Venous Thromboembolism — Anticoagulation
Definition
Anticoagulation for venous thromboembolism (VTE — encompassing DVT and PE) requires decisions across three dimensions: (1) agent choice (which anticoagulant), (2) dose (full vs. low-intensity), and (3) duration (fixed 3–6 months vs. extended/indefinite). Direct oral anticoagulants (NOACs) are first-line over vitamin K antagonists. The two most commonly used NOACs are apixaban and rivaroxaban; until COBRRA (2026), no head-to-head RCT had directly compared them. The 3-month provoked/unprovoked binary for duration decisions is challenged by HI-PRO (2025), which shows enduring risk factors convert "provoked" VTE into a high-recurrence state requiring extended therapy.
Key Concepts
Acute-Phase NOAC Dosing (First 3 Months)
Apixaban Regimen
- 10 mg BID × 7 days → 5 mg BID thereafter
- Higher initial dose over only 7 days (shorter loading vs rivaroxaban)
Rivaroxaban Regimen
- 15 mg BID × 21 days → 20 mg once daily thereafter
- Higher loading dose for 21 days; 50% above maintenance dose
- Loading phase associated with higher early bleeding risk (COBRRA)
NOAC Choice — Apixaban vs Rivaroxaban: COBRRA Trial
(sources/NOAC-VTE-COBRRA-NEJM-2026, rating: very high)
The COBRRA trial (n=2,760; PROBE design; 32 centres; Canada/Australia/Ireland; 3-month follow-up) is the first head-to-head RCT comparing apixaban vs rivaroxaban in acute VTE:
| Outcome | Apixaban | Rivaroxaban | RR (95% CI) | P |
|---|---|---|---|---|
| Clinically relevant bleeding | 3.3% | 7.1% | 0.46 (0.33–0.65) | <0.001 |
| Major bleeding | 0.4% | 2.4% | 0.16 (0.06–0.40) | — |
| CRNMB | 2.9% | 4.9% | 0.59 (0.40–0.86) | — |
| Recurrent VTE | 1.1% | 1.0% | 1.08 (0.52–2.23) | NS |
| All-cause death | 0.1% | 0.3% | 0.25 (0.03–2.26) | NS |
- Bleeding difference consistent across all prespecified subgroups
- Most bleeding difference occurs in first 21 days — correlates with rivaroxaban's higher loading dose period
- Medication adherence paradoxically lower with apixaban (65.7% vs 75.1%), yet VTE recurrence identical — the twice-daily dosing of apixaban may reduce adherence without affecting efficacy
- Excluded: active cancer, CrCl <30, weight >120 kg, AF
Duration of Anticoagulation — The Provoked/Unprovoked Binary
Standard Guidance
- Provoked VTE (surgery, trauma, immobility): 3 months anticoagulation; low recurrence risk (~2–3%/year) after stopping → no extended therapy in guidelines
- Unprovoked VTE: High recurrence risk (~6–10%/year) → extended/indefinite anticoagulation recommended
The ESC 2019 Recurrence-Risk Categorisation — a Third Framework
ESC 2019 abandoned the provoked/unprovoked binary explicitly: it avoids the terms "unprovoked" and "idiopathic" and instead categorises the index event's risk factor by estimated risk of long-term recurrence (its Table 11). This is a graded taxonomy rather than a binary, and it predates HI-PRO's challenge by six years. (sources/acute-pe-esc-2019, rating: very high)
| Estimated long-term recurrence | Risk factor category for index PE | Examples |
|---|---|---|
| Low (<3%/yr) | Major transient/reversible factors, >10-fold increased risk | Surgery with general anaesthesia >30 min; hospitalisation |
