2018 EHRA Practical Guide on NOACs in Atrial Fibrillation
Authors, Journal, Affiliations, Type, DOI
- Authors: Jan Steffel, Peter Verhamme, Tatjana S. Potpara, Pierre Albaladejo, Matthias Antz, Lien Desteghe, Karl Georg Haeusler, Jonas Oldgren, Holger Reinecke, Vanessa Roldan-Schilling, Nigel Rowell, Peter Sinnaeve, Ronan Collins, A. John Camm, Hein Heidbüchel (EHRA Writing Group)
- Journal: European Heart Journal 2018;39:1330–1393
- Affiliations: University Heart Center Zurich (lead), University of Leuven, Belgrade University, multiple European centres
- Type: Multi-society consensus / practical guide (second update of original 2013 EHRA guide)
- DOI: 10.1093/eurheartj/ehy136
Overview
The 2018 EHRA Practical Guide is the authoritative multi-society consensus on NOAC use in atrial fibrillation, organized around 20 concrete clinical scenarios. It provides unified practical guidance spanning from eligibility and start-up through perioperative management, cardioversion, acute stroke, and cancer-associated anticoagulation. Key contributions include the EHRA VHD Type 1/2 classification (defining true NOAC contraindications), drug-specific pharmacokinetic tables (renal clearance fractions), perioperative management without bridging, and the early evidence framework for dual over triple antithrombotic therapy post-PCI. Note: 2018 vintage means several areas have been updated by subsequent trials (AUGUSTUS, ENTRUST-AF PCI, EPIC-CAD, COBRRA).
Keywords
NOACs, DOACs, atrial fibrillation, stroke prevention, anticoagulation, pharmacokinetics, renal clearance, CKD, drug-drug interactions, perioperative management, cardioversion, bleeding management, reversal agents, idarucizumab, andexanet alfa, triple therapy, dual therapy, PCI
Key Takeaways
Topic 1: Eligibility for NOACs
- NOACs approved for non-valvular AF = AF in absence of mechanical prosthetic valve or moderate-to-severe mitral stenosis (usually rheumatic)
- EHRA VHD classification:
- Type 1 (VKA-mandatory): mechanical heart valve, moderate-to-severe mitral stenosis
- Type 2 (NOAC-eligible): all other native valvular stenoses and regurgitations, bioprosthetic valves (>3 months post-op), mitral valve repair, TAVI
- Severe aortic stenosis: limited data (excluded from RE-LY); most will undergo intervention before anticoagulation becomes primary question
- Bioprosthetic valve in rheumatic mitral stenosis: VKA may be preferred over NOAC despite >3 months post-op (large, diseased atria)
- TAVI + AF: no prospective NOAC data at time of writing; single or dual antiplatelet often required; combination with NOAC increases bleeding
- HCM + AF: NOACs appear eligible; AF in HCM resembles HFpEF-related AF (no mechanistic rationale for NOAC inferiority vs VKA); limited specific trial data but patients included in landmark trials
Topic 2: Practical Start-Up and Follow-Up
- NOAC choice based on: regulatory approval, CrCl, comorbidities, drug-drug interactions, patient preference
- ESC Class I/A: NOACs preferred over VKA in newly initiated stroke prevention
- SAMe-TT2R2 score (sex female +1, age <60 +1, medical history ≥2 conditions +1, treatment interactions +2, tobacco +2, race non-Caucasian +2): score ≥3 predicts poor INR control on VKA; useful in countries where VKA is required before NOAC access
- Follow-up schedule:
- Initial review at 1 month
- Then ≥every 3 months thereafter
