Author + information
- Received April 7, 2006
- Revision received September 29, 2006
- Accepted October 16, 2006
- Published online February 6, 2007.
- Nikolaos Dagres, MD⁎,⁎ (, )
- Robby Nieuwlaat, MSc†,
- Panos E. Vardas, MD, PhD, FACC‡,
- Dietrich Andresen, MD, FACC§,
- Samuel Lévy, MD, FACC∥,
- Stuart Cobbe, MD¶,
- Dimitrios Th. Kremastinos, MD, FACC⁎,
- Günter Breithardt, MD, FACC#,
- Dennis V. Cokkinos, MD, FACC⁎⁎ and
- Harry J.G.M. Crijns, MD†
- ↵⁎Reprint requests and correspondence:
Dr. Nikolaos Dagres, Second University Cardiology Department, Attikon Hospital, Rimini 1, 12462 Athens, Greece.
Objectives This study sought to investigate gender-related differences in patients with atrial fibrillation (AF) in Europe.
Background Gender-related differences may play a significant role in AF.
Methods We analyzed the data of 5,333 patients (42% female) enrolled in the Euro Heart Survey on Atrial Fibrillation.
Results Compared with men, the women were older, had a lower quality of life (QoL), had more comorbidities, more often had heart failure (HF) with preserved left ventricular systolic function (18% vs. 7%, p < 0.001), and less often had HF with systolic dysfunction (17% vs. 26%, p < 0.001). Among patients with typical AF symptoms (56% of women, 49% of men), there was no gender-related difference in the choice of rate or rhythm control. Among patients with atypical or no symptoms (44% of women, 51% of men), women less frequently underwent rhythm control (39% vs. 51%, p < 0.001) than did men. Women underwent less electrical cardioversion (22% vs. 28%, p < 0.001). Prescription of oral anticoagulants was identical (65%) in both genders. One-year outcome was similar except that women had a higher chance for stroke (odds ratio 1.83 in multivariable regression analysis, p = 0.019).
Conclusions Women with AF had more comorbidities, more HF with preserved systolic function, and a lower QoL than men. In the large group with atypical or no symptoms, women were treated appropriately more conservatively with less rhythm control than men. Women had a higher chance for stroke. Long-term QoL changes and other morbidities and mortality were similar.
Previous studies have reported significant gender differences in presentation and treatment of cardiovascular disorders, with a more conservative management in women (1,2).
In patients with atrial fibrillation (AF), gender differences in patient characteristics have been described in several reports (3–5). However, data regarding differences in treatment and outcome are scarce. An underuse of anticoagulants in women has been suggested, but reports are not consistent (3,6–8). The RACE (RAte Control versus Electrical cardioversion) study, which compared rhythm versus rate control in persistent AF, showed a gender-related difference in outcome (5): women with rhythm control had more cardiovascular morbidity and mortality compared with those with rate control, whereas this difference was not seen in men. Women also had a significantly lower quality of life (QoL) irrespective of strategy.
We examined gender differences in presentation, treatment, and outcome of patients with AF analyzing the data of the Euro Heart Survey on Atrial Fibrillation.
The design and main results of the survey have been described recently (9). Between September 2003 and July 2004, 182 centers from 35 European Society of Cardiology (ESC) countries enrolled 5,333 ambulant or hospitalized patients who had an electrocardiogram or Holter recording of AF during the qualifying admission or within the preceding 12 months.
Data were collected on patient characteristics and treatment. A clinical follow-up was performed at 1 year after discharge. Both at baseline and at 1-year follow-up, QoL was measured by the EuroQoL questionnaire, producing a sum score for 5 questions on health status (EQ-5D) and a score of self-reported health status on a visual analogue scale (EQ-VAS) (10,11).
Symptoms typical for AF were considered to be palpitations and syncope. Dyspnea, chest pain, dizziness, and fatigue were classified as “other symptoms” possibly related to AF.
