CHA₂DS₂-VASc Score Calculator

Atrial fibrillation lets blood pool in the left atrial appendage, where it can clot and embolise to the brain. The absolute risk of that varies enormously between patients, and anticoagulation carries its own bleeding risk — so the decision to treat needs a risk estimate rather than a blanket rule.

CHA₂DS₂-VASc was published in 2010 as a refinement of the older CHADS₂ score. Its central improvement was at the low end: CHADS₂ classified a large group of patients as intermediate risk, whereas CHA₂DS₂-VASc identifies a genuinely low-risk group — men scoring 0 and women scoring 1 — in whom anticoagulation can reasonably be withheld.

It is a decision aid for starting anticoagulation. It is not a diagnostic test, it does not measure bleeding risk, and it applies to non-valvular atrial fibrillation only.

The two doubled items — age 75 and over, and a prior stroke or TIA — are doubled because each roughly doubles stroke risk on its own. Prior stroke is the single strongest predictor in the score, which is why anyone with one is already at or above the treatment threshold before anything else is counted.

Age is scored in one band only. Someone aged 78 scores 2 for age and nothing for the 65–74 band. This calculator derives both age and sex points from the fields above rather than offering them as checkboxes, which removes the commonest scoring error.

The rates below are adjusted ischaemic stroke rates per year in patients not taking an anticoagulant, from the Swedish nationwide cohort of 182,678 patients reported by Friberg and colleagues in 2012.

Two caveats are worth stating plainly. The dip at score 8 is in the source data and reflects how few patients reach the highest scores, not a transcription error. And the original Lip 2010 derivation cohort reported materially different figures for the same scores — 1.3% at score 1 rather than 0.6%, for instance — which is why any published table should say which cohort it came from.

Contemporary cohorts often report lower absolute rates than either source, because background management of hypertension, diabetes and cholesterol has improved. Treat these numbers as a well-validated ranking rather than a precise prediction for one person.

Where anticoagulation is indicated, a direct oral anticoagulant is preferred over warfarin in non-valvular atrial fibrillation. Aspirin is not an adequate substitute — it offers little stroke protection in AF while still carrying bleeding risk, and guidelines no longer recommend it for this purpose.

The 2023 ACC/AHA/ACCP/HRS atrial fibrillation guideline moved somewhat away from treating CHA₂DS₂-VASc as a fixed threshold, framing the decision around estimated annual risk — broadly, anticoagulation where risk exceeds about 2% per year — and allowing other validated risk scores. The score remains the most widely used starting point, but it is explicitly a conversation rather than a cut-off.

Female sex behaves as a risk modifier rather than an independent risk factor. In a woman with no other risk factors, atrial fibrillation carries much the same stroke risk as it does in an equivalent man — but when other risk factors are present, being female amplifies their effect.

That is why a woman scoring 1 from the sex point alone is treated as low risk and not anticoagulated, while a woman scoring 2 has one genuine risk factor plus the modifier. In practice the rule is: count the score excluding the sex point, and treat the thresholds as identical from there.

Some guidelines now write the score as CHA₂DS₂-VA, dropping the sex point entirely and using one threshold for everyone, partly to avoid ambiguity for patients whose recorded sex does not reflect their situation. This calculator uses the original CHA₂DS₂-VASc definition, which is what the published stroke-rate tables were derived from.

• Bleeding risk. That needs a separate assessment — HAS-BLED and similar — and a high bleeding score is a prompt to address modifiable risk factors, not usually a reason to withhold anticoagulation.
• Valvular atrial fibrillation. Moderate-to-severe mitral stenosis and mechanical heart valves require anticoagulation with warfarin regardless of score, and direct oral anticoagulants are contraindicated in mechanical valves.
• AF burden. The score was derived in clinical AF; how it applies to brief episodes picked up by a wearable or implanted device is still an area of active study.
• Kidney function, which affects both stroke risk and the choice and dose of anticoagulant.
• Whether the patient will actually take the medication, which is often the decisive factor.