- Blood sampling (renal/liver function, Hb): yearly; 6-monthly if age ≥75 or frail; every CrCl/10 months if CrCl <60
- HAS-BLED score validated in NOAC patients; high bleeding risk does NOT automatically justify withholding OAC (stroke and bleeding risks track together)
- Proton pump inhibitors: consider for high GI bleeding risk (history of GI bleed/ulcer, concomitant antiplatelet)
Topic 3: Ensuring Adherence
- NOAC anticoagulant effect wanes within 12–24h of last dose; plasma levels cannot monitor adherence (reflect only last 24-48h)
- Real-world adherence 38–99%; NOACs have lower discontinuation rates vs VKA
- OD regimens generally achieve better adherence vs BID (but clinical thromboembolic benefit of dosing regimen differences unclear)
- Strategies: patient education, anticoagulation card, nurse-led AF clinic, pharmacy monitoring, medication apps (Medisafe, Dosecast), electronic blister packs
- If adherence poor despite all measures, VKA conversion may be considered (not necessarily superior)
Topic 4: Switching Between Anticoagulant Regimens
| Transition | Timing |
|---|---|
| VKA → NOAC | Start NOAC when INR <2.0; day of or day after if INR 2.0–2.5; wait if >2.5 |
| NOAC → VKA | Co-administer until INR in range (edoxaban: half-dose bridge until INR stable × 3 readings) |
| NOAC → IV UFH | Start UFH when next NOAC dose is due |
| NOAC → LMWH | Start LMWH when next NOAC dose is due |
| IV UFH → NOAC | Start NOAC 2–4h after stopping UFH |
| LMWH → NOAC | Start NOAC at next scheduled LMWH dose time |
| NOAC → NOAC | Start alternative at next dose due time (longer interval if impaired clearance expected) |
Topic 5: Pharmacokinetics and Drug–Drug Interactions
- P-gp (P-glycoprotein): transporter in gut, liver, kidney; P-gp inhibitors → ↑ all NOAC plasma levels; key inhibitors in AF: verapamil, dronedarone, amiodarone, quinidine, ticagrelor (loading dose only)
- CYP3A4: hepatic clearance of rivaroxaban (18%) and apixaban (25%); negligible for dabigatran and edoxaban; strong CYP3A4 inhibitors (ketoconazole, ritonavir) or inducers (rifampicin, carbamazepine, phenytoin) → rivaroxaban/apixaban significantly affected
- St. John's wort: strong CYP3A4+P-gp inducer; avoid with all NOACs
- Specific interactions:
- Dronedarone + dabigatran: contraindicated (strong P-gp effect)
- Dronedarone + rivaroxaban: avoid (moderate P-gp + CYP3A4)
- Verapamil + dabigatran: use 110mg BID; if immediate release, separate ≥2h; 60% AUC rise with slow-release
- Amiodarone + edoxaban: 40% AUC rise; no dose reduction recommended
- Diltiazem + apixaban: 40% AUC rise; caution (yellow)
- Ticagrelor 180mg loading + dabigatran: dabigatran Cmax +65%; stagger intake (ticagrelor 2h after dabigatran)
- Food interactions: rivaroxaban 15/20mg MUST be taken with food (AUC +39% → bioavailability ~100%); others: no food interaction; dabigatran capsules must NOT be opened (+75% bioavailability)
- Crushed tablets: apixaban, rivaroxaban, edoxaban can be crushed/given via NG tube; NOT dabigatran (capsule must remain intact)
- Pharmacokinetic table (renal clearance of absorbed dose): dabigatran 80%, edoxaban 50%, rivaroxaban 35%, apixaban 27%; plasma protein binding: apixaban 87%, rivaroxaban 95%, edoxaban 55%, dabigatran 35%
Topic 6: NOACs in Chronic Kidney Disease and Liver Disease
CKD:
- All NOACs partly renally eliminated; dabigatran most renal-dependent (80%)
