Advertisement

Does CHA2DS2-VASc Improve Stroke Risk Stratification in Postmenopausal Women with Atrial Fibrillation?

Published:October 18, 2013DOI:https://doi.org/10.1016/j.amjmed.2013.05.023

      Abstract

      Background

      Risk stratification of atrial fibrillation patients with a congestive heart failure (C), hypertension (H), age ≥ 75 (A), diabetes (D), stroke or transient ischemic attack (TIA) (S2) (CHADS2) score of <2 remains imprecise, particularly in women. Our objectives were to validate the CHADS2 and congestive heart failure (C), hypertension (H), age ≥ 75 (A2), diabetes (D), stroke, TIA or prior thromboembolic disease (S2)- vascular disease (V), age 65-74 (A), female gender (S) (CHA2DS2-VASc) stroke risk scores in a healthy cohort of American women with atrial fibrillation and to determine whether CHA2DS2-VASc further risk-stratifies individuals with a CHADS2 score of <2.

      Methods

      We identified a cohort of 5981 women with atrial fibrillation not on warfarin at baseline (mean age 65.9 ± 7.2 years) enrolled in the Women's Health Initiative and followed for a median of 11.8 years. Univariate and multivariate proportional hazards analyses were used to examine these 2 risk scores, with main outcome measures being annualized event rates of ischemic stroke or transient ischemic attack stratified by risk score.

      Results

      Annualized stroke/transient ischemic attack rates ranged from 0.36% to 2.43% with increasing CHADS2 score (0-4+) (hazard ratio [HR] 1.57; 95% confidence interval [CI], 1.45-1.71 for each 1-point increase) and 0.20%-2.02% with increasing CHA2DS2-VASc score (1-6+) (HR 1.50; 95% CI, 1.41-1.60 for each 1-point increase). CHA2DS2-VASc had a higher c statistic than CHADS2: 0.67 (95% CI, 0.65-0.69) versus 0.65 (95% CI, 0.62-0.67), P <.01. For CHADS2 scores <2, stroke risk almost doubled with every additional CHA2DS2-VASc point.

      Conclusions

      Although both CHADS2, and CHA2DS2-VASc are predictive of stroke risk in postmenopausal women with atrial fibrillation, CHA2DS2-VASc further risk-stratifies patients with a CHADS2 score <2.

      Keywords

      • CHADS2, and CHA2DS2-VASc, the most widely used stroke risk scores for patients with atrial fibrillation, guide decisions about anticoagulation.
      • CHADS2 classifies more than half of patients with atrial fibrillation as being at low or intermediate risk for stroke (score <2).
      • CHA2DS2-VASc further risk-stratifies CHADS2 <2 patients, which may help to guide clinical decisions about anticoagulation.
      Reported stroke rates in patients with atrial fibrillation not treated with anticoagulation range from <1% (low risk) to >18% (high risk) per year,
      • Gage B.F.
      • Waterman A.D.
      • Shannon W.
      • Boechler M.
      • Rich M.W.
      • Radford M.J.
      Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation.
      • Fang M.C.
      • Go A.S.
      • Chang Y.
      • Borowsky L.
      • Pomernacki N.K.
      • Singer D.E.
      Comparison of risk stratification schemes to predict thromboembolism in people with nonvalvular atrial fibrillation.
      • Wang T.J.
      • Massaro J.M.
      • Levy D.
      • et al.
      A risk score for predicting stroke or death in individuals with new-onset atrial fibrillation in the community: the Framingham Heart Study.
      and studies show that women are at higher risk.
      • Wang T.J.
      • Massaro J.M.
      • Levy D.
      • et al.
      A risk score for predicting stroke or death in individuals with new-onset atrial fibrillation in the community: the Framingham Heart Study.
      • Fang M.C.
      • Singer D.E.
      • Chang Y.
      • et al.
      Gender differences in the risk of ischemic stroke and peripheral embolism in atrial fibrillation: the AnTicoagulation and Risk factors In Atrial fibrillation (ATRIA) study.
      • Hart R.G.
      • Pearce L.A.
      • McBride R.
      • Rothbart R.M.
      • Asinger R.W.
      Factors associated with ischemic stroke during aspirin therapy in atrial fibrillation: analysis of 2012 participants in the SPAF I-III clinical trials. The Stroke Prevention in Atrial Fibrillation (SPAF) Investigators.
      • Lip G.Y.
      • Nieuwlaat R.
      • Pisters R.
      • Lane D.A.
      • Crijns H.J.
      Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
      • Chao T.F.
      • Liu C.J.
      • Chen S.J.
      • et al.
      Atrial fibrillation and the risk of ischemic stroke: does it still matter in patients with a CHA2DS2-VASc score of 0 or 1?.
      Although guidelines recommend treatment with anticoagulants in higher-risk and aspirin in lower-risk patients, the distinction between these 2 risk categories remains unclear.
      National Institute for Health and Clinical Excellence (NICE)
      NICE Clinical Guideline 36: Atrial Fibrillation: The Management of Atrial Fibrillation.
      • Fuster V.
      • Ryden L.E.
      • Cannom D.S.
      • et al.
      ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation—executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the European Society of Cardiology Committee for Practice Guidelines (Writing Committee to Revise the 2001 Guidelines for the Management of Patients With Atrial Fibrillation).
      Of the many scoring systems developed to predict stroke risk in atrial fibrillation, congestive heart failure (C), hypertension (H), age ≥ 75 (A), diabetes (D), stroke or transient ischemic attack (TIA) (S2) (CHADS2)
      • Gage B.F.
      • Waterman A.D.
      • Shannon W.
      • Boechler M.
      • Rich M.W.
      • Radford M.J.
      Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation.
      and congestive heart failure (C), hypertension (H), age ≥ 75 (A2), diabetes (D), stroke, TIA or prior thromboembolic disease (S2)- vascular disease (V), age 65-74 (A), female gender (S) (CHA2DS2-VASc)
      • Lip G.Y.
      • Nieuwlaat R.
      • Pisters R.
      • Lane D.A.
      • Crijns H.J.
      Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
      are the most widely used. CHADS2, developed from a combination of Atrial Fibrillation Investigators and Stroke Prevention in Atrial Fibrillation, was validated using the National Registry of Atrial Fibrillation comprising 1733 Medicare beneficiaries aged 65-95 years with nonrheumatic atrial fibrillation not on warfarin at hospital discharge.
      • Gage B.F.
      • Waterman A.D.
      • Shannon W.
      • Boechler M.
      • Rich M.W.
      • Radford M.J.
      Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation.
      In this score, whereas congestive heart failure (C), hypertension (H), age ≥75 years (A), and diabetes (D) receive 1 point each, stroke or transient ischemic attack (S2) receives 2 points.
      CHA2DS2-VASc, a modification of the 2006 Birmingham/National Institute for Health and Clinical Excellence scheme,
      National Institute for Health and Clinical Excellence (NICE)
      NICE Clinical Guideline 36: Atrial Fibrillation: The Management of Atrial Fibrillation.
      was validated in a cohort of 1084 hospitalized, ambulatory patients not anticoagulated at baseline from the Euro Heart Survey on atrial fibrillation.
      • Lip G.Y.
      • Nieuwlaat R.
      • Pisters R.
      • Lane D.A.
      • Crijns H.J.
      Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
      CHA2DS2-VASc expands on CHADS2 by 1) including a history of systemic thromboembolism (S2) in the stroke category and 2) adding vascular disease (defined as prior myocardial infarction, peripheral arterial disease, or aortic plaque [V]), age (65-74 years [A]), and female sex (S) as risk factors. All risk factors receive 1 point, except age ≥75 years and history of prior stroke/transient ischemic attack/thromboembolism, which receive 2 points each.
      As there are no validation studies comparing CHADS2 and CHA2DS2-VASc in an ambulatory US population of women with atrial fibrillation, our objectives were to validate and compare the predictive power of these scores, determine the annualized rates of stroke, and clarify the discriminatory ability of CHA2DS2-VASc in such a population.

