Prosthetic valve thrombosis

Article written and reviewed by Dr Luca Moretti

Prosthetic valve thrombosis consists of the deposition of thrombotic material on an artificial valve and may obstruct a mechanical occluder, thicken biological leaflets, cause regurgitation or embolize. The clinical spectrum is very broad, ranging from acute disc immobility with shock to the asymptomatic CT finding of hypoattenuated leaflet thickening after TAVI; the same definition therefore encompasses conditions with profoundly different prognoses and treatments.

In mechanical prostheses, thrombosis depends mainly on the balance between device thrombogenicity, valve position and anticoagulation intensity. Bioprostheses, long considered almost immune, can instead thrombose both early and late and mimic degeneration. Recognizing the mechanism is crucial because a biological thrombotic lesion may regress with treatment, whereas calcification and structural tearing do not respond to anticoagulation.

The diagnostic pathway integrates chronology, INR or antithrombotic therapy, changes in gradients and anatomical imaging. Echocardiography, fluoroscopy and CT are not competing modalities because they respectively describe the hemodynamic consequences, occluder kinematics and material morphology; in an unstable patient, this integration must occur rapidly enough not to delay surgery or fibrinolysis.

Prevention remains the most effective treatment. In mechanical prostheses, uninterrupted VKA therapy and competent management of interruptions prevent many events, whereas in bioprostheses a clear post-implant antithrombotic plan is essential. When thrombosis occurs, every decision must balance the immediate risk of death or embolism against bleeding risk and the likelihood that the observed tissue is truly thrombotic.

Pathogenesis, risk factors and prevention

Contact between blood and artificial surfaces activates plasma proteins, platelets and the coagulation cascade, while recirculation zones, non-physiological shear and stasis promote thrombus growth. Modern bileaflet valves have improved hemodynamics but have not eliminated thrombogenicity, and small washing jets help keep the hinges clean; when motion decreases, stasis increases and further fuels the thrombotic process.

Valve position substantially modifies risk. Mitral and especially tricuspid prostheses operate at lower velocities and pressures than aortic prostheses and are therefore more exposed to stasis; atrial fibrillation, left atrial dilatation, ventricular dysfunction, low cardiac output, previous embolism and prothrombotic states add further risk. Pregnancy, cancer, antiphospholipid syndrome and infection require even more individualized management.

In a mechanical prosthesis, the most important modifiable factor is insufficient or unstable anticoagulation. The INR target depends on model, position and patient characteristics, and there is therefore no universal target; what matters most is the time spent in range and treatment continuity, because repeated fluctuations between under- and over-anticoagulation increase thrombosis and bleeding simultaneously. Dietary variability, interactions, vomiting, malabsorption and non-adherence must be investigated systematically and without judgment.

Mechanical prostheses require lifelong treatment with a vitamin K antagonist, and direct oral anticoagulants cannot be considered interchangeable. The RE-ALIGN trial with dabigatran was stopped because of an excess of thromboembolic and bleeding events, and factor Xa inhibition has likewise not demonstrated that it can safely replace a VKA in this setting. In the treatment plan, writing simply “oral anticoagulant” is therefore dangerously ambiguous.

Interruptions for procedures require an individualized risk assessment, not automatic bridging or uniform discontinuation. Many minor procedures can be performed while maintaining the VKA within the therapeutic range, whereas when interruption is necessary, position, model and thrombotic factors determine whether heparin should be used; unnecessary bridging increases bleeding, but an excessively long unprotected interval can have catastrophic consequences.

In bioprostheses, thrombotic risk depends on a surface that is not yet endothelialized, leaflet geometry, degree of expansion, stasis in the neosinuses and flow state. The early antithrombotic plan therefore varies with surgical position, TAVI, sinus rhythm and any independent indications for anticoagulation, which is why it is incorrect to transfer a regimen automatically from a transcatheter aortic valve to a mitral bioprosthesis.

After TAVI, routine anticoagulation in the absence of another indication has not demonstrated a clear clinical benefit. In GALILEO, a rivaroxaban-based strategy increased death or thromboembolic events and bleeding compared with an antiplatelet strategy, showing that reducing CT findings does not justify a treatment capable of worsening clinical outcomes. Anticoagulation must therefore respond to a defined indication, not to the objective of “sterilizing” every image.