Heart failure (HF) was defined as the presence of signs and symptoms of either right (elevated central venous pressure, hepatomegaly, dependent edema) or left (exertional dyspnea, cough, fatigue, orthopnea, paroxysmal nocturnal dyspnea, cardiac enlargement, rales, gallop rhythm, pulmonary venous congestion) ventricular failure or both, and the diagnosis should be confirmed by noninvasive or hemodynamic measurements. Heart failure with left ventricular systolic dysfunction was defined as HF with a left ventricular ejection fraction <45% or end-diastolic diameter >55 mm in a recent echocardiographic study (within the preceding year). Heart failure with preserved systolic function was defined as HF in combination with a left ventricular ejection fraction ≥45% and end diastolic diameter ≤55 mm. For other definitions, we refer to the design paper (9). The American College of Cardiology (ACC)/American Heart Association (AHA)/ESC 2001 guidelines were used for patient classification into the different ACC/AHA/ESC stroke risk categories (12).
Categorical variables are reported as percentage within gender, and were tested for differences between men and women with the chi-square test. Age is reported as mean ± 1 SD, and was tested with the independent ttest. The QoL scores (i.e., [change in] EQ-5D and EQ-VAS) are reported as median with interquartile range, and were tested with the Mann-Whitney Utest because of a skewed distribution. Significant differences are indicated in the tables according to p < 0.001, p < 0.01, or p < 0.05; exact p values are reported in the Results section for differences not reported in the tables. Analyses were performed with the statistical software package SPSS (version 12.01, SPSS Inc., Chicago, Illinois). The large sample size may have caused differences that are statistically but not clinically significant. Therefore, the investigators focused on the findings that they considered clinically significant.
Multivariable stepwise logistic regression was performed to test the independent association of gender with occurrence of stroke and major bleeding during 1-year follow-up, correcting for differences in age, associated diseases, comorbidities, and treatment. Variables were removed stepwise from the model when the p value exceeded 0.10. Variables with a p value <0.05 in the final model were considered to be significant contributors and were kept in the model. The net odds ratio (OR) and its 95% confidence interval (CI) and p value are reported for gender regarding both analyses.
Of the 5,333 enrolled patients, 2,249 were female (42%) (Table 1).Women were older, had a lower QoL, and more frequently had hypertension, valvular heart disease, diabetes, and hyperthyroidism. Men more frequently had coronary artery disease (CAD) and idiopathic (lone) AF.
Approximately one-third of both men and women had HF. However, the HF type differed significantly: HF with preserved systolic function was far more frequent among women, HF with systolic dysfunction was more common among men. In both genders, HF with systolic dysfunction was equally often associated with CAD (both 49%, p = 0.976), and HF with preserved systolic function was equally often associated with hypertension (both 75%, p = 0.929). However, HF with preserved systolic function was more often associated with valvular heart disease in women (27% vs. 10%, p < 0.001). Women had more comorbidities (including more than 1 associated disease). Especially the higher prevalence of hypertension and valvular heart disease resulted in more women being at the highest risk for stroke according to the ACC/AHA/ESC 2001 guidelines, and in women more often having multiple stroke risk factors (54% vs. 39%, p < 0.001). At enrollment, women were more often symptomatic, mainly because of a higher prevalence of palpitations and dyspnea.
Diagnostic procedures are presented in Table 2.Exercise testing was less commonly performed in women (14% vs. 25%, p < 0.001), and when performed, was less often associated with CAD in women (51% vs. 44%, p = 0.023). Transesophageal echocardiography was less often performed in women to identify left atrial thrombus (73% vs. 81%, p = 0.007), and more often to evaluate (artificial) valve function (33% vs. 23%, p = 0.004).
Interventions for AF are given in Table 3.Women less often underwent electrical cardioversion or catheter ablation than did men. Other interventions showed no substantial difference.
Rhythm versus rate control related to symptoms is shown in Table 4.The treatment strategy did not differ between men and women with typical AF symptoms. In the remaining patients (44% of women and 51% of men) with symptoms other than palpitations or syncope or with asymptomatic AF, women were treated less aggressively, with a significantly lower percentage undergoing rhythm control and a higher percentage undergoing rate control compared with men.
Discharge therapy is shown in Table 5.Women were more commonly prescribed digitalis. The prescription of oral anticoagulants did not differ. Also, no significant differences were found in oral anticoagulation prescription rates among the different stroke risk categories according to the ACC/AHA/ESC 2001 guidelines and CHADS2stroke risk score.
