Audited ·Last updated 29 Jul 2026·4 citations·Tier 2·0 uses

HAS-BLED Score Calculator (Bleeding Risk in Atrial Fibrillation)

Score HAS-BLED bleeding risk in atrial fibrillation and see how much of the score comes from modifiable factors. A high score prompts review, not stopping.

HAS-BLED Score Calculator

over 65 = +1
yrs
Uncontrolled hypertension — systolic BP above 160 mmHg
Abnormal renal function — dialysis, transplant, or creatinine 200 µmol/L (2.26 mg/dL) or more
Abnormal liver function — cirrhosis, or bilirubin above 2× normal with AST/ALT/ALP above 3× normal
Previous stroke
Prior major bleeding, or a bleeding predisposition such as anaemia
Labile INR — unstable or high INRs, or time in therapeutic range below 60% (vitamin K antagonist only)
Concomitant antiplatelet agents or NSAIDs
Eight or more alcoholic drinks per week
HAS-BLED score
1
Clinical decision aid for clinicians — a high score prompts review, never withholding anticoagulation.
Bleeding risk band
Not high risk (0 to 2)
Modifiable points
0
Age points
1

Background.

HAS-BLED estimates the one-year risk of major bleeding in a patient with atrial fibrillation who is on, or being considered for, oral anticoagulation. The most important thing to know about it is what it is not for. The 2020 European Society of Cardiology guideline says plainly that a high bleeding risk score per se should not lead to withholding anticoagulation, and the 2023 ACC/AHA guideline agrees that a bleeding score is not a sole basis for that decision. A high HAS-BLED score is a prompt to fix what can be fixed and to see the patient sooner. It is a clinical decision aid for clinicians, not a diagnosis and not a treatment recommendation.

That is why this calculator reports a second number beside the total: how many of the points come from factors the ESC classes as modifiable. Four of the nine do — uncontrolled systolic blood pressure above 160, labile INR control on a vitamin K antagonist, concomitant antiplatelet or NSAID use, and eight or more alcoholic drinks a week. A score of 4 built entirely from those four is a very different clinical problem from a score of 4 built from prior stroke, prior bleed, renal impairment and age, and reading only the total hides the difference.

The acronym has seven letters but the score has nine points, and this trips up a surprising number of reproductions. The "A" covers abnormal renal function and abnormal liver function as two separately scored items, and the "D" covers drugs and alcohol as two more. Every item is worth exactly one point, so the total runs from 0 to 9. If you see HAS-BLED described as a seven-point score, that source is miscounting.

The operational definitions matter as much as the letters. Hypertension here means uncontrolled hypertension with a systolic pressure above 160 mmHg, not simply a diagnosis of hypertension — this is the one component that differs most from the way the same word is used in CHA₂DS₂-VASc, where any history of hypertension scores. Abnormal renal function means dialysis, transplant, or a serum creatinine of 200 µmol/L (2.26 mg/dL) or more. Abnormal liver function means cirrhosis, or bilirubin above twice the upper limit of normal together with transaminases or alkaline phosphatase above three times normal. Labile INR means unstable or high INRs, or a time in therapeutic range below 60 percent, and it has no meaning at all for a patient on a direct oral anticoagulant.

The age criterion is worth a specific warning. The derivation paper prints it as "Elderly (> 65 years)" — strictly greater than 65. A large number of reproductions render it as 65 or over. For an integer age those differ at exactly one value, and because the high-risk cut sits at 3, that one value can move a patient across the threshold. This calculator implements the derivation paper's own wording and shows the age point as its own output so you can see whether it is what carried the total over the line.

The calculator does not output an annual bleeding percentage. The derivation cohort was 3,978 patients in the Euro Heart Survey with 53 major bleeds in the first year, predominantly anticoagulated with vitamin K antagonists as prescribing stood in 2010. Absolute rates from that cohort do not describe a modern population on direct oral anticoagulants, and no defensible replacement figure could be sourced, so none is shown. The score, the band and the modifiable-factor count are what this page will give you.

