Guide détaillé à venir
Nous préparons un guide éducatif complet pour le HAS-BLED Score. Revenez bientôt pour des explications étape par étape, des formules, des exemples concrets et des conseils d'experts.
The HAS-BLED score is a validated clinical tool used to estimate the 1-year risk of major bleeding in patients with atrial fibrillation (AFib) who are receiving or being considered for oral anticoagulation therapy. Developed by Pisters and colleagues and published in Chest in 2010, it was validated in the Euro Heart Survey on Atrial Fibrillation, which followed over 3,000 patients across 35 European countries. The acronym stands for Hypertension (uncontrolled systolic BP > 160 mmHg), Abnormal renal or liver function (1 point each), Stroke history, Bleeding history or predisposition, Labile INR (if on warfarin, time-in-therapeutic-range < 60%), Elderly (age > 65 years), and Drugs (antiplatelet agents or NSAIDs) or alcohol use (1 point each). The score ranges from 0 to 9. A critical clinical principle: HAS-BLED is NOT a tool to withhold anticoagulation. Rather, it is designed to identify and correct modifiable bleeding risk factors before and during anticoagulation. The 2020 ESC guidelines explicitly state that a high HAS-BLED score should trigger active risk-factor management — addressing uncontrolled hypertension, improving INR control, stopping unnecessary antiplatelet agents or NSAIDs, and addressing alcohol use — rather than simply deciding against anticoagulation. Studies consistently show that the net clinical benefit of anticoagulation remains positive even at HAS-BLED scores of 3–4 when the CHA₂DS₂-VASc score is ≥ 2, because stroke prevention outweighs the bleeding risk in most patients.
HAS-BLED Score = H + A + S + B + L + E + D • H — Hypertension: uncontrolled systolic BP > 160 mmHg → 1 point • A — Abnormal renal function: dialysis, transplant, or creatinine > 200 µmol/L → 1 point • A — Abnormal liver function: chronic liver disease (cirrhosis) or biochemical evidence (bilirubin > 2× ULN + ALT/AST/ALP > 3× ULN) → 1 point (A criterion contributes up to 2 points if both renal AND liver dysfunction are present) • S — Stroke: prior stroke history → 1 point • B — Bleeding: prior major bleeding or predisposition to bleeding (e.g. bleeding diathesis, anaemia) → 1 point • L — Labile INR: unstable/high INR or time-in-therapeutic-range < 60% (relevant only if on warfarin) → 1 point • E — Elderly: age > 65 years → 1 point • D — Drugs: concomitant antiplatelet agents or NSAIDs → 1 point • D — Alcohol: ≥ 8 drinks per week → 1 point (D criterion contributes up to 2 points if both drug AND alcohol factors are present) Minimum: 0 | Maximum: 9 Score ≥ 3 = high bleeding risk — identify and correct modifiable factors
- 1Assess the H criterion first: measure resting blood pressure. Uncontrolled hypertension (systolic > 160 mmHg) is the most important modifiable bleeding risk factor — improving BP control can reduce bleeding risk substantially before or during anticoagulation.
- 2Evaluate both A sub-criteria independently: (1) abnormal renal function (dialysis, transplant, or creatinine > 200 µmol/L) and (2) abnormal liver function (cirrhosis, or bilirubin > 2× ULN with transaminases > 3× ULN). Each scores 1 point, so A can contribute 0, 1, or 2 points.
- 3Check S (prior stroke) and B (prior bleeding or predisposition). Prior stroke contributes independently — note that it is also a criterion in CHA₂DS₂-VASc, so these patients score points on both tools. Prior bleeding history includes GI haemorrhage, epistaxis requiring intervention, haematuria, or known coagulopathy.
- 4Assess L (labile INR) only in patients already on warfarin/vitamin K antagonists. Calculate time-in-therapeutic-range (TTR) from INR values over the past 3–6 months. TTR < 60% scores 1 point. Patients on NOACs score 0 for this criterion regardless of adherence.
- 5Evaluate E (elderly, age > 65) and both D sub-criteria: (1) antiplatelet drugs or NSAIDs (including low-dose aspirin, clopidogrel, ibuprofen) and (2) alcohol (≥ 8 units/week). Like A, D can contribute 0, 1, or 2 points.
- 6Sum all points for the total HAS-BLED score (0–9). A score of 0–2 represents low-to-moderate bleeding risk (~1.1–1.9 bleeds per 100 patient-years). A score ≥ 3 represents high risk (~3.7–17+ bleeds per 100 patient-years) and should trigger active modifiable risk-factor management.
- 7For each modifiable risk factor identified, implement corrective action: optimise BP control, switch from warfarin to a NOAC (improves labile INR), review the indication for antiplatelet agents, counsel on alcohol reduction. Reassess HAS-BLED after addressing modifiable factors — the score may decrease substantially.
Low risk — anticoagulate without additional risk-factor intervention
No identifiable bleeding risk factors. This patient's bleeding risk is low and comparable to the background population rate for their age. Anticoagulation is appropriate if CHA₂DS₂-VASc ≥ 1 (male) or ≥ 2 (female).