      Methods

      Study Population

      The study design has been described previously.
      • Hays J.
      • Hunt J.R.
      • Hubbell F.A.
      • et al.
      The Women's Health Initiative recruitment methods and results.
      • Anderson G.L.
      • Manson J.
      • Wallace R.
      • et al.
      Implementation of the Women's Health Initiative study design.
      Study participants were members of the Women's Health Initiative (WHI) cohort: a prospective, multiarm clinical trial and observational study that focused on the causes and prevention of cardiovascular disease, cancer, and osteoporosis in women. Major exclusion criteria were predicted survival <3 years, alcohol or drug dependency, dementia, severe mental illness, and participation in another clinical trial. WHI comprised an observational study and 4 randomized clinical trials: 1) estrogen plus progestin versus placebo, 2) estrogen alone versus placebo in hysterectomized women, 3) dietary modification trial, and 4) calcium/vitamin D versus placebo trial.
      Beginning in 1993, 161,809 postmenopausal women aged 50-79 years were prospectively enrolled in WHI. Events through September 2010 were used for this retrospective analysis. The initial study population consisted of women who reported a history of atrial fibrillation or had an electrocardiogram with documented atrial fibrillation at baseline (n = 7108). From this group, we excluded 291 with valvular heart disease or hyperthyroidism, 85 with missing values for either CHADS2 or CHA2DS2-VASc, and 790 on warfarin at WHI randomization or enrollment. There were 1127 excluded, leaving a final sample of 5981, of whom 2390 were participants in one of the clinical trials and 3591 were enrolled in the observational study; 5901 women with atrial fibrillation were identified by self-report, 24 by electrocardiogram, and 56 had both.

      Definition of Variables

      Congestive heart failure, diabetes mellitus, and prior stroke or transient ischemic attack were defined by self-report at initial examination. Hypertension was defined as systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or use of any antihypertensive medication. Vascular disease was defined as self-report of any of the following: myocardial infarction, percutaneous coronary intervention, coronary artery bypass surgery, or peripheral vascular disease. Information on aortic plaque and systemic thromboembolism, although included among the CHA2DS2-VASc risk factors,
      • Lip G.Y.
      • Nieuwlaat R.
      • Pisters R.
      • Lane D.A.
      • Crijns H.J.
      Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
      was not collected by WHI.