Patient education includes a written INR target, a contact for out-of-range values, knowledge of drug interactions, recognition of dyspnea or embolism, and the instruction never to stop treatment independently. In selected patients, self-monitoring may improve control, and after therapeutic changes, acute illness or antibiotic treatment, monitoring must be intensified. Prevention therefore does not consist of a single prescription, but of a system capable of avoiding prolonged subtherapeutic periods.

Mechanical prosthetic valve thrombosis: clinical presentation and diagnosis

Onset may be sudden, with dyspnea, pulmonary edema, hypotension or shock, or more subacute, with a progressive reduction in exercise tolerance. Cerebral or peripheral embolism may precede hemodynamic symptoms, and some patients report attenuation or irregularity of the valve click, whereas others perceive no change. An INR within range at the time of assessment does not exclude a previous subtherapeutic period and does not, by itself, prove that a mass is pannus.

Transthoracic echocardiography measures increased velocity and gradient, signal morphology, ventricular consequences and pulmonary pressure and, in the mitral position, must also report heart rate and flow. A single gradient is not sufficient: comparison with baseline, DVI, acceleration time and valve area together establish the diagnostic probability. Moreover, if the occluder remains stuck in the open position, regurgitation may predominate and the gradient may not be particularly high.

Transesophageal echocardiography makes it possible to define the presence, mobility, location and size of masses and to search for atrial thrombi or infectious complications. Prosthetic artifacts may nevertheless obscure part of the lesion and, although a soft and recent mass favors thrombus and a dense fixed formation favors pannus, the distinction is not absolute. Thrombus size and previous embolism also influence the risk of fibrinolysis and therapeutic choice.

Fluoroscopy documents the opening angle of the discs with high precision relative to the expected values for the model and is particularly useful when echocardiography visualizes the occluders poorly. It can rapidly confirm an immobile or hypomobile disc, but it does not distinguish thrombus from pannus and cannot by itself predict the effectiveness of fibrinolysis; when possible without dangerous delays, it should therefore be supplemented with morphological information.

ECG-gated CT shows peri-occluder material and measures its attenuation. Very dense, small tissue adherent to the ventricular side of the aortic sewing ring is more consistent with pannus, whereas lower-attenuation, bulkier material suggests thrombus; mixed lesions are nevertheless common and Hounsfield-unit thresholds depend on the protocol. The finding must therefore be interpreted by experienced operators while accounting for metallic artifacts.

The differential diagnosis includes prosthesis-patient mismatch, high flow, Doppler errors, endocarditis and non-structural dysfunction. Mismatch is present from baseline with preserved motion, whereas thrombosis produces a new change; fever and dehiscence point toward infection, although thrombus and vegetation may coexist. If endocarditis is plausible, blood cultures must be obtained before any fibrinolysis.

The distinction between obstructive and non-obstructive thrombosis is based on the hemodynamic consequences and not solely on the presence of a mass, because even a mobile non-obstructive lesion may embolize. Left-sided location, size, growth during treatment and previous events contribute to defining risk, whereas functional class and hemodynamic stability guide urgency but may change rapidly.

Bioprosthetic thrombosis, HALT and clinical significance

Bioprosthetic valve thrombosis may present with increased gradients, leaflet thickening, dyspnea or embolism and may occur both in the first months and many years after implantation, in surgical, transcatheter or valve-in-valve prostheses. Mitral position, low cardiac output and previous thrombosis increase suspicion; when gradients regress with anticoagulation, the therapeutic response retrospectively supports the diagnosis.

Four-dimensional CT has made it possible to describe hypoattenuated leaflet thickening, defined as HALT, often associated with reduced leaflet motion. The finding may involve one or more leaflets without significantly changing the gradient and is defined as subclinical when it causes no manifestations; however, it is not equivalent to valve failure and does not prove that every case causes neurological events. Prevalence also depends on the timing and protocol of the examination.