Finally, HAS-BLED answers only half of the anticoagulation question. Stroke risk is the other half, and the two are meant to be read together — a patient can be high risk on both, which is common and is usually an argument for anticoagulating carefully rather than not at all. A companion CHA₂DS₂-VASc calculator is linked from this page.

What is has-bled score calculator?

HAS-BLED is a nine-point bleeding risk score derived by Pisters and colleagues in 2010 from 3,978 patients in the Euro Heart Survey on Atrial Fibrillation. It predicts major bleeding within one year, defined in the derivation study as intracranial bleeding, bleeding requiring hospitalisation, a haemoglobin decrease of more than 2 g/L, or bleeding requiring transfusion. Its C-statistic in the overall derivation population was 0.72.

The acronym expands to Hypertension, Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile INR, Elderly, and Drugs/alcohol concomitantly. Because two of those letters cover two items each, the maximum score is 9 rather than 7.

Its role in modern guidelines is narrow and specific: identify patients at high bleeding risk so they can be reviewed earlier and more often, and identify the modifiable contributors so they can be treated. It is explicitly not a gate on anticoagulation. Patients with the highest bleeding scores often have the highest stroke risk as well, and withholding anticoagulation from them on the basis of a bleeding score generally trades a smaller harm for a larger one.

What HAS-BLED is not: a nuisance-bleeding predictor, a reason to choose one anticoagulant over another, a tool validated for patients without atrial fibrillation, or a number that means anything for a patient on no antithrombotic therapy at all.

How to use this calculator.

  1. Enter the patient's age. The score adds a point above 65 — strictly above, following the derivation paper's own wording.
  2. Answer the hypertension item on uncontrolled systolic pressure above 160 mmHg, not on whether the patient carries a hypertension diagnosis. This is the component most often over-scored.
  3. Score renal and liver function as two separate items. Both can apply to the same patient.
  4. Answer the labile INR item only for a patient on a vitamin K antagonist. On a direct oral anticoagulant it does not apply and should be left at No.
  5. Score drugs and alcohol as two separate items — antiplatelet agents or NSAIDs for the first, eight or more drinks per week for the second.
  6. Read the modifiable-points output next to the total. It tells you how much of the risk you can actually do something about.
  7. Read the score alongside a stroke risk score, not instead of one, and remember that a high bleeding score is a prompt for closer follow-up rather than for stopping anticoagulation.

The formula.

HAS-BLED = H + A(renal) + A(liver) + S + B + L + E + D(drugs) + D(alcohol), each worth 1

HAS-BLED is an unweighted count. Nine items each score one point: uncontrolled hypertension with systolic BP above 160 mmHg; abnormal renal function; abnormal liver function; previous stroke; prior major bleeding or a bleeding predisposition; labile INR; age over 65; concomitant antiplatelet agents or NSAIDs; and eight or more alcoholic drinks per week. The total runs from 0 to 9. Nothing subtracts.

Rounding stage: there is none. All nine weights are integers and the total is an exact integer sum. The only numeric input is age, and it is compared as entered against 65 without rounding, so an age of 65.0 scores 0 and 65.1 scores 1 — the literal reading of "greater than 65 years".

The age comparison is a documented point of disagreement between sources. The derivation paper prints "Elderly (> 65 years)". Many reproductions render the same criterion as "65 or over". This calculator uses the strict form from the derivation paper, so a 65-year-old scores 0 age points and a 66-year-old scores 1. Because the high-risk cut is at 3, a patient with two other points sits below the threshold at 65 and above it at 66, which is why the age point is exposed as its own output rather than folded silently into the total.