      Follow-Up and End-Point Determination

      Intensity of follow-up visits varied based on enrollment arm, ranging from every 6 months (clinical trials) to every 3 years (observational study). When a potential outcome was identified, medical records were obtained and stroke (including self-reports) and transient ischemic attack (only the first event) were centrally adjudicated.
      • Wassertheil-Smoller S.
      • Hendrix S.L.
      • Limacher M.
      • et al.
      Effect of estrogen plus progestin on stroke in postmenopausal women: the Women's Health Initiative: a randomized trial.
      No bleeding end points were collected. We lost 2.3% of our cohort to follow-up and 4.2% stopped follow-up early.

      Statistical Analysis

      We summarized baseline characteristics with means and SDs for continuous variables and frequencies and percentages for categorical variables. Annualized percentages for each construct are presented, calculating the percentage for each CHADS2 and CHA2DS2-VASc level as the total number of events divided by the total at-risk follow-up time.
      Proportional hazards modeling was used to evaluate the relationship between the 2 constructs and stroke. Both constructs were evaluated in a continuous and categorical form, modeling the stroke outcome by each construct, with nonstroke participants censored at death or when lost to follow-up. Hazard ratios (HR) and corresponding P-values are presented for each model.
      For each model, Harrell's c statistic was calculated to quantify the discriminatory ability of the constructs, and 95% confidence intervals (CI) for each c statistic and the difference between c statistics using bootstrapping with 1000 replications were computed.
      Components of each construct as predictors of stroke were evaluated using both univariate and multivariate modeling. Events and annualized rates for each component level are presented with their corresponding univariate P-value from a model evaluating each component individually. All components were put into a single model with resulting HRs and corresponding P-values presented.
      All proportional hazards models were adjusted for aspirin use and stratified within the model by WHI hormone trial arm (not randomized, active, placebo), dietary modification trial arm (not randomized, intervention, comparison), and calcium/vitamin D arm (not randomized, active, placebo). Analyses were completed using SAS version 9.1 (SAS Institute Inc., Cary, NC).
      To compare CHADS2 and CHA2DS2-VASc on stroke risk prediction, we used the Net Reclassification Improvement Index (NRI), which identifies how many participants are correctly and incorrectly reclassified into different risk categories (upward for events, downward for nonevents).
      • Pencina M.J.
      • D'Agostino Sr., R.B.
      • D'Agostino Jr., R.B.
      • Vasan R.S.
      Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond.
      Because 3.5% of our atrial fibrillation participants had a stroke/transient ischemic attack event over the first 5 years of follow-up, we examined the <3%, 3%-<6%, 6%-<9%, and ≥9% risk groups, excluding participants censored before 5 years of follow-up.

      Results

      Baseline Characteristics

      In the WHI cohort with atrial fibrillation not on warfarin at baseline, 64.9% were hypertensive, 3.7% had congestive heart failure, 9.2% had diabetes mellitus, 2.6% had prior stroke, 4.9% had prior transient ischemic attack, 10.3% had prior coronary artery disease, and 5.2% had peripheral vascular disease (Table 1). Warfarin users (not included in this analysis) had a significantly higher prevalence of CHADS2 and CHA2DS2-VASc risk factors than nonusers. The WHI did not collect information about why patients were not on anticoagulation.
      Table 1Demographic and Clinical Characteristics of Participants with Baseline Atrial Fibrillation by Warfarin Use at Baseline
      Baseline CharacteristicNonusers (n = 5981)Warfarin Users (n = 753)P-Value
      n%n%
      Congestive heart failure2223.714519.3<.001
      Hypertension (BP ≥140/90 mm Hg or meds use)387964.961581.7<.001
      Age, years: mean (SD)65.85 (7.18)69.03 (5.97)<.001
       <65246941.316121.4
       65-74278946.644759.4
       ≥7572312.114519.3
      Diabetes mellitus5539.210213.5<.001
      Prior stroke1562.610213.5<.001
      Prior TIA2964.912716.9<.001
      Clopidogrel (Plavix) use10.000.0.723
      Aspirin use (≥80 mg/day)204134.1496.5<.001
      Hormone use<.001
       Never245541.039452.3
       Past118419.814719.5
       Current233639.121228.2
      History of coronary artery disease (MI/CABG/PTCA)61610.310914.5<.001
      History of MI5118.58110.8.043
      History of CABG/PTCA2594.3577.6<.001
      History of PVD3125.2597.8.012
      BP = blood pressure; CABG = coronary artery bypass graft; MI = myocardial infarction; PTCA = percutaneous transluminal coronary angioplasty; PVD = peripheral vascular disease; TIA = transient ischemic attack.

      Stroke Rate

      During a median of 11.8 years of follow-up (interquartile range 8.0-13.6), there were 457 stroke/transient ischemic attack events (137 of which were transient ischemic attacks) among the 5981 women. Annualized stroke/transient ischemic attack rates ranged from 0.36% to 2.43% with increasing CHADS2 score (0-4+), and 0.20% to 2.02% with increasing CHA2DS2-VASc score (1-6+, because all were female) (Table 2). Kaplan-Meier curves for survival free from stroke/transient ischemic attack, stratified by CHADS2 and CHA2DS2-VASc score, demonstrate a progressive decrease in event-free survival over time as the respective scores increase (Figure).
      Table 2Annualized Rates of Stroke/TIA by CHADS2 and CHA2DS2-VASc Scores
      Construct, LevelStroke†/TIA
      nEventsAnnual %
      CHADS2 score
       01760710.36
       128792190.72
       29221061.27
       3299381.45
       4+121232.43
      CHA2DS2-VASc score
       1951220.20
       21794960.48
       317991520.82
       49111081.30
       5343511.71
       6+183282.02
      TIA = transient ischemic attack.
      Does not include hemorrhagic stroke.
      Figure thumbnail gr1
      FigureKaplan-Meier curves of stroke/transient ischemic attack (TIA)-free survival stratified by CHADS2 and CHA2DS2-VASc score.