Echocardiography may remain normal in subclinical forms because partial restriction of mobility does not yet alter overall flow. When the gradient increases, DVI, valve area and baseline must be compared and thrombosis distinguished from structural valve deterioration: calcification and tearing favor the latter, whereas non-calcific thickening and reduced motion point more strongly toward thrombosis. The two conditions may nevertheless coexist.

The prognostic significance of isolated HALT remains more uncertain than that of clinical thrombosis. Some observational studies have associated the finding with neurological events or deterioration, but randomized analyses have not demonstrated that routine anticoagulation improves the overall clinical balance; the imaging finding must therefore be interpreted together with symptoms, embolic events, gradient, extent of the finding and bleeding risk.

CT is not indicated for indiscriminate serial screening of all TAVI recipients, but is used when there is a specific question, such as an unexplained increase in gradient, reduced leaflet motion, an embolic event or clinical suspicion. Radiation exposure, contrast and incidental findings entail costs and risks; the main surveillance strategy therefore remains echocardiographic, with CT targeted to anatomical definition.

Valve-in-valve procedures create a geometry and neosinuses that may increase stasis, and flow also depends on commissural orientation and frame expansion. This does not, however, imply an automatic indication for permanent anticoagulation: the strategy must take into account position, rhythm, previous thrombosis, bleeding risk and other indications. In patients with atrial fibrillation, moreover, anticoagulation also addresses an embolic risk independent of the prosthesis.

Therapeutic response must be demonstrated by imaging rather than simply assumed. After an appropriate interval, echocardiography and, when necessary, CT are repeated to verify reduced thickening, recovery of mobility and normalization of gradients; an incomplete response requires reconsideration of the diagnosis, adherence and possible concomitant degeneration before indefinitely prolonging a treatment burdened by bleeding risk.

Treatment of mechanical prosthetic valve thrombosis

Obstructive thrombosis of a left-sided mechanical prosthesis associated with instability requires an immediate decision between surgery and fibrinolysis. European guidelines favor urgent surgery in critically ill patients when it can be performed without delay and with acceptable risk, but local experience, contraindications, valve position, probability of pannus and actual availability may alter the choice. An excessively long transfer of a patient in shock may in fact negate the theoretical advantage of the surgical option.

Surgery makes it possible to remove thrombus and pannus, free or replace the prosthesis, and simultaneously treat endocarditis or other lesions. It is particularly indicated when pannus or infection is likely, the thrombus is bulky, fibrinolysis is contraindicated or another correction is required; operative risk increases in shock and in reoperations, which is why assessment must proceed alongside stabilization rather than after a prolonged ineffective pharmacological attempt.

Fibrinolysis can restore prosthetic valve motion without reoperation, but carries risks of bleeding, embolization and failure, especially when the mass is bulky or the tissue is not truly thrombotic. Low-dose protocols with slow or ultraslow infusion, guided by echocardiography, have achieved favorable results in experienced centers and selected patients, but they do not represent a universal formula and require exclusion of contraindications and availability of surgical rescue.

Therapeutic choice integrates functional status, thrombus size and mobility, history of stroke, pregnancy, bleeding risk, previous operations and center expertise. Suspected endocarditis is a strong reason against fibrinolysis; when time allows, the urgent decision should be documented jointly, making clear that both strategies carry substantial risks but that no treatment is often the most dangerous option.

In non-obstructive thrombosis, optimization of VKA therapy and a short course of heparin may be appropriate for small thrombi not associated with embolism, provided imaging is repeated shortly thereafter. A large, mobile mass that persists despite therapy or is associated with embolic events may instead require surgery; size, location and morphology must be considered together because progression from a non-obstructive to an obstructive form can be rapid.

After successful treatment, it is necessary to understand why thrombosis occurred. An inappropriate target, excessively spaced monitoring, interactions, interruptions, malabsorption or prothrombotic states require correction, whereas indiscriminately increasing the INR above target exposes the patient to bleeding and the addition of antiplatelet agents is not automatic. The new plan must include an explicit target, procedural management and early reassessment of motion and gradients.

