopathy; baseline body mass index; low-density lipoprotein; and use of aspirin, β-blockers, and statins. Since many of these comorbidities contribute to incidence of AF, this study did not provide critical information regarding the role of TTh on the incidence of AF. In addition, as acknowledged by the authors “We cannot determine how the diagnosis of hypogonadism was done, and we did not have the dose of T, the way of administration nor the adherence to TTh by each participant over the follow-up” [1]. Therefore, the findings of this study do not unequivocally conclude that TTh contributes to higher incidence of AF.

Furthermore, this study did not provide laboratory T values and no data on patients’ adherence to TTh or if the TTh normalized physiological T levels as discussed by Sharma et al [2], it is clear that this study lacks critical information necessary to ascertain the relationship between TTh and incidence of AF. In addition, the authors concluded that TTh in transmen did not increase the risk of AF and went further to state that transmen treated with TTh had a lower rate of AF than untreated cis-males and lower APE/DVT rate than cis-females using oral contraceptives. It should be noted that transmen are biologically women (female-to-male transsexuals); thus, unless the pathophysiology of AF is different in women and men, this conclusion is considered baseless and at best unwarranted. Interestingly, however, a recent study by Bonnet et al [3] reported that TTh in diabetic men was associated with reduced risk of AF (hazard ratio [HR] 0.91, 95% confidence interval [CI] 0.85 to 0.98, p=0.01) instead of increased risk of AF as they reported previously. Thus, it is difficult to reconcile such findings by the same group of investigators.

Data from the recent TRAVERSE trial [4] had also indicated that TTh may increase the risk of AF (Table 1). Although the incidence of AF in the T arm was 91 (3.5%) compared to 63 patients (2.4%) in the placebo group, this increased risk, albeit small, was attributed to TTh. It should be noted that in the TRAVERSE trial, in men with or without coronavirus disease (Covid) showed no significant differences between T treated and the placebo group. In men with Covid AF was noted in 3/178 in the T arm and 3/180 in the placebo group. Similarly, in men who did not get Covid, but received T treatment there were 32 cases of AF out of 2,406 (1.4%). In the placebo group in men without Covid, there were 26 cases of AF out of 2,410 (1.1%). This shows only 6 more cases of AF in the T treated group when compared with the placebo group [5]. It is unclear why the values provided in these two publications by the same group of authors differed significantly form one another. These discrepancies do not align with original findings of Lincoff et al [4] suggesting that such cases may be due to chance alone. In addition, it should be noted that in this trial, no precise information was provided on patients’ adherence to TTh or achieving normal physiological levels of T over the entire course of treatment. The fact that a total of 61.4% of the patients in the T group discontinued TTh, and 61.7% of the patients in the placebo group discontinued placebo, suggests poor but normal compliance for a study of this sort. Furthermore, since all patients in this trial had either vascular disease or a history of vascular disease, it is not surprising that some of these patients are likely to have AF over the course of follow-up period, since many of the vascular disease risk factors contribute to onset of AF [6].

Tran et al [7] analyzed findings from the ASPirin in Reducing Events in the Elderly (ASPREE) study with 4,570 participants with mean follow-up of 4.4 years and suggested that higher circulating T levels are independently associated with an increased risk of AF in healthy older men. Similarly, data from the prospective Atherosclerosis Risk in Communities (ARIC) study [8] (Table 1) identified incident AF among 9,282 participants who had plasma total T measured by liquid chromatography tandem mass spectrometry at Visit 4. After a mean of 13.7 years of follow-up, 1,664 incident cases of AF were identified. There was a positive association between plasma total T and incident AF in men. It was concluded that a higher plasma total T concentration was associated with a modestly greater incidence rate of AF in men.

O'Neal et al [9] analyzed 4,883 participants from the Multi-Ethnic Study of Atherosclerosis. Over a median follow-up of 10.9 years, higher bioavailable T levels were associated with increased AF risk in men. The authors concluded that higher levels of endogenous bioavailable T contribute to AF development in men. Based on the aforementioned studies it appears that higher T levels in circulation or TTh in men with T deficiency may increase the incidence of AF.