| Intermediate (3–8%/yr) | Transient/reversible factors with ≤10-fold risk; non-malignant persistent risk factors; no risk factor | Minor surgery (GA <30 min); admission <3 days with acute illness; inflammatory bowel disease; active autoimmune disease |
| High (>8%/yr) | — | Active cancer; ≥1 previous VTE episode without a major transient/reversible factor; antiphospholipid antibody syndrome |
- "No risk factor" sits in the intermediate tier (3–8%/yr), not the high tier. This is a notable departure from the common assumption that unprovoked VTE carries 6–10%/yr — on the ESC taxonomy the high tier is reserved for active cancer, previous VTE, and APS. (sources/acute-pe-esc-2019, rating: very high)
- Duration recommendations built on that taxonomy: therapeutic anticoagulation ≥3 months for all patients (Class I/A); discontinue after 3 months for a first PE from a major transient/reversible factor (I/B); indefinite duration for recurrent VTE unrelated to a major transient/reversible factor (I); indefinite VKA for APS (I/B); extended indefinite anticoagulation should be considered for a first PE with no identifiable risk factor (IIa/A), a persistent risk factor other than APS (IIa/C), or a minor transient/reversible risk factor (IIa/C). Reduced-dose apixaban 2.5 mg b.i.d. or rivaroxaban 10 mg o.d. should be considered after 6 months of therapeutic anticoagulation (IIa) — the same agents and doses the later trials validated. (sources/acute-pe-esc-2019, rating: very high)
- Thrombophilia testing (antiphospholipid antibodies and lupus anticoagulant) may be considered in patients with VTE at a young age (e.g. <50 years) and no otherwise identifiable risk factor, especially when it occurs against a background of thrombosis. (sources/acute-pe-esc-2019, rating: very high)
- Bleeding-risk anchors: ~3%/year major bleeding on VKAs in cohorts more than 15 years old; in extended-treatment phase III trials, major bleeding may reach ~1% while clinically relevant non-major bleeding reaches as high as 6%. (sources/acute-pe-esc-2019, rating: very high)
- Cancer-specific duration (ESC 2019): LMWH should be considered for the first 6 months over VKAs (IIa/A); edoxaban (IIa/B) or rivaroxaban (IIa/C) are alternatives only outside gastrointestinal cancer; extended anticoagulation beyond 6 months should be considered indefinitely or until the cancer is cured (IIa/B). (sources/acute-pe-esc-2019, rating: very high)
The HI-PRO Challenge to the Binary
(sources/Apixaban-VTE-HIPRO-NEJM-2025, rating: high)
HI-PRO (n=600; double-blind; Brigham and Women's Hospital; 12-month extension after ≥3 months standard treatment) tested apixaban 2.5 mg BID vs placebo in patients with provoked VTE + ≥1 enduring risk factor:
| Outcome | Apixaban 2.5 mg BID | Placebo | HR (95% CI) | P |
|---|---|---|---|---|
| Recurrent VTE | 1.3% | 10.0% | 0.13 (0.04–0.36) | <0.001 |
| Major bleeding | 0.3% | 0% | — | NS |
| CRNMB | 4.8% | 1.7% | 2.68 (0.96–7.43) | 0.06 |
| All-cause death | 0.3% | 1.0% | — | NS |
Key finding: The 10% placebo recurrence rate in "provoked" VTE patients with enduring risk factors equals or exceeds the ~6–10%/year threshold used to justify indefinite anticoagulation in unprovoked VTE. Enduring risk factors tested: obesity (BMI ≥30), chronic inflammatory/autoimmune disorder, chronic lung disease, ASCVD.