      Stroke Scheme Performance

      Although both risk scores were predictive of stroke risk according to univariate proportional hazards analysis (P <.001 for both), the discriminatory ability of CHA2DS2-VASc was higher than that of CHADS2 (P <.01). For every 1-point increase, while CHADS2 had a HR of 1.57 (95% CI, 1.45-1.71, c statistic: 0.65 [95% CI, 0.62-0.67]), CHA2DS2-VASc had a HR of 1.50 (95% CI, 1.41-1.60, c statistic: 0.67 [95% CI, 0.65-0.69]) (Table 3).
      Table 3Univariate Proportional Hazards Analyses
      All models are adjusted for aspirin use and stratified within the model by WHI hormone trial arm (not randomized, active, placebo); WHI dietary modification arm (not randomized, intervention, comparison); WHI calcium/vitamin D arm (not randomized, active, placebo).
      of Stroke/TIA by CHADS2 and CHA2DS2-VASc Constructs
      Construct, LevelStroke/TIA (n = 5981 )
      HR (95% CI)P-valuec statistic (95% CI)
      CHADS2 score
       Continuous - 1 pt increase1.57 (1.45-1.71)<.0010.65 (0.62-0.67)
       Categorical<.0010.65 (0.62-0.67)
      01.00 (ref)
      12.00 (1.53-2.61)
      23.55 (2.62-4.81)
      34.02 (2.70-5.99)
      4+6.69 (4.16-10.78)
      CHA2DS2-VASc score
       Continuous - 1 pt increase1.50 (1.41-1.60)<.0010.67 (0.65-0.69)
       Categorical<.0010.67 (0.65-0.69)
      11.00 (ref)
      22.45 (1.54-3.90)
      34.16 (2.65-6.52)
      46.67 (4.20-10.60)
      58.92 (5.38-14.78)
      6+10.34 (5.87-18.23)
      CI = confidence interval; HR = hazard ratio; TIA = transient ischemic attack.
      All models are adjusted for aspirin use and stratified within the model by WHI hormone trial arm (not randomized, active, placebo); WHI dietary modification arm (not randomized, intervention, comparison); WHI calcium/vitamin D arm (not randomized, active, placebo).
      All individual construct components of the risk scores were predictive of stroke in univariate analysis, but congestive heart failure was not significant in multivariate analysis (HR 1.05; 95% CI, 0.68-1.64, P = .827) (Table 4). The most predictive variables by multivariate analysis were age ≥75 years (HR 2.91; 95% CI, 2.19-3.86, P <.001) and history of stroke or transient ischemic attack (HR 2.13; 95% CI, 1.60-2.82, P <.001).
      Table 4Univariate and Multivariate Proportional Hazards Analyses of Stroke/TIA by Risk Factor
      Risk FactorGroup Event (Ann %)Univariate P-ValueMultivariate HR (95% CI)Multivariate P-Value
      Hx of CHF22 (1.14).0421.05 (0.68-1.64).827
      Hypertension351 (0.90)<.0011.55 (1.24-1.94)<.001
      Age, years (ref <65)<.001<.001
       65-74248 (0.88)1.95 (1.56-2.45)
       75+93 (1.46)2.91 (2.19-3.86)
      Hx of diabetes58 (1.15)<.0011.38 (1.04-1.83).027
      Hx of stroke/TIA59 (1.80)<.0012.13 (1.60-2.82)<.001
      Hx of CAD/PVD99 (1.29)<.0011.46 (1.15-1.85).002
      CAD = coronary artery disease; CHF = congestive heart failure; HR = hazard ratio; Hx = history; PVD = peripheral vascular disease; TIA = transient ischemic attack.
      To determine the possible effect of hormone replacement therapy on the prognostic ability of these risk scores, we evaluated a subgroup interaction between both the continuous CHADS2 and CHA2DS2-VASc scores and hormone therapy, defined by intervention assignment if enrolled in the hormone therapy trial and by baseline hormone use if not. For both CHADS2 (P = .70) and CHA2DS2-VASc (P = .43), the association of the score was not significantly modified by hormone use.

