Prosthetic Valves and Anticoagulation

Meet the patient

A 34-year-old woman with a mechanical mitral valve lands at 3am, 26 weeks pregnant, suddenly breathless and saturating 84 percent, her mitral clicks muffled, lungs full of water. Her dose-adjusted LMWH was halved last week for a nosebleed.[1]

Two questions now decide the next hour and the next decade: is the valve clotted? (echo now) and why is she on LMWH instead of warfarin? (the pregnancy question). Hold both, and everything below slots into place.[1][2]

The trade-off that decides everything — durability versus anticoagulation

Every prosthetic valve is one bargain, struck once, on the operating table. Roughly 100,000 valve replacements are done each year in the United States, and the count climbs as degenerative aortic stenosis overtakes rheumatic disease as the dominant indication.[1]

Mechanical valves are machined from pyrolytic carbon and titanium — essentially indestructible, outliving the patient — but every non-biological surface activates clotting on every heartbeat, so the price is lifelong warfarin. Bioprosthetic (tissue) valves are built from bovine pericardium or porcine leaflets, barely thrombogenic, anticoagulated for only the first three months — but the tissue calcifies and tears, failing at 10 to 15 years and demanding a re-operation.[1][2]

The bargain is irreversible, which is why the valve consultation matters more than the operation. TAVI/TAVR added a third door — a tissue valve delivered through the groin for the elderly or inoperable, no sternotomy, no bypass.[5]

Etymology for viva gold: prosthesis, from the Greek prosthesis, "an addition" — a part added to replace what disease took away. The word survived two and a half millennia because the concept never changed.[1]

Three families, one decision — the classification

The classification runs along two axes — how it is built (mechanical, bioprosthetic, transcatheter) and where it sits (aortic, mitral, tricuspid, pulmonary). Both axes set the INR target and the thrombosis risk.[1][2]

[1] [1] [5] [1]

The position rule that earns marks: mitral and tricuspid valves thrombose far more than aortic, because the low-pressure chambers leave slow flow across the leaflets. The high-velocity aorta is protective. That single haemodynamic fact explains every INR difference on this page.[1]

The 65-year mirage — how to choose the valve

Two decades ago, half of all implants were mechanical; today only 20 to 25 percent are, swept aside by tissue and transcatheter options as patients trade durability for freedom from the needle.[1]

[1]

The cluster rule — "young: mechanical; old: tissue; everyone in between: a conversation." Under 50, mechanical usually wins on durability (a redo at 60 is brutal); over 70, bioprosthetic or TAVI (a redo in 15 years may never come); in between, a shared decision around anticoagulation burden, re-operation risk, comorbidity, and patient preference. Modern guidelines have dropped the rigid age cut-off in favour of individualised choice.[1][2]

Mechanical valves — hear the click, mind the INR

The bileaflet (St. Jude, CarboMedics, On-X) is the modern standard — over 80 percent of mechanical implants, two semicircular pyrolytic-carbon leaflets pivoting in a housing, near-physiological laminar flow, durable for decades. Its one soft spot is the hinge, where stasis breeds thrombus the moment the INR dips.[1]

The tilting disc (Medtronic Hall, Bjork-Shiley) and caged ball (Starr-Edwards, the first successful valve, 1960) are history you will still meet in long-term follow-up clinics. Know them by their clicks; do not implant them. The Bjork-Shiley even earned its own infamy — strut fracture with outlet-strut embolisation.[1]

Bioprosthetic valves — buy time, not a lifetime of warfarin

The tissue valve's gift is three months of warfarin and then freedom. Bovine pericardial (Carpentier-Edwards PERIMOUNT) leads; stented porcine (Hancock II, Mosaic) follows; stentless (Freestyle, Toronto SPV) and sutureless (Perceval, Intuity) variants serve the small annulus and the high-risk surgical patient.[1]

Everyone forgets: structural valve deterioration is not a valve defect — it is the expected failure mode. Calcium hydroxyapatite stiffens the glutaraldehyde-fixed collagen scaffold, then the constant flexing tears the leaflet at the stent-post base, producing a mixed stenotic-and-regurgitant lesion over months.[4]

