New, potentially important concerns have been raised about bioprosthetic aortic valves, both those implanted during surgery and those during a catheter-based procedure (TAVR).
Investigators from three separate groups reported on their troubling findings in a paper published in the New England Journal of Medicine. The article was accompanied by both an editorial and a perspective from the FDA. The FDA authors offer reassurance, stating that the problem uncovered in the main paper “is an early signal of unknown clinical significance.” Based on the currently available data “the FDA believes that the benefits of using these devices for the currently approved indications continue to outweigh the risks.”
The chief finding is a high rate of reduced leaflet mobility with bioprosthetic valves. Although the numbers are still quite small, the researchers report rates of reduced leaflet mobility ranging from 7% to as high as 40%. The clinical significance of the finding is far from clear at this point.
The leaflet problem first emerged during the ongoing PORTICO IDE trial comparing the investigational St. Jude Portico valve to TAVR valves already on the market. In that trial 22 out of 55 TAVR patients (40%) were found to have reduced aortic-valve leaflet motion. (As a result of the finding the trial was paused in September 2014. The trial resumed in June of this year.) In response to the finding investigators then initiated two registries, RESOLVE at Cedars-Sinai Medical Center in Los Angeles and SAVORY at the Rigshospitalet in Denmark. Reduced leaflet motion was found in 13% (17 out of 132) of patients in the registries. The rate of reduced leaflet motion was 7% (2 out of 27) in patients with surgical valves.
The investigators believe that the finding is likely due to subclinical leaflet thrombosis. In addition to imaging findings suggestive of thrombus formation on the affected leaflets, there was a significant correlation of reduced leaflet motion and anticoagulant dose. 51% of patients (21 out of 41) receiving subtherapeutic or no anticoagulation were found to have reduced leaflet motion, compared with no patients (0 out of 8) receiving warfarin (p=0.007). Reduced leaflet motion persisted in 9 out of 10 patients who did not receive anticoagulation but was resolved in 11 patients who started or continued to receive therapeutic anticoagulation.
The clinical implications are unclear at this point. There was no significant association between leaflet motion and aortic-valve gradient. The potentially highly troubling finding of an increased risk of stroke in these patients was deemed “preliminary and inconclusive.” Overall 6 strokes occurred in the studies, but all occurred within 24 hours of implantation, suggesting that “these first-day strokes were related to the procedural aspects of TAVR rather than to leaflet thrombosis.” But, the investigators acknowledge the possibility that patients receiving bioprosthetic valves may be at higher risk for “leaflet thrombosis and consequent embolic stroke than has perviously been recognized.”
In an accompanying editorial, David Holmes and Michael Mack write that “given the rapid clinical adoption of TAVR” the new findings are “extremely important,” though “clear scientific answers are still lacking.” They list a number of important questions raised by the paper, including the true incidence of reduced leaflet motion and its clinical significance. Finally, they ask, in the context of continued broad and widening acceptance of TAVR, whether this issue needs “to be fully resolved before the expansion of Food and Drug Administration approval of TAVR for lower-risk patients.”
The FDA officials said in their perspective that the agency “is closely monitoring this signal and working with the clinical community and the device manufacturers to elucidate it.”
Responding to the papers, Sanjay Kaul (Cedars-Sinai Medical Center) provided the following comment:
Overall, the quality and quantity of data are insufficient to evaluate the clinical relevance of the findings or to justify a hasty action by the regulatory agencies. However, it is appropriate to raise the awareness and increase vigilance. If replicated across other valve programs and non-thrombotic causes ruled out, the data would suggest initiating anticoagulation treatment routinely in all TAVR cases unless contraindicated. The choice and timing of anticoagulation is unclear but the prothrombotic adverse events seen with dabigatran in the RE-ALIGN trial in patients with mechanical prosthetic valves makes me lean in favor of warfarin over the newer anticoagulants. This, however, should be an active area of investigation.
Ajay Kirtane (Columbia University) said that until now this issue has not been on the radar screens of most interventional cardiologists, largely because it is difficult to detect and has not been linked to date with any clinical sequelae. He was also reassured by the fact that the strokes occurred on the first day after implantation, “whereas this phenomenon of subclinical leaflet thrombosis appears to be more subacute… I do think it’s important to note that the TAVR population as a whole is at greater risk for strokes just based upon demographics and risk factors alone. And even with longer term follow-up in the major TAVR trials, there hasn’t appeared to be an adverse signal of high late stroke events or valve deterioration, especially when comparing either TAVR vs. medical therapy or TAVR vs. SAVR.”
Kirtane said that the editorial by the FDA authors “is very much on point. While this phenomenon must be understood, it must also be placed in the clinical context of the overall beneficial results of aortic valve replacement. The issue of whether to anticoagulate after aortic valve replacement remains controversial, with varying practice throughout the country even after surgical AVR. Whether patients ought to be screened by 4D CT at some interval is something that will now certainly be debated as well. But before any new therapeutic interventions are recommended, it is important to gain more information on the overall true incidence of this with contemporary valves, and the clinical sequelae that can be attributed directly to it.”