      Risk Score Comparison

      Although CHADS2 is helpful for scores ≥2, clinical decision-making becomes more ambiguous when the score is <2. Because the 2 highest frequency CHADS2 scores in our population were 1 (n = 2879, 48%) and 0 (n = 1760, 29%), we compared the classification of these patients in both schemes to see if CHA2DS2-VASc would have an added benefit (Table 5). Table 5 demonstrates that within any CHADS2 score column, a higher CHA2DS2-VASc score corresponds to a higher event rate, but the same is not true when stratifying risk within any given CHA2DS2-VASc score row. For CHADS2 scores <2, stroke risk almost doubles with every additional CHA2DS2-VASc point. Further, when using CHA2DS2-VASc at 5-year follow-up, the NRI was 21.1%, z-stat = 4.70, P <.001: 65 of 212 patients with a stroke/thromboembolic event (30.7%) and 897 of 5372 patients without an event (16.6%) were reclassified to a higher risk category.
      Table 5Event Totals/Participant Totals and (Annualized Rates of Stroke/TIA) by Risk Score
      Total ParticipantsCHADS2
      CHA2DS2-VASc0123≥4
       122/951 (0.20)
       244/755 (0.53)52/1039 (0.44)
       35/54 (0.91)136/1601 (0.82)11/144 (0.77)
       431/239 (1.43)71/610 (1.29)6/62 (0.98)
       524/168 (1.68)26/169 (1.73)1/6 (1.61)
       ≥66/68 (1.19)22/115 (2.49)
      TIA = transient ischemic attack.