2) Studies suggesting that T deficiency and not higher T levels increases the risk of AF

On the contrary to the studies discussed above, Magnani et al [10] identified a significant interaction between age and T, in a cohort of 1,251 men of whom 275 developed incident AF and therefore stratified men into age 55 to 69 (n=786), 70 to 79 (n=351), and ≥80 (n=114). In men 55 to 69 each 1-standard deviation (SD) decrease in T was associated with HR 1.30 (95% CI, 1.07 to 1.59) for incident AF (Table 2). The association between T and 10-year incident AF in men 70 to 79 did not reach statistical significance. However, in men ≥80 years a 1-SD decrease in T was associated with HR 3.53 (95% CI 1.96 to 6.37) for AF risk. It was concluded that T is associated with incident AF in a cohort of older men. In study assessing the link of sex hormones with AF, it was suggested that T deficiency is associated with increased atrial fibrosis, impaired connexin-mediated gap junctional communication, and heightened autonomic imbalance, all of which promote conduction heterogeneity and re-entry circuits that facilitate AF initiation and maintenance [11].

Table 2. Studies demonstrating that lower T levels increase the risk of AF and TTh is not associated with increased incidence of AF.

Study	Type of study	No. of subjects	Duration of follow-up (y)	Comments
Lai et al, 2009 [20]	Prospective clinical study	58		T levels were significantly lower in subjects with lone AF when compared with controls
Magnani et al, 2014 [10]	The Framingham Heart Study	1,251	10	T deficiency in men ≥80 is strongly associated with AF risk
Sharma et al, 2017 [2]	Retrospective study	76,639		Normalization of total T levels after TTh is associated with a significant decrease in the incidence of AF
Zeller et al, 2018 [12]	Prospective clinical study	3,876	13.8	Low T levels are associated with increased risk of future atrial fibrillation
Rosenberg et al, 2018 [13]	Prospective clinical study	1,019	9.5	Low free DHT is associated with an increased risk of incident AF
Hasegawa et al, 2021 [23]	Prospective clinical study	149		Medical castration prolongs the QT/QTc intervals in most patients with prostate cancer, and it could cause TdP/VFs even in patients with no risk of QT prolongation before the therapy
Hu et al, 2022 [21]	Meta-analyses			Total T and was not associated with AF risk
Cho et al, 2024 [14]	Retrospective study	179,988	11.8	Low T levels are associated with increased risk of AF
Xu et al, 2024 [17]	Retrospective study	173,498	11	Lower levels of T and cFT were associated with increased risk of AF. Higher SHBG levels were associated with increased risk of AF
Greenberg et al, 2024 [18]	Retrospective study	2,134		TTh is not associated with a significantly increased risk of new onset of AF
Sood et al, 2024 [19]	Meta-analyses	10,941		TTh is not associated with increased risk of atrial fibrillation, stroke, pulmonary embolism, heart failure

T: testosterone, AF: atrial fibrillation, TTh: T therapy, DHT: dihydrotestosterone, TdP/VFs: torsades de pointes/ventricular fibrillations, cFT: calculated free T, SHBG: sex hormone binding globulin.

Sharma et al [2] investigated a national cohort of 76,639 veterans with low total T (TT) levels who were divided into 3 groups. Group 1 had TTh resulting in normalization of TT levels (normalized T), group 2 had TTh without normalization of TT levels (nonnormalized T), and group 3 did not receive T (no T) (Table 1). Normalization of T levels after TTh is associated with a significant decrease in the incidence of AF. This is contrary to the studies cited above and suggests caution should be exercised prior to assigning a false premise that TTh contributes to increased AF risk.

Zeller et al [12] reported on a population cohort of 7,892 subjects (3,876), aged 25 to 74 years, with median follow-up of 13.8 years, indicating that increasing levels of T were associated with lower risk of AF in men and noted that low T levels are associated with increased risk of AF in men (Table 2). In a study of 1,019 men with median follow-up of 9.5 years in which plasma dihydrotestosterone (DHT) levels were measured, Rosenberg et al [13] reported that 304 (30%) men developed AF (Table 1). A non-linear association with risk of incident AF was noted with both free and total DHT, in which subjects with the lowest levels of DHT had a higher risk of incident AF. The authors concluded that among older men, low free DHT is associated with an increased risk of incident AF. More recently, Cho et al [14] reported on data from the UK Biobank, with a population-based prospective cohort, focusing on 179,988 men who had complete serum TT level data and no AF history (Table 2). During the median follow-up of 11.8 years the authors noted 13,847 incidents of AF (7.7%). Participants with normal TT levels (≥300 ng/dL) showed a 15% reduced risk of AF compared to those with low TT (<300 ng/dL). Furthermore, each 1-SD increase in TT levels was associated with a reduced AF risk. These findings suggest that low TT levels are associated with reduced left atrial (LA) function and increased risk of AF.