- Use Cockcroft-Gault (not CKD-EPI) for NOAC dosing decisions (matches trial methodology)
- CKD staging/dosing table:
- CrCl ≥30 mL/min (mild-moderate): all NOACs appropriate with standard or reduced dose per criteria
- CrCl 15–29 mL/min (severe): apixaban, rivaroxaban, edoxaban approved with dose reduction; dabigatran NOT approved
- CrCl <15 mL/min: no NOAC officially indicated
- CrCl >95 mL/min (high normal): edoxaban may have reduced efficacy vs warfarin (FDA 2015 warning; EMA caution); consider alternative NOAC
- Warfarin may worsen renal function (vascular calcification, warfarin nephropathy); NOAC may delay CKD progression (RE-LY subanalysis)
- Apixaban bleeding benefit vs warfarin increases as CrCl falls; dabigatran 110mg bleeding benefit lost at CrCl <50
- Dialysis: dabigatran 50-60% dialysable; FXa inhibitors <14% dialysable (high protein binding)
Liver disease (Child-Pugh dosing):
| Child-Pugh | Dabigatran | Apixaban | Edoxaban | Rivaroxaban |
|---|---|---|---|---|
| A (5–6 pts) | Standard | Standard | Standard | Standard |
| B (7–9 pts) | Caution | Caution | Caution | Avoid |
| C (10–15 pts) | Avoid | Avoid | Avoid | Avoid |
- All NOACs contraindicated in Child-Pugh C
- Rivaroxaban additionally contraindicated in Child B (>2-fold drug exposure)
Topic 7–8: Measuring NOAC Anticoagulant Effect and Plasma Levels
Available assays:
- Dabigatran: dTT (diluted thrombin time) or ECA (ecarin chromogenic assay) — linear, quantitative; TT (thrombin time) — qualitative only (normal TT excludes even low dabigatran levels); aPTT — qualitative, curvilinear
- FXa inhibitors (apixaban, rivaroxaban, edoxaban): anti-FXa chromogenic assay with calibrators — quantitative; PT — limited (reagent-dependent; cannot use INR)
- Standard PT/aPTT unreliable for any NOAC quantitative assessment
Expected plasma levels on AF doses:
| NOAC | Peak (ng/mL) | Trough (ng/mL) |
|---|---|---|
| Dabigatran | 64–443 | 31–225 |
| Apixaban | 69–321 | 34–230 |
| Edoxaban | 91–321 | 31–230 |
| Rivaroxaban | 184–343 | 12–137 |
Indications for level measurement (rare):
- Emergency: bleeding, urgent surgery, acute stroke (thrombolysis decision)
- Pre-elective surgery when CrCl/timing uncertain
- Very obese or underweight patients (BMI extremes)
- Complex drug-drug interactions
- Long-term monitoring discouraged (no outcome data to guide dose adjustment)
Topic 9–10: Dosing Errors and Overdose
Missed dose:
- BID (12h interval): can take up to 6h late; skip if later (high stroke risk: may extend to next dose)
- OD (24h interval): can take up to 12h late; skip if later (high stroke risk: may extend further)
Double dose:
- BID: skip next scheduled dose; restart normal 24h after double dose
- OD: continue normal schedule (no skip needed)
Uncertainty about dose:
- BID: do NOT take extra; continue regular schedule
- OD + CHA₂DS₂-VASc ≥3: take another dose and continue; ≤2: wait for next scheduled dose
Overdose without bleeding:
- Activated charcoal within 2–3h of ingestion
- Maintain diuresis
- Dialysis: effective for dabigatran (50-60% removal); NOT for FXa inhibitors (high protein binding)
- Specific reversal agents (see Topic 11) if high plasma levels confirmed
Topic 11: Management of Bleeding Under NOAC
Three-tier severity:
Nuisance/minor:
- Delay intake or withhold ≤1 dose
- Treat underlying cause (PPI for GI, antibiotics for UTI)
- Epistaxis/gum: local anti-fibrinolytics (tranexamic acid mouthwash)