      Discussion

      This study validated CHADS2 and CHA2DS2-VASc for predicting stroke risk in a US population of postmenopausal women with atrial fibrillation followed for a median of 11.8 years. While both scores have modest but similar discriminative accuracy, CHA2DS2-VASc appears to be useful in further risk-stratifying women with CHADS2 <2.
      In the original validation study, CHADS2 had a c statistic of 0.82 (95% CI, 0.80-0.84), while the Atrial Fibrillation Investigators and Stroke Prevention in Atrial Fibrillation studies had c statistics of 0.68 (95% CI, 0.65-0.71) and 0.74 (95% CI, 0.71-0.76), respectively.
      • Gage B.F.
      • Waterman A.D.
      • Shannon W.
      • Boechler M.
      • Rich M.W.
      • Radford M.J.
      Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation.
      Largely based on this study, current US guidelines
      • Fuster V.
      • Ryden L.E.
      • Cannom D.S.
      • et al.
      2011 ACCF/AHA/HRS focused updates incorporated into the ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines developed in partnership with the European Society of Cardiology and in collaboration with the European Heart Rhythm Association and the Heart Rhythm Society.
      recommend risk-stratification and anticoagulation treatment as follows: aspirin 81-325 mg daily, CHADS2 = 0; aspirin or anticoagulation with warfarin or dabigatran, CHADS2 = 1; and anticoagulation, CHADS2 ≥2.
      Although CHADS2 is a useful risk-stratification scheme that has been adopted by the American College of Cardiology, American Heart Association, European Society of Cardiology, and Heart Rhythm Society,
      • Fuster V.
      • Ryden L.E.
      • Cannom D.S.
      • et al.
      ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation—executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the European Society of Cardiology Committee for Practice Guidelines (Writing Committee to Revise the 2001 Guidelines for the Management of Patients With Atrial Fibrillation).
      • Fuster V.
      • Ryden L.E.
      • Cannom D.S.
      • et al.
      2011 ACCF/AHA/HRS focused updates incorporated into the ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines developed in partnership with the European Society of Cardiology and in collaboration with the European Heart Rhythm Association and the Heart Rhythm Society.
      • Camm A.J.
      • Kirchhof P.
      • Lip G.Y.
      • et al.
      Guidelines for the management of atrial fibrillation: the Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC).
      one drawback is that it places a large percentage of patients (up to 47% by some estimates; 48% of our cohort)
      • Conen D.
      • Chae C.U.
      • Glynn R.J.
      • et al.
      Risk of death and cardiovascular events in initially healthy women with new-onset atrial fibrillation.
      in an intermediate risk category (CHADS2 = 1) with subsequent ambiguity as to the appropriate anticoagulation strategy. If aspirin, dabigatran, rivaroxaban, apixaban, and warfarin all offered equivalent protection from stroke, this recommendation would be uncontroversial. However, while warfarin decreases stroke risk in atrial fibrillation patients by approximately 68%, aspirin has been shown to provide only a 20%-36% reduction in ischemic stroke risk.
      Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation. Analysis of pooled data from five randomized controlled trials.
      Atrial Fibrillation Investigators. The efficacy of aspirin in patients with atrial fibrillation. Analysis of pooled data from 3 randomized trials.
      • Aguilar M.I.
      • Hart R.
      • Pearce L.A.
      Oral anticoagulants versus antiplatelet therapy for preventing stroke in patients with non-valvular atrial fibrillation and no history of stroke of transient ischemic attacks.
      • Connolly S.J.
      • Ezekowitz M.D.
      • Yusuf S.
      • et al.
      Dabigatran versus warfarin in patients with atrial fibrillation.
      Moreover, aspirin is often used even when warfarin is indicated, because of its ease of administration and the widely held view that it provides a decreased risk of major bleeding, studies demonstrating equivalent risk notwithstanding.
      Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation. Analysis of pooled data from five randomized controlled trials.
      • Mant J.
      • Hobbs F.D.
      • Fletcher K.
      • et al.
      Warfarin versus aspirin for stroke prevention in an elderly community population with atrial fibrillation (the Birmingham Atrial Fibrillation Treatment of the Aged Study, BAFTA): a randomised controlled trial.
      • Roy B.
      • Desai R.V.
      • Mujib M.
      • et al.
      Effect of warfarin on outcomes in septuagenarian patients with atrial fibrillation.
      Lip et al (2010)
      • Lip G.Y.
      • Nieuwlaat R.
      • Pisters R.
      • Lane D.A.
      • Crijns H.J.
      Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
      suggested that a significant advantage of CHA2DS2-VASc was that it classified very few women as intermediate risk (score of 1 = 15.1%) compared with CHADS2 (1 score = 34.9%). Also, there was a very low event rate in the CHA2DS2-VASc 0-1 group (0 score = 0%; 1 score = 0.6%). In their analysis, CHADS2 and CHA2DS2-VASc had c statistics of 0.586 and 0.606, respectively. Based on these findings, the European Society of Cardiology guidelines recommend that CHA2DS2-VASc be used when CHADS2 ≤1 and that anticoagulation be prescribed when CHA2DS2-VASc ≥2, aspirin 75-325 mg or anticoagulation when CHA2DS2-VASc = 1 (anticoagulation preferred), and aspirin only when CHA2DS2-VASc = 0.
      • Camm A.J.
      • Kirchhof P.
      • Lip G.Y.
      • et al.
      Guidelines for the management of atrial fibrillation: the Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC).
      Several studies have validated and compared CHA2DS2-VASc to CHADS2 in recently hospitalized populations not on anticoagulation. In a cohort of Danish patients, Olesen et al (2011)
      • Olesen J.B.
      • Lip G.Y.
      • Hansen M.L.
      • et al.
      Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study.
      found that both were predictive of stroke risk (CHADS2: c statistic of 0.812; CHA2DS2-VASc: c statistic of 0.885). Friberg et al (2012)
      • Friberg L.
      • Rosenqvist M.
      • Lip G.Y.
      Evaluation of risk stratification schemes for ischaemic stroke and bleeding in 182 678 patients with atrial fibrillation: the Swedish Atrial Fibrillation cohort study.
      evaluated stroke risk schemes in a cohort from the National Swedish Drug registry over a median of 1.4 years and found that both schemes were predictive of stroke risk (CHADS2: c statistic of 0.62; CHA2DS2-VASc: c statistic of 0.67). In both studies, CHA2DS2-VASc was better at identifying patients assessed as truly low risk for thromboembolic stroke (0 score). Olesen et al (2012)
      • Olesen J.B.
      • Torp-Pedersen C.
      • Hansen M.L.
      • Lip G.Y.
      The value of the CHA2DS2-VASc score for refining stroke risk stratification in patients with atrial fibrillation with a CHADS2 score 0-1: a nationwide cohort study.
      validated CHA2DS2-VASc in a population with a CHADS2 score <2 and, as in our study, found that CHA2DS2-VASc improved on the predictive ability of CHADS2 (NRI = 14.2; P <.001).
      Our study offers unique data in a large prospective cohort of American women by validating CHADS2 and CHA2DS2-VASc in a population of healthy, ambulatory, postmenopausal women with more than 11 years of follow-up. This is particularly important because all validations to date of CHA2DS2-VASc have been performed in recently hospitalized, non-US cohorts. Second, it provides additional evidence that the risk factors of age 65-74 years and history of vascular disease included in CHA2DS2-VASc may help to further risk-stratify those patients who would otherwise fall into either a low or intermediate risk category (CHADS2 ≤1).
      A first limitation of this study involves the inherent confines of the WHI cohort. Because it did not include men, no conclusions can be made about the effect of female sex on stroke risk, and because information on aortic plaque and systemic thromboembolism was not available, these variables were not able to be included in the evaluation of CHA2DS2-VASc.
      Another limitation is that the designation of atrial fibrillation (and of many of the risk factors) was made by self-report in the majority of participants (5901 of 5981), which could have led to participant misclassification and lower-than-expected event rates (Table 2).
      • Gage B.F.
      • Waterman A.D.
      • Shannon W.
      • Boechler M.
      • Rich M.W.
      • Radford M.J.
      Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation.
      • Lip G.Y.
      • Nieuwlaat R.
      • Pisters R.
      • Lane D.A.
      • Crijns H.J.
      Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
      However, a recent study suggests that self-report of atrial fibrillation is as predictive of stroke risk as documentation by electrocardiogram.
      • Soliman E.Z.
      • Howard G.
      • Meschia J.F.
      • et al.
      Self-reported atrial fibrillation and risk of stroke in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study.
      An alternative explanation for the relatively low stroke/transient ischemic attack rates may be the relative health of our cohort. A recent comparison of stroke risk schemes in a cohort of 13,559 patients with atrial fibrillation by Fang et al
      • Fang M.C.
      • Go A.S.
      • Chang Y.
      • Borowsky L.
      • Pomernacki N.K.
      • Singer D.E.
      Comparison of risk stratification schemes to predict thromboembolism in people with nonvalvular atrial fibrillation.
      found similar annual stroke rates in low- and intermediate-risk patients. It also is possible that some of the women may have started warfarin during the follow-up period, which may have influenced (presumably by decreasing) the event rate of stroke/transient ischemic attack. However, at year 3, only 340 of the 5021 women on whom we have medication data had started warfarin. Further, re-analysis after excluding these women did not substantially change the stroke rates. It also is possible that patients in our cohort died of causes other than a stroke, because the annualized death rates in our cohort more closely approximate the annualized stroke/transient ischemic attack rates in the original studies (Table 6).
      • Gage B.F.
      • Waterman A.D.
      • Shannon W.
      • Boechler M.
      • Rich M.W.
      • Radford M.J.
      Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation.
      • Lip G.Y.
      • Nieuwlaat R.
      • Pisters R.
      • Lane D.A.
      • Crijns H.J.
      Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
      Table 6Annualized Rates of Death by CHADS2 and CHA2DS2-VASc Scores
      Construct, LevelnDeath
      EventsAnnual %
      CHADS2 Score
       017601750.87
       128794941.57
       29223093.53
       3299973.49
       4+121636.13
      CHA2DS2-VASc Score
       1951560.50
       217942041.00
       317993451.80
       49113073.53
       53431334.18
       6+183936.28
      A further finding of our study that does not correlate with some of the prior cohorts is that a history of congestive heart failure was not predictive of stroke/transient ischemic attack in multivariate analysis. This could be due to the low incidence of congestive heart failure in our cohort, possibly because women are more likely to have diastolic heart failure, thereby attenuating the impact of systolic heart failure in this cohort or because these women underreported diagnoses of heart failure. Nonetheless, at least 2 other studies have failed to support congestive heart failure as a risk factor for stroke in atrial fibrillation.
      • Friberg L.
      • Rosenqvist M.
      • Lip G.Y.
      Evaluation of risk stratification schemes for ischaemic stroke and bleeding in 182 678 patients with atrial fibrillation: the Swedish Atrial Fibrillation cohort study.
      • Hart R.G.
      • Pearce L.A.
      Current status of stroke risk stratification in patients with atrial fibrillation.
      In conclusion, both CHADS2, and CHA2DS2-VASc are predictive of stroke risk in ambulatory, postmenopausal female patients with atrial fibrillation. This study provides valuable information on the added value of using CHA2DS2-VASc to further risk-stratify women who have a CHADS2 score <2, which, when combined with an assessment of bleeding risk, may have clinical implications for identifying patients who should be offered anticoagulation instead of aspirin therapy. Future studies are needed to clarify anticoagulation recommendations for CHADS2 <2 individuals, particularly in the era of new oral anticoagulants.

