stroke prevention in atrial fibrillation
Left atrial appendage stasis → thrombus; risk scales with the substrate (age, HTN, HF, diabetes, prior stroke), not with AF pattern, symptoms or rhythm outcome. CHADS-65: age ≥65 or any CHADS₂ factor → OAC, DOAC over warfarin. Warfarin only for mechanical valves and moderate–severe mitral stenosis. HCM and cardiac amyloidosis → OAC regardless of score. Bleeding risk scores flag fixable factors; they do not decide who gets OAC. Rate/rhythm management → atrial fibrillation.
the CCS algorithm (CHADS-65)
graph LR
A[AF or flutter<br/>any pattern] --> B{Mechanical valve<br/>or mod–severe MS?}
B -->|yes| W[Warfarin]
B -->|no| C{Age ≥65?}
C -->|yes| O[OAC<br/>DOAC preferred]
C -->|no| D{CHF, HTN,<br/>diabetes or<br/>stroke/TIA?}
D -->|yes| O
D -->|no| E{CAD<br/>or PAD?}
E -->|yes| F[ASA 81 mg]
E -->|no| G[No<br/>antithrombotic]
- Reassess yearly — patients age into OAC
- OAC regardless of score: HCM, cardiac amyloidosis → hypertrophic cardiomyopathy, cardiac amyloidosis
- International scores reach the same patients by a different route: CHA₂DS₂-VA ≥2 → OAC, 1 → consider (ESC 2024; sex removed); CHA₂DS₂-VASc ≥2 men / ≥3 women (≥2%/year) → OAC (ACC/AHA 2023)
- Flutter → same algorithm as AF
- ASA is not an alternative to OAC in anyone who qualifies for OAC (less protection, still causes bleeding). If OAC is declined or truly contraindicated, ASA 81 mg may be considered
- Bleeding scores (HAS-BLED) identify modifiable factors — BP, NSAIDs, alcohol, unneeded antiplatelets, labile INR. They do not decide whether to anticoagulate: high-bleeding-risk patients are usually the high-stroke-risk ones
agent and dose
- DOAC over warfarin (less ICH, no monitoring) — including bioprosthetic valves and all valve disease other than moderate–severe MS
- Warfarin (INR 2–3): mechanical valves; moderate–severe or rheumatic MS — rivaroxaban was inferior to VKA in rheumatic AF in INVICTUS (2022). Warfarin TTR <70% → switch to DOAC if eligible
- Frail ≥75 y stable on warfarin with polypharmacy → switching to a DOAC ↑ bleeding in FRAIL-AF (2023) → reasonable to stay on warfarin (ESC IIb)
- Dose by Cockcroft–Gault CrCl; recheck Cr at least yearly (q6 mo if CrCl 30–60, q3 mo if 15–30)
- No off-label dose reduction — underdosing ↑ stroke without ↓ bleeding
| Standard | Reduced dose if | |
|---|---|---|
| apixaban | 5 mg BID | 2.5 mg BID if ≥2 of: age ≥80, weight ≤60 kg, Cr ≥133 µmol/L |
| dabigatran | 150 mg BID | 110 mg BID if age ≥80, or ≥75 with bleeding risk (incl. CrCl 30–50); contraindicated CrCl <30 |
| edoxaban | 60 mg daily | 30 mg daily if CrCl 15–50, weight ≤60 kg, or potent P-gp inhibitor (not amiodarone/verapamil) |
| rivaroxaban | 20 mg daily with food | 15 mg daily if CrCl 30–49 (15–29: very limited data) |
CrCl <30, extremes of weight → DOAC dosing in renal impairment and obesity. Drug detail → direct oral anticoagulants, warfarin.
special populations
| Setting | Approach |
|---|---|
| CKD 4 (CrCl 15–30) | OAC per CHADS-65 (weak); apixaban or warfarin |
| dialysis / CrCl <15 | no proven benefit, high bleeding; CCS suggests not routinely; ACC/AHA: warfarin or apixaban may be reasonable → individualise |
| cirrhosis | Child-Pugh C or coagulopathy → no routine OAC; Child-Pugh B → avoid rivaroxaban |
| frail elderly, falls | OAC (strong) — greatest absolute benefit; falls rarely outweigh stroke risk |
| active cancer | individualise; DOAC over warfarin (interactions, INR instability) |
| secondary AF that resolved | CCS: not routinely; ESC: long-term OAC if elevated risk. Recurrence is common → follow indefinitely |
| post-op AF | reassess need for OAC, rate and rhythm therapy at 6–12 weeks; after cardiac surgery consider holding OAC for the first 72 h (ACC/AHA: OAC for 60 days once surgical bleeding risk allows) |
| AF in acute illness or non-cardiac surgery | high recurrence risk → counsel; benefit of starting OAC in hospital uncertain (septic critical illness: ACC/AHA 2b) → CHADS-65 and OAC decision at outpatient follow-up (ACC/AHA 2a), with AF surveillance |
| pregnancy | LMWH, or VKA outside the first trimester and after week 36 → heart disease in pregnancy |
device-detected AF (atrial high-rate episodes)
| Episode duration | Approach |
|---|---|
| <5 min | no OAC |
| 5 min–24 h | benefit unclear (CCS); DOAC may be considered if high stroke risk and not high bleeding risk (ESC IIb; ACC/AHA 2b if CHA₂DS₂-VASc ≥3) |
| ≥24 h | treat as clinical AF → CHADS-65 (CCS weak; ACC/AHA 2a) |
- Confirm with an ECG strip where possible (device artefact)
- ARTESIA (2024): apixaban vs ASA ↓ stroke/systemic embolism, ↑ major bleeding. NOAH-AFNET 6 (2023): edoxaban did not ↓ the composite, ↑ bleeding. Absolute stroke rate on these short episodes ~1%/year — far below clinical AF
stroke, ESUS and bleeding
- ESUS: ≥24 h ambulatory ECG, then prolonged monitoring (loop recorder) in selected older patients; no empiric OAC without documented AF → cryptogenic and young stroke
- OAC after ischaemic stroke: CCS timing 1 day (TIA), 3 (mild, NIHSS <8), 6 (moderate, 8–15), 12–14 (severe). Earlier DOAC (≤48 h minor/moderate, day 6–7 major) was safe in ELAN (2023) → acute ischaemic stroke
- Stroke despite OAC: check adherence, dose and non-cardioembolic causes; adding antiplatelet or switching DOAC does not help
- Bleeding on OAC: find the source (GI/GU bleeding often unmasks a lesion), then restart in high-risk patients once corrected → anticoagulant reversal
- Procedures: no bridging when interrupting a DOAC → periprocedural anticoagulation
left atrial appendage occlusion
- Percutaneous LAAO: moderate–high stroke risk + absolute contraindication to long-term OAC (CCS weak; ACC/AHA 2a; ESC IIb). Needs short-term antithrombotics after implant
- Surgical LAA closure during cardiac surgery for another reason → adjunct to OAC, not a replacement: ↓ stroke by a third in LAAOS III (2021). Continue OAC after
AF with CAD, PCI or ACS
OAC + clopidogrel, ASA ≤30 days, OAC alone after 12 months → antithrombotic therapy in AF with CAD.
Prepared 2026-10-19. Not yet verified — check every number against the sources.