TAVI — the valve through the groin

The balloon-expandable Edwards SAPIEN and the self-expanding Medtronic CoreValve Evolut have marched TAVI from inoperable (PARTNER 1B) to high- (CoreValve US Pivotal), intermediate- (PARTNER 2, SURTAVI), and now low-risk severe aortic stenosis — non-inferior or superior to surgery across the spectrum.[5]

TAVI is now the commonest aortic-valve procedure in the elderly, and its anticoagulation footprint is light: lifelong aspirin 75 to 100 mg, clopidogrel 75 mg for 1 to 6 months, and no routine anticoagulation without another indication.[7]

Why mechanical valves clot, why tissue valves fail

A mechanical valve is a permanent intravascular foreign body, and blood does what blood does to foreign bodies — it clots on them. Factor XII meets the negatively charged carbon surface (contact activation, the intrinsic pathway); monocytes on the sewing ring express tissue factor; platelets adhere and release procoagulant microparticles; and the abnormal flow leaves stagnation in the hinges and the minor orifice — the classic address of valve thrombosis.[1]

A tissue valve fails by calcifying. Glutaraldehyde kills the cells but leaves the collagen scaffold, and over the years calcium hydroxyapatite stiffens the leaflets until they tear — usually at the base near the stent post — producing a mixed stenotic-and-regurgitant lesion that worsens over months. That is structural valve deterioration, and it comes faster in the young, in the mitral position, in chronic kidney disease, and in pregnancy.[4]

[1]

Thrombus versus pannus — the face-off every stem is built on

Both obstruct the valve; only one lyses. This is the single most repeated comparison in prosthetic-valve exams, and examiners check every discriminator.[1]

[1]

The discriminator line: thrombus is acute, soft, subtherapeutic, and dissolves with lysis; pannus is chronic, dense, therapeutic, and needs the knife. Cardiac CT settles it — thrombus enhances with contrast, pannus stays low-attenuation.[1]

Read the prosthetic valve like the echo does — presentation and bedside

Most replaced-valve patients are well; the click is the only sign of the metal inside them. Symptoms mean a complication has arrived, and the tempo tells you which one.[1]

• Acute (hours): sudden severe dyspnoea, pulmonary oedema, syncope, or cardiogenic shock in a mechanical-valve patient — acute valve thrombosis until proven otherwise.
• Subacute (days to weeks): fever, sweats, malaise, new murmur, embolic phenomena — prosthetic valve endocarditis.
• Gradual (months): slowly worsening dyspnoea and fatigue with a new murmur — structural deterioration, pannus, or patient-prosthesis mismatch.
• Insidious: anaemia, jaundice, dark urine from intravascular haemolysis — paravalvular leak.
• Embolic: transient ischaemic attack, stroke, or systemic embolus — thrombus or vegetation.
• Bleeding: epistaxis, gastrointestinal bleed, bruising, intracranial haemorrhage — over-anticoagulation.[1]

The click is the bedside anchor. A mechanical valve snaps a sharp, metallic, high-pitched click at closure — loud at the apex for mitral, at the right second intercostal space for aortic. A muffled or absent click is an alarm for thrombosis with restricted leaflet motion; a tissue valve closes with a soft, native-sounding sound you will not separate from S1.[1]

Everyone forgets: a normal TTE does not exclude prosthetic mitral pathology — the metal casts an acoustic shadow that hides thrombus and vegetations behind it. Suspect the mitral, order the TEE. TTE catches only 25 to 50 percent of prosthetic vegetations; TEE lifts that to 82 to 96 percent.[1][8]

The INR ladder — name the number, name the valve

These are the numbers a prosthetic-valve prescriber must recite from memory, in the dark, at 3am.[1][2]

[1]

The full ladder, in the order the examiner expects it:[1]

