The Rise and Fall of BVS and the FDA Device-Approval Process

John M. Mandrola, MD; Sanjay Kaul, MD


February 26, 2018

John M. Mandrola, MD: Hello. I'm John Mandrola for and Medscape Cardiology. I'm here at Transcatheter Cardiovascular Therapeutics (TCT) 2017 in Denver and joined by cardiologist Sanjay Kaul, from Cedars-Sinai Medical Center. Dr Kaul is a known voice of reason. Welcome, Sanjay.

Sanjay Kaul, MD: Thank you, John.

Regulatory criteria for device approval are reasonable assurance of safety and effectiveness. In my opinion, the device failed to meet those criteria.

First-Generation Absorb BVS: Why Was It Approved?

Dr Mandrola: We are going to discuss the rise and fall of the first-generation Absorb bioresorbable vascular scaffold (BVS). By way of background, the promise of the BVS was the potential for reduced late stent thrombosis and neoatherosclerosis and restoration of basal motor activity.

The Absorb BVS received CE Mark approval in Europe in 2011 and US Food and Drug Administration (FDA) approval in July 2016. At that time, I was pretty shocked because every study I had read showed numerically more complications with the BVS. For instance, the 3-year results from ABSORB II[1] found that the BVS did not meet its mechanistic co-primary endpoint of superior vasomotor reactivity. The 2-year results from ABSORB III[2] led to an FDA warning in March of 2017 about higher major adverse cardiac event rates with the BVS. By May, Abbott had restricted use of the BVS in Europe; and in September, Abbott halted sales of the first-generation Absorb BVS in the United States.

Sanjay, my first question to you is: Should the so-called "disappearing stent" have been approved in the first place?

Dr Kaul: With the caveat that hindsight is always 20/20, my answer is no, it should not have been. Regulatory criteria for device approval are reasonable assurance of safety and effectiveness. In my opinion, the device failed to meet those criteria, even though the panel voted overwhelmingly in favor of approval (9-0), and the FDA approved the device in 2016.

It was a noninferiority study. Noninferiority studies are typically done to show that your new device or drug is no worse than your standard control treatment. One is willing to accept any degree of inferiority provided the new device or the drug offers ancillary advantages in terms of safety, convenience, or cost. The new drug has to offer something before we will consider any degree of inferiority for it to be clinically acceptable.

Here, they used a 4.5% absolute risk difference as the noninferiority margin.[3] The pivotal trial ABSORB III[4] showed that the upper bound of the difference between the BVS scaffold and the Xience drug-eluting stent (DES) was less than the noninferiority margin. Therefore, noninferiority was established.

The expected event rate with the DES arm was 7%, but the observed event rate was actually 6%. What that means is that had they fixed the noninferiority margin in terms of risk ratio rather than risk difference, noninferiority would not have been inferred. And all of the other patient-oriented endpoints were leaning in the wrong direction.

Dr Mandrola: What do you mean by patient-oriented?

Dr Kaul: Stent thrombosis was numerically greater; it was not statistically significant at 1-year follow-up. The target-vessel myocardial infarction (TVMI) was also numerically greater in the BVS arm. There was no evidence of any angiographic procedural success or advantage in terms of minimal luminal diameter or percent stenosis. There was no advantage in secondary endpoints that were powered such as target-vessel revascularization or restenosis.

With no evidence of any ancillary benefits with the device, why would a clinician accept any degree of inferiority?

Dr Mandrola: Again, I'm not a trialist; I'm a beginner. Why was it appropriate to choose a noninferiority trial? I can see a noninferiority trial for a new device that is something special or different. Is a different kind of stent that much different from the stent we have now?

Dr Kaul: I would argue that a superiority trial would have been the right design, but then the sample-size requirement would have been much higher. Therefore, for feasibility reasons, a noninferiority design is considered reasonable.

The outcomes were based on 1-year data, but theoretical advantages of the new device would not accrue until 2 years, 3 years, or even later. One of the criticisms I have for the FDA's decision was that they did they not wait for 2-year outcome data or 3-year outcome data. At the time of the approval of the BVS scaffold, they already had access to the 2-year data from the ABSORB II study,[5] which is a smaller study. All of the patient-oriented and device-oriented endpoints were going in the wrong direction. It seems to me that the device was approved on a promissory note. I would argue that for regulatory agencies, drugs and devices need to be approved on the basis of the evidence at hand.

How Do You Balance Evaluation and Innovation?

