HAS-BLED Score Calculator

Uncontrolled hypertension is a modifiable risk factor and one of the most impactful targets for bleeding risk reduction. Sustained systolic blood pressure above 160 mmHg increases the risk of intracerebral haemorrhage in anticoagulated patients by impairing cerebrovascular autoregulation and promoting small vessel disease.

Aggressive blood pressure management to below 140/90 mmHg (or lower in high-risk patients) can reduce both bleeding and stroke risk simultaneously. The ESC recommends addressing uncontrolled hypertension as a first-line intervention when HAS-BLED is elevated.

Key point: This criterion refers to uncontrolled blood pressure, not a history of hypertension. A patient with well-controlled hypertension on medication should score 0 for this criterion.

Abnormal renal function is defined as chronic dialysis, renal transplantation, or serum creatinine ≥200 µmol/L (≥2.26 mg/dL). Impaired renal clearance affects the metabolism of anticoagulants — particularly DOACs such as dabigatran, which is ~80% renally cleared — and disrupts platelet function through uraemic toxins.

CKD stage 4–5 is associated with a 2–3 fold increase in bleeding events during anticoagulation. Dose adjustments are required for most DOACs in renal impairment: rivaroxaban and apixaban have specific CrCl thresholds, while dabigatran is contraindicated below CrCl 30 mL/min in some regions. Regular monitoring of renal function (at least every 6–12 months) is recommended.

This criterion is scored independently from abnormal liver function — a patient may receive 0, 1, or 2 points for the “A” in HAS-BLED.

Abnormal liver function is defined as chronic hepatic disease (e.g., cirrhosis) or biochemical evidence of significant hepatic derangement: bilirubin >2× the upper limit of normal (ULN) in conjunction with AST/ALT/alkaline phosphatase >3× ULN. The liver synthesises coagulation factors (II, VII, IX, X), anticoagulant proteins (protein C, protein S, antithrombin), and fibrinolytic components.

Hepatic dysfunction creates a complex coagulopathy that may paradoxically increase both bleeding and thrombotic risk. Patients with significant liver disease may have unpredictable responses to warfarin (due to impaired factor synthesis) and altered DOAC metabolism (particularly rivaroxaban and apixaban, which undergo hepatic clearance).

Portal hypertension and oesophageal varices represent additional anatomical bleeding sources in cirrhotic patients that amplify the risk of major haemorrhage on anticoagulation.

A history of prior stroke or transient ischaemic attack (TIA) contributes to bleeding risk through several mechanisms: underlying cerebral small vessel disease, prior haemorrhagic transformation, and microbleeds visible on susceptibility-weighted MRI. Notably, stroke history is also a major criterion in CHA₂DS₂-VASc (scoring 2 points), meaning these patients derive the greatest net benefit from anticoagulation despite elevated bleeding risk.

For patients with recent ischaemic stroke, timing of anticoagulation initiation is critical. Current ESC guidance recommends waiting at least 3–12 days after an acute ischaemic stroke before starting or restarting anticoagulation, depending on stroke severity and imaging findings. Brain MRI with susceptibility-weighted imaging should be considered to assess microbleed burden, particularly in elderly patients.

This criterion captures prior major bleeding events (intracranial haemorrhage, gastrointestinal bleeding requiring hospitalisation or transfusion), anaemia, or any inherited or acquired predisposition to bleeding (e.g., thrombocytopenia, von Willebrand disease, platelet disorders).

Gastrointestinal bleeding is the most common major bleeding complication of anticoagulation. In patients with a history of GI bleeding, addressing the underlying cause (e.g., eradication of H. pylori, treatment of peptic ulcer disease, endoscopic management of angiodysplasia) and co-prescribing a proton pump inhibitor may reduce recurrent bleeding risk. DOACs such as apixaban may be associated with lower GI bleeding rates than warfarin.

Prior intracranial haemorrhage warrants specialist haematology or neurology input before restarting anticoagulation, with a risk-benefit discussion documented clearly.

Labile INR is defined as unstable or high INR values, or poor time in therapeutic range (TTR <60%) in patients on vitamin K antagonists (VKAs) such as warfarin. This criterion applies only to patients on VKAs — for patients on DOACs or those not yet on anticoagulation, this should be scored as 0.

Poor INR control is the single strongest modifiable predictor of both bleeding and thromboembolism in warfarin-treated patients. A TTR below 60% is associated with significantly worse outcomes, while TTR above 70% is associated with outcomes comparable to DOACs. Strategies to improve TTR include patient education, pharmacist-led anticoagulation clinics, self-monitoring, and attention to drug and dietary interactions.