• Mechanical aortic valve, bileaflet, all low-risk modifiers (sinus rhythm, no LV dysfunction, no LA enlargement, no prior thromboembolism): INR target 2.5 (range 2.0 to 3.0). Add aspirin 75 to 100 mg daily.
• Mechanical aortic valve with any high-risk modifier (older-generation valve, atrial fibrillation, prior thromboembolism, hypercoagulable state, LV dysfunction, LA enlargement): INR target 3.0 (range 2.5 to 3.5). Add aspirin 75 to 100 mg.
• Mechanical mitral valve: INR target 3.0 (range 2.5 to 3.5). Add aspirin 75 to 100 mg.
• Mechanical tricuspid valve: INR target 3.0 (range 2.5 to 3.5). Bioprosthetic is preferred when feasible because right-sided mechanical valves thrombose at unacceptable rates.
• On-X mechanical aortic valve, selected patients (the only mechanical valve with evidence for a lower INR): INR target 1.5 to 2.0 from 3 months post-implant, per PROACT. Add aspirin 75 to 100 mg. This is the one exception, and it requires the specific On-X valve.[9]
• Bioprosthetic valve (any position): warfarin INR 2.5 (range 2.0 to 3.0) for the first 3 months, then lifelong aspirin 75 to 100 mg daily. If atrial fibrillation coexists, anticoagulate long-term for AF.
• TAVI: lifelong aspirin 75 to 100 mg, with clopidogrel 75 mg for 1 to 6 months (no role for routine dual antiplatelet beyond 6 months). If AF coexists, anticoagulate for AF — but GALILEO showed rivaroxaban post-TAVI without AF caused harm.[7]

The cluster rule — "3 mitral, 2.5 aortic low-risk, 3 aortic high-risk, 1.5 to 2.0 only On-X": say it as one breath and you have the spine of every INR question. Add aspirin 75 to 100 mg across the mechanical set.[1]

The one line that must never be crossed — DOACs in mechanical valves

The classic trap: a junior reaches for a DOAC because it is "easier" than INR monitoring — and converts a controlled patient into a stroke. Check the valve type before you sign any anticoagulant prescription. Dabigatran, rivaroxaban, apixaban, edoxaban — none of them belong in a mechanical valve.[3]

Bioprosthetic plus AF, and the TAVI anticoagulation trap

The picture flips completely for tissue. For a bioprosthetic mitral valve with atrial fibrillation, RIVER (Guimarães, NEJM 2020) showed rivaroxaban non-inferior to warfarin for death or major bleeding at one year — so a DOAC is acceptable for bioprosthetic valves with AF after the first 3 months. The same logic extends to the bioprosthetic aortic valve, supported by sub-analyses of ARISTOTLE and ROCKET-AF.[6]

The TAVI trap is the mirror image. GALILEO (Dangas, NEJM 2020) tested rivaroxaban 10 mg plus aspirin versus clopidogrel plus aspirin after TAVI in patients with no other anticoagulation indication, and was stopped early for harm — rivaroxaban raised death, thromboembolic events, and major bleeding. Do not routinely anticoagulate a TAVI valve unless there is a clear indication such as AF.[7]

[6]

On-X — the one mechanical valve that bends the INR rule

The On-X bileaflet aortic valve is the sole mechanical valve licensed for a lower INR. Its smoother pure-carbon surface and more laminar flow profile cut thrombogenicity enough that PROACT proved INR 1.5 to 2.0 (with aspirin) non-inferior to standard from 3 months post-implant, with reduced bleeding.[9]

The named trap: PROACT Xa then asked whether apixaban could replace warfarin even on the On-X platform — and was discontinued early for excess thromboembolic events. Even On-X needs warfarin. The DOAC ban is absolute.[9][3]

Acute valve thrombosis — the 3am emergency

This is the time-critical emergency of prosthetic-valve medicine, and untreated mortality approaches 50 percent. A mechanical-valve patient (especially mitral) who suddenly becomes breathless, oedematous, syncopal, or shocked has valve thrombosis until proven otherwise.[1]

[1]

A small non-obstructive thrombus in a stable patient often melts with a slow UFH infusion and re-imaging — not every clot needs lysis or the knife. The other emergencies — endocarditis with heart failure, a haemodynamically significant paravalvular leak, acute mechanical failure — all converge on urgent or emergency surgery, with antibiotics, support, and transfusion as temporising measures.[1][8]

Surgery, lysis, or heparin — the thrombosis decision

Obstructive plus shock goes to surgery; obstructive with no surgeon goes to lysis; non-obstructive and stable goes to heparin and watch. That is the whole decision tree in one line.[1]

