π©Έ HAS-BLED Bleeding Risk Calculator
Bleeding risk to balance against stroke risk.
For informational purposes only, not medical advice. Consult a healthcare professional.
HAS-BLED sums bleeding risk factors for anticoagulated patients: a 72-year-old with hypertension, prior bleed, and NSAID use scores 4, about 8.7 major bleeds per 100 patient-years. Check the applicable factors for the score, annual risk, and what to do about it.
How to Use This Calculator
Pisters et al. derived HAS-BLED in 2010 for atrial fibrillation patients on anticoagulation; guidelines use it alongside stroke-risk scores to balance the decision. Check every factor that applies.
Check hypertension only if uncontrolled (SBP >160). Controlled hypertension does not score.
Check renal and liver dysfunction by the strict definitions (dialysis/transplant/creatinine β₯2.26 mg/dL; cirrhosis or bilirubin >2Γ with enzymes >3Γ).
Count drugs and alcohol separately (up to 2 points). Read the score: β₯3 is high risk and triggers review of modifiable factors.
The Score and Bleeding Rates
The rates come from Lip's validation cohort on anticoagulation: risk climbs gently to 2 points then steeply, which is why 3 is the action threshold. The acronym's clinical genius is that half its points are modifiable: control the blood pressure, stabilize the INR (or switch to a DOAC), stop the NSAID, counsel the alcohol. A HAS-BLED of 4 built from modifiable factors is a to-do list; the same 4 built from age, prior stroke, and renal failure is a prognosis. The score predicts bleeding on anticoagulation specifically; off anticoagulation the numbers do not apply.
Worked Example
72 years old (+1), SBP 170 (+1), GI bleed two years ago (+1), daily ibuprofen (+1): total 4, ~8.7% annual major bleeding risk, high risk. Modifiable fixes: control BP (removes 1), stop NSAID (removes 1): new score 2, ~1.9%: the review paid off. Contrast: 68-year-old (+1) with labile INRs on warfarin (+1): score 2, moderate; switching to a DOAC removes the labile-INR point entirely. I verified the calculator applies the 0-9 range and the published rates. The pattern: score first, then attack every reversible point before deciding anything about the anticoagulant.
The Decision It Informs
| HAS-BLED | Annual major bleed | Action |
|---|---|---|
| 0-2 | ~1-2% | Anticoagulate; routine follow-up |
| β₯3 | ~4-12%+ | Address modifiable factors; closer follow-up; usually still anticoagulate |
The critical guideline point: high HAS-BLED alone is not a reason to withhold anticoagulation. Net benefit usually persists because the stroke being prevented (often disabling or fatal) outweighs the bleed risk, and many high HAS-BLED patients are also high stroke risk. Withholding leaves the stroke risk fully exposed while the bleeding risk it "avoids" was partly modifiable anyway. Document the score, fix what is fixable, choose the anticoagulant with the better bleeding profile for the patient (DOACs beat warfarin on intracranial bleeding), and follow closely. For the other side of the balance, our Framingham score is primary prevention while AF stroke risk needs CHAβDSβ-VASc; our blood pressure checker tracks the modifiable point, and our GFR calculator defines the renal one.
Modifiable vs Fixed Points
Modifiable: uncontrolled hypertension, labile INRs (switch agent or improve TTR), NSAIDs/antiplatelets (stop or add PPI), alcohol excess. Partially modifiable: anemia workup, fall risk. Fixed: age, prior stroke, prior bleed, renal/liver disease. The clinical visit after a high score should walk this list explicitly: each modifiable point removed is bleeding risk genuinely lowered, unlike the score's fixed points which only inform vigilance. Re-score after interventions; the number should move.
Limits
HAS-BLED was derived in warfarin-era AF cohorts; DOAC-era absolute rates run lower, though ranking holds. It underweights some real risks (falls, frailty, thrombocytopenia) and its definitions (e.g. SBP >160) are stricter than everyday "uncontrolled." It does not apply to non-AF anticoagulation (VTE treatment, valves) where bleeding calculus differs. Use it as the structured review it is, not a stop/go gauge.
Bleeding on anticoagulation also needs an action plan alongside the score: know which bleeds go to the emergency department, carry an updated medication list, and discuss reversal options with your clinician before you need them.
HAS-BLED in the DOAC Era
HAS-BLED was born in the warfarin era, and direct oral anticoagulants changed the bleeding landscape: DOACs roughly halve intracranial bleeding versus warfarin with similar or lower major bleeding overall, which shifts absolute risk down across every score. The score's ranking still works (a 5 still bleeds more than a 2), but the published rates overstate modern DOAC bleeding somewhat. Two practical consequences: the labile-INR point vanishes on a DOAC, often dropping the score by a full point at the switch, and the threshold for "acceptable" bleeding risk arguably loosens since the catastrophic bleeds (intracranial) fell most. What has not changed: the modifiable-factor review. NSAIDs still cause GI bleeds on apixaban, alcohol still raises risk on rivaroxaban, and uncontrolled hypertension still predicts intracranial events. Score every anticoagulated AF patient yearly regardless of agent, fix the fixable, and let the DOAC's better profile do the rest.
For informational purposes only, not medical advice. Consult a healthcare professional.
Frequently Asked Questions
What is the HAS-BLED score?
A 9-point bleeding-risk score for anticoagulated AF patients: hypertension, renal/liver dysfunction, stroke, bleeding history, labile INR, elderly, drugs/alcohol. 3+ is high risk.
What is a high HAS-BLED score?
3 or more: ~3.7% annual major bleeding at 3, rising to ~12.5% at 5+. It triggers review of modifiable factors and closer follow-up, not automatic avoidance of anticoagulation.
Should anticoagulation be stopped if HAS-BLED is high?
Usually not: guidelines say high HAS-BLED flags caution and modifiable-factor review, because the stroke risk being treated typically outweighs the bleed risk. The decision is individualized.
How can HAS-BLED be lowered?
Control blood pressure, stabilize INR or switch to a DOAC, stop NSAIDs, reduce alcohol, work up anemia. These modifiable points genuinely lower bleeding risk when fixed.
What is labile INR in HAS-BLED?
Unstable INRs, frequent high values, or under 60% of time in therapeutic range on warfarin. Switching to a DOAC removes this point entirely.
Does HAS-BLED apply to DOACs?
It was derived in warfarin-era cohorts; absolute bleeding rates are lower with DOACs, but the score still ranks patients usefully. The modifiable-factor review applies regardless of agent.
What counts as abnormal renal function in HAS-BLED?
Dialysis, transplant, or creatinine β₯200 Β΅mol/L (2.26 mg/dL). Everyday mild CKD does not score; the definition is strict.
HAS-BLED vs CHA2DS2-VASc: which matters more?
They answer different questions: CHAβDSβ-VASc estimates stroke risk (reason to anticoagulate), HAS-BLED estimates bleeding risk (reason for caution and review). Guidelines use both together.
Can HAS-BLED be used for DVT treatment?
It was validated in atrial fibrillation; applying it to VTE anticoagulation or mechanical valves is off-label use of the tool, though clinicians sometimes do so cautiously.
How often should HAS-BLED be reassessed?
At least annually and whenever something changes: new bleed, new interacting drug, worsening renal function, or loss of INR control. Modifiable factors should be rechecked after intervention.