Patients with persistently labile INR despite optimisation efforts are strong candidates for switching to a DOAC, which effectively eliminates this risk factor from the HAS-BLED score.

Age over 65 years is a non-modifiable risk factor that reflects age-related vascular fragility, increased fall risk, polypharmacy, declining renal and hepatic function, and cerebral amyloid angiopathy. The annual incidence of major bleeding on anticoagulation approximately doubles for each decade above 65.

However, age is also one of the strongest risk factors for stroke in atrial fibrillation (it contributes to CHA₂DS₂-VASc as well), meaning elderly patients typically have a high net clinical benefit from anticoagulation despite elevated bleeding risk. Falls are commonly cited as a reason to withhold anticoagulation, but studies consistently show that a patient would need to fall approximately 295 times per year for the bleeding risk from falls to outweigh the stroke prevention benefit of warfarin.

Drugs (1 point): Concomitant use of antiplatelet agents (aspirin, clopidogrel, prasugrel, ticagrelor) or non-steroidal anti-inflammatory drugs (NSAIDs) significantly increases bleeding risk when combined with anticoagulation. Dual antiplatelet therapy plus anticoagulation (“triple therapy”) carries the highest risk and should be limited to the shortest necessary duration after coronary stenting. Clinicians should regularly review whether concurrent antiplatelet therapy is still indicated.

Alcohol (1 point): Excessive alcohol consumption (≥8 standard drinks per week) increases bleeding risk through hepatotoxicity, thrombocytopenia, impaired platelet aggregation, and erratic warfarin metabolism. Alcohol also increases the risk of trauma-related bleeding through falls and injuries. This is a modifiable risk factor — counselling on alcohol reduction and referral for alcohol use disorder support should be considered.

The drugs and alcohol components are scored independently — a patient may receive 0, 1, or 2 points for the “D” in HAS-BLED.

This is the most dangerous and most common misapplication of the score. HAS-BLED was designed to flag patients who need closer monitoring and risk factor modification, not to serve as a contraindication to anticoagulation. In the vast majority of patients with AF and CHA₂DS₂-VASc ≥2, the net clinical benefit of anticoagulation is preserved even with HAS-BLED ≥3.

ESC 2020 guidelines explicitly state that a high HAS-BLED score alone should not be used to deny anticoagulation. Studies have shown that patients inappropriately denied anticoagulation due to high bleeding scores experience significantly higher rates of ischaemic stroke with worse outcomes than those who experience major bleeding while anticoagulated.

The labile INR criterion was developed for patients on vitamin K antagonists and reflects the risk associated with time spent outside the therapeutic INR range. For patients on DOACs, this criterion does not apply and should always be scored as 0. Erroneously assigning this point to DOAC patients artificially inflates the score and may lead to inappropriate clinical decisions.

Similarly, for patients who are not yet on any anticoagulation (i.e., the score is being used to assess risk before initiating therapy), the labile INR criterion should be scored as 0 unless there is historical evidence of poor INR control from previous warfarin use.

Several HAS-BLED components are modifiable, and their identification should prompt action rather than simply a numerical score. Uncontrolled hypertension can be treated; unnecessary NSAIDs and antiplatelets can be discontinued; alcohol consumption can be reduced; and labile INR can be improved through better monitoring or a switch to DOACs.

A re-scored HAS-BLED after addressing modifiable factors often drops by 1–3 points. For instance, switching a patient from warfarin with TTR <60% to a DOAC eliminates the labile INR point, while stopping a concurrent NSAID removes the drug interaction point. This “dynamic” use of HAS-BLED is much more clinically useful than a one-time static assessment.

Bleeding risk is not static — it changes with age, new comorbidities, new medications, and changes in renal or liver function. A patient with a low HAS-BLED at age 60 may have a significantly higher score by age 70 with worsening CKD and the addition of aspirin for a new coronary stent. HAS-BLED should be recalculated at regular intervals (at least annually) and whenever there is a significant change in clinical status.

This is particularly important for patients on warfarin, where TTR should be reassessed at each clinic visit to determine whether the labile INR criterion applies.

The HAS-BLED score was derived and validated primarily in European populations enrolled in the Euro Heart Survey. While subsequent validation studies have demonstrated reasonable performance in other populations (including East Asian and North American cohorts), the absolute bleeding rates at each score level may differ across ethnicities and healthcare systems.

East Asian patients on warfarin may have higher bleeding risk at equivalent INR levels compared to European patients, possibly reflecting pharmacogenomic differences in warfarin metabolism (CYP2C9 and VKORC1 polymorphisms). The score has not been extensively validated in Sub-Saharan African or Middle Eastern populations.