Major non-life-threatening:
- Causal therapy (endoscopy, compression, surgery)
- Maintain diuresis; dialysis option for dabigatran only
- Tranexamic acid 1g IV q6h
- Desmopressin 0.3 mg/kg IV (max 20 mg) if associated coagulopathy/thrombopathy
Life-threatening:
- Specific reversal:
- Dabigatran: idarucizumab 5g IV (2×2.5g) — humanized antibody fragment (RE-VERSE AD trial)
- FXa inhibitors: andexanet alfa — recombinant FXa analogue (approved post-2018); ciraparantag (in development)
- Non-specific (if specific reversal unavailable): (activated) prothrombin complex concentrates (aPCC or PCC)
- NOT useful: fresh frozen plasma (large volume required; NOACs inhibit added factors); vitamin K (no role); protamine (only for concomitant heparin)
- NOAC-related bleeding: opportunity to review correct dose, choice, and modifiable bleeding risk factors (hypertension, NSAIDs, alcohol, antiplatelet agents)
Topic 12: Planned Invasive Procedures, Surgery, and Ablation
No bridging with LMWH/UFH: predictable NOAC waning avoids need for bridging; bridging increases bleeding without reducing events (BRIDGE trial evidence for VKA; same principle applied to NOACs)
Perioperative interruption timing:
| Dabigatran (Low/High risk) | FXa Inhibitors (Low/High risk) | |
|---|---|---|
| CrCl ≥80 | ≥24h / ≥48h | ≥24h / ≥48h |
| CrCl 50–79 | ≥36h / ≥72h | ≥24h / ≥48h |
| CrCl 30–49 | ≥48h / ≥96h | ≥24h / ≥48h |
| CrCl 15–29 | Not indicated | ≥36h / ≥48h |
Resumption: ≥24h post low-risk; 48–72h post high-risk interventions
Dental surgery: usually no NOAC interruption needed; local hemostasis (oxidized cellulose, tranexamic acid mouthwash); avoid NSAIDs
Device implantation (BRUISE-CONTROL 2 data): last NOAC dose morning before procedure; restart the following day; no bridging
Neuraxial anaesthesia/lumbar puncture (high-risk): interrupt ≥3 days (FXa inhibitors, 5 half-lives) to ≥4–5 days (dabigatran); resume 24h after
AF catheter ablation:
- Uninterrupted strategy (truly) or last dose 12h before
- IV heparin ACT 300–350s during procedure
- Resume NOAC 3–5h after sheath removal (if hemostasis adequate, no pericardial effusion)
- If >36h without NOAC, rule out LAA thrombus with TOE before ablation
Topic 13: Urgent Surgical Intervention
Three urgency tiers:
- Immediate (life/limb-threatening, minutes): operate without delay; specific reversal if available; prefer general over spinal anaesthesia if no reversal agent; (a)PCC if no specific reversal
- Urgent (hours): defer ≥12h (ideally 24h) if possible; check coagulation; (a)PCC as needed
- Expedite (days): follow elective rules (Topic 12)
Topic 14: AF + Coronary Artery Disease
Antithrombotic framework:
- DAPT alone: insufficient for stroke prevention
- OAC alone: insufficient for stent thrombosis prevention in acute/subacute period
- Triple therapy (OAC + aspirin + P2Y12): needed short-term; avoid long-term due to excessive bleeding
RCT evidence for NOAC post-PCI (2018 data):
- PIONEER AF-PCI (rivaroxaban): Two rivaroxaban arms vs VKA triple therapy in AF + PCI patients — both rivaroxaban arms reduced clinically significant bleeding; rivaroxaban 15mg OD + P2Y12 = sustained (non-standard dose); both underpowered for ischemic efficacy
- RE-DUAL PCI (dabigatran): Dabigatran 110mg or 150mg BID + P2Y12 (no aspirin) vs VKA triple therapy — both dabigatran arms reduced major bleeding; non-inferior for composite ischemic endpoint; 150mg non-inferior for stroke vs warfarin