Moderate risk — anticoagulate, optimise BP, review INR control
Age is a non-modifiable risk factor contributing 1 point. Blood pressure is borderline — tighter control would further reduce bleeding risk. TTR of 65% is above threshold but switching to a NOAC would eliminate the labile-INR concern entirely.
High risk — do NOT stop anticoagulation; correct modifiable factors
Multiple modifiable risk factors are present: uncontrolled hypertension (treat to < 130/80), labile INR (switch to a NOAC), unnecessary aspirin use in AFib without recent ACS/stent (review and likely stop), and alcohol excess (counsel to reduce). Addressing these 4 factors could reduce the score from 7 to ~3.
High risk — but anticoagulation benefit greatly outweighs risk at CHA₂DS₂-VASc 6+
HAS-BLED 3 = high bleeding risk, but this patient's CHA₂DS₂-VASc score is likely ≥ 6 given age, renal disease, and prior stroke. The absolute stroke risk (~10–12%/year) far exceeds the major bleeding risk (~3.7%/year). Apixaban is appropriate here with dose adjustment for renal function (apixaban 2.5mg bd if 2 of 3 criteria: age ≥ 80, weight ≤ 60kg, creatinine ≥ 133 µmol/L).
Paired with CHA₂DS₂-VASc at every AFib patient review to give a complete picture of stroke prevention benefit versus bleeding risk, supporting shared decision-making about anticoagulation
Generating a personalised modifiable risk-factor action plan at each clinic visit — documenting specific interventions for each modifiable criterion identified
Monitoring outcome of interventions over time — if BP was optimised last visit, confirm H criterion has resolved; if patient switched from warfarin to NOAC, confirm L criterion is now 0
Emergency department assessment of AFib patients who present with bleeding while on anticoagulation — HAS-BLED retrospective scoring helps contextualise whether this was a foreseeable high-risk event
Population-level AFib registry analysis to identify and prioritise high-risk patients for clinical pharmacist-led anticoagulation management and monitoring programmes
Prior intracranial haemorrhage
Patients with a history of intracranial haemorrhage (ICH) score 1 point for B (prior bleeding) and 1 point for S (cerebrovascular disease). ICH is the most feared complication of anticoagulation and warrants specialist input from a stroke physician or haematologist before any anticoagulation decision. Left atrial appendage occlusion (LAAO) devices or aspirin alone may be discussed in selected cases.
Switching from warfarin to a NOAC reduces score
Patients with a high HAS-BLED score partly driven by labile INR (L criterion, TTR < 60%) should be considered for switch to a NOAC. This eliminates the L criterion (saving 1 point) and reduces overall bleeding risk. NOACs also have lower intracranial bleeding rates than warfarin in landmark trials, providing additional safety benefit.
Dual antiplatelet therapy with anticoagulation ('triple therapy')
Patients with AFib and recent acute coronary syndrome or coronary stenting may temporarily require triple therapy (anticoagulant + 2 antiplatelet agents). This dramatically increases bleeding risk and should be duration-limited (typically 1–4 weeks), transitioning to dual therapy (anticoagulant + one antiplatelet) and eventually single anticoagulant therapy as soon as clinically appropriate.
Alcohol use — units versus cultural variation
The alcohol criterion (≥ 8 drinks per week) uses the original Euro Heart Survey definition. The underlying concern is both pharmacodynamic (ethanol inhibits coagulation and platelet function) and indirect (alcohol excess leads to falls, liver disease, and poor medication adherence). Alcohol unit definitions vary internationally — the clinician should use clinical judgment based on alcohol volume consumed.
HAS-BLED in valvular atrial fibrillation
HAS-BLED was validated in non-valvular AFib populations. Patients with valvular AFib (moderate-to-severe mitral stenosis or mechanical prosthetic valves) require anticoagulation with warfarin regardless of HAS-BLED score — NOACs are contraindicated. Bleeding risk assessment in these patients should use clinical judgment alongside HAS-BLED, acknowledging the tool was not validated in this group.
| Score | Risk Category | Bleeds per 100 pt-years | Action |
|---|---|---|---|
| 0 | Low | ~1.13 | Anticoagulate — low bleeding risk |
| 1 | Low | ~1.02 | Anticoagulate — low bleeding risk |
| 2 | Moderate | ~1.88 | Anticoagulate — address modifiable factors |
| 3 | High | ~3.74 | High risk — correct modifiable factors, anticoagulate |
| 4 | High | ~8.70 | Very high risk — urgently correct modifiable factors |
| 5 | Very High | ~12.50 | Specialist input recommended before anticoagulation |
| ≥6 | Very High | >15.0 | Specialist haematology/cardiology review essential |
What is the HAS-BLED score used for?
The HAS-BLED score estimates the 1-year risk of major bleeding in patients with atrial fibrillation who are being considered for or are receiving oral anticoagulation therapy. It is primarily used to identify modifiable bleeding risk factors that should be corrected before or during anticoagulation — NOT to decide whether to anticoagulate. The 2020 ESC AFib guidelines recommend using HAS-BLED alongside CHA₂DS₂-VASc for complete stroke and bleeding risk assessment.
Does a high HAS-BLED score mean I should stop anticoagulation?
No. A high HAS-BLED score (≥ 3) should prompt identification and correction of modifiable risk factors, not discontinuation of anticoagulation. Modifiable factors include uncontrolled hypertension, labile INR (switch to NOAC), unnecessary antiplatelet/NSAID use, and excess alcohol consumption. Studies show that for most patients with CHA₂DS₂-VASc ≥ 2, the stroke prevention benefit of anticoagulation outweighs the bleeding risk even at high HAS-BLED scores.
What score is considered high risk?
A HAS-BLED score ≥ 3 is generally considered high risk, corresponding to approximately 3.7+ major bleeds per 100 patient-years. However, the absolute risk varies widely — a score of 3 carries ~3.74 bleeds/100 patient-years while a score of 5+ carries 12.5–17+ bleeds/100 patient-years. The score should always be interpreted in the clinical context of the patient's stroke risk (CHA₂DS₂-VASc) and individual circumstances.
What does 'abnormal renal function' mean in HAS-BLED?
Abnormal renal function in HAS-BLED includes any of the following: dialysis-dependent renal failure, renal transplantation, or serum creatinine concentration ≥ 200 µmol/L (≈ 2.26 mg/dL). This does not include mild CKD (stage 1–2) or creatinine values below this threshold. Note that creatinine affects NOAC dosing independently — always check manufacturer dosing criteria before prescribing a NOAC in CKD.
How is HAS-BLED different from other bleeding scores?
HAS-BLED was specifically developed and validated for atrial fibrillation patients on anticoagulation. Other bleeding scores include ORBIT (which uses haemoglobin and platelet count) and HEMORR₂HAGES (which includes 11 criteria). Comparative studies suggest HAS-BLED and ORBIT have similar predictive accuracy. HAS-BLED is recommended by the ESC and is the most widely used because it explicitly identifies modifiable risk factors that can be addressed clinically.
Should I score L (labile INR) for patients on NOACs?
No. The L criterion (labile INR, TTR < 60%) applies only to patients taking vitamin K antagonists such as warfarin. Patients on direct oral anticoagulants (NOACs/DOACs) — apixaban, rivaroxaban, dabigatran, edoxaban — score 0 for this criterion because NOACs do not require INR monitoring. This means switching from warfarin to a NOAC can reduce the HAS-BLED score by 1 point if the patient previously had a labile INR.
What counts as 'bleeding predisposition' for the B criterion?
Bleeding predisposition includes any condition that increases intrinsic bleeding risk: prior major bleeding episode (gastrointestinal, intracranial, or haematuria), known bleeding diathesis (haemophilia, von Willebrand disease, platelet disorders), chronic anaemia (often a marker of occult GI blood loss), or current thrombocytopenia (platelet count < 100 × 10⁹/L). Prior intracranial haemorrhage is the highest-risk bleeding history and warrants specialist input before any anticoagulation decision.
How often should HAS-BLED be reassessed?
HAS-BLED should be reassessed at every follow-up visit for AFib patients on anticoagulation — typically every 3–6 months. Several components are dynamic: blood pressure control may improve or worsen, INR stability changes over time, renal function may deteriorate in elderly patients, and patients may start or stop antiplatelet agents or NSAIDs. Reassessment allows proactive identification of newly developed risk factors and verification that modifiable factors remain corrected.
Conseil Pro
Use HAS-BLED as a modifiable risk-factor checklist, not a decision gate. For every high-risk criterion identified, create an action plan: (1) Hypertension → optimise antihypertensive therapy targeting SBP < 130 mmHg; (2) Labile INR → switch to a NOAC; (3) Drugs → review necessity of each antiplatelet/NSAID; (4) Alcohol → brief intervention and counselling. Documenting this action plan in the patient record demonstrates structured clinical decision-making.
Le saviez-vous?
The HAS-BLED acronym was refined from an 11-factor score (HEMORR₂HAGES) down to the 7 most clinically actionable predictors. The key insight of Pisters et al. was that several bleeding risk factors are modifiable — this makes HAS-BLED not just a prediction tool but a treatment guide. In the original validation cohort, a score of 3 corresponded to 3.74 bleeds per 100 patient-years — comparable in absolute terms to the stroke rate at CHA₂DS₂-VASc score 2–3, emphasising that anticoagulation benefit still outweighs risk at this threshold.
Références
- ›Pisters R et al. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation — Chest 2010
- ›Lip GY et al. Comparative validation of a novel risk score for predicting bleeding risk in anticoagulated patients — J Am Coll Cardiol 2011
- ›ESC 2020 Guidelines for the diagnosis and management of atrial fibrillation — European Heart Journal 2021
- ›Roldán V et al. Predictive value of the HAS-BLED and ATRIA bleeding scores for the risk of serious bleeding — Circulation 2013
- ›Lane DA, Lip GYH. Use of the CHA₂DS₂-VASc and HAS-BLED scores to aid decision making for thromboprophylaxis in nonvalvular AF — Circulation 2012