Atrial fibrillation‐related stroke in women: Evidence and inequalities in epidemiology, mechanisms, clinical presentation, and management

Abstract Background Atrial fibrillation (AF) is the most common clinical arrhythmia and one of the major causes of stroke, heart failure, sudden death, and cardiovascular morbidity. Despite substantial advances in (interventional) rhythm control treatment during the last decade, anticoagulation for stroke prevention remains a major component of AF treatment. Hypothesis There are important sex‐specific differences in AF‐related stroke, resulting from sex‐specific mechanisms and therapeutic differences. Methods This review summarizes available data on sex differences in risk assessment and prevention of stroke and highlights current knowledge gaps in AF‐related stroke mechanisms, prevention and management that warrant further research. Results Increased thrombotic risk in women is multifactorial, involving hormonal changes after menopause, structural, endocrine and lifestyle/social factors and their interactions. It is clear from randomized studies that women benefit from anticoagulant treatment and that their bleeding risk is similar to men. Women should therefore receive equivalent treatment to men, based on the validated criteria for anticoagulation therapy. However, women are not represented equally in the large randomized studies and sex‐related information in many fields is lacking. Conclusions Female sex is an established risk factor for stroke in AF patients. The evidence for sex‐specific differences in stroke risk assessment and stroke prevention is accumulating. However, the underlying biological mechanisms remain incompletely understood and further studies are required in order to decrease AF‐related morbidity and mortality.

Results: Increased thrombotic risk in women is multifactorial, involving hormonal changes after menopause, structural, endocrine and lifestyle/social factors and their interactions. It is clear from randomized studies that women benefit from anticoagulant treatment and that their bleeding risk is similar to men. Women should therefore receive equivalent treatment to men, based on the validated criteria for anticoagulation therapy. However, women are not represented equally in the large randomized studies and sex-related information in many fields is lacking.
Conclusions: Female sex is an established risk factor for stroke in AF patients. The evidence for sex-specific differences in stroke risk assessment and stroke prevention is accumulating. However, the underlying biological mechanisms remain incompletely understood and further studies are required in order to decrease AF-related morbidity and mortality.

| INTRODUCTION
Atrial fibrillation (AF) is the most common clinical arrhythmia and one of the major causes of stroke, heart failure, sudden death, and cardiovascular morbidity worldwide. 1 The incidence of AF has been rising progressively during the last decade as a result of longer total life expectancy, increased prevalence of comorbidities, lifestyle changes, and improved detection. 1 The association between stroke and AF increases significantly with age. Approximately 1/3 of total ischemic strokes and more than 50% of those in patients above 80 years of age are associated with AF. 1 Moreover, patients with AF who suffer an ischemic stroke appear to have a worse outcome, more disability, and greater mortality, than those who have an ischemic stroke in the absence of AF. 2 Oral anticoagulation (OAC) therapy is an important component of AF treatment, providing significant clinical benefit by preventing ischemic strokes and prolonging life. 1

| ATRIAL FIBRILLATION IN WOMEN
Although North American and European women at all ages have a lower prevalence of AF compared to men, all-cause mortality is higher and AF is independently associated with a 2-fold increase in risk of death in women compared to a 1.5-fold increase in men. 1 In the ATRIA study 3 the annual thromboembolism rate in patients not taking warfarin was 3.5% for women vs 1.8% for men. Females with additional stroke risk factors, particularly older age (>65 years), are at greater risk of stroke, even when adequately anticoagulated, whereas the bleeding risk on anticoagulation was similar in both sexes. 3,4 Women with AF are typically older, have more comorbidities than men, are more often symptomatic and experience worse quality of life (QOL) and more severe strokes. 3,4 European guidelines 1 and a recent review on sex differences in arrhythmias by the European Heart Rhythm Association (EHRA) 5 underline these sex-specific issues and recommend offering effective diagnostic tools and therapeutic options to women and men equally. However, in practice, females appear less likely to receive specialist care and when they do, a more conservative approach is followed. 1,6 The reasons for this are incompletely understood but appear to involve both health provider and patient characteristics. This is a review on sex differences in stroke risk assessment and prevention in patients with AF. It focuses on current knowledge gaps and areas in AF stroke mechanisms and prevention that warrant further research.