Dr Mandrola: That's a really good segue into what I wanted to ask you about next. I struggle with this as an electrophysiology doctor who uses high-tech devices. How do we balance safety and progress? We want innovation; we want to be on that cutting edge. On the other hand, devices like a BVS or the Sprint Fidelis lead in the implantable cardiac defibrillator world are getting implanted into patients without a lot of premarket evaluation. How do you balance that? How should doctors think about it?

Dr Kaul: Yes, the FDA is always caught in a bind. Their two core missions are to promote public safety and to promote innovation. Sometimes it can be very difficult to reconcile the two.

As a clinician, I lean towards first establishing that the drug or the device is safe; "first do no harm" is the guiding principle. There was already evidence that stent thrombosis rates were going in the wrong direction. I think that the FDA should have waited. Now at longer follow-ups, we have data from the ABSORB II trial, 3-year follow-up data[1] that were presented at TCT in 2016. Even the so-called theoretical advantage in terms of vasomotion and less luminal loss did not materialize.

We also have the 2-year follow-up of ABSORB III[2] that was presented earlier in the year. The differences that were not inferior at 1-year follow-up, both for the primary endpoint as well as stent thrombosis, were now statistically significantly inferior in the BVS arm.

At the TCT meetings, the 3-year follow-up data[4] confirmed the trend. There was a three-fold increase in risk for stent thrombosis and a 2.6% absolute risk increase in TVMI. It is a cautionary tale about getting ahead of the evidence in terms of the approval.

Challenges With Device Approval

Dr Mandrola: Devices seem to be coming out sooner in Europe than in the United States, and we seem to be in a catch-up game with them. For instance, it looked like it was good to be ahead with transcatheter aortic valve replacement (TAVR), but perhaps with the BVS and maybe left atrial appendage occlusion, maybe it's not so beneficial to be ahead.

Dr Kaul: Right. You can extend that to renal denervation [for hypertension]. It was approved in Europe, and hundreds and thousands of people got [treated with] the device. When we did the appropriate sham-controlled trial,[6] we did not see the benefits; we only saw the harms.

Faster access of devices to the market can only be justified if we can ensure reasonable assurance of efficacy and safety in the postapproval phase.

I'm very conservative about how the FDA should approve devices. The current trend is to shift the burden of proof from the preapproval phase to a lifecycle approach. The emphasis is that you accept some evidence during the preapproval phase with some degree of uncertainty that is acceptable, but you conduct postapproval studies to ensure that benefits and risks are fully considered in the postapproval phase.

The problem is that this lifecycle approach is predicated on a timely completion of the postapproval commitments. Recently, a letter published in the New England Journal of Medicine[7] addressed this issue. Only 50% of the postapproval commitments were completed 5-6 years after the drugs or devices were approved. Twenty-five percent had not even been initiated, and 25% were delayed.

In 2007, Congress legislated authority to the FDA with the Prescription Drug User Fee Act (PDUFA) IV that if companies did not complete their postapproval commitments, their approval could be revoked or rescinded. I've not seen that happen.

A transparent accounting of the postapproval commitments is essential to justify a lifecycle approach.

Dr Mandrola: Maybe we will improve with that. In the absence of that now, what do you think of the idea of approving devices or even reimbursing for devices on the basis of a controlled registry or some kind of controlled trial?

One example I like to cite is the clot-removing device for stroke. A big study was done in a northern European country because they could not get the device unless they enrolled patients in a clinical trial. I've always wondered about the Watchman [left atrial appendage closure device], for instance. If we had put these devices in patients enrolled in a clinical trial, would we be well along in knowing the answer?

Dr Kaul: Yes, reimbursement with evidence development. It's a right step. You can extend that to the patent foramen ovale (PFO) closure devices. We implanted the device in hundreds of thousands of patients, and it took us 9 years to enroll a few hundred patients to do the trial. Even after the trial was done, it did not meet its primary endpoint, and yet the device was approved by the FDA based on fulfillment of postapproval commitments.

I would say that faster access of devices to the market can only be justified if we can ensure reasonable assurance of efficacy and safety in the postapproval phase. For that, the postapproval process has to be tightened up.

John Mandrola: Sanjay, it's an honor to talk to you. Thank you for helping us understand these important concepts.

Dr Kaul: Delighted to be here.

Dr Mandrola: To our listeners, we invite your input. Please leave comments and your ideas for the use and approval of novel medical devices. Until next time, this is John Mandrola for | Medscape Cardiology.


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.
Post as: