Dr Toby Rogers (Medstar Washington Hospital Center, Washington, DC, US) discusses anticoagulation versus antiplatelet therapy after transcatheter aortic valve replacement in low risk patients.
This feasibility and hypothesis generating study assessed whether anticoagulation in a low risk population after TAVR for one month reduces subclinical leaflet thrombosis warfarin reduced this phenomenon without causing excessive bleeding.
1. Why did you conduct this research? Do low risk patients provide unique challenges?
2. What was the study design, technology and drugs used and endpoints?
3. What did you find and what conclusions can be made?
Filmed on location at CRT 2020.
Interviewer: William Cadden
Videographer: Charlie McClanahan
Question 1 : Why did you conduct this research? Do low risk patients provide unique challenges?
[Rogers] A number of clinical trials now and observational registries have shown that thrombosis can develop on the leaflet of bioprosthetic valves, both surgical and transcatheter. It was first detected in transcatheter valves, but was subsequently shown to affect all bioprosthetic valves. We still don't really know what it means. Does it cause more strokes? No one's every been able to really show that. Could it impact the longevity or the durability of these valves? Possibly, but the only way we'll find that out is by following these patients for many many years. So in the short term, we're stuck with this phenomenon which is we see a stuck leaflet with thrombus on it, so the leaflet's not opening and closing, and it's very hard for us to believe that that's a good thing. And the other observation from some of the registries in the past was that patients who happened to be on oral anticoagulation, so Warfarin or a DOACs, a direct oral anticoagulant drug, seemed to have a lot less of this leaflet thrombosis, which kind of makes sense. If you're on a bloodthinner, you're less likely to develop thrombosis. And so the rationale of this study was to take a low risk patient population, meaning a younger TAVR population and randomise these patients to receive either antiplatelet drugs or anticoagulation with Warfarin for one month after the TAVR, and then we'd look at them all at 30 days with a CT scan to see whether they have leaflet thrombosis or not, and whether the oral anticoagulation reduced or prevented this phenomenon from occurring.
Question 2 : What was the study design, technology and drugs used and endpoints?
[Rogers] So this was a randomised control trial, multi-center, in the US, and the original study design was for 200 patients to be randomised, one to one, to receive either antiplatelets, aspirin, or Warfarin after TAVR. So these were low risk patients with an STS score of less than three. Typical for the low risk studies that you've seen published just in the last year, which led to the approval of low risk TAVR in the US. And so that was the randomised cohort, and then patients who are low risk, but who for example, had another indication for anticoagulation, they had atrial fibulation, they had a DVT, or patients who had a bleeding problem during the TAVR, meaning it wouldn't be appropriate to randomise them and potentially end up on Warfarin, were enrolled in a separate registry arm as well. So we have the randomised cohorts at the centre of the trial, and a registry cohort on the side.
Question 3 : What did you find and what conclusions can be made?
[Rogers] So first things first, zero mortality, zero disabling stroke, at 30 days, which is consistent with all of the low risk data we've see so far, which is, if you take a low risk group of patients and you do TAVR, you get excellent results. So, zero mortality, zero disabling stroke, but also very good hemodynamics, very low rates of in-hospital and 30-day complications in terms of vascular access complications, low pacemaker rate. They're sort of the typical findings for a clinical trial in TAVR. One of the most reassuring things was that we didn't see excess bleeding in the patients who were randomised to anticoagulation. So that's one of the key safety endpoints that we were looking at in the study was are we causing more bleeding by randomising, putting patients on oral anticoagulation, which we wouldn't normally do as standard of care, at the moment at least, in TAVR. It is important to remember that the current surgical guidelines recommend that every patient who receives a bioprosthetic surgical valve should have three months of Warfarin afterwards, and it's specifically to prevent these sorts of phenomenon. Many surgeons, both in the US and abroad, don't do that. They just use aspirin because they worry about bleeding. And certainly I think bleeding is something to be concerned about in an older population, in your typical 80 year old TAVR patient, bloodthinners are often a problem. But I think these are younger patients, average age in this study was 73, and it was very similar to all of the other low risk studies, these are younger patients with fewer comorbitities and fewer bleeding risk factors. So we were very pleased to see that there was no excess bleeding. But then specifically in terms of the leaflet thrombosis, so in the randomised cohort, the instance of leaflet thrombosis in the aspirin arm was 16.3%, and in the anticoagulation arm was 4.7%. So that's quite a dramatic reduction. Unfortunately, partway through our study, low risk was approved in the US, and almost overnight, that made enrolling in a randomised trial in the low risk patients very difficult. And so we locked the data with only 94 patients randomised, as opposed to the 200 that we initially tried to get. And so despite having that quite remarkable difference in absolute value, 16.3 and 4.7, the p-value was just missed statistical significance, 0.07. But if we then take the patients who are in the registry arm, of which there were 30, and we put them into the groups depending on whether they were on oral anticoagulation or antiplatelet therapy, we then ended up with a rate of hypertonicity for thickening, or leaflet thrombosis, in the aspirin arm of 16.4% and in the oral anticoagulation of 3.1%. And that p-value is very positive, 0.01, which suggests that there really is a significance, that Warfarin is an effective strategy to reduce leaflet thrombosis at 30-days. So if I could summarise what we found, essentially Warfarin did effectively reduce, if not completely prevent the phenomenon of leaflet thrombosis, with 16% down to 3% incidence, without causing excessive bleeding. The big question is what do we do with this data? Where do we go next? Should every TAVR patient be given Warfarin? I think we should treat this as a feasibility study, because that's what it was, and as a hypothesis generating study. And I think we will need to obviously follow these patients long term, and there are other data sets from the low risk TAVR studies who have been followed long term. And if this phenomenon of leaflet thrombosis does turn out to be a predictor of premature deterioration or failure of these bioprosthetic TAVR valves, then I think there will definitely be a move within the community to be more aggressive with anticoagulation in these younger, low risk patients, who are going to live much longer with their transcatheter valve. For now, I think we also need to think carefully about whether we need more studies on this and certainly I'm looking forward to discussion at the session tomorrow.