An evidence-based approach for North Wales, built on NICE guidance and the European SERENITY programme.
The patient and family are not part of the decision.
Nobody holds the review once the patient leaves hospital.
Continuation is never revisited or measured.
Local context: service lead, Haematology (North Wales).
of people on antithrombotics when poor-prognosis cancer was diagnosed continued until death. n = 25,783; median survival 145 days; 32% on treatment at diagnosis.
Aldridge 2025, Br J Haematol
on antithrombotics at terminal illness declaration (median survival 57 days). Most continued until death: 58.8% of DOAC and 61.6% of VKA users.
Søgaard 2025, J Thromb Haemost · n = 86,732
median time from stopping to death in GP records. Only 22.1% of users discontinued at all; the commonest reason was recognising the terminal phase.
Abbel 2026, Ann Fam Med · n = 2,860
of VKA users with a life-limiting disease continued until the last month of life (median 14 days from stopping to death). Bleeding and thrombosis rates were similar on and off treatment.
Kempers 2026, Thromb Haemost · n = 18,145 · SERENITY
Who am I to say that I'm not going to take it
Title of the SERENITY patient interview study: 60 patients, 13 relatives, 4 countries (Baddeley 2025, Thromb Res). Patients were reluctant to stop alone; many deferred to a clinician whose role they valued highly.
Blurred roles and responsibilities, a "culture of continuation", and passivity about review.
We're in a silo medicine, where every specialist is in his corner.
FR12, geriatrician · Baddeley 2026, Res Pract Thromb Haemost · SERENITY
Deference to the original prescriber, unclear ownership, and fear that patients will see stopping as giving up.
Jagosh 2025, PLoS Med · realist synthesis of 91 papers · SERENITY
Shevlin 2025, Clin Med (Lond)
20% had never deprescribed antithrombotics in end-of-life care.
Martens 2025, Thromb Res · SERENITY WP1
Tardy 2017 · Abbel 2026 (SERENITY; median follow-up ~6 weeks) · White 2019 · Søgaard 2025 · bars drawn to one 0–30% scale
Discontinuing after home palliative care began · Ontario, n = 98,089
Chin-Yee 2022, CMAJ · adjusted HR (thrombosis, bleeding) and adjusted OR (home death) vs continuing
Independent predictors of clinically relevant bleeding · RHESO
Tardy 2017, J Thromb Haemost · Fine–Gray multivariable HR; previous bleeding = within 4 weeks
recurrent VTE per 100 person-years in the first 3 months after stopping CAT anticoagulation. Measured over months, not weeks.
van Hylckama Vlieg 2023, EClinicalMedicine
Palliative patients found LMWH acceptable, and many said it gave reassurance.
Noble 2005, Palliat Med · Noble 2006, BMJ
A recent symptomatic clot, a mechanical valve or a recent stent are different decisions from primary prevention.
Gurumurthy 2026, Res Pract Thromb Haemost
| Phase | Prophylaxis | Treatment of established cancer VTE |
|---|---|---|
| Months | "Consider pharmacological VTE prophylaxis for people who are having palliative care" NG89 1.7.1 · "Review VTE prophylaxis daily" 1.7.4 | "Offer … anticoagulation treatment for 3 to 6 months" NG158 1.3.15 · consider a DOAC 1.3.17 · tumour site, interactions, bleeding risk 1.3.16 |
| Weeks | Daily review with person, family and MDT NG89 1.7.4 | "Assess and discuss the benefits and risks of continuing, stopping or changing" NG158 1.4.1 |
| Last days | "Do not offer VTE prophylaxis to people in the last days of life" NG89 1.7.2 | "stop any previously prescribed medicines that are not providing symptomatic benefit or that may cause harm" NG31 1.5.1 |
NICE NG89 (updated 2019), NG158 (updated 2023), NG31 (2015) · quoted verbatim, checked on nice.org.uk 11 Oct 2026
Redrawn from Gurumurthy et al. Res Pract Thromb Haemost 2026;10:103459, Figure 1, "Framework for decision making on antithrombotic therapy at the end-of-life" · CC BY 4.0 · interpret by indication: expected benefit differs between AF and VTE
| Situation | Practical point |
|---|---|
| Warfarin | Avoid where possible: interactions, monitoring burden, a labile INR as intake falls |
| Poor appetite | Therapeutic rivaroxaban is better absorbed with food; apixaban can be given regardless of meals |
| Feeding tube | Apixaban and rivaroxaban can be crushed into a tube ending in the stomach; dabigatran must not be opened or crushed |
| Falling eGFR | DOAC and LMWH exposure rises; review dose, or whether to continue |
| Nausea, pill burden | Some prefer once-daily LMWH, which bypasses the gut |
| Reversal | Agree in advance whether a bleed would prompt hospital transfer |
Redrawn from Gurumurthy et al. 2026, Figure 2, "Suggested approach when a decision is made to continue anticoagulation at the end-of-life" · CC BY 4.0 · table from the same review, §8
Palliative care, oncology and general practice should be involved in deprescribing decisions, prompted by a clinical trigger or an estimated 3-month prognosis. Eleven outcomes were chosen for the trial.