| METHODS
Medical literature was reviewed to identify articles investigating atrial fibrillation and stroke risk in terms of pathophysiology and anticoagulation in women. The OVID-MEDLINE, PubMed, and Cochrane CENTRAL databases were searched for English-language articles without time restriction. Further selection was based on abstracts and clinical relevance. When available, we focused on randomized controlled trials; if unavailable, important observational studies were used. Our review represents a selection of the main studies published on this topic and is not exhaustive.

| Mechanisms of increased AF and stroke risk in women
The association between female sex and stroke in AF is incompletely understood and a number of clinical and non-clinical factors have been proposed (Figure 1).

| Clinical and cardiometabolic risk factors
Several of the validated clinical predictors of thromboembolism differ between men and women. 3,4 Women with AF are on average older than male AF patients, in part due to longer life expectancy. In a large retrospective Swedish study 7 the risk of ischemic stroke was 1.5-fold higher in women compared to men but this association appeared to be the result of confounding by age. This study showed that in the low risk end, the CHA 2 DS 2 -VASc risk score underestimated the ischemic stroke risk conferred by age 65 to 74 years, while it overestimated the risk conferred by female sex.
In addition, women with AF are more often hypertensive but have less coronary artery disease. 3,4 Although age differences between males and females can be taken into account in the CHA 2 DS 2 -VASc scheme, 1 differences in disease severity and quality of treatment, which may also influence stroke risk, are not incorporated, since both hypertension and vascular disease are represented as binary variables.
Renal failure is also a strong predictor of stroke, but is not included in CHA 2 DS 2 -VASc and no differences between both sexes have been identified.
The association between cardiometabolic factors, AF-related stroke and sex is incompletely understood. Women with AF seem to be more obese than men but whether this truly increases stroke risk is not clear. 1,8 A correlation between serum high-density lipoprotein (HDL), total cholesterol and stroke has been suggested, although these findings are controversial and non-specific for AF. Surprisingly, in the INTERSTROKE study, 8 increased HDL was associated with a significantly increased risk of non-fatal stroke and cerebral infarction in women, whereas men showed an inverse relationship. Similarly, thyroid disorders with increased risk of stroke such as hyperthyroidism are more common in women. 9 It is however unknown whether females with thyroid disease have a higher thrombotic risk compared to men with thyroid disease. The risk for systemic embolization in the setting of thyrotoxicosis is not precisely known and issues regarding anticoagulation of patients with hyperthyroidism and AF remain unclear.

| Endogenous sex-related hormones
Natural estrogens have many functions apart from maintaining the female reproductive system. Estradiol, the most powerful estrogen has very potent endothelial effects that promote vasodilatation and blood flow through a direct effect on nitric oxide and prostacyclin. 10 Nitric oxide and prostacyclin, being potent vasorelaxants, inhibit platelet aggregation and adhesion. In the event of an ischemic stroke, estrogens increase energy reserves and decrease reactive oxygen species during reperfusion. Vasoprotection occurs through the preservation of the blood-brain barrier and the minimization of edema and inflammation. 10 Estrogens also have direct effects on platelet activity and on the coagulation system through factor VIII and von Willebrand Factor. 11,12 Menopause leads to a loss of these vasoprotective functions and to increased thrombogenesis as assessed through thromboxane, a product of activated platelets. 11,13 However, it is not clear whether menopause itself is an independent risk factor for ischemic stroke since higher thrombotic risk has been documented only for women with early menopause. 13 F I G U R E 1 Mechanisms of increased stroke risk in women