• Valve thrombosis, obstructive with shock: emergency surgery; fibrinolysis if surgery unavailable.
• Valve thrombosis, non-obstructive or stable: slow UFH infusion and re-image; many small thrombi resolve with anticoagulation alone.
• Structural valve deterioration: redo surgery for younger surgical candidates; valve-in-valve TAVI for elderly or high-risk patients with a degenerated bioprosthetic aortic valve.
• Paravalvular leak: percutaneous closure (Amplatzer vascular plug) for symptomatic or haemolysing leaks; surgical repair for large leaks or those with infection.
• Pannus: surgical excision or valve replacement — pannus does not respond to anticoagulation or fibrinolysis.
• Endocarditis: organism-targeted IV antibiotics for 4 to 6 weeks; surgery for heart failure, uncontrolled infection, large mobile vegetation, periannular extension, or resistant organism.
• Patient-prosthesis mismatch: prevention at implantation; valve-in-valve or surgical replacement for severe symptomatic PPM.[1]

The surgical threshold in prosthetic-valve endocarditis is lower than in native-valve IE. Heart failure from regurgitation, periannular extension (abscess, fistula, heart block), uncontrolled infection (persistent bacteraemia, fungal or resistant organism), and a large mobile vegetation with embolic risk are all indications to operate.[8]

Bridging for surgery — who, when, and who to leave alone

Bridging is an examiner favourite and a daily ward decision — and the modern answer is "less than you think."[1]

[1]

Who needs a bridge? Only the high thrombotic risk patient — mechanical mitral, older-generation aortic (tilting disc, caged ball), any mechanical valve with prior thromboembolism, or any mechanical valve with multiple risk factors (AF, hypercoagulable state). The low-risk bileaflet aortic in sinus rhythm does NOT need bridging — the BRIDGE and PERIOP2 trials showed that omitting it was non-inferior for thromboembolism and significantly reduced bleeding.[1]

The classic trap: bridging every mechanical-valve patient "to be safe." You are not being safe — you are causing bleeds for zero thrombosis benefit in the low-risk group. Reserve the bridge for the high-risk ladder above.[1]

Pregnancy with a mechanical valve — the highest-risk clinic

Pregnancy in a mechanical-valve patient is high-risk by definition. The maternal thromboembolism rate runs 5 to 15 percent even with optimal care, and warfarin is teratogenic. The strategy is built before conception, in a joint obstetric-cardiac-haematology clinic.[1]

[1]

Emergency reversal — vitamin K and PCC

Match the reversal to the bleed — never nuke a slightly high INR. The options, in order of speed and completeness:[1]

• Vitamin K 5 to 10 mg intravenously (slow infusion over 20 to 30 minutes to avoid anaphylactoid reaction) — restores functional clotting-factor synthesis in 6 to 12 hours, full effect up to 24 hours; the reversal is sustained.
• Prothrombin complex concentrate (PCC) 25 to 50 IU/kg intravenously — reverses within minutes and is preferred for life-threatening bleeding or immediately before emergency surgery; combine with vitamin K for sustained reversal.
• Fresh frozen plasma 15 mL/kg — slower and less complete than PCC but widely available.
• Avoid abrupt reversal in a mechanical-valve patient without life-threatening bleeding — the thrombotic risk is high.[1]

Dental prophylaxis — the 2-gram habit

Every prosthetic valve needs antibiotic cover before invasive dental work — there are no exceptions. Extractions, scaling, periodontal surgery, and implant placement all seed the sewing ring with oral streptococci.[1][8]

[1]

The dose you carry for life: amoxicillin 2 g orally, 30 to 60 minutes before. Penicillin-allergic: clindamycin 600 mg, or azithromycin or clarithromycin 500 mg. Unable to take oral: ampicillin 2 g IV or IM, or cefazolin or ceftriaxone 1 g IV or IM.[1][8]

Prosthetic valve endocarditis — Modified Duke and early versus late

Fever plus a new murmur in a prosthetic valve is endocarditis until proven otherwise. Three sets of blood cultures, TTE and TEE (TEE mandatory), then organism-targeted IV antibiotics for 4 to 6 weeks.[8]

Modified Duke — reproduced verbatim, because examiners check every line.[8]