Mahé 2026, Thromb Res · 24-member steering group
Tests the SERENITY shared decision support tool against usual care. Running in the Netherlands, Spain, Poland and the UK; first patients included.
serenity-research.eu, news 21 Apr 2026 and research pages, read 11 Oct 2026
| Question | Consensus (Mahé 2026, Table 1) |
|---|---|
| When | At a triggering circumstance, or at a 3-month life expectancy. A 6-month time point is not recommended outside other triggers. |
| Triggers | High bleeding risk · low platelets · low haemoglobin · high fall risk · patient preference · transition to palliative care · oral or subcutaneous route difficulty · anticancer drugs stopped · low thrombotic risk |
| Who decides | Primary responsibility: oncologist, palliative care specialist and GP. Other relevant physicians, including haematology and cardiology, are contacted to take part. |
| Who is told | The patient, family, nurse and the original prescriber. |
| Reassess | After an intercurrent event, at a patient or carer request, and at follow-up visits. Not routinely every 4 weeks. |
| Before the MDT | Collect the information systematically, for example in a checklist. |
"I think it has to be a joint decision." SP08, oncologist · Baddeley 2026
Holds the prescription decision for patients under oncology care, as SERENITY's Delphi recommends.
Prognosis, goals and the conversation; NG31 shared decision-making and a named lead.
Complex cases, stewardship screen, interactions and renal dosing, and follow-through in the community.
Delphi triggers: ~3-month prognosis, transition to palliative care, bleeding risk, low platelets or Hb, route difficulty, anticancer drugs stopped, patient preference.
Checklist before the MDT (Delphi): every antithrombotic, purpose, event date, dose, renal, interactions.
Oncology, palliative care, haematology, nursing, pharmacy apply the decision aid.
Goals, preferences, reversal intent, what would happen after a bleed or a clot.
Decision, end or review date, owner and safety-net in the letter to GP, DN and hospice.
Intercurrent event, patient or carer request, follow-up visit (Delphi); prophylaxis daily (NG89 1.7.4).
Clear near-term benefit; route, monitoring, reversal plan and review trigger written down.
Benefit uncertain or a reversible risk; end date plus explicit restart or stop criteria.
No benefit in time, harm outweighs it, or the person declines; symptoms to watch for and who to call.
Synthesis of SERENITY Delphi (Mahé 2026), Gurumurthy 2026 framework, NICE NG31 1.3 and NG158 1.4.2 safety-net model · teaching pathway, not a validated protocol
| Family meeting | Anchor |
|---|---|
| Level of involvement wanted; honesty about the plan | NG31 1.3.1 |
| Informed or involved? Patients range from deferring entirely to deciding alone; relatives may speak as one unit with the patient | Baddeley 2025 |
| Advance statement, ADRT or health and welfare LPA; goals and wishes | NG31 1.3.2 |
| Named lead, with out-of-hours contacts | NG31 1.3.3 |
| Why the drug was started; benefit expected, and when; bleeding and burden | Gurumurthy 2026 |
| Would a bleed or a clot prompt hospital transfer? | Gurumurthy 2026 |
| Discharge letter | Anchor |
|---|---|
| Indication, prognosis band, bleed-risk amplifiers | Gurumurthy 2026 §8.4 |
| Decision: continue, end date, or stopped | Pathway |
| Route feasibility, e.g. tube-compatible agent | Gurumurthy 2026 §8.4 |
| If stopped: symptoms to watch for, thrombosis contact, out-of-hours | NG158 1.4.2 model |
| Plan shared with patient, family and the whole team | NG31 1.3.6 |
of recommendations from a physician–pharmacist stewardship team were accepted; 64% of 553 reviews generated one, mostly dose and interaction changes.