      Acknowledgment

      Women's Health Initiative (WHI) Program Office: (National Heart, Lung, and Blood Institute, Bethesda, MD) Jacques Rossouw, Shari Ludlam, Joan McGowan, Leslie Ford, Nancy Geller; WHI Clinical Coordinating Center (Fred Hutchinson Cancer Research Center, Seattle, WA): Garnet Anderson, Ross Prentice, Andrea LaCroix, Charles Kooperberg; WHI Investigators and Academic Centers: (Brigham and Women's Hospital, Harvard Medical School, Boston, MA) JoAnn E. Manson; (MedStar Health Research Institute/Georgetown-Howard Center for Clinical and Translational Sciences) Barbara V. Howard; (Stanford Prevention Research Center, Stanford, CA) Marcia L. Stefanick; (The Ohio State University, Columbus, OH) Rebecca Jackson; (University of Arizona, Tucson/Phoenix, AZ) Cynthia A. Thomson; (University at Buffalo, Buffalo, NY) Jean Wactawski-Wende; (University of Florida, Gainesville/Jacksonville, FL) Marian Limacher; (University of Iowa, Iowa City/Davenport, IA) Robert Wallace; (University of Pittsburgh, Pittsburgh, PA) Lewis Kuller; (Wake Forest University School of Medicine, Winston-Salem, NC) Sally Shumaker.

