Should anticoagulation (anticoagulant therapy) be held in patients with cerebral amyloid angiopathy (CAA)?

Should Anticoagulation Be Held in Patients with Cerebral Amyloid Angiopathy?

Yes, anticoagulation should generally be held (or avoided) in patients with cerebral amyloid angiopathy, particularly those with lobar intracerebral hemorrhage, multiple cortical microbleeds, or cortical superficial siderosis, as the risk of recurrent catastrophic intracranial bleeding substantially outweighs the benefits of stroke prevention in most cases. 1

The Evidence Against Anticoagulation in CAA

The American Heart Association/American Stroke Association explicitly identifies lobar ICH suggestive of CAA as a bleeding risk that outweighs the benefits of anticoagulation in patients with atrial fibrillation 1. This recommendation is based on the exceptionally high recurrence rate of intracerebral hemorrhage in CAA patients:

• Patients with CAA and prior lobar ICH have an approximately 7% annual recurrence risk, compared to about 1% for those not fulfilling CAA criteria 2
• CAA with cortical subarachnoid hemorrhage carries an even higher annual lobar ICH risk of 19% 3
• Decision analysis modeling demonstrates that in elderly patients with lobar hemorrhage likely due to amyloid angiopathy, avoiding warfarin increases quality-adjusted life years by 1.9 years compared to continuing anticoagulation 1

Critical Distinction: Lobar vs. Deep Hemorrhage Location

The location of hemorrhage fundamentally determines anticoagulation decisions:

• Lobar hemorrhages in elderly nonhypertensive patients characteristically indicate CAA and carry the highest recurrence risk, making anticoagulation contraindicated 1
• Deep hemorrhages (basal ganglia, thalamus, pons, cerebellum) typically result from hypertensive arteriopathy and carry different, lower recurrence risks 1
• For patients with small deep ICH, the risk of restarting versus withholding warfarin is similar, and the 2018 CHEST guidelines suggest the balance of net benefit from anticoagulation may be more favorable in those with deep ICH 2, 1

MRI Findings That Strengthen the Contraindication

You must obtain MRI with gradient-echo or susceptibility-weighted imaging to detect CAA biomarkers—CT alone is insufficient 1:

• Multiple juxtacortical microhemorrhages (≥4 microbleeds <10 mm) are highly specific for CAA and highly predict future bleeding risk, contraindicating anticoagulation 1
• Lobar macrohemorrhages >10 mm in diameter significantly increase recurrent hemorrhage risk with anticoagulation 1
• Cortical superficial siderosis indicates severe CAA and represents a strong contraindication to anticoagulation 3
• Patients with ≥2 cortical microbleeds require in-depth risk-benefit analysis, though most should avoid anticoagulation 3

When Anticoagulation Might Be Reconsidered (Rare Exceptions)

The 2018 CHEST guidelines provide a narrow window for reconsidering anticoagulation, but only after careful consideration of specific criteria 2:

• After acute spontaneous ICH in patients with AF and high ischemic stroke risk, anticoagulation with a NOAC (not warfarin) may be considered only if:
  • The hemorrhage was deep (not lobar) 2, 1
  • There is no neuroimaging evidence of CAA (no multiple microbleeds, no cortical superficial siderosis) 2, 1
  • At least 4 weeks have elapsed since the acute hemorrhage 2, 1
  • The patient has very high ischemic stroke risk (CHA₂DS₂-VASc ≥4) 3

This exception does NOT apply to patients with lobar ICH or MRI evidence of CAA—these patients should not receive anticoagulation 1.

The Preferred Alternative: Left Atrial Appendage Occlusion

For CAA patients with atrial fibrillation requiring stroke prevention, left atrial appendage closure (LAAC) is the preferred strategy over oral anticoagulation 1, 4:

• LAAC reduces ischemic stroke and systemic embolism from AF without exposing patients to long-term bleeding risk from anticoagulation 2
• Small observational studies demonstrate that LAAC appears safe and tolerable in CAA patients, with no documented ischemic strokes or symptomatic ICH during the 30 days after device implantation 5
• In a cohort of 26 CAA patients (13 with prior lobar hemorrhage) followed for an average of 25 months after LAAC, only one ischemic stroke occurred, with no other thromboembolic events 5
• Periprocedural management uses antiplatelet therapy (clopidogrel and/or aspirin for 6 weeks) rather than anticoagulation in patients with symptomatic ICH 5

Common Pitfalls to Avoid

Do not assume all ICH carries equal recurrence risk: Deep hemorrhages from hypertension have fundamentally different risk profiles than lobar hemorrhages from CAA 1. Making this distinction requires both clinical context (age, hypertension history) and neuroimaging.

Do not rely on CT alone: MRI with gradient-echo or susceptibility-weighted imaging is mandatory to detect microhemorrhages and superficial siderosis that indicate CAA 1. CT will miss these critical findings that determine bleeding risk.

Do not use bleeding risk scores to withhold anticoagulation in non-CAA patients: The 2024 ESC guidelines explicitly state that bleeding risk scores should not be used to decide on withdrawing oral anticoagulation in AF patients to avoid under-use 2. However, CAA with lobar ICH or multiple microbleeds represents a specific, validated contraindication that supersedes general bleeding risk assessment 1.

Do not restart anticoagulation early if you do reconsider it: If anticoagulation is considered after ICH in a patient without CAA features, delay at least 4 weeks beyond the acute phase (approximately 48 hours) and preferably use NOACs over warfarin 2, 1.

Management Algorithm for CAA Patients

Step 1: Confirm CAA diagnosis using validated clinico-radiological criteria 1, 4:

• Obtain MRI with gradient-echo or susceptibility-weighted imaging
• Assess for lobar microbleeds, cortical superficial siderosis, white matter changes
• Determine hemorrhage location (lobar vs. deep)

Step 2: Stratify bleeding risk 1, 3:

• Highest risk (avoid anticoagulation): Prior lobar ICH, cortical superficial siderosis, cortical subarachnoid hemorrhage, ≥4 cortical microbleeds
• High risk (generally avoid anticoagulation): 2-3 cortical microbleeds, lobar macrohemorrhages
• Lower risk (may consider anticoagulation): Deep ICH only, no CAA biomarkers on MRI

Step 3: Assess ischemic stroke risk 3:

• Calculate CHA₂DS₂-VASc score
• Assess left atrial characteristics on echocardiography

Step 4: Choose management strategy 1, 4, 5:

• For lobar ICH or multiple CAA biomarkers: Permanently avoid anticoagulation; consider LAAC if AF present
• For deep ICH without CAA biomarkers and high stroke risk: May consider NOAC after ≥4 weeks
• For CAA without prior ICH but multiple microbleeds: Generally avoid anticoagulation; consider LAAC if AF present

Step 5: Implement strict blood pressure control 4:

• Target systolic BP 130-150 mmHg acutely, <140/90 mmHg long-term (<130/80 with diabetes/CKD)
• This is the only proven modifiable intervention to reduce CAA-related hemorrhage recurrence 4