ACCEL: Assessing Stroke Risk: CHADS2 versus CHA2DS2-VASc

  • Atrial fibrillation significantly increases risk of stroke.
  • Risk is not homogenous, as it is dependent upon several risk factors.
  • While the CHADS2 score is now widely used, CHA2DS2-VASc appears to provide more accurate risk assessment with little more difficulty in obtaining it than the CHADS2 score.
  • The current ACCF/AHA/HRS guidelines for AF management offer a three-tier set of risk factors meant to guide antithrombotic therapy and reduce stroke risk.1 The list goes beyond general risk factors to include issues such as mitral stenosis, a prosthetic heart valve, and LVEF <35%. The guidelines rank whether the factors are of high or moderate risk, or are less validated and considered weaker factors. The latter category includes female gender, coronary artery disease, thyrotoxicosis, and age between 65 and 74 years.

    According to the expert writing committee, an individual with no risk factors is of low risk but aspirin 81–325 mg daily is recommended. An individual with one moderate-risk factor is a candidate for warfarin therapy, while a patient with a high-risk factor, such as prior stroke, should receive warfarin.

    About a decade ago, in search of a user-friendly risk stratification scheme, Gage et al. integrated risk factors shown to independently predict risk in two previous studies: Atrial Fibrillation Investigators (AFI) and Stroke Prevention in Atrial Fibrillation (SPAF).2 They created the CHADS2 index which takes its acronym from both the factors and scores upon which it is based (Table 1), with each factor counting as 1 point, except prior stroke, which—as the strongest risk factor—gets 2 points.

    After creating CHADS2 scoring, they assembled the National Registry of Atrial Fibrillation (NRAF), consisting of 1,733 Medicare patients, ages 65–95 years, who had nonrheumatic AF and were not on warfarin at hospital discharge. Stroke rate per 100 patient-years rose by a factor of 1.5 for each 1-point increase in CHADS2 score. In other words, an individual with no risk factors and a CHADS2 score of 0 still had some risk—an annual stroke rate of 1.9—while an individual with every risk factor scored a 6 and had a stroke rate of 18.2.

    As for level of risk, a CHADS2 score of 0 is considered low risk; a 1 or 2, moderate risk; and higher than 2 was indicative of high risk. A CHADS2 score of 0—similar to a low-risk patient from the ACCF/AHA/HRS guideline stratification—is not a candidate for warfarin therapy because of the bleeding risk, while a score of 1 may warrant warfarin. In CHADS2, the authors said high-risk patients (scores of 2 or more) should always be considered for anticoagulation unless contraindicated.


    While applauded, and widely adopted, for its simplicity and broad applicability, some investigators were concerned that CHADS2 failed to consider important risk factors. The Stroke in AF Working Group concluded that only four clinical factors—prior stroke/transient ischemic attack (TIA), advancing age, hypertension, and diabetes—were consistent independent factors.3 However, other data delineated additional variables that suggest greater risk of stroke or thromboembolic events (TEs) in patients with AF.

    Table 1
    CHADS2 Risk Factors Score
    Congestive heart failure 1
    Hypertension (BP > 140/90 mm Hg or treated HTN) 1
    Age ≥75 years 1
    Diabetes mellitus 1
    Stroke/TIA (prior) 2
    Maximum Score 6
    For example, in the EuroHeart Survey (as well as other studies), female sex increased TE risk. Similarly, vascular disease increases risk of thrombosis for AF patients. In the United Kingdom, Lip et al. took another look at an older score, known as the Birmingham scheme, to transform it into a risk factor–based approach, reclassifying existing and adding new risk factors thought to be relevant. They identified definitive risk factors (previous stroke/TIA/TE and age >75 years) and combination risk factors (heart failure/moderate-severe cardiac dysfunction, hypertension, diabetes, vascular disease, female sex, and age 65–74 years).

    They also assigned numeric values to each risk factor and called the point-based scoring system CHA2DS2-VASc with a top score of 9 (Table 2). Age can account for 0 (<65 years), 1 (65–74 years), or 2 points (>75 years). Individuals with a score of 0 are low risk, a score of 1 of intermediate risk, and anyone scoring 2 or more are at high risk. In other words, women are already at intermediate risk due to their sex even if they have no other risk factors.