In a retrospective observational study, in which 133,584 with hypogonadism were included after propensity matching who were prescribed TTh vs. no T treatment. Blackwell et al [15] reported a decreased risk of AF (3.6% vs. 4.0%; relative risk 0.900; p<0.001) for patients treated with T. In a systematic review and metanalysis, Corona et al [16] reported that out of 8,126 patients with hypogonadism treated with TTh and 7,310 patients who remained untreated (placebo), the incidence of non-fatal arrhythmias and AF was increased only in one study. In addition, the authors pointed out that “both non-fatal arrhythmia and AF were not trial endpoints of the TRAVERSE study but only investigator-reported adverse events, and hence they should be interpreted with caution.” This, however, was not confirmed when all other studies were considered. Furthermore, no relationship between endogenous T levels and AF incidence was observed after the adjustment for confounders.

Xu et al [17] investigated a cohort of 173,498 men (aged 37 to 73 years, followed for 11 years) from the UK Biobank. Among these men, 11,368 had incident AF. Compared with the third quartiles, the lowest category of serum T was associated with increased risks of AF (Table 2). Similar associations were found between calculated free T (cFT) values and AF events. Interestingly however, higher levels of cFT were associated with increased risks of AF. In addition, higher SHBG concentrations were associated with increased risks of AF. Therefore, lower T concentrations were an independent risk predictor for incident AF in men.

A recent study utilizing data from the TriNetX Research Network compare hypogonadal men who had a prescription for topical TTh with men who did not [18] (Table 2). There were 2,134 men included in each cohort after propensity score matching. The findings of this study suggested that TTh was not associated with a significantly increased risk of new-onset AF. Finally, a meta-analysis of randomized clinical trials by Sood et al [19] demonstrated that TTh was not associated with a statistically increased or decreased risk of AF (Table 2). It should be noted that data from several other observational studies [13,20] and meta-analyses [14,19,21] also suggested that low T is a contributor to incidence of AF in older men.

3) Findings from animal studies suggesting that T deficiency increases the risk of AF

Findings from animal studies by Tsuneda et al [22] had demonstrated that repetitive atrial responses induced by electrical stimuli significantly increased in castrated rats without changes in other electrophysiological properties and were abolished by administration of T. In men, medical castration prolongs the QT/QTc intervals in most patients with prostate cancer treated with androgen deprivation therapy and it could cause torsades de pointes/ventricular fibrillation (TdP/VF) even in patients with no risk of QT prolongation before the therapy [23]. Thus, it appears that physiological T levels do not increase the incidence of AF while low T contribute to the pathophysiology of AF in animals and humans. Similar studies by Tsai et al [24] have suggested that T deficiency plays a role in the pathophysiology of AF. Androgen regulates cardiac electrophysiology and calcium (Ca²⁺) homeostasis, and findings from androgen receptor knock out (ARKO) animal model demonstrated that that loss of the androgen receptor alters atrial electrophysiology with increased atrial arrhythmogenesis. Furthermore, Zhang et al [25] demonstrated that castration in male mice increased the incidence of AF. The number of AF episodes in castrated mice increased by 13.5-fold when compared with sham mice. The authors demonstrated that androgen treatment alleviated the occurrence of AF. The authors concluded that TD contributed to the increased late sodium current, prolonged action potential repolarization, and increased susceptibility to AF.

2. Do endogenous vs. exogenous T levels have significantly different effects on the incidence of AF?

It is paramount to distinguish between the effects of endogenous circulating T levels vs. those assessed after TTh on the incidence of AF. Several population studies have reported on the incidences of AF in relation to low and high T levels [1,3,7,8,9,10,11,12,13,14,17,20]. A common denominator was that low T levels are often associated with increased incidence of AF. Interestingly, however, few population studies have suggested that high circulating T levels may also be associated with increased incidence of AF. This has led to postulating the non-linear U-shaped relationship between T levels and incidence of AF. Moreover, several clinical studies in which the effects of TTh on the incidences of AF were assessed, similar observations regarding the incidence of AF were noted [2,4,5,18,19]. The fact that low T is associated with increased risk of AF is a common observation in most studies. However, the effects of high T levels endogenously or exogenously derived, is still open for debate. Indeed, to address this real question adequately, additional research is needed to evaluate such pharmacodynamics of endogenous and exogenous T on the physiology and pathophysiology of the cardiovascular system.

