Which validated risk‑assessment scores are used in internal medicine to determine when and how to anticoagulate patients?

Risk Assessment Scores for Anticoagulation in Internal Medicine

Primary Scores by Clinical Indication

Atrial Fibrillation: Stroke Risk Assessment

Use the CHA₂DS₂-VASc score to determine anticoagulation need in all patients with non-valvular atrial fibrillation. 1

CHA₂DS₂-VASc Components (1 point each, except where noted):

• Congestive heart failure (1 point) 1
• Hypertension (1 point) 1
• Age ≥75 years (2 points) 1
• Diabetes mellitus (1 point) 1
• Stroke/TIA/thromboembolism history (2 points) 1
• Vascular disease (prior MI, peripheral artery disease, aortic plaque) (1 point) 1
• Age 65-74 years (1 point) 1
• Sex category (female) (1 point) 1

Anticoagulation Thresholds:

• Score 0 (men) or 1 (women): No anticoagulation recommended 1
• Score 1 (men) or 2 (women): Anticoagulation should be considered 1
• Score ≥2 (men) or ≥3 (women): Anticoagulation strongly recommended 1

The annual stroke risk ranges from 0.6-1.3% with a score of 1, up to >15% with a score of 9. 1

---

Atrial Fibrillation: Bleeding Risk Assessment

Calculate the HAS-BLED score in all AF patients receiving anticoagulation to identify modifiable bleeding risk factors and guide monitoring intensity—never to withhold anticoagulation. 1, 2

HAS-BLED Components (1 point each):

• Hypertension (uncontrolled systolic BP >160 mmHg) 2
• Abnormal renal function (dialysis, transplant, creatinine >200 µmol/L) OR abnormal liver function (cirrhosis, bilirubin >2× normal, AST/ALT/ALP >3× normal) (1 point each) 2
• Stroke history 2
• Bleeding history or predisposition (prior major bleed, anemia, thrombocytopenia) 2
• Labile INRs (time in therapeutic range <60% for warfarin patients) 2
• Elderly (age >65 years) 2
• Drugs (concomitant antiplatelet agents or NSAIDs) OR alcohol (≥8 drinks/week) (1 point each) 2

Risk Stratification:

• Score 0: 0.6-1.1% annual major bleeding risk 2
• Score 1: 1.0-1.5% annual major bleeding risk 2
• Score 2: 1.9-3.2% annual major bleeding risk 2
• Score ≥3: 3.7-21.4% annual major bleeding risk (high risk) 2

Patients with HAS-BLED ≥3 require more frequent clinical reviews and aggressive management of modifiable risk factors, but anticoagulation should NOT be withheld. 1, 2 The net clinical benefit of anticoagulation is often greatest in high bleeding risk patients because they typically also have high stroke risk. 2, 3

Alternative Bleeding Score: ORBIT

The ORBIT score is an alternative validated tool with superior discriminative ability (C-index 0.66) that does not require INR data. 4, 3

ORBIT Components:

• Older age (≥75 years): 1 point 4
• Reduced hemoglobin/Reduced hematocrit/anemia: 2 points 4, 3
• Bleeding history: 2 points 4, 3
• Insufficient kidney function (eGFR <60 mL/min): 1 point 4, 3
• Treatment with antiplatelet agents: 1 point 4, 3

Risk Categories:

• Low risk (0-2 points): 2.4% annual bleeding 4, 3
• Intermediate risk (3 points): 4.7% annual bleeding 4, 3
• High risk (≥4 points): 8.1% annual bleeding 4, 3

For warfarin-treated patients specifically, HAS-BLED is preferred because it incorporates labile INR/time in therapeutic range. 2, 3

---

Perioperative Thromboembolism Risk

Use a three-tiered risk stratification scheme (low, moderate, high) based on the indication for chronic anticoagulation to guide bridging decisions. 1

Mechanical Heart Valves:

• High TE risk: Caged-ball or tilting disk valve in any position, bileaflet valve in mitral position, recent stroke/TIA (within 3 months) 1
• Moderate TE risk: Bileaflet valve in aortic position with additional stroke risk factors 1
• Low TE risk: Bileaflet valve in aortic position without other risk factors 1

Atrial Fibrillation (Perioperative):

• High TE risk: CHA₂DS₂-VASc ≥7 OR recent stroke/TIA (within 3 months) 1
• Moderate TE risk: CHA₂DS₂-VASc 5-6 1
• Low TE risk: CHA₂DS₂-VASc ≤4 1

The CHADS₂ score (older version) shows stepwise 30-day post-operative stroke risk: 1.0-2.0% (score 0-2), 2.6-3.6% (score 3-4), and 3.6-7.3% (score 5-6). 1

Venous Thromboembolism:

• High TE risk: VTE within 3 months (especially within 1 month), severe thrombophilia, vena cava filter 1
• Moderate TE risk: VTE 3-12 months ago 1
• Low TE risk: VTE >12 months ago without recurrence 1

---

Surgical VTE Prophylaxis Risk

Use the Caprini Risk Assessment Model to estimate VTE risk and guide thromboprophylaxis for most non-orthopedic surgical patients. 5

The Caprini score stratifies patients into risk categories based on cumulative points from multiple risk factors including age, BMI, surgical type, mobility status, and medical comorbidities. 5, 6 Patients undergoing major orthopedic surgery should receive short-term anticoagulation regardless of score. 5

---

Hospitalized Medical Patients: VTE Risk

Risk assessment models for hospitalized medical patients have generally weak predictive accuracy (C-statistics often <0.7), with the Padua prediction score and IMPROVE models being most widely validated. 6

The Padua prediction score and IMPROVE VTE model are the most commonly used tools for medical inpatients, though sensitivity and specificity vary markedly across validation studies (sensitivity 12-100%, specificity 7.2-100%). 6

---

COVID-19 Patients: Thrombosis Risk

Risk assessment models to estimate thrombotic risk have been validated in hospitalized COVID-19 patients (critically or non-critically ill), with modest prognostic performance. 1

No specific RAM has been definitively recommended, and the 2025 ASH guidelines suggest prophylactic-intensity anticoagulation for most hospitalized COVID-19 patients without confirmed VTE, regardless of RAM score. 1 Higher-intensity anticoagulation may be preferred for patients judged to be at low bleeding risk and high thrombotic risk through individualized assessment. 1

---

Critical Clinical Principles

When to Use Each Score:

1. CHA₂DS₂-VASc: All AF patients to determine if anticoagulation is needed 1
2. HAS-BLED or ORBIT: All AF patients on anticoagulation to identify modifiable bleeding risk factors and guide monitoring frequency 1, 2, 3
3. Perioperative TE risk stratification: Patients on chronic anticoagulation undergoing elective procedures to guide bridging decisions 1
4. Caprini: Non-orthopedic surgical patients to determine VTE prophylaxis need 5
5. Padua/IMPROVE: Hospitalized medical patients to determine VTE prophylaxis need 6

Universal Caveats:

Bleeding risk scores must never be used as standalone reasons to withhold or discontinue anticoagulation. 1, 2, 3 The purpose is to identify modifiable risk factors (uncontrolled hypertension, excess alcohol, unnecessary antiplatelet agents, anemia, labile INRs) and increase monitoring frequency. 1, 2

Reassess all risk scores at every patient contact because risk is dynamic and changes over time. 2, 3 Serial "delta HAS-BLED" assessments are more predictive of major bleeding than a single baseline score. 2

Do not withhold oral anticoagulation solely because of fall risk—patients would need to fall 295 times per year for bleeding risk to outweigh stroke prevention benefit. 1