    Compared to other risk stratification schemes, CHA2DS2-VASc was the most likely to classify an individual from the EuroHeart Survey as high risk (75.7%) and least likely to classify a patient as low risk (9.2%).4 In contrast, Framingham found only 10.2% of patients to be high-risk and 48.3% to be low-risk.

    Table 2
    CHA2DS2-VASc Risk Factors Score
    Congestive heart failure/LV dysfunction 1
    Hypertension 1
    Age ≥75 years 2
    Diabetes mellitus 1
    Stroke/TIA/thromboembolism 2
    Vascular disease 1
    Age 65–74 years 1
    Sex category (i.e., female sex) 1
    Maximum Score 9

    Which approach is more accurate? Lip et al. found that the patients classified at low risk by CHA2DS2-VASc were truly low-risk (with no recorded TEs), whereas TEs occurred in 1.4% of low-risk CHADS2 patients and 1.8% of low-risk SPAF patients.

    When CHA2DS2-VASc was compared to other risk stratification schemes as predictors of TE risk, categorization varied greatly. CHA2DS2-VASc classified 94.2% as high risk; most of the other schemes categorized about two-thirds as high risk. However, of the 184 TE events that occurred in the patient population, 181 (98.4%) occurred in patients who were identified as high risk per CHA2DS2-VASc.

    NCDR Data

    Recently at the annual meeting of the National Cardiovascular Data Registry® (NCDR), Piyaskulkaew et al. demonstrated that patients with AF are heterogeneous, consisting of patients with a wide range of stroke risk within the same CHADS2 score. In a single-center pilot study, conducted at St. John Hospital and Medical Center, Wayne State University, Detroit, they found that patients with a CHADS2 score of 0 have annual stroke risk of 1.9%—the same as that reported earlier by the NRAF. When CHA2DS2-VASc was calculated, 52.3% had a score of 1 (stroke risk of 1.75%). The CHADS2 score of 0 overestimated stroke risk in 27.3% and underestimated the risk in 20.5% of patients.

    ACCF/AHA guidelines recommend anticoagulation for CHADS2 score >2 (stroke risk of 4.0%) and it is optional for those with score of 1. Warfarin is not strongly indicated in any of the 161 patients in this study with a CHADS2 score of 0 or 1. Calculating CHA2DS2-VASc score, however, and 61.5% with the score >2 qualify for warfarin (per 2010 ESC guidelines), resulting in a change of anticoagulation strategy in more than half of the studied patients based on their CHADS2 score. (The ESC Task Force on AF is chaired by A. John Camm, MD, and his committee recently updated the guidelines.5)

    Guidelines differ as to what level risk of stroke warrants anticoagulation. In European guidelines, a CHA2DS2-VASc score of 2 with annual stroke risk of 2.2% warrants warfarin therapy. In contrast, North American guidelines recommend either warfarin or aspirin for a CHADS2 score of 1, with its 2.8% annual risk of stroke.

    Whatever level of annual risk of stroke is chosen as warranting anticoagulation, the NCDR team concluded that CHA2DS2-VASc score will provide more accurate data for a decision. If confirmed in a larger North American population, the authors said the NCDR analysis suggests that CHA2DS2-VASc should replace CHADS2 for assessment of anticoagulation strategy.

    After the 2012 European Society of Cardiology meeting, Christophe Leclerque, MD, PhD, Professor of Cardiology at Rennes University Hospital, Rennes, France said, “New [ESC] guidelines suggest that for risk assessment, CHA2DS2-VASC is better than CHADS2.5 But the main message is that for the patient with a CHA2DS2-VASc score of 0, there is no indication for any anticoagulation or any antithrombotic therapy, so nothing to prevent embolism, not even aspirin, which increases risk of bleeding.”


    1. Anderson JL, et al. J Am Coll Cardiol. 2013;61:1935-44.
    2. Gage BF, et al. JAMA. 2001;285:2864-70.
    3. Stroke Risk in Atrial Fibrillation Working Group. Neurology. 2007;69:546-54.
    4. Dagres N, et al. J Am Coll Cardiol. 2007;49:572-7.
    5. Camm AJ, et al. Eur Heart J. 2012;33:2719-47.

    Keywords: Stroke, Coronary Artery Disease, Ischemic Attack, Transient, Warfarin, Risk Factors, Registries, Heart Failure, Mitral Valve Stenosis, Embolism, Medicare, Hypertension, Diabetes Mellitus

    < Back to Listings