Regarding 1-year outcome, the univariable analysis yielded the following results. Women had a significantly higher rate of stroke (2.2% vs. 1.2% in men, p = 0.011) and major bleeding (2.2% vs. 1.3%, p = 0.028). No significant differences between genders were observed regarding mortality (5.1% in women vs. 5.4% in men, p = 0.567) or HF (11.2% vs. 11.4%, p = 0.878). The QoL changes from baseline were minimal and did not differ significantly between genders (EQ-5D: 0.00 [−0.10 to 0.15] in women vs. 0.00 [−0.07 to 0.11] in men, p = 0.385; EQ-VAS: 0 [−5 to 15] vs. 0 [−5 to 10], p = 0.125). When correcting for differences in age and stroke risk in multivariable logistic regression, women still had a higher chance for stroke (OR 1.83 [95% CI 1.10 to 3.03], p = 0.019). Of note, a prior stroke/transient ischemic attack was the only acknowledged stroke risk factor that was more significantly associated with occurrence of stroke than female gender. The higher chance for major bleeding in women became nonsignificant in multivariable analysis (OR 1.28 [95% CI 0.76 to 2.13], p = 0.351). The difference in stroke rate was largely attributable to a higher stroke rate in women who did not receive oral anticoagulation, who had at least 1 stroke risk factor, and who underwent rhythm control (5.0% vs. 1.2%, p = 0.006). When stratifying to the number of stroke risk factors, this difference was only significant when 1 solitary stroke risk factor was present (6.9% vs. 0.5%, p = 0.001). In the latter group, of 9 strokes in women, 6 were associated with hypertension, 2 with HF, and 1 with mitral stenosis, whereas the only stroke in men was associated with hypertension. In contrast with the rhythm control group, there was no difference in outcome between men and women undergoing rate control.
Women with AF had a worse cardiovascular risk profile and a lower QoL than men. They were more symptomatic, had more comorbidities, and more often had HF with preserved systolic function. Women received less often rhythm control treatment if they had atypical or no symptoms. The use of antithrombotic drugs was similar in both genders, also when correcting for the higher prevalence of stroke risk factors in women. However, women had a higher risk for stroke, and withholding oral anticoagulation from patients at risk mainly seemed to affect women. In this regard, women undergoing rhythm control and women with hypertension as a solitary stroke risk factor seem to be at high risk for stroke. No substantial changes of QoL over time were seen in either gender. Our results suggest that female gender should be reconsidered as an independent risk factor for stroke. In addition, women may benefit from consistent anticoagulation as well as tailored HF management.
Baseline characteristics and evaluation
Female patients were older and more frequently had hypertension, valvular heart disease, diabetes, and thyroid disease; male patients more frequently had CAD and idiopathic AF. In total, women had more comorbidities, were more often at the highest risk for stroke, and were more often symptomatic than men. These findings agree with those of previous reports (3–5). The higher frequency of exercise tests in men cannot be ascribed fully to a higher prevalence of CAD. Apparently, women have a lower chance of undergoing an exercise test in general, and when they do undergo one, it is less often in association with CAD.
The overall frequency of HF was similar in both genders. However, HF with preserved systolic function was more frequent among women, HF with systolic dysfunction was more frequent among men. This is a novel finding in AF patients, and agrees with previous observations in the total population of HF patients (13). The higher prevalence of HF with systolic dysfunction in men partly can be explained by the higher prevalence of CAD, but not exclusively, because in both genders only half of the HF cases with systolic dysfunction were associated with CAD. The higher prevalence of HF with preserved systolic function in women seems mainly attributable to a higher prevalence of hypertension and valvular heart disease.
Rhythm and rate control
We found significant treatment differences between genders regarding rhythm or rate control in relation to symptoms. Approximately half of the patients had typical AF symptoms (palpitations or syncope). In these, the proportion of patients assigned to rhythm or rate control did not differ between genders. In the remaining half of the patients who had either atypical or no AF symptoms, treatment was more conservative in women, with significantly less rhythm control than in men. Consistently, fewer women underwent electrical cardioversion. These findings are also novel in the field of AF treatment and corroborate the previous reports on a generally more conservative management of women. This conservative approach may have precluded women from appropriate therapy, i.e., rhythm control of (atypical) AF symptoms. On the other hand, in asymptomatic patients, this conservative approach might have been beneficial for women, because potential adverse effects of rhythm control interventions and drugs are avoided. Appropriately withholding rhythm control might even be more beneficial in women than in men, because we found an excess of strokes among women on rhythm control and because rhythm control in women was associated with worse outcome than rate control in the RACE study.