The band is a single cut. A total of 3 or more is high bleeding risk in the 2020 ESC framing, which recommends earlier and more frequent clinical review for those patients and the active treatment of modifiable factors. Totals of 0 to 2 are reported as not high risk. No finer stratification is offered below 3 because no source defines one, and inventing intermediate bands would give the score a precision it does not have.

The modifiable-points output counts how many of the scored points come from the four items the ESC classes as modifiable: uncontrolled systolic blood pressure, poor INR control on a vitamin K antagonist, concomitant antiplatelet or NSAID use, and excess alcohol. It is always a subset of the total and can never exceed it.

Worked through the example on this page: uncontrolled hypertension 1, abnormal renal function 1, bleeding history 1, labile INR 1, and age 74 giving 1 sum to 1 + 1 + 1 + 1 + 1 = 5. Five is 3 or more, so the band is high bleeding risk. Two of those five points — the hypertension and the labile INR — are modifiable.

A worked example.

Example

A 74-year-old on warfarin for non-valvular atrial fibrillation attends for anticoagulation review. Clinic systolic blood pressure is 172 mmHg on three readings. Serum creatinine is 240 µmol/L on a background of chronic kidney disease. There was an upper gastrointestinal bleed two years ago that required admission and transfusion. Time in therapeutic range over the last six months is 48 percent. Liver function is normal, there is no history of stroke, no antiplatelet or NSAID is prescribed, and alcohol intake is around three units a week. Five items score. Uncontrolled hypertension contributes 1. Abnormal renal function contributes 1. Bleeding history contributes 1. Labile INR contributes 1. Age 74 is over 65 and contributes 1. Liver function, stroke history, concomitant drugs and alcohol all score 0. The total is 1 + 1 + 1 + 1 + 1 = 5, and 5 is 3 or more, so the band is high bleeding risk. The more useful number is the second one. Two of those five points — the uncontrolled blood pressure and the poor INR control — are modifiable. Treating the hypertension and either improving INR control or moving off a vitamin K antagonist addresses them directly. The remaining three points reflect age, kidney function and a past bleed, none of which can be reversed. Nothing in this result argues for stopping anticoagulation. It argues for treating the blood pressure, reviewing the anticoagulant choice, and seeing this patient again sooner than routine. If the same patient's stroke risk score is also high — which at 74 with vascular comorbidity it very likely is — then the two scores together describe a patient who needs anticoagulation delivered more carefully, not withdrawn.

abnormal Renal Functionyes
concomitant Drugsno
hypertension Uncontrolledyes
abnormal Liver Functionno
alcohol Useno
stroke Historyno
labile Inryes
bleeding Historyyes
age74

Frequently asked questions.

