John Mandrola, MD


March 22, 2017

Two years ago, Dr Raj Makkar (Cedars Sinai, Los Angeles, CA) and colleagues stunned the cardiology community with their discovery of subclinical leaflet thrombosis in bioprosthetic aortic valves—both surgical and transcatheter types.

In a 2015 paper published in the New England Journal of Medicine, they reported reduced leaflet mobility due to subclinical thrombosis in 13% of 132 patients in the RESOLVE and SAVORY registries and 40% of 55 patients in the PORTICO IDE clinical trial.[1] The investigators used 4D computed tomography (CT) imaging to detect reduced leaflet motion.

This was provocative data, but the sample size was small. At the time, the investigators wrote that further systematic study of this issue was needed.

Now we have it. In a late-breaking clinical trial here at the at American College of Cardiology (ACC) 2017 Scientific Sessions, Makkar presented follow-up work. The observational study, which was simultaneously published in the Lancet, involved a much larger group of patients from the SAVORY and RESOLVE registries.[2]

From this larger cohort, the authors were able to determine the prevalence of subclinical thrombosis, compare effects in surgical vs transcatheter valves, assess the impact of anticoagulation, and observe associations between subclinical thrombosis and valve hemodynamics and clinical events.

The study design included 890 patients who had interpretable 4D CT scans; 752 with transcatheter valves and 138 with surgical valves. Enrolled patients had multiple valve types—findings were listed for six surgical valves and 10 transcatheter valves (including ones we don't have in the US). The authors defined subclinical thrombosis by the presence of hypoattenuated leaflet thickening and reduced leaflet motion of at least 50%. Then they correlated radiographic findings with clinical data, such as the use of antiplatelet vs anticoagulant drugs, clinical events, and echocardiography.

Seven Major Findings

Prevalence of Subclinical Thrombosis. Researchers observed subclinical thrombosis in 11.9% of valves overall with 13.4% (101 out of 752) in transcatheter valves and 3.6% (five out of 138) in surgical valves. The severity of reduced leaflet motion, measured by leaflet thickness and percentage of leaflet motion restriction, was worse in transcutaneous valves.

Antithrombotic Regimen Made a Big Difference. One-quarter of the cohort were taking anticoagulants at the time of their first CT scan—about half on warfarin and half on the new direct oral anticoagulants (NOACs/DOACs). The prevalence of reduced leaflet motion was lower among patients receiving anticoagulation (4%) compared with those on dual antiplatelet (15%), single antiplatelet (16%), or no anticoagulant therapy (15%). These differences easily reached statistical significance.

NOACs Looked Equivalent to Warfarin. In 107 patients on NOACs and 117 patients on warfarin, the prevalence of reduced leaflet motion was 3% and 4%, respectively.

Anticoagulation Resolved the Problem of Reduced Leaflet Motion. Follow-up imaging was done in 58 patients who had reduced leaflet motion. A 3-month regimen of anticoagulation in 36 of these patients restored normal motion. Of the remaining 22 patients who did not receive anticoagulation, the reduced leaflet motion worsened or persisted in 20 of them.

What Happened After Stopping Anticoagulation? After restoration of normal leaflet motion with anticoagulation, reduced leaflet motion recurred in four of eight patients in whom anticoagulation was discontinued compared with none of the 15 patients who remained on anticoagulation. This observation was from the RESOLVE registry, where decisions to stop or start anticoagulation were based on bleeding risk and physician/patient preferences.

Reduced Leaflet Motion and Valve Function. Patients with reduced leaflet motion were much more likely to have AV gradients >10 to 20 mm Hg (on echo) and were more likely to have an increase of AV gradient on follow-up echos. If these patients were treated with 3 months of anticoagulation, the mean decrease in AV gradient was -7.9 mm Hg while those not on anticoagulation incurred a 0.92-mm-Hg increase in gradient.

Association With Clinical Events. During a mean follow-up of 540 days, the authors observed no differences in rates of death, MI, or stroke. Reduced leaflet motion was associated with a threefold increased rate of transient ischemic attack (TIA) (hazard ratio 3.3, 95% CI 1.45–7.50).

Context and Takeaways

Dr Makkar was quick to emphasize the limitations of the paper: crucially, this was an observational study, and unmeasured confounding is possible. He cautioned that the time separation between clinical events and the CT scans makes it difficult to say leaflet thrombosis was the cause of the TIAs.

On the higher prevalence and severity of reduced leaflet motion in transcatheter valves relative to surgical valves, Makkar offered plausible reasons—crimping during deployment, incomplete expansion or overexpansion, retained calcified native valve components—but also pointed to baseline differences in the surgical and transcatheter groups. Namely, transcatheter valves could have sustained more subclinical thrombosis because they were implanted in sicker, older patients.

Limitations aside, I think this is seminal work on a clinically relevant topic. We would not have known that slightly more than one in 10 patients with tissue valves could have subclinical thrombosis and reduced leaflet mobility if not for these researchers.

The observation that anticoagulation prevents and resolves reduced leaflet motion (but antiplatelet drugs do not) challenges current guideline recommendations to use dual antiplatelet therapy after transcatheter-valve replacement. That's not surprising, since the recommendation was based mostly on expert opinion.

The favorable performance of NOAC drugs in tissue valves is reassuring. This is consistent with increasing evidence that NOAC drugs perform similarly to warfarin for stroke prevention in patients with atrial fibrillation and bioprosthetic valves.[3,4]

The SAVORY/RESOLVE findings also reframe our thinking about small increases in valve gradients measured by echo. Typically, a 10- to 20-mm-Hg increase in gradient in patients with tissue valves is passed off as business as usual. The observation that small increases in gradient may herald subclinical thrombosis—which anticoagulants can resolve—is relevant to clinical practice now.

Like all good research, this study raises questions for future studies.

Which patients with tissue valves should be screened with 4D CT scans? Or maybe we should screen all tissue valves for reduced leaflet mobility.

A randomized trial testing anticoagulation vs antiplatelet drugs vs no antithrombotic drugs seems like a vital next step. That study should assess both valve movement and clinical outcomes.

Avoiding anticoagulation is the great advantage for tissue valves. Will this thinking hold up to future study? Adding anticoagulation is far from a freebie. In the same late-breaking session at ACC, French researchers reported a 23% incidence of new MRI-detected cerebral microbleeds in patients who had recent TAVR.[5] More anticoagulation might worsen this sort of lesion.

Another immediately relevant clinical question is whether patients with tissue valves who have TIA or stroke should undergo 4D CT scan imaging to look for reduced leaflet motion.

The most provocative and difficult-to-answer question is whether or how subclinical thromboses affect valve durability. Perhaps tissue valves have trailed mechanical valves in durability because imaging technology has not allowed us to see small clots that reduce leaflet motion and create turbulent flow. (Transthoracic echo cannot detect these small clots.)

As TAVR is offered to younger patients, uncertainty surrounding durability and use of anticoagulants adds complexity to an already-tough decision.

Yes, the discovery of leaflet thrombosis is bad news, but the silver lining is that it could prompt engineers to design better valves and it should prompt us to be more thoughtful in how we proceed with new technologies.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.