      References

        • Gage B.F.
        • Waterman A.D.
        • Shannon W.
        • Boechler M.
        • Rich M.W.
        • Radford M.J.
        Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation.
        JAMA. 2001; 285: 2864-2870
        • Fang M.C.
        • Go A.S.
        • Chang Y.
        • Borowsky L.
        • Pomernacki N.K.
        • Singer D.E.
        Comparison of risk stratification schemes to predict thromboembolism in people with nonvalvular atrial fibrillation.
        J Am Coll Cardiol. 2008; 51: 810-815
        • Wang T.J.
        • Massaro J.M.
        • Levy D.
        • et al.
        A risk score for predicting stroke or death in individuals with new-onset atrial fibrillation in the community: the Framingham Heart Study.
        JAMA. 2003; 290: 1049-1056
        • Fang M.C.
        • Singer D.E.
        • Chang Y.
        • et al.
        Gender differences in the risk of ischemic stroke and peripheral embolism in atrial fibrillation: the AnTicoagulation and Risk factors In Atrial fibrillation (ATRIA) study.
        Circulation. 2005; 112: 1687-1691
        • Hart R.G.
        • Pearce L.A.
        • McBride R.
        • Rothbart R.M.
        • Asinger R.W.
        Factors associated with ischemic stroke during aspirin therapy in atrial fibrillation: analysis of 2012 participants in the SPAF I-III clinical trials. The Stroke Prevention in Atrial Fibrillation (SPAF) Investigators.
        Stroke. 1999; 30: 1223-1229
        • Lip G.Y.
        • Nieuwlaat R.
        • Pisters R.
        • Lane D.A.
        • Crijns H.J.
        Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation.
        Chest. 2010; 137: 263-272
        • Chao T.F.
        • Liu C.J.
        • Chen S.J.
        • et al.
        Atrial fibrillation and the risk of ischemic stroke: does it still matter in patients with a CHA2DS2-VASc score of 0 or 1?.
        Stroke. 2012; 43: 2551-2555
        • National Institute for Health and Clinical Excellence (NICE)
        NICE Clinical Guideline 36: Atrial Fibrillation: The Management of Atrial Fibrillation.
        National Institute for Health and Clinical Excellence, London2006
        • Fuster V.
        • Ryden L.E.
        • Cannom D.S.
        • et al.
        ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation—executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the European Society of Cardiology Committee for Practice Guidelines (Writing Committee to Revise the 2001 Guidelines for the Management of Patients With Atrial Fibrillation).
        J Am Coll Cardiol. 2006; 48: 854-906
        • Hays J.
        • Hunt J.R.
        • Hubbell F.A.
        • et al.
        The Women's Health Initiative recruitment methods and results.
        Ann Epidemiol. 2003; 13: S18-S77
        • Anderson G.L.
        • Manson J.
        • Wallace R.
        • et al.
        Implementation of the Women's Health Initiative study design.
        Ann Epidemiol. 2003; 13: S5-S17
        • Wassertheil-Smoller S.
        • Hendrix S.L.
        • Limacher M.
        • et al.
        Effect of estrogen plus progestin on stroke in postmenopausal women: the Women's Health Initiative: a randomized trial.
        JAMA. 2003; 289: 2673-2684
        • Pencina M.J.
        • D'Agostino Sr., R.B.
        • D'Agostino Jr., R.B.
        • Vasan R.S.
        Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond.
        Stat Med. 2008; 27 (discussion 207-112): 157-172
        • Fuster V.
        • Ryden L.E.
        • Cannom D.S.
        • et al.
        2011 ACCF/AHA/HRS focused updates incorporated into the ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines developed in partnership with the European Society of Cardiology and in collaboration with the European Heart Rhythm Association and the Heart Rhythm Society.
        J Am Coll Cardiol. 2011; 57: e101-e198
        • Camm A.J.
        • Kirchhof P.
        • Lip G.Y.
        • et al.
        Guidelines for the management of atrial fibrillation: the Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC).
        Eur Heart J. 2010; 31: 2369-2429
        • Conen D.
        • Chae C.U.
        • Glynn R.J.
        • et al.
        Risk of death and cardiovascular events in initially healthy women with new-onset atrial fibrillation.
        JAMA. 2011; 305: 2080-2087
      1. Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation. Analysis of pooled data from five randomized controlled trials.
        Arch Intern Med. 1994; 154: 1449-1457
      2. Atrial Fibrillation Investigators. The efficacy of aspirin in patients with atrial fibrillation. Analysis of pooled data from 3 randomized trials.
        Arch Intern Med. 1997; 157: 1237-1240
        • Aguilar M.I.
        • Hart R.
        • Pearce L.A.
        Oral anticoagulants versus antiplatelet therapy for preventing stroke in patients with non-valvular atrial fibrillation and no history of stroke of transient ischemic attacks.
        Cochrane Libr. 2011; 4: 1-34
        • Connolly S.J.
        • Ezekowitz M.D.
        • Yusuf S.
        • et al.
        Dabigatran versus warfarin in patients with atrial fibrillation.
        N Engl J Med. 2009; 361: 1139-1151
        • Mant J.
        • Hobbs F.D.
        • Fletcher K.
        • et al.
        Warfarin versus aspirin for stroke prevention in an elderly community population with atrial fibrillation (the Birmingham Atrial Fibrillation Treatment of the Aged Study, BAFTA): a randomised controlled trial.
        Lancet. 2007; 370: 493-503
        • Roy B.
        • Desai R.V.
        • Mujib M.
        • et al.
        Effect of warfarin on outcomes in septuagenarian patients with atrial fibrillation.
        Am J Cardiol. 2012; 109: 370-377
        • Olesen J.B.
        • Lip G.Y.
        • Hansen M.L.
        • et al.
        Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study.
        BMJ. 2011; 342: d124
        • Friberg L.
        • Rosenqvist M.
        • Lip G.Y.
        Evaluation of risk stratification schemes for ischaemic stroke and bleeding in 182 678 patients with atrial fibrillation: the Swedish Atrial Fibrillation cohort study.
        Eur Heart J. 2012; 33: 1500-1510
        • Olesen J.B.
        • Torp-Pedersen C.
        • Hansen M.L.
        • Lip G.Y.
        The value of the CHA2DS2-VASc score for refining stroke risk stratification in patients with atrial fibrillation with a CHADS2 score 0-1: a nationwide cohort study.
        Thromb Haemost. 2012; 107: 1172-1179
        • Soliman E.Z.
        • Howard G.
        • Meschia J.F.
        • et al.
        Self-reported atrial fibrillation and risk of stroke in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study.
        Stroke. 2011; 42: 2950-2953
        • Hart R.G.
        • Pearce L.A.
        Current status of stroke risk stratification in patients with atrial fibrillation.
        Stroke. 2009; 40: 2607-2610