Koolian 2022, Res Pract Thromb Haemost
Indication and event date · dose for renal function · interactions with anticancer drugs NG158 1.3.16 · stop non-essential antiplatelets and NSAIDs · route feasibility · documentation.
| Palliative prophylaxis reviewed daily | NG89 1.7.4 |
| Prophylaxis stopped when dying recognised | NG89 1.7.2 |
| Medicine review once dying recognised | NG31 1.5.1 |
| Purpose and event date documented | local |
Treatment beyond 3–6 months with a shortening prognosis and two bleeding flags. NG158 1.4.1: assess and discuss continuing, stopping or changing.
Trajectory phase, purpose of each drug, time since VTE, platelet count, bleeding flags, swallowing and injections, what matters most.
A suggested direction with reasons, the verbatim NICE text that applies, and a documentation checklist for the family meeting and discharge letter.
Named leads in each discipline; agree triggers; one-page proforma; decision aid in the MDT.
On antithrombotics at SPC referral; documented purpose; continued to death; NICE standards.
Weekly virtual MDT slot; pharmacist stewardship screen; family-meeting template; discharge fields.
Re-audit; patient and family feedback; adopt the SERENITY tool when published; explore CoClarity links.
Let's make continuing anticoagulation a decision, not a default.
Abbel D, et al. Ann Fam Med 2026;24(2):104-110. PMID 41876113 · SERENITY
Aldridge SJ, et al. Br J Haematol 2025;207(5):2080-2090. PMID 40890090 · SERENITY
Baddeley E, et al. Res Pract Thromb Haemost 2026;10(3):103427. PMID 42017173 · SERENITY
Baddeley E, et al. Thromb Res 2025;253:109399. PMID 40651261 · SERENITY
Burnett AE, et al. A call to action for anticoagulation stewardship. Res Pract Thromb Haemost 2022;6(5):e12757. PMID 35865732
Chin-Yee N, et al. CMAJ 2022;194(35):E1198-E1208. PMID 36096505
Falanga A, et al. EHA guidelines on antithrombotic treatments in thrombocytopenic patients with cancer. HemaSphere 2022 (Zotero BMVUKAPQ)
Goedegebuur J, et al. Thromb Res 2023;228:54-60. PMID 37276718 · SERENITY protocol
Gurumurthy G, et al. Res Pract Thromb Haemost 2026;10(3):103459. PMID 42027311
Jagosh J, et al. PLoS Med 2025;22(8):e1004663. PMID 40853973 · SERENITY
Kempers EK, et al. Thromb Haemost 2026;126(7):714-728. PMID 39855271 · SERENITY
Koolian M, et al. Res Pract Thromb Haemost 2022;6(6):e12758. PMID 36186105
Mahé I, et al. Thromb Res 2026;258:109573. PMID 41548313 · SERENITY Delphi
Martens ESL, et al. Thromb Res 2025;245:109205. PMID 39667100 · SERENITY flash mob
NICE NG31 (2015); NG89 (2018, updated 2019); NG158 (2020, updated 2023)
Noble S, Banerjee S, Pease NJ. BMJ Support Palliat Care 2022;12(e6):e834-e837. PMID 31239260
Noble SI, Finlay IG. Palliat Med 2005;19(3):197-201. PMID 15920933
Noble SI, et al. BMJ 2006;332(7541):577-80. PMID 16459340
SERENITY consortium. serenity-research.eu (research, publications, news), read 11 Oct 2026
Shevlin T, et al. Clin Med (Lond) 2025;25(5):100505. PMID 40854421
Søgaard M, et al. J Thromb Haemost 2025;23(1):190-200. PMID 39393778 · SERENITY
Tardy B, et al. J Thromb Haemost 2017;15(3):420-428. PMID 28035750 · RHESO
van Hylckama Vlieg MAM, et al. EClinicalMedicine 2023;64:102194. PMID 37731937
White C, et al. Lancet Haematol 2019;6(2):e79-e88. PMID 30709436 · HIDDen
All evidence on antithrombotic continuation at the end of life is observational. Kowalewska 2017 (family and discharge slide) is quoted from Gurumurthy 2026 and was not re-verified. The pathway and decision aid are teaching syntheses awaiting local governance.