Gender-Related Differences In Treatment Patterns And Outcomes of Patients With Atrial Fibrillation: Insights From The MISOAC-AF Trial


 PurposeTo assess the gender-related differences in the treatment patterns of patients with atrial fibrillation (AF), and their prognostic value.MethodsIn this post-hoc analysis of a randomized controlled trial, 1140 hospitalized patients with comorbid AF were followed-up for a median of 2.6 years. Kaplan-Meier and multivariable Cox-regression analyses assessed the adjusted hazard ratios (aHRs) for outcomes in males and females, according to oral anticoagulation (OAC) type (vitamin K antagonist or non-vitamin K antagonist oral anticoagulants), rhythm or rate control treatment. The primary outcome was all-cause mortality and the secondary outcomes were stroke and the composite of any hospitalization or cardiovascular death. ResultsAmong 622 males and 518 females, use of OAC (61% vs 62%), rate control (56% vs 57%), and rhythm control (31% vs 28%) treatments was similar (all p>0.05). In males, use of rate control, as compared with rhythm control, was independently associated with higher rates of all-cause mortality (aHR=2.06; 95% confidence interval [CI] 1.24-3.41) and the composite of hospitalization or cardiovascular death (aHR=1.34, 95% CI 1.01-1.85). In females, use of rhythm control was significantly associated with higher rates of hospitalization-or cardiovascular mortality (aHR=1.74, 95% CI 1.03-2.94). Among genders, stroke rates were similar regardless of OAC type, rate or rhythm control treatment.ConclusionsIn patients discharged from the hospital with comorbid AF, the use of OAC, rhythm or rate control treatment was similar among genders. However, males seemed to benefit more from rhythm, whereas females from rate control treatment.


Introduction
Atrial brillation (AF) has developed into a true 21st -century pandemic among both female and male individuals, claiming a disproportionate amount of the physicians' devotion [1,2]. Treatment strategies in these patients mainly include oral anticoagulation (OAC) agents, in combination with a rate or rhythm control drug. Whether OAC type, rate, or rhythm control treatment translate into different outcomes in the real-world is an active research eld. In theory, these treatments apply similarly to male and female patients.
Nevertheless, one can name more than a few medical conditions, such as coronary heart disease [3], hypertension [4], or heart failure [5], in which there are surprising differences between the two genders, as far as the therapeutic paths followed and the subsequent clinical outcomes are concerned. Distinct disparities in the clinical presentation among men and women have also been reported in AF, which could in uence decisions on discharge treatment [6,7]. However, data on gender-related differences in AF treatment patterns, along with prospective outcomes, remain sparse.
In the current study, we analyzed a contemporary cohort of patients with AF. We attempted to depict gender-related disparities in the main axes of AF treatment, namely OAC, rate and rhythm control. We further aimed to investigate the association of these treatment modalities with fatal and non-fatal clinical outcomes in males and females.

STUDY DESIGN AND DATA SOURCES
This is a retrospective cohort study using data from the MISOAC-AF (Motivational Interviewing to Support Oral Anti Coagulation adherence in patients with non-valvular AF) randomized controlled trial. The protocol, as well as the baseline characteristics and main results of the MISOAC-AF registry have been previously published elsewhere. [8,9]. Brie y, the objective of the MISOAC-AF trial was to assess the impact of a motivational intervention on the adherence of patients to OAC medication. Study protocols were approved by the appropriate Institutional Review Board at the Aristotle University of Thessaloniki and every trial procedure conformed to the Declaration of Helsinki [10].
All clinical baseline characteristics, as well as demographic, laboratory, echocardiographic and medication data of the patients were provided by the MISOAC-AF database.

DEFINITION OF COVARIATES
Atrial brillation was de ned as an electrocardiographically con rmed irregular heart rate that lasted more than 30 seconds, with an absence of p waves [11]. Patients denoted on rhythm control treatment at discharge were those treated with amiodarone, propafenone, or sotalol, with or without concomitant ratecontrol treatment. Patients on rate control were those treated with any combination of b-blocker, nondihydropyridine calcium channel blocker, or digoxin, without antiarrhythmic drug use.