[8]

Clinical Definite IE requires two major, or one major plus three minor, or five minor criteria.[8]

[8] [8]

Possible IE requires one major plus one minor, or three minor. Rejected when a firm alternative diagnosis is made, resolution occurs on antimicrobial therapy for under four days, there is no pathological evidence at surgery or autopsy, or the criteria above are not met.[8]

[8]

The organism clock: early (under 1 year) — coagulase-negative Staph, then Staph aureus; late (over 1 year) — viridans streptococci, then Staph aureus, enterococci, HACEK. The echo findings that satisfy the major criterion are an oscillating intracardiac mass on the valve, an abscess, or new partial dehiscence of a prosthetic valve — and for prosthetic material, TEE is more sensitive than TTE.[1][8]

Investigations — echo first, always against the baseline

An echocardiography-first strategy, every time, read against the post-operative baseline.[1][2]

[1]

Follow-up — how often to scan

Baseline TTE before discharge; TTE at 6 to 12 months; then TTE every year for a mechanical valve, and annually for a bioprosthetic after 5 years. INR every 4 weeks when stable, more often after a dose change or a new drug — antibiotics and amiodarone move it fastest.[1][2]

• Baseline TTE before discharge after implantation — the reference for every future comparison.
• TTE at 6 to 12 months post-implant — catches early dysfunction, patient-prosthesis mismatch, paravalvular leak.
• TTE annually for mechanical valves from then on — asymptomatic thrombus, endocarditis surveillance, ventricular function.
• TTE annually for bioprosthetic valves after 5 years (or sooner if symptomatic); before 5 years, every 1 to 2 years is acceptable.
• TTE any time the clinical status changes — new murmur, fever, dyspnoea, embolic event.
• INR monitoring for mechanical valves — every 4 weeks when stable; more often after dose changes, new medications, or intercurrent illness. Self-monitoring devices improve time in therapeutic range.[1]

[1]

How patients come to harm — the preventable list

• A DOAC prescribed to a mechanical-valve patient — the single most preventable catastrophe. RE-ALIGN harm, reproducible in a single signature.[3]
• An acute mitral-valve thrombosis read as "heart failure" while the muffled clicks went unexamined, delaying echo and surgery.[1]
• A normal TTR trusted to exclude a prosthetic mitral vegetation, when TEE was mandatory.[1][8]
• Dental prophylaxis omitted for a scaling in a prosthetic-valve patient, with endocarditis weeks later.[1][8]
• A young woman with a mechanical valve started on warfarin in pregnancy without a pre-conception plan — embryopathy in weeks 6 to 12.[1]
• A low-risk bileaflet aortic valve bridged "to be safe" for elective surgery, with a major bleed and no thrombosis prevented.[1]
• High-dose vitamin K given for a trivially high INR and a nosebleed, triggering valve thrombosis from rebound hypercoagulability.[1]
• A tissue valve chosen for a 35-year-old to avoid warfarin, failing structurally at 8 years and demanding a high-risk redo.[4][1]
• Routine rivaroxaban after a TAVI without AF — GALILEO harm, avoidable.[7]

The trials that changed practice

[3] [6] [7] [9] [5]

The mantra, and the mnemonic

[1]

The mantra: mechanical means warfarin for life, tissue means warfarin for three months, and no DOAC ever touches a mechanical valve.[1][3]

Ward-round test — three stems

Stem 1 — the pregnant woman from the top of the topic (answer)

The 34-year-old with a mechanical mitral valve, 26 weeks pregnant, suddenly breathless with muffled clicks and pulmonary oedema. Her LMWH was halved for a nosebleed. What happened, and what do you do? Model: This is acute prosthetic valve thrombosis precipitated by a subtherapeutic anticoagulant dose — the halved LMWH dropped her below anti-Xa 0.8 to 1.2 U/mL. Confirm with TTE then TEE plus fluoroscopy (restricted leaflet motion, raised gradient, visible thrombus). She is in cardiogenic shock, so emergency surgery is the treatment of choice; if no surgeon is available, fibrinolysis (alteplase 10 mg bolus then 90 mg over 90 minutes for a left-sided valve). Restore full anticoagulation with UFH post-stabilisation. The deeper lesson: her anticoagulation was a pre-conception decision that should never have been halved for a nosebleed without anti-Xa rechecking.[1]

