Article Text

Download PDFPDF

Subclinical leaflet thrombosis after transcatheter aortic valve implantation
  1. Satoshi Nakatani
  1. Correspondence to Professor Satoshi Nakatani, Department of Health Sciences, Osaka University Graduate School of Medicine, 1-7 Yamada-oka, Suita, Osaka 565-0871, Japan; nakatani{at}


Although clinically significant valve thrombosis after transcatheter aortic valve implantation (TAVI) is rare, the incidence of subclinical leaflet thrombosis has been reported to be up to about 10%–15%. It is mostly found 1–3 months after procedure in any type of transcatheter heart valve. Leaflet thrombosis is detected by high-resolution CT in the form of limited valve opening/closure and hypoattenuated leaflet thickening. Transthoracic or transesophageal echocardiography is capable of detecting limitations of valve motion, leaflet thickening, increased flow velocity across the valve. However, CT seems to be more sensitive than echocardiography to detect leaflet thrombosis. It can occur under dual antiplatelet therapy with aspirin and a thienopyridine but rarely occurs with anticoagulation with a vitamin K antagonist. A vitamin K antagonist is also helpful to resolve leaflet thrombosis. Several studies are ongoing to determine the effect of new oral anticoagulants (NOACs) in preventing major cardiovascular events. They will also provide useful information on whether NOACs prevent leaflet thrombosis.

  • cardiac imaging and diagnostics
  • aortic stenosis
  • transcatheter valve interventions

Statistics from

Transcatheter aortic valve implantation (TAVI) has been widely performed as a treatment for patients with symptomatic severe aortic stenosis (AS) that is inoperable or at high risk from surgery. Recently, there has been a movement to expand the scope of indications to include patients who are less compromised (at intermediate or low risk) on the basis of improved safety and favourable interim results. TAVI may be established as a first-line treatment for severe AS in the near future. However, TAVI is associated with some problems, including bleeding and vascular complications, paravalvular leaks, atrioventricular block, AS and/or aortic regurgitation (AR) due to structural valve degeneration, stroke and endocarditis. Valve thrombosis has also recently been problematic, although it is considered uncommon.1 It may progress and produce symptoms of heart failure and systemic thromboembolism. Moreover, it could affect valve durability.2 There is increasing evidence that asymptomatic thrombosis is not at all rare. It can occur even under antiplatelet drugs. This paper reviews the literature on the present status of subclinical leaflet thrombosis of transcatheter heart valves (THVs) used in TAVI.

Imaging features of post-TAVI leaflet thrombosis

Traditionally, prosthetic valve thrombosis has been diagnosed by identifying reductions in the opening/closure angle of the valve at fluoroscopy in the case of mechanical valves, limitation of valve opening/closure, abnormal echoes near the valve, increases in flow velocity across the valve on transthoracic or transesophageal echocardiography and other approaches. It is often after the development of clinical symptoms that these signs are detected; we have been unable to diagnose valve thrombosis in an asymptomatic patient early. As TAVI has been widely performed, it has become common practice to examine the aortic valve using CT, making it possible to make the diagnosis before the development of haemodynamic abnormalities. Post-TAVI leaflet thrombosis is detected by high-resolution CT in the form of limited valve opening/closure3 and hypoattenuated leaflet thickening (HALT) (figure 1).3 4 Although no direct evidence is available, it seems certain that these signs truly indicate the adhesion of a thrombus for two reasons: both signs are ameliorated by anticoagulant therapy5 and CT delineates thrombi as low attenuation while pannus as high attenuation.6 Echocardiography too is capable of detecting limitations of valve opening/closure, leaflet thickening, increased flow velocity across the valve and increased mean pressure differences, and shows such changes.2 7 8 Leaflet thicknesses of 2 mm or more and the presence of abnormal cusp motion are recognised as signs of valve thrombosis that are found on echocardiography.8 However, even with the use of transesophageal echocardiography, it is not always possible to clearly delineate leaflet motion after TAVI, and many patients with mild opening limitation will not necessarily have increased pressure differences. Leetmaa et al, for example, studied four patients who had CT-detected valve thrombosis but were asymptomatic, reporting aortic valve peak pressure differences of 13, 15, 29 and 18 mm Hg and effective valve areas of 1.6, 2.3, 1.3 and 1.4 cm2, which were not substantially different from the levels obtained at the time of hospital discharge.7 These patients had mobility limitations involving one or more valve cusps with the use of transesophageal echocardiography, whereas clearly mobile masses were identified in only one patient. Therefore, CT is considered to be more able to detect early leaflet thrombosis than echocardiography. A recent study reported that paucity of colour Doppler flow within a region of the valve stent can be used as an indicator of leaflet restriction and probable thrombosis.9 Paucity of colour Doppler flow may be a useful sign that can be detected by transthoracic echocardiography, although its capacity for HALT detection seems to be lower than that of CT.