Practical recommendations:
- Triple therapy ≤1 week, then dual (OAC + P2Y12 inhibitor)
- Clopidogrel preferred P2Y12 in dual therapy; ticagrelor appears safe in subsets
- Prasugrel + ticagrelor as part of triple therapy: Class III/C (discouraged)
- No switch from NOAC to VKA post-PCI (higher risk)
- Elective PCI: NOAC stopped ≥12–24h before; UFH 70IU/kg or bivalirudin periprocedural (not enoxaparin)
- STEMI: primary PCI preferred; add UFH/enoxaparin/bivalirudin regardless of NOAC timing; avoid routine GP IIb/IIIa; no fibrinolysis if NOAC levels above reference range
- Stable CAD + AF in chronic phase: emerging evidence for OAC monotherapy (AFIRE pending at 2018; EPIC-CAD 2024 subsequently confirms)
Topic 15: NOAC Dosing Across Indications
Stroke prevention in AF (standard/reduced dose criteria):
| NOAC | Standard | Dose Reduction Criteria |
|---|---|---|
| Apixaban | 5mg BID | 2.5mg BID if ≥2 of: age ≥80, weight ≤60kg, creatinine ≥133 μmol/L (or CrCl 15–29) |
| Dabigatran | 150mg BID (preferred) / 110mg BID | No pre-specified criteria; 110mg for age >75, high bleed risk, CrCl 30–49 |
| Edoxaban | 60mg OD | 30mg OD if: weight ≤60kg, CrCl ≤50, strong P-gp inhibitor |
| Rivaroxaban | 20mg OD | 15mg OD if CrCl ≤50 |
Key principle: Never use lower doses outside tested dose-reduction algorithms; underdosing associated with higher stroke rates without safety benefit
Topic 16: Cardioversion in NOAC-Treated Patients
AF ≥48h (or unknown duration):
- Strategy A (delayed): Anticoagulate ≥3 weeks before cardioversion, then cardiovert; continue OAC ≥4 weeks post-CV
- Strategy B (early, TOE-guided): Perform TOE to exclude LAA thrombus; if negative, cardiovert with single NOAC dose ≥4h before (or ≥2h after apixaban 10mg loading dose); continue OAC ≥4 weeks post-CV
- Both strategies comparable in X-VeRT (rivaroxaban), ENSURE-AF (edoxaban), EMANATE (apixaban) — no significant difference vs VKA
AF <48h:
- Lower thromboembolic risk but not zero (especially CHA₂DS₂-VASc ≥2, AF >12h)
- Current institutional practice with LMWH ± TOE acceptable
- Single NOAC dose 2–4h before CV may be used as alternative to LMWH; in high-risk patients (CHA₂DS₂-VASc ≥4), TOE-guided strategy recommended
LAA thrombus on TOE: do NOT cardiovert; treat with NOAC (rivaroxaban 20mg achieves thrombus resolution in 41.5%; X-TRA study)
Post-CV long-term anticoagulation:
- CHA₂DS₂-VASc ≥2 (men) or ≥3 (women): long-term OAC regardless of CV success
- Low CHA₂DS₂-VASc + AF ≥48h: OAC ≥4 weeks post-CV; decision after 4 weeks by score
- Even "triggered" AF (PE, sepsis, surgery): OAC decision based on underlying structural/vascular risk, not trigger
Topic 17: Acute Stroke While on NOAC
- Ischemic stroke + thrombolysis consideration: thrombolysis only if NOAC plasma level below reference range (normal dTT/ECA for dabigatran; normal PT for rivaroxaban/edoxaban); mechanical thrombectomy not contraindicated by NOAC
- Restart after ischemic stroke: individualize; generally 3–14 days based on infarct size and hemorrhagic transformation risk
- ICH while on NOAC: discontinue NOAC immediately; specific reversal if available; neurosurgical evaluation; restart OAC highly individualized (LAA occlusion as alternative if OAC cannot be restarted)
Topic 18–19: Special Situations and Malignancy
- Elderly/frail: reassess dosing; falls alone ≠ reason to withhold OAC; education is key; renal function monitoring more frequent