Does a HAS-BLED score of 3 or more mean anticoagulation should be stopped?
No. The 2020 ESC guideline states that a high bleeding risk score per se should not lead to withholding anticoagulation, and the 2023 ACC/AHA guideline agrees that a bleeding score is not a sole basis for that decision. A score of 3 or more identifies a patient for earlier and more frequent review and for active treatment of modifiable bleeding factors. Patients with high bleeding scores usually have high stroke scores too, and withholding anticoagulation from them typically trades a smaller harm for a larger one.
Why is the maximum 9 when the acronym has seven letters?
Two letters cover two items each. The "A" is abnormal renal function and abnormal liver function, scored separately, and the "D" is drugs and alcohol, also scored separately. Every item is worth one point, so a patient with both renal and liver impairment scores 2 for the single letter A. Sources describing HAS-BLED as a seven-point score have collapsed those pairs. This calculator presents all nine as separate inputs so the total can be reconstructed from the page.
Does the hypertension item mean any diagnosis of hypertension?
No, and this is the most commonly over-scored component. HAS-BLED asks about uncontrolled hypertension, defined as a systolic blood pressure above 160 mmHg. A patient with well-controlled treated hypertension scores 0 here. That is the opposite of CHA₂DS₂-VASc, where any history of hypertension scores a point. Applying the CHA₂DS₂-VASc reading to HAS-BLED inflates the bleeding score for a large fraction of patients, and because the high-risk cut sits at 3 it moves people across it.
What if the patient is on a direct oral anticoagulant rather than warfarin?
Leave the labile INR item at No — there is no INR to be labile, and the criterion has no meaning outside vitamin K antagonist therapy. The other eight items still apply. Be aware that the score was derived in a cohort anticoagulated predominantly with vitamin K antagonists, so its absolute calibration in a DOAC population is not established, even though it continues to be used and recommended. This is one reason the page does not print an annual bleeding percentage.
Is the age criterion over 65 or 65 and over?
The derivation paper prints "Elderly (> 65 years)" — strictly over 65 — and this calculator implements that. Many reproductions render it as 65 or over. For an integer age the two differ only at 65 itself, but with a high-risk cut at 3 that single point can change the band. The age contribution is shown as a separate output so you can see whether it is what carried the total across the threshold, and so you can adjust if your local protocol uses the inclusive form.
Why does this page not show a percentage risk of bleeding?
Because no defensible figure could be sourced for a modern population. The derivation study followed 3,978 patients and recorded 53 major bleeds (1.5 percent) in the first year, in a cohort anticoagulated as prescribing stood in 2010 — overwhelmingly with vitamin K antagonists. Per-score rates from that cohort do not transfer cleanly to patients on direct oral anticoagulants, and quoting them as though they did would give the page a precision it has not earned. The score, the band and the modifiable count are what it reports.
How should HAS-BLED and CHA₂DS₂-VASc be used together?
As two halves of one conversation, not as a subtraction. CHA₂DS₂-VASc estimates the risk of stroke without anticoagulation; HAS-BLED estimates the risk of major bleeding with it. Guidelines expect both to be assessed, with the bleeding score used to identify and treat modifiable factors and to set the follow-up interval. There is no published rule that says a HAS-BLED score above a CHA₂DS₂-VASc score means do not anticoagulate, and treating the two as directly comparable numbers is a misuse of both.

References& sources.

  1. [1]Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJGM, Lip GYH. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest. 2010;138(5):1093–1100. Source of the acronym, the "Elderly (> 65 years)" wording, the major-bleeding endpoint definition, and the derivation figures (n = 3,978; 53 major bleeds; C-statistic 0.72). (Abstract retrieved verbatim 2026-07-29; publisher full text paywalled.)
  2. [2]Hindricks G, Potpara T, Dagres N, et al. 2020 ESC Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with EACTS. Eur Heart J. 2021;42(5):373–498. Source of the ≥3 high-risk threshold with earlier and more frequent review, of the modifiable bleeding risk factor classification used by the modifiable-points output, and of the statement that a high bleeding risk score per se should not lead to withholding anticoagulation. (Retrieved 2026-07-29.)
  3. [3]Joglar JA, Chung MK, Armbruster AL, et al. 2023 ACC/AHA/ACCP/HRS Guideline for the Diagnosis and Management of Atrial Fibrillation. Circulation. 2024;149(1):e1–e156. Independent agreement that a bleeding risk score is not a sole basis for withholding anticoagulation. (Bibliographic — ahajournals.org blocks automated retrieval; consulted 2026-07-29.)
  4. [4]StatPearls, HAS-BLED Scores. Operational thresholds used for the component definitions on this page: uncontrolled hypertension systolic > 160 mmHg; creatinine ≥ 200 µmol/L (2.26 mg/dL) or dialysis/transplant; cirrhosis or bilirubin > 2× ULN with AST/ALT/ALP > 3× ULN; labile INR / time in therapeutic range < 60 %; ≥ 8 alcoholic drinks per week. (Retrieved 2026-07-29; secondary point-of-care reference, used only to pin operational thresholds already implied by the derivation paper.)

In this category

Embed

Quanta Pro

Paid features are coming later.

  • All 762 calculators remain free
  • No billing is enabled
Coming soon