During treatment, neurological signs, bleeding, hemodynamics and valvular response are monitored, and recovery of the valve click cannot replace imaging. If fibrinolysis fails or the prosthesis remains dysfunctional, repeating cycles without a new mechanistic assessment may delay necessary surgery; pannus, organized thrombus and mixed lesions account for some failures.

Treatment of bioprosthetic valves and follow-up

In clinically significant bioprosthetic valve thrombosis, a vitamin K antagonist generally represents first-line anticoagulant treatment before considering reintervention, provided there is no progressive acute heart failure or hemodynamic instability requiring an urgent strategy. Choice and duration depend on position, severity, concomitant indications and bleeding risk, and the goal is not only to reduce the gradient but also to restore leaflet mobility and prevent recurrence.

A severe obstructive form with shock or acute regurgitation caused by thrombus may not allow time to await a response to anticoagulation. Surgery or transcatheter procedures are therefore discussed according to the type of bioprosthesis, anatomy and risk; valve-in-valve may be useful in a selected failed valve, but implanting a second device on unresolved thrombosis risks trapping material, maintaining gradients and missing a potentially reversible diagnosis.

Incidental HALT, in the absence of symptoms, a significant increase in gradient or embolism, does not automatically constitute an indication for anticoagulation. Any independent indications are first identified and bleeding risk is assessed; if treatment is chosen, anatomical or hemodynamic follow-up must be planned in advance to avoid indefinite prescribing based on a finding that might resolve or remain clinically irrelevant.

Monitoring records velocity, mean gradient, DVI, valve area, regurgitation and ventricular function under comparable conditions, and CT is repeated when documentation of leaflet mobility may modify the duration or type of therapy. Normalization of the gradient together with newly mobile leaflets confirms the response, whereas progressive calcification suggests that the prosthetic valve dysfunction includes a different component.

Recurrence requires reassessment of adherence, cancer, selected thrombophilias, atrial fibrillation, low cardiac output and prosthetic geometry. Not all patients require an extensive thrombophilia panel: age, personal and family history and the circumstances of the event guide further investigations. In some cases permanent therapy may be reasonable, but it must always be balanced against bleeding risk and reassessed over time.

Before discharge, the patient must receive clear instructions on the name and dose of the drug, target and duration, monitoring schedule, management of missed doses and contacts to use before procedures. Sudden dyspnea, neurological deficit, acute limb pain or changes in the valve click require urgent medical attention. Structured communication thus becomes part of secondary prevention and reduces the risk that an initially treatable event progresses to a catastrophic complication.

Prognosis is better when diagnosis and treatment precede hemodynamic instability and organ damage. In mechanical prostheses, the main modifiable determinant is anticoagulation stability, whereas in biological prostheses serial comparison makes it possible to recognize a reversible component before it is labeled as degeneration. In both cases, thrombosis is not merely an image or an INR value, but the result of the interaction between device, flow and patient characteristics.

Pregnancy requires a dedicated pathway. In women with a mechanical prosthesis, thrombotic risk increases and no regimen simultaneously eliminates maternal and fetal risks: VKA therapy is effective for the valve but crosses the placenta, whereas low-molecular-weight heparin requires correct dosing and rigorous anti-Xa monitoring. Transitions early in pregnancy and near delivery are particularly vulnerable phases and must be planned before conception by a Pregnancy Heart Team.

When thrombosis is associated with a neurological event, management must be coordinated with neurologists and brain imaging. The urgency of restoring valve function must be balanced against the risk of hemorrhagic transformation, which may contraindicate or delay fibrinolysis and alter surgical timing; infarct size, presence of hemorrhage, prosthetic valve stability and risk of further embolism must be reassessed sequentially. Rules applicable to thrombosis without stroke cannot be transferred automatically to this scenario.

The safety of VKA treatment is also measured by time spent in the therapeutic range and not only by the most recent INR. A diary recording doses, values and concomitant medications makes fluctuations and interactions visible and, if control remains unstable, intensive education, appropriate self-monitoring and review of concomitant therapy are preferable to episodic corrections. The aim is to reduce thrombotic recurrence and bleeding together, both of which are promoted by prolonged periods outside the therapeutic range.