- Extremes of weight (<50kg or >120kg): underrepresented in trials; plasma trough level monitoring may be warranted
- Cancer + AF: LMWH traditional preference; NOACs emerging as option but limited data; GI tumors at highest GI bleeding risk; carefully weigh thrombocytopenia
- Transplant: no data; select NOAC dose per renal function; multiple drug interactions with immunosuppressants
- Antiepileptics: phenytoin and carbamazepine are strong CYP3A4 inducers — markedly reduce rivaroxaban and apixaban; avoid or use with extreme caution (Table 5)
Topic 20: VKA Optimization
- TTR >70% needed for clinical benefit; patients below target may benefit from switching to NOAC
- SAMe-TT2R2 score ≥3 predicts poor TTR; useful to identify patients likely to need NOAC
- Loading doses: NOT recommended for warfarin/acenocoumarol; recommended for phenprocoumon
- VKA half-lives: acenocoumarol 8–24h; warfarin 36–48h; phenprocoumon 120–200h (6 days)
Limitations of the Document
- 2018 vintage: several recommendations have been superseded or strengthened by subsequent trials — AUGUSTUS (apixaban post-PCI), ENTRUST-AF PCI (edoxaban post-PCI), EPIC-CAD (OAC monotherapy in stable CAD), COBRRA (apixaban vs rivaroxaban head-to-head in VTE), API-CAT (reduced-dose apixaban in cancer VTE)
- Reversal agents: andexanet alfa described as "in development" — now approved for FXa inhibitors; ciraparantag still investigational
- VHD/TAVI: no prospective NOAC data for TAVI; subsequent GALILEO (rivaroxaban post-TAVI) and ATLANTIS trials now provide data
- Dose-monitoring gap: no hard outcome data to support plasma level-guided dose adjustment
- Cancer anticoagulation: pre-dates ADAM VTE, SELECT-D, CARAVAGGIO, Hokusai-VTE Cancer showing DOAC non-inferiority or superiority to LMWH (with higher GI bleeding caveats)
- Industry advisory support: writing group assisted by manufacturer medical experts; final responsibility with EHRA
Key Concepts Mentioned
- concepts/NOAC-AF-Management — core NOAC use framework in AF
- concepts/NOAC-Perioperative — perioperative interruption and bridging
- concepts/AF-Stable-CAD-Antithrombotic — dual vs triple antithrombotic therapy post-PCI
- concepts/AF-CARE — broader AF management framework (EHRA/ESC)
- concepts/CHA2DS2-VA — stroke risk stratification
- concepts/Subclinical-AF — NOAC timing in low-burden AF
- concepts/Venous-Thromboembolism-Anticoagulation — cross-indication NOAC dosing
Key Entities Mentioned
- entities/Dabigatran — direct thrombin inhibitor; 80% renal; idarucizumab reversal
- entities/Apixaban — FXa inhibitor; 27% renal; lowest renal dependence
- entities/Rivaroxaban — FXa inhibitor; 35% renal; CYP3A4+P-gp substrate; food interaction
- entities/Edoxaban — FXa inhibitor; 50% renal; reduced efficacy at high CrCl
- entities/Idarucizumab — dabigatran-specific reversal agent (antibody fragment)
- entities/Andexanet-Alfa — FXa inhibitor reversal agent (recombinant FXa analogue)
Wiki Pages Updated
- wiki/sources/noac-ehra-2018.md — created (this file)
- wiki/concepts/NOAC-AF-Management.md — created
- wiki/concepts/NOAC-Perioperative.md — created
- wiki/concepts/AF-CARE.md — added NOAC connection
- wiki/concepts/AF-Stable-CAD-Antithrombotic.md — added PIONEER/RE-DUAL PCI historical context
- wiki/sourceindex.md — new entry prepended
- wiki/wikiindex.md — new concept entries prepended