Figure 1

Leaflet thrombosis found in a man aged 83 years after 1 week of Symetis valve implantation. The valve was asymmetric and hypoattenuated leaflet thickening (white arrow) was observed at the short- and long-axis view of the aortic valve. This patient had developed transient cerebral ischaemic attack 2 days before CT examination.


The incidence of valve thrombosis after aortic valve replacement (AVR) with a bioprosthetic valve is not so high. A meta-analysis of 5837 patients undergoing AVR with stented porcine bioprostheses showed that the annual incidence of postoperative valve thrombosis was 0.38%.10 According to a report from the Mayo Clinic, 0.74% of patients undergoing valve replacement with a bioprosthesis needed repeat surgery for valve thrombosis.8 On the other hand, a pathological study on 397 bioprosthetic valves removed from the body for some reason detected thrombosis in 46, of which 29 were in the aortic position.8 However, most of these patients had increased valvular pressure differences by echocardiography, or had paroxysmal atrial fibrillation, or had a poorly controlled international normalised ratio (INR). Therefore, the incidence of subclinical leaflet thrombosis remains unknown but may be higher than previously expected.8 In fact, one study reported that CT revealed reduced leaflet motion in 5 of 138 patients (4%) after AVR.3

Because clinically significant valve thrombosis was not observed in a multicentre trial of the CoreValve or in the PARTNER trial of the Sapien valve, the incidence of significant valve thrombosis after TAVI does not seem to be high.11 12 However, the incidence of subclinical thrombosis remains unknown. Recent published articles that mention the incidence of clinical or subclinical valve thrombosis is summarised in table 1. The frequency of leaflet thrombosis, including symptomatic cases, varied widely from 0.19% to 40%. According to one case report, time of onset varies widely from 3 days to 2 years after procedure.13 When these literature references are combined, subclinical leaflet thrombosis seems to be observed 1–3 months after TAVI with the incidence up to about 10%–15%.14 15 In addition, leaflet thrombosis has been reported in various kinds of THVs.16

Table 1

Recent studies on leaflet thrombosis

Clinical findings

As leaflet thrombosis progresses, dyspnoea develops because of valve stenosis13; however, most patients are asymptomatic when HALT is found incidentally after repeated CT examinations. Pressure differences naturally increase as the valve opening becomes more limited; one study reported that 16% of patients who were asymptomatic but had reduced leaflet motion on CT had a peak pressure difference of 20 mm Hg or greater.14 However, there are some cases where echocardiography does not demonstrate any significant increase in valvular pressure difference.4 Yanagisawa et al reported that the D-dimer level was significantly higher in a group of patients with HALT observed 6 months postoperatively than in the group without (2.3 vs 1.1 mg/mL, p=0.002).4 Elevated D-dimer levels may serve as a reference indicator in cases of suspected HALT. Elevated N-terminal pro-B-type natriuretic peptide levels may be another sign.17 One report showed that the incidence rate of transient ischaemic attacks was higher in a HALT group than in a non-HALT group.14

Why does subclinical leaflet thrombosis occur?

The Edwards Sapien valve is a bovine pericardial tissue valve, whereas the Medtronic CoreValve is made of porcine pericardial tissue. These materials have traditionally been used in bioprosthetic valves, and yet they seem to be associated with a higher incidence of thrombosis than conventional bioprosthetic valves.14 According to Mylotte et al, there are five likely causes of post-TAVI thrombosis: (i) the elderly TAVI population is more likely to have coexisting prothrombotic conditions (eg, cancer), (ii) the metallic THV frame could potentially provide a nidus for thrombosis, (iii) incomplete THV expansion can create leaflet folds and potential recesses for thrombus formation, (iv) incomplete THV apposition to the aortic wall may delay endothelialisation and (v) the native leaflets may overhang balloon-expandable systems creating areas of diminished blood flow and stagnation.18 Other possibilities include failure to achieve normal structuring of leaflet tissue making thrombosis likely to occur because the valve is folded before insertion, the valve is pushed and expanded by the balloon in the balloon-expandable type and in some cases, post-dilatation is added using the balloon.16 One report suggested the possibility of association of hypersensitive reaction to metal.19 The following patients are also reported to be likely to have thrombosis: patients with large valves,2 4 patients with an ejection fraction of 35% or lower,7 patients with a valve-in-valve placement,17 20 males,4 patients with a large sinus of Valsalva,4 patients not receiving anticoagulants2 and patients in whom incomplete or asymmetrical expansion occurred.20 These facts suggest the involvement of blood flow stagnation near the valve, regional low flow rates and other factors important for thrombus formation, but these have not definitively been established as risk factors. Although it has been suggested that leaflet thrombosis occurred more commonly with balloon-expandable THV,17 it can occur in all kinds of THVs (table 1).


Bioprosthetic valves are made of porcine valves or materials derived from bovine pericardium. Following AVR with a bioprosthetic valve, low dose aspirin is usually performed for the first 3 months after implantation until completion of endothelialisation, even in patients in sinus rhythm, after which thrombosis does not pose a substantial clinical problem even without the use of antithrombotic drugs.21 22 However, there are some studies that showed the superiority of anticoagulation to aspirin23 and anticoagulation is reasonable in patients at low risk of bleeding in the recent American Heart Association/American College of Cardiology (AHA/ACC) guidelines.22

After TAVI, currently available guidelines recommend that aspirin alone or dual antiplatelet therapy with aspirin and a thienopyridine (DAPT) should be administered for 1–6 months, after which aspirin should be taken for life (table 2). Because patients undergoing TAVI are mainly octogenarians with multiple comorbidities, some investigators have advocated the use of aspirin monotherapy which is safer than DAPT.24 25 Anticoagulation with a vitamin K antagonist (VKA) is recommended in patients at low risk of bleeding in AHA/ACC guidelines in 2017 as class IIb21 22 26 (table 2). Under these recommendations, stroke occurs in about 3% of patients, although most cases are probably procedure related.27 Taking into account patient’s age and the current situation that the incidence of life-threatening bleeding is high at 10%–15%,27 the use of antithrombotic drugs should be minimised. There are several retrospective studies showing that single antiplatelet therapy may be an acceptable treatment compared with DAPT in elderly patients at high risk of bleeding.25 A recent prospective randomised study (Aspirin Versus Aspirin+ClopidogRel Following Transcatheter Aortic Valve Implantation, NCT01559298 and NCT02640794) also showed aspirin alone reduced the risk for major or life-threatening events while not increasing the risk for myocardial infarction or stroke.24 However, if we consider leaflet thrombosis as an end point, aspirin monotherapy or DAPT may be ineffective in completely preventing leaflet thrombosis. Symptomatic patients with thrombus-based valve dysfunction are treated by thrombolysis, surgical thrombectomy, AVR or valve-in-valve placement. However, these treatments are invasive and are problematic if used in patients with few symptoms. On the other hand, it is noteworthy that the majority of patients with leaflet thrombosis improve with anticoagulant therapy.3 5 16 28 Chakravaty et al reported anticoagulation (VKA or new oral anticoagulants (NOAC)) for 3 months was associated with restoration of normal leaflet motion in all patients,14 while Jose et al reported the median time to reduction of peak pressure difference with oral anticoagulation was 14 days.17 Therefore, desirable treatment for valve thrombosis seems to be early detection and use of anticoagulants.