Stem 2 — the dabigatran prescription (answer)

A 60-year-old with a mechanical bileaflet aortic valve is referred from the GP on dabigatran 150 mg twice daily for "stroke prevention," INR not checked. What is the error, and what do you do? Model: This is the single most preventable catastrophic error. DOACs are absolutely contraindicated in any mechanical valve — RE-ALIGN showed dabigatran caused more valve thrombosis, stroke, and bleeding than warfarin. Stop the dabigatran immediately, bridge with UFH, and switch to warfarin INR 2.0 to 3.0 (target 2.5) plus aspirin 75 to 100 mg (a low-risk bileaflet aortic in sinus rhythm). Image with TTE and TEE for any thrombus the DOAC may have allowed to form. The rule is absolute: no DOAC, no exception, not even On-X.[3][1]

Stem 3 — the bioprosthetic mitral valve and new AF (answer)

A 72-year-old, 4 months after a bioprosthetic mitral valve replacement, develops atrial fibrillation. The registrar plans to start warfarin lifelong. What is the modern alternative, and what is the trial? Model: A DOAC is acceptable here. RIVER (Guimarães, NEJM 2020) showed rivaroxaban non-inferior to warfarin for death or major bleeding at one year in patients with a bioprosthetic mitral valve and AF. She is past the 3-month post-operative window, so rivaroxaban (or apixaban, by extension) is a reasonable choice. The contrast to remember: a DOAC is fine in bioprosthetic plus AF after 3 months, but forbidden in any mechanical valve, and not routine after TAVI without AF (GALILEO harm).[6][7]

References

1. [1]Otto CM, Nishimura RA, Bonow RO, et al. 2020 ACC/AHA Guideline for the Management of Patients With Valvular Heart Disease: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines Circulation, 2021.PMID 33332150
2. [2]Vahanian A, Beyersdorf F, Praz F, et al. 2021 ESC/EACTS Guidelines for the management of valvular heart disease Eur Heart J, 2022.PMID 34453165
3. [3]Eikelboom JW, Connolly SJ, Brueckmann M, et al. Dabigatran versus warfarin in patients with mechanical heart valves N Engl J Med, 2013.PMID 23991661
4. [4]Capodanno D, Petronio AS, Prendergast B, et al. Standardized definitions of structural deterioration and valve failure in assessing long-term durability of transcatheter and surgical aortic bioprosthetic valves: a consensus statement from the European Association of Percutaneous Cardiovascular Interventions (EAPCI) endorsed by the European Society of Cardiology (ESC) and the European Association for Cardio-Thoracic Surgery (EACTS) Eur Heart J, 2017.PMID 29020344
5. [5]Adams DH, Popma JJ, Reardon MJ, et al. Transcatheter aortic-valve replacement with a self-expanding prosthesis N Engl J Med, 2014.PMID 24678937
6. [6]Guimarães HP, Lopes RD, de Barros E Silva PGM, et al. Rivaroxaban in Patients with Atrial Fibrillation and a Bioprosthetic Mitral Valve N Engl J Med, 2020.PMID 33196155
7. [7]Dangas GD, Tijssen JGP, Wöhrle J, et al. A Controlled Trial of Rivaroxaban after Transcatheter Aortic-Valve Replacement N Engl J Med, 2020.PMID 31733180
8. [8]Habib G, Lancellotti P, Antunes MJ, et al. 2015 ESC Guidelines for the management of infective endocarditis: The Task Force for the Management of Infective Endocarditis of the European Society of Cardiology (ESC). Endorsed by: European Association for Cardio-Thoracic Surgery (EACTS), the European Association of Nuclear Medicine (EANM) Eur Heart J, 2015.PMID 26320109
9. [9]Jawitz OK, Wang TY, Lopes RD, et al. Rationale and design of PROACT Xa: A randomized, multicenter, open-label, clinical trial to evaluate the efficacy and safety of apixaban versus warfarin in patients with a mechanical On-X Aortic Heart Valve Am Heart J, 2020.PMID 32693197