| Hormone replacement therapy (HRT)
In contrast to the protective effect of endogenous female hormones, estrogens prescribed as replacement treatment increase markers of hypercoagulability (eg, factor VIIa) and are known to increase thrombotic risk even in younger groups with additional risk factors. 14 Earlier studies 14 showed that HRT was associated with a higher risk of ischemic stroke. However, this has not been confirmed in subsequent studies. In a small group of women on estrogen replacement treatment in the ATRIA study 3 there was no increase in stroke risk during follow up. Similarly, in a recent sub-analysis of the AFFIRM study 15  There is accumulating evidence about the AF-promoting role of inflammation and oxidative stress through functional and structural atrial remodeling, for example, from patients with post-operative AF 20 or after catheter ablation procedures. 21 Several anti-inflammatory therapies have been evaluated, but a systematic decrease in AF burden has not been demonstrated. 21 However, the underlying molecular mechanisms, as well as the relationship between systemic indices of inflammation and local atrial inflammation remain incompletely understood. [20][21][22] Systemic CRP levels were independently associated with AF in men, but not in women. 22 On the other hand, recent data suggest that both female sex and IL-6 levels are associated with worse survival after stroke. 18 Accumulating preclinical data also suggest differential microRNA-mediated regulation of cardiac remodeling as well as responses to cerebral ischemia between males and females, although their role in AF-related stroke in female patients is still largely unknown. 23

| Autonomic dysfunction
Dysfunction of the autonomic nervous system (ANS) has been implicated in cardiac arrhythmogenesis as well as stroke. 24 Autonomic triggers can be sympathetic, vagal or mixed. Sympathetic stimulation can trigger AF by promoting ectopic activity. For example, in patients undergoing coronary artery bypass grafting, increased circulating noradrenaline levels and higher resting heart rates due to increased sympathetic tone can induce AF which responds well to peri-operative beta-blockade. 20,25 Vagal AF, on the other hand, refers to a subset of AF patients, who experience episodes in relation to established vagal triggers, including eating, alcohol, nocturnal symptoms or those associated with bradycardia or heart block. 24 Vagal AF has been suggested to occur primarily in athletes with structurally normal hearts and limited co-morbidities. 24 In a Euro Heart Survey substudy, autonomic triggers were found in approximately 1/3 of patients, 41% of them being women. 25 Surprisingly, vagal AF often occurred in elderly men whose age did not differ from that in the general population and with significant underlying disease. No further sex-specific analysis was provided in this study, but electrophysiological studies in healthy volunteers have identified male/female differences in the electrophysiological response to activation or blockade of the ANS, with more pronounced shortening of atrial refractoriness in response to ANS activation in women. Additionally, autonomic dysfunction, characterized by altered heart rate variability, was an independent determinant of AF episodes in type 2 diabetic patients 26 Women and men were equally represented in this study and no sex-specific differences were shown. Nonetheless, sex-specific AF-promoting mechanisms related to autonomic dysfunction remain an interesting topic for future studies.

| Symptoms, life style, and quality of treatment
Women with AF have increased perceived stress, lower QOL, and a higher tendency to develop depression compared to men. 27 As a result, they more often present with atypical symptoms that can confound the diagnosis and delay treatment. Interestingly, women with AF are more frequently prescribed antidepressants compared to men. 27 The multi-center randomized controlled trial SAFETY 28 showed that women, who were more likely to be older and alone,  high risk patients, in particular women, by physicians has been raised.
On the other hand, different studies have shown that although treatment is offered equally, women much more frequently opt for a conservative approach, especially in terms of rhythm management. 6,29 The AFFIRM trial 30 and the PINNACLE registry 6 showed lower quality in OAC treatment in women, suggesting that compliance might be an issue, although this was not confirmed in the ORBIT-AF registry, 29 which did not show differences in coagulation rates or times in therapeutic range.
How relevant these issues are in every day practice and whether these differences are due to the healthcare providers, delayed diagnosis, or patients' preferences and compliance represents a real gap in our knowledge. ized ratio (INR) range (40 ± 0.7% vs 37 ± 0.5%, P = .0001) and even more importantly, they were more often sub-therapeutic than men (29 ± 0.7% vs 26 ± 0.5%; P = .0002). Moreover, when patients with a fairly high time in therapeutic range (≥66%) were compared, 29,30 as in the ORBIT study 29 women still had a higher residual risk of stroke (log rank P = .009) and female sex remained an independent risk factor.

