October 9th, 2014

FDA Panel Gives Cautious Endorsement to Novel Boston Scientific Device

The FDA’s Circulatory System Devices advisory panel gave an extremely cautious endorsement on Wednesday to Boston Scientific’s Watchman device, a novel catheter-delivered left atrial appendage closure device for people with atrial fibrillation. They signaled that although they thought the device should be made available they also thought that there should be significant restrictions on its use.

The panel wrestled throughout the day with a fundamental problem: combined data from the two clinical trials, PROTECT AF and PREVAIL, showed clearly that Watchman was not equivalent to warfarin for the chief indication of stroke reduction in the study population. Furthermore, in clinical trials patients who received the Watchman were compared to patients taking warfarin. But there was a strong sentiment among panel members that the device should be available only to AF patients who are eligible for warfarin but don’t want to take it and, perhaps also, for patients who are not eligible for warfarin therapy. The problem, of course, is that neither of these patient groups was studied in the clinical trials.

Panel members delivered a complex message to the FDA regulators who will ultimately decide the fate of the device. They unanimously agreed that the device was safe, but split 6-6 on the question of efficacy. The panel’s chair, Richard Page, broke the tie with a no vote. On the final and third vote on whether the risks outweighed the benefits, 6 panel members voted yes, 5 no, and 1 (former ACC president Ralph Brindis) abstained. But panel members were highly vocal in stating that their vote did not indicate in any way support for the actual proposed indication for stroke reduction in warfarin-eligible patients.

Panel members said that they wanted the FDA to craft a narrow and highly restricted indication. Summarizing the panel’s sentiment, Page said it was incumbent on the sponsor and the FDA to maximize the degree to which only appropriate patients would get the device. “It’s not an alternative to every warfarin eligible patient,” he said. Panel members and FDA staff discussed several possible measures to achieve this goal, including patient guides, a controlled rollout with limited device availability, the requirement that an independent physician sign off before the procedure, and a national registry.

Bram Zuckerman, the director of the FDA’s Division of Cardiovascular Devices, agreed that “if approved it would require continuous monitoring of appropriate use.”

Panel members were acutely aware of their difficult position. Said Page: “The problem is that the proposed indication — prevention of thromboemboslim — is no longer realistic, so a likely real world use is for the device to be used in a different population and not as a replacement for warfarin.” Patricia Kelly suggested that the Watchman “could be offered to patients who don’t want to take warfarin, but they would need to be told that it is not considered equivalent to warfarin.”

The panel struggled with the differences between first-line therapy and second-line therapy as well as warfarin-eligible, warfarin-intolerant, and people who just don’t want to take warfarin. Said Page: “Every patient in Watchman clinical trials was warfarin eligible but now the only possible indication may be for warfarin-ineligible patients.”

David Kandzari said that even a “second line indication would be hard when both primary efficacy endpoints were missed.” Page noted that it would be difficult to make sure that it was available only as a second line.

Given the failure of the trials to demonstrate noninferiority, Rick Lange commented: “I’m just a simple guy. If something is not noninferior then it’s inferior.” He then asked “how do we recommend an inferior therapy to patients?”

Ultimately, however, the panelists agreed with Page that the “totality of data argues in favor of clinical choice,” although they were unable to define that choice with any precision.

Watchman has been under development for more than a decade and its approval has twice been postponed by the FDA. Most of the controversy over the device has centered around two clinical trials. Although the initial PROTECT AF trial was technically a positive trial, it had a number of important weaknesses and in 2010 the FDA required the company to perform another clinical trial to demonstrate the long-term safety and effectiveness of the drug. The PREVAIL trial was then designed to supplement PROTECT AF. The initial results appeared somewhat promising, but the complicated trial design, with three co-primary endpoints and a novel and difficult-to-understand Bayesian trial design, led to considerable controversy. Based on complete results from the trial, FDA reviewers and panel members agreed that it failed to show that Watchman was an acceptable alternative to warfarin.



4 Responses to “FDA Panel Gives Cautious Endorsement to Novel Boston Scientific Device”

  1. Is the fundamental issue that we over simplified and based too much on the assumption that stroke risk in a fib patients is tightly linked to LAA thrombus generation and embolization? Neurological events in this patient population can be from a variety of mechanisms and LAA occlusion treats one mechanism. There are parallels to what we are dealing with in the PFO-cryptogenic stroke area. It has been presumed that the causality linkage with LAA thrombus risk would be tighter in this arena.

  2. I wonder if a better trial design would have been to take patients deemed too high-risk for oral anticoagulation by their clinicians (estimated at up to 40-50% of elderly patients in some studies) and randomized them to usual care versus the Watchman device. We have drugs that are already effective (warfarin and novel anticoagulants) but the issue in practice is often how to apply stroke reduction therapies to our frail elderly patients who are at high-risk for events such as falls and intracranial bleeding. These high-risk patients may be the most relevant candidates for the device – but its use in them will have to be based on individual judgment rather than trial effectiveness.

  3. John-
    The design you propose was considered by almost every device manufacturer in the space over the last decade. And since it would have been a superiority study, it would have been far easier to design (superiority studies are almost always simpler than non-inferiority trials) and possibly even smaller than the studies we have now. The problem was that the FDA could never agree on a meaningful definition of “too high risk for oral anticoagulation” and therefore, the study design was never really pursued in earnest. It is ironic, then, that the population that seems to be being targeted for these devices is exactly the population that you are describing.

  4. David Powell , MD, FACC says:

    This is indeed a difficult sell. Patients who are “not eligible” for anticoagulation are not only difficult to classify, but also likely older with more comorbidities. They may be more susceptible to procedural complications. Any absolute benefit of the device over aspirin or nothing is increased by this elderly population’s increased annual stroke risk but also in the longer run limited by their shorter expected survival.
    Is the other device (Lariat?) trying to study this “anticoagulation ineligible” group?