Sex Differences in Stroke Risk Among Older Patients with Recently Diagnosed Atrial Fibrillation

Ambar Afshar, M.D. (Disclosure)
Brian F. Gage, M.D. (Disclosure)

September 13, 2012

Editor's Note: This Article of the Month is based on Avgil Tsadok M, Jackevicius CA, Rahme E, et al. Sex Differences in Stroke Risk Among Older Patients With Recently Diagnosed Atrial Fibrillation. JAMA 2012;307:1952-1958. doi:10.1001/jama.2012.3490.

Introduction

Atrial fibrillation (AF) is the most common cardiac arrhythmia. Patients with AF have a 5-fold increase in the risk of stroke. Some, but not all, prior AF studies have found that female sex is an independent risk factor for stroke.

Methods

A retrospective, inception cohort of administrative data elucidated patterns of warfarin use and stroke incidence in men and women with AF.

Tsadok et al. analyzed Canadian administrative data from January 1, 1998 to March 31, 2007. Patients were elderly (65 years or older) residents of Quebec, who were discharged from the hospital with a primary diagnosis of AF. Information on outpatient prescriptions, and inpatient and outpatient diagnostic and therapeutic procedures was available. Patients who developed AF as a post-admission complication, had perioperative AF, or had hyperthyroidism or thyrotoxicosis were excluded. Patients were also excluded if they were 105 years or older, or resided in chronic care facilities. The Med-Echo database was used to obtain information on patient characteristics in order to calculate the CHADS₂ score (congestive heart failure, hypertension, age ≥ 75 diabetes, prior stroke or transient ischemic attack). The primary outcome was incident (ischemic) stroke; rates of transient ischemic attack and/or retinal infarct also were reported. Adherence to warfarin therapy was assessed by the medication possession ratio, which is defined as the number of days of tablets supplied divided by 365 days.

Continuous variables were compared using the t test or Mann-Whitney U test. Dichotomous variables were compared using the chi squared test. Incidence rates of stroke and mortality were calculated as the number of events per 100 person-years of follow-up. Cox proportional hazards regression were used to identity independent determinants of stroke. Follow-up was until the first diagnosis of stroke, death, or the end of the study period.

Results

The cohort comprised of 39,398 (47.2%) men and 44,115 (52.8%) women. Women were on average 80.2 years old and men were 77.2 years old. The mean CHADS₂ score in women was 1.99 (standard deviation (SD) 1.10) versus 1.74 (SD 1.13; p < .001) in men. Warfarin prescription rates were higher among men, but women tended to have more prescriptions filled for warfarin within 30 days post-discharge and had a greater medication possession ratio (indicating better compliance).

In both univariate and adjusted analyses, stroke incidence was greater in women, especially elderly women. The overall incidence (95% confidence interval [CI]) of stroke per 100 person-years was 2.02 (1.95-2.10) in women and 1.61 (95% CI, 1.54-1.69) in men (p < 0.001). Older women had significantly higher rates of stroke than older men, regardless of warfarin use or adherence level. Women had higher risk of stroke compared to men (hazard ratio 1.14 [95% CI, 1.07-1.22]), even after adjusting for comorbid conditions, CHADS₂ score and warfarin treatment. The increased stroke risk from female gender was insignificant in patients younger than 75 years, but significant in older patients (p = 0.02 for interaction with age). There was no difference in rates of intracerebral hemorrhage between men and women.

Conclusion

Among elderly patients with AF, women have a higher risk of stroke than men. After controlling for comorbid conditions, women have 1.14-fold higher risk of stroke than men.

Perspective

Some,^{5, 7, 8, 10, 13, 20, 22, 23, 27} but not all, prior studies (Table) have found female gender to be an independent risk factor for stroke in AF. Tsadok et al. have quantified this effect more precisely than prior studies. The significant interaction between female sex and age greater than 75, partially explains the inconsistencies between prior studies. The SPAF investigators first reported that among elderly (age > 75) patients with AF, females had higher risk of stroke than men.¹³ However, because the biological basis for this association remains speculative, validation of this relationship, as done in this study, was needed.

Overall, this study supports current ACCP¹⁵ recommendations to offer anticoagulant therapy to patients with a CHADS₂ score of 1 or greater. Tsadok et al. found that the risk of stroke correlated with CHADS₂ score among men and women and that the annual rate of stroke was too low (1% per year) in men and women with a CHADS₂ score of 0 to justify routine prescription of anticoagulant therapy.

Variations in risk perception can lead to differences in the use of stroke prophylaxis. Women with AF tend to be older and have more hypertension and valvular heart disease, than their male counterparts.⁷ Other AF studies have shown that women are less likely than men to receive warfarin²⁴, sometimes with devastating consequences.^{8, 21} Disparities in treatment of women may exist due to a misperception of low stroke risk in these patients.²⁵ However, atrial remodeling, increased left atrial size, and thrombotic tendency (e.g. from estrogen replacement therapy) may elevate the risk of stroke in elderly women with AF.^{4, 27}

In prior ACC/AHA/ESC guidelines¹¹, female gender was included as a "less validated or weaker risk factor," for stroke in AF. That classification reflected those studies that have failed to demonstrate that female gender is an independent risk factor for stroke in AF.^{1, 2, 3, 9, 16, 17, 18, 19, 26} In contrast, the 2011 ESC guidelines⁶ feature the CHA₂DS₂-VASc score, which assigns an additional point to Vascular disease, Age > 75 years, and female Sex. However, in Tsadok et al, vascular disease was not an independent predictor of stroke and female sex was a relatively weak predictor of stroke.

Based on Tsadok et al. and other studies (Table), we conclude that female gender is a validated, but relatively weak risk factor for stroke in patients with AF unless their age is > 75 years. Because AF guidelines already recommend anticoagulant therapy to elderly patients with AF regardless of gender, and the association between female gender and stroke remains weak, the guidelines do not need to advocate anticoagulant therapy for all women.

Likewise, the present study does not support assigning a full point to female gender, as advocated by the CHA₂DS₂-VASc score, because its hazard ratio of 1.14 is less than a typical CHADS₂ point, which has a hazard ratio of 1.4-1.5.^{1, 3, 12, 27}