- 87% relative risk reduction in VTE recurrence
- Very low major bleeding risk (1 event: minor subdural haematoma, no clinical consequence)
- CRNMB trend (HR 2.68; P=0.06): clinically meaningful but trial underpowered for this endpoint; absolute rate 4.8% — nuanced risk-benefit, especially since all recurrent VTEs were nonfatal
- Consistent benefit across major vs minor transient provoking factors (post hoc)
Extended-Phase Dosing Options (After ≥3 Months)
- Apixaban 2.5 mg BID — validated in AMPLIFY-EXT (unprovoked VTE) and HI-PRO (provoked + enduring risk)
- Rivaroxaban 10 mg OD — validated in EINSTEIN-CHOICE (unprovoked VTE; Einstein-Choice also showed rivaroxaban 10 mg OD superior to aspirin for recurrence)
- Apixaban 5 mg BID (full dose) vs 2.5 mg BID (reduced dose): API-CAT (NEJM 2025) showed 2.5 mg BID noninferior for VTE recurrence with superior bleeding in cancer-associated VTE (see concepts/Cancer-Associated-VTE)
Risk Stratification for Extended Duration
- VTE-PREDICT score (Eur Heart J 2023): integrates clinical variables to estimate VTE recurrence and bleeding risk on extended anticoagulation; assists shared decision-making
- Polygenic risk scores and AI — emerging tools, not yet validated for routine clinical use (referenced in HI-PRO)
- Enduring risk factors as a practical clinical trigger for extended therapy in provoked VTE: obesity, autoimmune disease, chronic lung disease, ASCVD
Anticoagulation Exclusions Common to Both Trials
- Active cancer (separate management, see concepts/Cancer-Associated-VTE)
- Atrial fibrillation (separate indication, different dosing)
- CrCl <30 ml/min (renal impairment requiring dose adjustment)
Contradictions / Open Questions
-
COBRRA: mechanism of bleeding difference unexplained — whether the rivaroxaban loading-dose hypothesis is correct requires a trial comparing rivaroxaban 10 mg OD (extended dose) vs apixaban in the acute phase; pharmacodynamic differences between the two factor Xa inhibitors cannot be excluded. (sources/NOAC-VTE-COBRRA-NEJM-2026)
-
COBRRA: adherence paradox — apixaban adherence 65.7% vs rivaroxaban 75.1%, yet VTE recurrence identical (1.1% vs 1.0%); raises questions about whether subtherapeutic adherence in the apixaban arm underestimates its true efficacy advantage. (sources/NOAC-VTE-COBRRA-NEJM-2026)
-
HI-PRO: CRNMB signal (HR 2.68; P=0.06) — trial not powered for this endpoint; whether the ~3% absolute excess in CRNMB with apixaban outweighs the 8.7% absolute reduction in recurrent VTE depends on patient values and clinical context. (sources/Apixaban-VTE-HIPRO-NEJM-2025)
-
HI-PRO: single-center limitation — generalizability to diverse health systems uncertain. (sources/Apixaban-VTE-HIPRO-NEJM-2025)
-
Duration beyond 12 months in provoked + enduring risk: HI-PRO only covers 12 months of extended treatment; no data exist on whether anticoagulation should continue indefinitely or can be safely stopped
-
Which enduring risk factors matter most: HI-PRO enrolled a heterogeneous enduring-risk population (obesity, autoimmune, lung disease, ASCVD); subgroup analyses were underpowered; it is unknown whether all confer equivalent recurrence risk
-
Aspirin co-administration: ~20% of HI-PRO patients were on low-dose aspirin; discontinuation may reduce CRNMB without affecting VTE protection
-
The recurrence rate assigned to "no risk factor" differs materially by source. This page's standard-guidance framing puts unprovoked VTE at ~6–10%/year; the ESC 2019 taxonomy places "no risk factor" in the intermediate tier at 3–8%/year and reserves the high tier (>8%/year) for active cancer, previous VTE, and APS. The bands overlap but the central estimates differ, and the thresholds were derived from different cohort eras. Which figure is quoted changes how strongly extended anticoagulation is justified to a patient, so the provenance should travel with the number. (sources/acute-pe-esc-2019, rating: very high)
-
The vocabulary itself is contested. ESC 2019 deliberately avoids "unprovoked" and "idiopathic", describing the index event's risk factor instead; this page, HI-PRO, and AHA/ACC 2026 all use the provoked/unprovoked binary as their organising frame. The ESC position is that the binary is the source of the imprecision HI-PRO later exposed — but because the trial literature is written in binary terms, mapping between vocabularies is unavoidable when applying results. (sources/acute-pe-esc-2019, rating: very high)
-
ESC 2019 predates the extended-phase dose-reduction trials it anticipated. Its IIa recommendation for reduced-dose apixaban or rivaroxaban after 6 months was made on the basis of the same phase III extension programmes later analysed by AMPLIFY-EXT, EINSTEIN-CHOICE and API-CAT — so it is an early guideline endorsement of a practice subsequently confirmed, not an independent validation. (sources/acute-pe-esc-2019, rating: very high)
Connections
- Related to concepts/Cancer-Associated-VTE — cancer VTE excluded from both trials; has separate DOAC evidence (API-CAT, Caravaggio)
- Related to concepts/Acute-PE-Clinical-Categories — acute management precedes anticoagulation choice decisions
- Related to concepts/Pulmonary-Hypertension — chronic thromboembolic PH as long-term sequela of VTE (CTEPH)