| NOACs
Landmark NOAC studies 32,33 did not primarily focus on sex differences and women were underrepresented (37% of 42 411 patients in total, were associated with reduced risk of heart failure admissions and allcause mortality in both sexes. The ARISTOTLE trial showed similar rates of thromboembolism in the apixaban and warfarin groups between men and women but less clinically relevant bleeding in women. Although firm data is lacking and studies with direct comparisons of NOACs in women and men are not available, the existing information indicates that NOACs can be used equally in women and men.

| Cardioversion and ablation
Controlled data and evidence on sex-related outcomes with regard to cardioversion and ablation are scarce and sex differences are rather absent. There are few randomized controlled trials (RCTs) on cardioversion and women are underrepresented in most (Table 3). 37

| Acute coronary syndromes-Percutaneous interventions and major surgery
Although women are underrepresented in randomized clinical trials, they are believed to be more susceptible to cardiovascular events with higher mortality and increased bleeding complications after PCI and full antiplatelet treatment (Yentl syndrome). However, in the four major studies, RE-DUAL-PCI, 43 PIONEER AF-PCI, 44 WOEST, 45 and ATLAS ACS II-TIMI 51 trial, 46 in which 20% to 25% of patients were female, no sex differences were identified in terms of safety and efficiency of dual or triple therapy ( cross the placenta and are therefore discontinued after the first trimester. 1 Information on potential consequences of NOACs during pregnancy is sparse. Currently, only one observational study is available assessing risks in women inadvertently exposed to rivaroxaban during early pregnancy, but the small cohort does not allow to refute an increased malformation risk. Beyer-Westendorf et al, 47 recently reviewed 233 unique cases of NOACs in pregnancy. Outcome was available in 137 (58.8%) cases. Seven (5.1%) showed abnormalities from which 3 (2.2%) could be potentially interpreted as embryopathy.
The authors concluded that data was incomplete and women exposed to NOACs should therefore not be directly advised to terminate pregnancy but change to LWMH.

| Chronic kidney disease
Limited data are available for OACs in patients with renal disease and sex-specific information is lacking.

| OUTCOME AFTER STROKE
Women with AF present more frequently with stroke compared to men and have worse outcome due to multiple factors and characteristics, including smaller body mass index and older age. 1 Intracranial hemorrhage is the most feared complication of OAC treatment and is associated with high mortality and morbidity. In patients with previous intracranial hemorrhage, restarting OAC treatment as early as clinically feasible, was associated with significantly lower risk of ischemic stroke and all-cause mortality, without statistically significant risk of recurrent bleeding. 1 Unfortunately, these patients are usually not included in RCTs, and the problem of the best time window for restarting the OAC after intracranial hemorrhage, taking into account potential sex differences, remains an issue of debate.

| LIMITATIONS
This review includes multiple studies of non-homogenous design and representation of females (randomized non randomized and observational). Females are underrepresented in most studies and drawing conclusions incorporates a bias risk. In addition, a major limitation of current studies is the fact that most of them were conducted in Europe or North America. To what extend sex-specific aspects are relevant for diagnosis and treatment in other parts of the world is at present largely unknown.

| CONCLUSION
Female sex is an established risk factor for stroke in AF patients. The evidence for sex-specific differences in stroke risk assessment and stroke prevention, especially OAC treatment is accumulating. However, the underlying biological mechanisms remain incompletely understood and there are many gaps in evidence that need to be addressed. Advancing the knowledge in this field will likely help to decrease AF-related morbidity and mortality and to further personalize treatment.