Atrial fibrillation (AF) is the most common clinically relevant arrhythmia and increases the risk of thromboembolism and stroke; however, these risks are not the same for women and men. This review examines the evidence and clinical significance of increased thromboembolic risk in women with AF. The balance of results from over 30 recent studies suggests that female gender is an independent stroke risk factor in AF, and the inclusion of female gender in stroke risk stratification models, such as CHA 2 DS 2 -VASc, has improved risk assessment. Reasons for the increased thrombogenicity in women remain incompletely elucidated, but biological factors including increased hypertension, renal dysfunction, and hyperthyroidism in female patients with AF; cardiovascular remodeling; increased hypercoagulability, and estrogen hormone replacement therapy in women have been proposed. More importantly, gender differences exist in medical management of patients with AF, and compared with men, women have been found to have greater thromboembolic risk when not on anticoagulants, but may benefit from greater risk reduction when systemically anticoagulated. In conclusion, increased clinician awareness of these gender differences may help to improve the management of patients with AF.
Atrial fibrillation (AF), the most common clinically relevant arrhythmia, affects 2.7 to 6.1 million Americans, with prevalence projected to double by the year 2050. The prevalence of AF is 3.2% of the population aged ≥20 years and reaches 20% at age 80. Men have a greater risk of developing AF than women by a factor of 1.5 after adjusting for other risk factors. However, the absolute numbers of men and women with AF are roughly equal because of the higher average life expectancy of women. Women make up about 60% of the population with AF aged >75, the median age of AF onset.
AF is associated with a fivefold increased risk of stroke and is attributed with at least 50% of strokes occurring in subjects aged 80 years and older. Many risk stratification models have been proposed to quantify the risk of stroke in AF. The inclusion of female gender as an independent risk factor has been the subject of recent examination. AF is more frequently noted in women presenting with stroke than in men. In addition, women have a worse poststroke outcome than men in terms of motor and cognitive function and activities of daily living. AF is an independent stroke predictor of in-hospital mortality for women but is not for men.
Thus, these gender differences are clinically relevant to make accurate estimations of inherent stroke risk in patients with AF. This is important because patients with AF with the highest stroke risk derive the greatest absolute benefit from systemic anticoagulation. As such, clinician awareness of such gender differences becomes useful when a decision regarding anticoagulation is needed and few or no other risk factors exist. Current European Society of Cardiology (ESC) guidelines recommend that no systemic anticoagulation is required for female patients aged <65 years with lone AF (CHA 2 DS 2 -VASc = 1) because these patients are considered low risk for stroke, which stands in contrast to other subgroups with CHA 2 DS 2 -VASc = 1. The primary objective of this review is to provide an updated overview of the existing evidence for gender differences in thromboembolic risk and to discuss the clinical importance of such differences.
The PubMed database was used to review the English language reports addressing gender differences and thromboembolic risk in AF from 1994 to the present. The search used combinations of terms including “atrial fibrillation,” “gender OR sex OR female OR women,” and “thromboembolism OR stroke.” References of retrieved studies were further reviewed in detail for additional relevant studies and reviews.
Studies were selected for inclusion if they published stroke incidence data in men and in women. The number of women, number of total study participants, mean age of men and women, percent incidence of stroke in men and women, and relative risk (RR) for stroke for women were collected from each study when available. Difference in stroke risk was evaluated by examining the reported RR values, and if these were unavailable, by examining the p -value for statistically significant differences in stroke rates between men and women.
No extramural funding was used to support this work. The investigators are solely responsible for the design and conduct of this study, all study analyses, the drafting, and editing of the study and its final contents.
Evidence for Gender Differences in Thromboembolic Risk
We compiled over 30 studies published since 1999 that examine gender and thromboembolic risk, including 5 randomized controlled trials (RCTs) and 24 observational studies ( Tables 1 and 2 ). Of these 30 studies, 17 studies reported that female gender is a significant risk factor, 12 studies reported that female gender is not significant, and only 1 study reported that male gender is a significant risk factor. Four additional RCTs compared novel oral anticoagulant drugs (NOAC) and warfarin, providing further data on gender differences ( Table 3 ). However, 1 RCT and 1 observational study no longer found a significant difference after multivariate analysis. Four studies, all reporting insignificant gender differences, only reported univariate risk estimates associated with female gender.