STUDY POPULATION
The study involved adult individuals with primary or secondary diagnosis of non-valvular AF. All patients had been hospitalized in the cardiology ward of AHEPA University Hospital of Thessaloniki. Exclusion criteria in our study included: 1) patients whose end-stage disease made impossible the collection of follow-up data and 2) patients with unobtainable or unknown data concerning their rhythm or rate control treatment.

Page 5/18
The primary outcome was all-cause mortality, de ned as death from any cause. The secondary outcomes were the composite of any hospitalization-or cardiovascular (CV) death, and stroke, during the follow-up period.
Follow-up data were collected via in-person or telephone interviews at 12 and 24 months from the enrollment in our database. Patients were considered censored at the time of an outcome event, or death. The follow-up process was completed in April 2020. All deaths were searched in the Greek web-based civil registration insurance system.

STATISTICAL ANALYSIS
Patient demographics, physical and laboratory ndings, medical history, and medication were summarized using frequencies and percentages for categorical variables and mean ± standard deviation (SD) values for continuous ones. Differences between groups were tested using a chi-square or Fisher's exact test for descriptive variables and Students T-test for continuous ones. To examine the association between treatment decision (OAC, rate or rhythm control) and study outcomes, cumulative incidence was calculated based on Kaplan-Meier estimates and differences between groups were tested using a logrank test. Cox regression analysis for the primary outcome and the secondary outcomes was performed.
To account for potential confounding by characteristics in uencing treatment decisions and outcome of AF patients, we adjusted our analysis with the following, clinically relevant, covariates: gender, age, body mass index (BMI), diabetes mellitus (DM), hypertension, alcohol consumption, use of OAC agents, nterminal pro-hormone blood natriuretic peptide (NT-proBNP) and estimated glomerular ltration rate (eGFR). The adjusted hazard ratios (aHR) along with the respective con dence intervals (CI) are presented. The performed tests were two-sided, while the probability value of less than p = 0.05 was considered statistically signi cant. The SPSS Statistics for Windows, Version 26.0 (Armonk, NY: IBM Corp) and Stata statistical software 13.0 were used for our analysis.

Baseline clinical characteristics
Of 1140 patients included in this post-hoc analysis, 622 (57%) were male and 518 (43%) were female.
The baseline clinical characteristics of male and female individuals are presented in Table 1. Women were signi cantly older, with a higher left ventricular ejection fraction, and were less often bearing a pacemaker or an implantable cardioverter de brillator. A higher percentage of men had a history of smoking, alcohol use, and a lower percentage had a history of thyroid disease.