In contrast with earlier studies (3,6), there were no differences in the prescription of antithrombotic medication, and also not when correcting for the higher prevalence of stroke risk factors in women. This could mirror a better guideline implementation, as also suggested by the higher rate of anticoagulant prescription in the present survey (65%) compared with prior reports (6–8).
Gender and risk for stroke
In the total population, women had a higher chance for stroke than men, but no further differences were found regarding 1-year outcomes. Previous investigations on outcomes in men and women provided contradictory results (3,5). The higher stroke rate in women in our survey was mainly observed when withholding oral anticoagulation from women undergoing rhythm control and having at least 1 stroke risk factor, mainly hypertension. Because this difference was not exclusively attributable to a higher prevalence of stroke risk factors in women, these results might indicate a subgroup of women that is at high risk for stroke.
These findings agree with the higher stroke rate among women in the Framingham study (14), which led to the incorporation of female gender in the Framingham stroke risk scheme (15). These findings also agree with the results of the SPAF (Stroke Prevention in Atrial Fibrillation) trials (16) and the ATRIA (AnTicoagulation and Risk factors In Atrial fibrillation) study, which showed a higher risk for AF-related thromboembolism in women not receiving oral anticoagulation (17). Interestingly, in the ATRIA study the reduction in rates of thromboembolism with anticoagulation was larger in women than in men with similar rates of major bleeding (17). However, female gender is not incorporated in the most recent stroke risk classification scheme, CHADS2(18), and the recently revised ACC/AHA/ESC guidelines classify female gender into the “less validated or weaker risk factors” (19). Although the higher stroke rate did not affect mortality in our survey, we speculate that female gender should be reconsidered as an independent stroke risk factor.
Results from surveys like the present one could be influenced by a potential center effect because a limited number of centers, of which a large proportion were university related and highly specialized, were recruited among the 35 participating countries. The older age of women may have affected some of the observed differences, especially in comorbidities. Obviously, in clinical practice these differences should be taken into account. In addition, some patients had a first detected episode of AF; others had paroxysmal, persistent, or permanent AF. However, because the patient proportions with these different AF types were almost identical in both genders, this heterogeneity should not have affected our main results.
Women with AF are older and sicker than men, with a lower QoL. They also seem to be at a higher risk of stroke, especially when left unprotected. Although applied equally among genders, our data suggest that antithrombotic treatment must receive more attention in women. The higher prevalence of HF with preserved systolic function is intriguing and should trigger special management approaches in women. The generally less aggressive management in women comes to their advantage with far less rhythm control in female AF patients with absent or atypical AF symptoms. However, this did not translate to a higher QoL, presumably because of higher age and higher disease burden.
- Abbreviations and Acronyms
- American College of Cardiology
- atrial fibrillation
- American Heart Association
- coronary artery disease
- EuroQoL Questionnaire–Visual Analogue Scale
- EuroQoL Questionnaire–Five-Dimension Score
- European Society of Cardiology
- heart failure
- implantable cardioverter-defibrillator
- odds ratio
- quality of life
- Received April 7, 2006.
- Revision received September 29, 2006.
- Accepted October 16, 2006.
- American College of Cardiology Foundation
- Wong C.C.,
- Froelicher E.S.,
- Bacchetti P.,
- et al.
- Humphries K.H.,
- Kerr C.R.,
- Connolly S.J.,
- et al.
- Levy S.,
- Maarek M.,
- Coumel P.,
- et al.
- Rienstra M.,
- van Veldhuisen D.J.,
- Hagens V.E.,
- et al.
- Gage B.F.,
- Boechler M.,
- Doggette A.L.,
- et al.
- Brass L.M.,
- Krumholz H.M.,
- Scinto J.M.,
- Radford M.
- Nieuwlaat R.,
- Capucci A.,
- Camm A.J.,
- et al.
- Fuster V.,
- Ryden L.E.,
- Asinger R.W.,
- et al.
- Lenzen M.J.,
- Scholte op Reimer W.J.M.,
- Boersma E.,
- et al.
- Wolf P.A.,
- Abbott R.D.,
- Kannel W.B.
- Hart R.G.,
- Pearce L.A.,
- McBride R.,
- Rothbart R.M.,
- Asinger R.W.,
- Stroke Prevention in Atrial Fibrillation (SPAF) Investigators
- Fang M.C.,
- Singer D.E.,
- Chang Y.,
- et al.
- Fuster V.,
- Ryden L.E.,
- Cannom D.S.,
- et al.