• Atrial fibrillation
• Epidemiology
• Stroke
• Atrial arrhythmia ablation procedures

The deleterious health effects of atrial fibrillation (AF), including impaired quality of life and significantly increased risks of stroke, heart failure and all-cause mortality, can be attenuated using the therapies for AF symptoms management and/or reduction in adverse cardiovascular outcomes. As certain consistently reported sex-related differences in the epidemiology, pathophysiology, clinical presentation and prognosis of AF (table 1)1–3 may affect ultimate effectiveness of AF treatment, these differences should be well appreciated in the personalised, individual patient-centred approach to AF management in clinical practice.

In this issue, Schnabel et al 4 describe sex-related differences in clinical presentation and 1-year outcomes in a prospective industry-sponsored registry of 6412 patients with AF (39.7% female) across seven Western European countries (France, Germany, Austria, Switzerland, Italy, Spain and the UK), mostly managed by cardiologists (89% of study participants). The study addressed a number of clinically relevant sex-related issues in patients with AF.

For instance, the extensively debated contribution of female sex to AF-related stroke risk has been eventually acknowledged in most of the international AF guidelines recommending the CHA₂DS₂-VASc score (where Sc is for sex, and being female scores 1 point) as the tool for stroke risk assessment. Although numerous observational studies of different patient populations with variable follow-up inconsistently reported female sex as a multivariable AF-related stroke risk factor (as systematically reviewed in a recent paper),3 the overall balance of evidence suggests that female sex is an independent thromboembolic risk factor. In the Swedish nationwide non-anticoagulated AF cohort (n=1 00 802), for example, female sex was associated with an 18% greater risk of stroke compared with male sex on a multivariable analysis adjusting for 35 cofactors (HR: 1.18; 95% CI, 1.12 to 1.24).5

However, available evidence suggests that there is an age dependency of the female sex-related stroke risk. Although the risk tends to be higher in females compared with males in all age groups, the difference between sexes reaches statistical significance only at the age of ≥65 years (eg, in the Swedish cohort, the annual stroke rates in female and male patients with AF <65 years old without additional stroke risk factors were 0.7% and 0.5%, respectively; p=0.09). Although females with AF and no additional stroke risk factors will have the CHA₂DS₂-VASc score of 1 by virtue of their sex, they have such low annual stroke rates (as observed in multiple studies) that the use of oral anticoagulant therapy (OAC) is not warranted.

Several potential mechanisms have been proposed to explain the observed higher risk of AF-related stroke in females, including sex-related differences in the vasculature and myocardial structure (which may predispose to alterations in blood flow, shear stress and altered endothelial function), or a potential sex-related increase in systemic inflammatory and procoagulant markers, thrombogenic particles, and platelet aggregation (especially in postmenopausal females).6 Although sex-related differences in stroke risk and outcomes may result from variations in stroke risk profile, management or underutilisation of anticoagulant therapy in females with AF, it has been shown that an increased stroke risk in females persists despite baseline anticoagulant use.3 6 The study by Schnabel et al 4 showed no significant sex-related difference in thromboembolic events, likely due to a relatively short follow-up, small total number of events and equal utilisation of antithrombotic therapies among male and female patients.

Indeed, along with sex-related differences listed in table 1, increasing evidence shows that the utilisation and outcomes of AF-directed therapies may differ in males compared with females (see table 1).1 2 7 8 For example, multiple observational studies reported that female patients with AF were less likely to receive rhythm control therapies compared with males. Among those receiving a rhythm control treatment, females were less likely to undergo electrical cardioversion or AF catheter ablation, the reason of which may be related to a higher frequency of AF ablation-related vascular complications, bleeding or pericardial effusion, and recurrent AF post ablation.8 Further studies are urgently needed as detailed analysis of these gender differences are lacking and may merely reflect a selection of a sicker female patient population related to the reluctance to refer female patients for AF ablation unless severely symptomatic.

However, in contrast to earlier observations on the lower use of OAC in females with AF, more recently published reports show that the sex-related differences in the utilisation of OAC for stroke prevention have been diminishing,1 2 7 8 but there may be some differences in the treatment outcomes, such as a higher residual stroke risk in female patients taking warfarin or better safety of non-vitamin K oral anticoagulants in female patients with AF (see table 1). Of note, the new European Society of Cardiology (ESC) AF Guidelines, published in 2016, changed the decision algorithm for the use of oral anticoagulation, using separate risk stratification scoring for female and male patients with AF and stating that OAC is indicated for males with CHA₂DS₂-VASc score of 2 but requiring the score of 3 for females (and should be considered in males with a score of 1 and females with a score of 2). Because this way female gender as such is no longer regarded as a risk factor, this new approach with higher CHA₂DS₂-VASc scores for females could lead to an inherent risk for under-treatment of female patients with AF.

The study by Schnabel et al 4 broadly confirms the observations discussed above, including sex-related differences in the utilisation of healthcare resources, such as electrical cardioversion and AF ablation. However, the study does not strictly differentiate between true sex-related disparities as opposed to possible biases in the healthcare use.

In general, when assessing sex-related differences beyond the observational level, the first step would be to establish the clinical equity approach (ie, to provide equal treatment based on the corresponding clinical need regardless of the sex of the patient). Thereafter, it should be documented that observed difference(s) associated with worse clinical outcome(s) are not due to patient-related factors such as eligibility, the presence of contraindications to treatment, patient preferences or other confounding, and only then the difference can be regarded as a true sex-related disparity. Otherwise, observed treatment differences should be referred to as stereotypes (ie, biases) and/or unconscious biases, whereas variations in treatment, which are observed in the absence of evidence for worse clinical outcomes or without adjusting for patient-related factors, should be termed sex-related differences rather than disparities.3

Unfortunately, female individuals have generally been under-represented in randomised AF trials. Randomisation best diminishes potential bias, but stringent inclusion and exclusion criteria in randomised trials (which are necessary for obtaining the high-quality scientific evidence) may affect the ‘real-world’ applicability of the trial findings. On the other hand, observational studies of sex-related differences in AF often do not (or cannot) provide the differentiation between true sex-related disparities and biases in healthcare use.

Hence, as Schnabel et al 4 rightly noted, further sex-specific research in AF is very much needed, and better understanding of mechanisms underlying sex-related differences among patients with AF may provide the opportunity to develop sex-specific preventive and therapeutic strategies for AF management, thus improving the ultimate outcomes.