Discussion
In this study of AF patients who were discharged following an acute hospitalization, the rates of essential medication use at discharge concerning OAC agents, rhythm or rate control, were similar between the two genders. Signi cant gender-related differences in outcomes according to treatment modality emerged by multivariate Cox regression analysis. In speci c, use of rhythm control, as compared with rate control, was associated with lower rates of all-cause mortality in males. On the contrary, rhythm control treatment showed a correlation with higher rates of all-cause hospitalization or cardiovascular death in females.
No signi cant gender-related disparity was noted regarding the use of OAC medication. Nonetheless, use of OAC agents has been reported to be underused in female patients with AF, when compared to their male counterparts, a worrying prospect considering the fact that female gender has been linked to an increased risk of stroke in the setting of other coexisting stroke-related factors [12][13][14][15]. However, Lip et al have suggested that thromboprophylaxis is underused in AF patients of both genders and more so in those with higher CHADS2-VASc scores [16,17]. It has been suggested that acutely hospitalized patients could be at the center of this paradox, as their severe main illness may divert physicians from prescribing OAC treatment [18].
In addition, we found no signi cant difference between men and women, concerning use of rate or rhythm control medication at discharge. Nevertheless, some international studies concur on the fact that women with AF tend to receive signi cantly more rate, than rhythm control therapy, when compared to their male counterparts [12,13,19]. Euro Heart Survey researchers proposed that physicians may have considered a rate control strategy more advantageous for women, especially asymptomatic ones, as potential antiarrhythmic drug adverse effects occur more often in them [13]. No such treatment trend emerged among patients in our study.
Rate control medication in our study was associated with worse survival prospects in males. This nding comes in contrast to the results of previous trials, such as the ROCKET-AF [20], AFFIRM [21] and PIAF [22] ones, who showcased that there is no signi cant difference between use of rate and rhythm control medication in AF patients, regarding all-cause mortality. A possible explanation is that the increased comorbidity burden of males in our study may have may have in uenced survival rates. Males in our study were found to have a lower LVEF when compared to female counterparts, were bearing more often a pacemaker or ICD, and a higher percentage of men had a history of smoking or alcohol use (see Table 1). Interestingly, it has been suggested that b-blockers are associated with a reduced effect on AF males who have a lower systolic LVEF, when compared to sinus rhythm ones [23], a nding that possibly reinforces our comorbidity hypothesis In our cohort, rhythm control was associated with an increased rate of hospitalization or CV death in females. Our ndings concur with those of the RACE trial, which indicated that rhythm control in females with AF leads to three times higher cardiovascular morbidity and mortality [24]. It has been suggested that a higher rate of antiarrhythmic drug adverse effects in female patients could constitute the underlying cause [24][25][26], as well as their increased risk for torsades de pointes when treated with sotalol and bradyarrhythmias under antiarrhythmic medication [27]. However, Roy et al. demonstrated that rhythm-control did not result in a reduction in all-cause mortality [28]. Interestingly, Ionescu et al. have suggested that rhythm control is associated with higher long-term survival rates when handling the arrhythmia regardless of gender, in a population-based sample of hospitalized patients with AF [29].
Neither a rate, nor a rhythm control strategy was signi cantly associated with reduced stroke rates in males or females with AF. No international consensus has been reached on which treatment choice is correlated with better outcomes, concerning stroke. Tsadok et al proposed a superiority of rhythm control in minimizing stroke incidence in both genders [30], a nding not reinforced by a large meta-analysis [31]. Moreover, the ATHENA trial showcased a notable stroke risk reduction in high-risk patients using the antiarrhythmic drug dronedarone [32]. Nonetheless, the above are in contrast with data from the RACE study, as well as the EuroHeart Survey, who claimed that female AF patients under rhythm control had a higher risk of developing thromboembolic complications such as stroke [13,24].

Limitations
Our study is an observational retrospective one, with all the inherent limitations this entails, however we minimized de nition errors or missing data everywhere possible, and we conducted adequate follow-up. Furthermore, one cannot claim to adjust for all possible confounders, especially when keeping in mind the different baseline characteristics of each gender. As far as the drugs administered are concerned, data such as drug compliance or therapeutic range are missing, when it comes to rate or rhythm control medication. The patients' preferences and their effect upon the choice of treatment could not be assessed either. Finally, our study represents the treatment strategies followed in daily clinical practice in a single-center in Greece; however, deductions about other countries always entail the risk that the healthcare system there is fundamentally different than ours.

Conclusion
Among patients being discharged from the hospital with comorbid AF, the use of OAC, rate, and rhythm control treatment did not differ signi cantly between the two genders. Use of OAC treatment at discharge was not associated with a signi cant difference in mortality rates in either gender. A rhythm control strategy was associated with lower all-cause mortality and any hospitalization-or CV mortality in males, whereas it was associated with higher hospitalization-or CV death rates in females.

Con icts of interest
No such con ict to be declared.

Availability of data and material
Data are available from George Giannakoulas (e-mail: ggiannakoulas@auth.gr) upon reasonable request and with permission of AHEPA University Hospital.

Code availability
Not applicable

Ethics approval
The Institutional Review Board of Aristotle University of Thessaloniki approved the protocol of our study. and

Consent to participate
Every individual bestowed written and informed consent before taking part in the study.

Consent for publication
The patients participating in the study gave their written informed consent for publication.