|Cohort||Total n||# of |
|Age (years)||Stroke (%)||Relative Risk|
|Males||Females||p Value||Males||Females||p Value||Females||p Value|
|Inoue (2000) ∗||Japan||740||234||31.6||56 †||NR||NR||NR||NR||RR 0.5 M||0.0291|
|Humphries (2001) ∗||CARAF (Canada)||1097||339||30.9||60.5±0.6||65.4±0.7||<0.001||6.5 ‡||7.8 ‡||NS||NR||NR|
|Wang (2003)||Framingham Heart Study (USA)||705||336||47.7||75 †||NR||NR||NR||NR||HR 1.92 M||NR|
|Friberg (2004)||Copenhagen City Heart Study||276||110||39.9||67±8.4||69±6.8||NR||7.8||20||NR||HR 2.6 M||NR|
|4.7 years §|
|Dagres (2007) ∗||Euro Heart Survey on AF||5333||2249||42.2||64±13||70±12||<0.001||1.2||2.2||0.011||OR 1.83 M||0.019 M|
|1 year §|
|Poli (2009) ∗||University of Florence (Italy)||780||275||35.3||74||76||<0.001||1.2 ¶||2.43 ¶||0.042||HR 2.3 M||<0.01|
|Ruigomez (2009) ∗||UK General Practice Research||831||426||51.3||61.2% subjects aged ≥70 †||NR||NR||NR||NR||RR 1.0 M||NS|
|Lin (2011)||Taiwan NHI research database||7920||3633||45.9||63.3% subjects aged ≥65 †||NR||NR||NR||NR||OR 0.942 U||0.512 U|
|Oleson (2011)||Denmark national register||73538||37651||51.2||59.7% subjects aged ≥75 †||NR||NR||NR||NR||HR 1.6 M||0.04 M|
|van Staa (2011)||United Kingdom||79844||39704||49.7||73.3 †||NR||1.2||1.9||NR||RR 1.05 M||NS|
|4 years §|
|Chao (2012) ∗||Taiwan NHI research database||829||320||38.6||45.4 ± 12 †||1||1.6||4.4||0.014||HR 2.48 M||0.042 M|
|57.4 ± 35.7 months §|
|Friberg (2012)||Sweden||100802||50667||50.3||74.7||80.9||NR||4.2 ‖||6.2 ‖||<0.0001||HR 1.47 U; 1.18 M||<0.001|
|Mikkelsen (2012)||Denmark||87202||44744||51.3||71||78.2||<0.0001||3.7 ¶||5.43 ¶||NR||HR 1.04 M||NR|
|Potpara (2012)||Belgrade AF Study (Serbia)||862||315||46.5||49.6||56.7||<0.001||6.9||7||0.579||HR 1.11 U||0.579 U|
|10.1±6.1 years §|
|Tsadok (2012) ∗||Quebec (Canada)||83513||44115||52.8||77.2||80.2||NR||4.3||5.8||<0.001||HR 1.14 M||<0.001 M|
|30, 90, 365 days §|
|Bosch (2013) ∗||Germany||2742||1021||37.2||67.5±9.9||71.2±9.3||<0.001||3.4||3.6||0.74||NR||NR|
|6, 12 months §|
|Disertori (2013) ∗||GISSI-AF subset||1234||487||39.5||66.75 paroxysmal AF; 68.78 persistent AF †||NR||3||9||NR||NR||NR|
|1 year §|
|Guo (2013)||Chinese PLA General Hospital||1034||281||27.2||78||71||<0.0001||8.1||6.05||0.267||NR||NR|
|1.9 years §|
|Poli (2013) ∗||EPICA study (Italy)||3015||1654||54.9||82.6 ∗∗||83.1 ∗∗||0.001||1.3 ¶ §||1.6 ¶||0.25||OR 1.2 U||0.3 U|
|Salam (2013) ∗||Hamad General Hospital (Qatar)||3849||1417||36.8||54.5 ± 15.7||59 ± 15||0.001||0.4||0.4||0.8||NR||NR|
|20 years §|
|Aakre (2014)||Olmsted County, Minnesota (USA)||2720||1320||48.5||73.33±14.57 †||NR||NR||NR||NR||HR 1.45 M||0.0015 M|
|Inoue (2014)||J-RHYTHM||7406||2165||29.2||69±10||73±9||<0.001||1.8||1.6||0.576||OR 0.89 U||0.576 U|
|2 years §|
|Shroff (2014)||US Medicare patients 2010||80314||40879||50.9||NR||NR||NR||1.4 ‡||1.9 ‡||NR||NR||NR|
|Siu (2014)||Queen Mary Hospital, Hong Kong||9727||5064||52.1||76.9±12.5 †||NR||NR||NR||NR||HR 1.16 U; 1.03 M||0.026 U; 0.723 M|
|Yang (2014) ∗||Chinese AF registry||2016||1104||54.8||68.5||NR||NR||NR||NR||HR 1.359 U; 1.419 M||0.073 U; 0.048 M|
|# of |
|Age (years)||Stroke (%)||Relative Risk|
|Males||Females||p Value||Males||Females||p |
|Hart (1999)||SPAF I-III||1853||514||27.7||68||71||NR||2.1 ∗||4.4 ∗||NR||RR 1.8 U; 1.6 M||0.03 U; 0.01 M|
|Fang (2005)||ATRIA cohort||13559||5795||42.7||NR||NR||NR||1.8 ∗||3.5 ∗||NR||RR 1.6 M||NR|
|Rienstra (2005) †||RACE substudy||522||192||36.8||67±9||71±8||<0.001||6.7||6.8||NS||NR||NR|
|2.3 years ‡|
|Gomberg-Maitland (2006)||SPORTIF III and V||7329||2257||30.8||69.8±9||73.4±8||<0.0001||1.44 ∗||2.08 ∗||0.016||HR 1.44 U; 1.27 M||0.0161 U; 0.16 M|
|Sullivan (2012)||AFFIRM substudy||4060||1594||39.3||68.3±8.3||71.3±7.5||<0.0001||3||5||0.002||OR 1.6 M||0.002 M|
|2000 days ‡|