Indeed, a non-linear, U-shaped relationship between serum T levels as well as free T and incidences of AF has been postulated and may in fact exist. Such conceptual framework may have some merit based on the evidence in the contemporary literature. However, at the present time, it is not feasible to establish a clear-cut clinical “therapeutic window” with an established lower cut-off and an upper cut-off T value that would be used clinically to manage patients and reduce the incidence of AF. Such therapeutic window would have to consider (a) the pharmacogenetics of individuals and (b) the differences in the androgen receptor CAG repeats which may contribute to variations that affects this non-linear relationship. For this reason, we feel unjustified, at present, to put forth a hypothetical and impractical “therapeutic window” that may be flawed at best and would have no clinical value.

It should be noted that Mason et al [26] argued that the presence of comorbidities and factors connected to lifestyle risk factors themselves affect the incidence of AF and makes it particularly complex to determine the exact mechanisms underlying the higher incidence of AF in men. It should be noted that normalizing T levels in hypogonadal men has been associated with a decrease in AF incidence [2]. Furthermore, the under-reporting of individual T levels in patients who developed AF in most of the studies and the relationship between T levels achieved under therapy and incidence of AF poses an open and difficult question. Because of the aforementioned reasons, we would hesitate to introduce a “therapeutic window” that would be applicable to a very diverse and heterogenous groups of patients that may have little or no clinical value.

SUMMARY AND CONCLUSIONS

Based on the findings of the contemporary clinical literature, T deficiency is a real contributor to increased risk of AF in older men and normalization of T levels with TTh may decrease the rates of AF [2]. However, the impact of TTh on the onset or incidence of AF remains equivocal and requires further studies. The studies described in Table 1 advocating that higher circulating T or TTh increase the risk of AF remind us of the purported studies on the adverse effects of TTh on increasing the risk of cardiovascular disease [27,28,29,30] which were proven false by findings from the TTrials [31] T4DM trial [32] and the TRAVERSE trial [4]. The findings of the studies described in Table 2 suggest that T deficiency and not TTh may be associated with cardiometabolic co-morbidities and in turn may contribute to the increased incidence of AF in older men.

We should note that a key unresolved issue is not simply whether “TTh” increases or had no effect on the incidence of AF, but rather a host of variables come into play including achieving and maintaining physiological levels of T subsequent to treatment. At present, we do not fully understand the underlying molecular mechanisms as suggested recently by Mason et al [26]. For example, in studies evaluating sex differences of AF, data are often collected from population cohorts that are largely heterogeneous in terms of varying risk factors, especially as these confounding factors often increase in number and severity with age. The presence of comorbidities and factors connected to lifestyle risk factors for AF makes it particularly complex to determine the likely mechanisms underlying the higher incidence of AF in men, irrespective of T levels.

However, the findings from the various studies may be reconciled via a U-shaped relationship in which high and low T levels may promote arrhythmogenic remodeling through potentially distinct mechanisms. As proposed by Mason et al [26] low T levels may result in reduced binding of FKBP12.6 to Ryanodine Receptor type 2 (RyR2), which could induce calcium leakage from the sarcoendoplasmic reticulum. On the other hand, supraphysiological T levels may induce reduction in the inward rectifier potassium current (IK1) thus, contributing to electrical pathophysiology and arrhythmia [26].

The relationship between circulating T levels or TTh and increased incidence of AF remains controversial and hotly debated. It is important to ask the key question: Do normal physiological or higher T levels represent a risk for AF? If this were the case, then we expect younger men between ages of 18 and 60 years old to be at the highest risk of AF than older men, since they have normal or high physiological T levels compared to older men. However, the incidence of AF is higher in older men where T levels are often low compared to younger men. Therefore, in absence of real scientific and clinical evidence for the etiology and the pathophysiological mechanisms of T action on the onset of AF, these claims remain unjustified and at, best, anecdotal. It must be noted that almost all studies suggesting that TTh increases the risk of AF do not provide key information on T levels or patients’ adherence to TTh and other known co-morbidities that contribute to onset of AF. Hence, it is our opinion that adequate TTh (achieving physiological T levels with good patient compliance) either has no effect or a beneficial effect in reducing incidence of AF.