Table 2

Recommendations for antithrombotic therapy following TAVI

In recent years, some articles have shown NOACs are effective and may be a useful treatment without a need for INR control.14 According to a study of TAVI for AS complicated by atrial fibrillation, safety end points specified in the VARC-2 criteria (a composite of all-cause mortality, all strokes, life-threatening bleeding, acute kidney injury, coronary obstruction, major vascular complications and valve dysfunction requiring reintervention) were reached by a smaller number of patients using apixaban postoperatively than by those using warfarin.29 An international, randomised, open-label, event-driven, phase III trial, the GALILEO study (Global Study Comparing a rivaroxaban based Antithrombotic Strategy to an antiplatelet based Strategy After Transcatheter aortic valve rEplacement to Optimise Clinical Outcomes, NCT02556203) is currently underway in an attempt to determine whether rivaroxaban is capable of preventing major cardiovascular events. The primary end point is the composite of all-cause death, stroke, myocardial infarction, symptomatic valve thrombosis, pulmonary embolism, deep vein thrombosis and systemic embolism. We anticipate that the study may provide useful information on whether NOACs prevent leaflet thrombosis, although the focus of this review is subclinical thrombosis. The ATLANTIS study (Anti-Thrombotic Strategy After Trans-Aortic Valve Implantation for Aortic Stenosis, NCT02664649) is a randomised study to identify which of the following improve prognosis: apixaban, warfarin or antiplatelet therapy. Valve thrombosis is an end point, and outcomes are of special relevance. The POPular-TAVI Study (Antiplatelet Therapy for Patients Undergoing Transcatheter Aortic Valve Implantation, NCT02247128) was conducted as a registry-based prospective study to assess the safety of post-TAVI monotherapy by comparing outcomes of aspirin monotherapy against a combination therapy of aspirin and clopidogrel in one cohort; and outcomes of warfarin monotherapy against a combination of warfarin and clopidogrel in the other cohort, where patients required warfarin treatment for atrial fibrillation or other reasons. Although the objective of this study was not to determine the effects on subclinical leaflet thrombosis, it may provide additive information from subgroup analysis.

Unanswered questions

There is no doubt that subclinical leaflet thrombosis occurs after TAVI. Many questions however remain to be answered, including when to suspect this disorder and how to diagnose it.


If the patient is dyspnoeic or symptoms of heart failure worsen after TAVI, valvular dysfunction must be suspected, and various tests including echocardiography should be performed. However, if we wish to diagnose valve thrombosis before clinical symptoms develop, which of the available imaging tools should we choose? CT seems to be more sensitive than echocardiography in diagnosing valve thrombosis. However, many questions remain to be answered, including when and on whom to perform CT examinations, whether to perform CT examinations on a regular basis or how to determine the risk if CT examinations are performed on high-risk patients only. Fundamentally, we need to establish CT diagnostic criteria for valve thrombosis and its severity. Recently reported systematic CT methodology for the evaluation of subclinical leaflet thrombosis devised during the Portico trial investigation may be useful.30 Echocardiography is non-invasive and permits repeated use, thus being suitable for follow-up. In this sense, it may be reasonable to perform repeated, careful echocardiography, and to do CT scans when the pressure difference starts to increase or valve mobility starts to decrease.


In many cases, subclinical leaflet thrombosis is known to improve with anticoagulation. Some questions then arise: Which types of patient should receive anticoagulation? Are there cases where thrombus disappears spontaneously as suggested by Bax and Stone?15 Should anticoagulation be used in all patients after TAVI, or be given only in high-risk patients? How do we determine the risk? What is an appropriate duration of treatment for complete endothelialisation to occur? In addition, use of NOAC is convenient as there is no need to check the INR periodically. There are several studies reporting the disappearance of valve thrombosis with the use of NOAC.14 We would like to know which is better, NOAC or VKA after TAVI? The GALILEO study and the ATLANTIS study will give us the answer.


Here too, there are many questions to answer, for example, assuming resolution of valve thrombosis following anticoagulation, how long should anticoagulation be continued and how do we check for recurrence of valve thrombosis?

Thus, there are many questions to be answered. To answer these questions, large-scale randomised longitudinal trials are needed.



  • Contributors SN is the sole author.

  • Competing interests None declared.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles