This article is a collaboration between 鶹ý and:
GAITHERSBURG, Md. -- An FDA advisory committee struggling with limits of the regulatory process delivered a split vote Wednesday along with a nuanced recommendation -- once again leaving the Watchman left atrial appendage closure device's future in serious doubt.
The Circulatory System Devices Advisory Panel convened to deliberate the Watchman device's risk-benefit profile, after updated clinical trial data from the PREVAIL trial showed a higher incidence of ischemic stroke in Watchman subjects who were given short-term warfarin compared with a control group given warfarin alone.
At the end of the day, the panel voted unanimously that the Watchman device was safe for the indicated patient group.
A second vote, on whether the product is effective in preventing clot formation and stroke, was initially split 6-6. Committee chairman , a cardiologist at the University of Wisconsin's School of Medicine and Public Health and president of the Heart Rhythm Society, who only votes in case of ties, cast his ballot in the No column.
A third vote regarding the benefit-risk profile was again split with six members in favor of a positive balance, five against, and one abstaining.
Normally the FDA regards split votes as unfavorable to the product in question, although it is not bound by panel recommendations.
After much reflection and confusion over a Bayesian meta-analysis pooling two clinical trials, PREVAIL and the PROTECT-AF, the panel appeared to struggle with a regulatory bind: how to recommend the device as a second-line therapy when the patient group indicated by the voting question, as well as the data, did not match the select subset of patients to whom panel members might hypothetically offer the device.
The Watchman device, produced by Boston Scientific, specifically blocks off the left atrial appendage, where clots have been known to form, with the goal of preventing the possibility of such clots lodging in an artery leading to the brain resulting in stroke. The company is seeking its approval for patients who, owing to their CHADS2 or CHA2DS2-VASc scores, would be recommended for warfarin therapy to lower their risk of stroke and systemic embolism.
This was the third time that the device had come before the committee. In 2009, it split on whether it should be approved, and then a year ago it came down definitively on the side of approval. Yet the FDA has continued to demand additional data before granting its O.K., setting the stage for Wednesday's meeting.
, a cardiac electrophysiologist from the Community Medical Center in Missoula, Montana told 鶹ý she voted No to the question concerning the device's efficacy and to approving the current risk-benefit profile. "After the last meeting, I assumed that when we had more patients and more information we would have even stronger data in support of the device and we actually don't. It's beginning to diverge," said Kelly, referring to differences in the two major trials, PREVAIL and PROTECT-AF (which were later pooled and differentially weighted).
"I would love to have this device for use when I needed it, but I'm less comfortable having it approved for blanket use in anybody that would otherwise be eligible for warfarin," Kelly said.
The day's discussion focused on balancing the benefit of oral anticoagulation compared with its bleeding risk and the stroke risks that may be elevated in the Watchman device group.
An underlying theme was the suggestion of an "unmet need" for an alternative to warfarin for certain populations who either aren't offered the medication because of the high risk of major bleeding or who are very dissatisfied with the quality of life the medication offers, as patients who commented during the public session commented. The vote prompted some intricately worded responses and many caveats.
Admitting that it might not be "kosher," , a professor and statistician at the University of Alabama at Birmingham said he voted Yes to the third question concerning the risk-benefit profile. Rather than the current indications, he said, "I voted on what I think the indications are going to be based on the discussion."
Several members of the panel, including some of those who dissented, concurred with Naftel's logic.
, an electrophysiologist at Long Island Jewish Medical Center who voted No said, "I felt that I had to take the questions literally." He later said his vote was "changeable" depending on the final labeling.
Kelly told 鶹ý that the "absolute" vote wasn't necessarily significant. "The numbers I don't think matter all that much if you told them I voted Yes, but I only mean Yes if it's this select population."
As to how the FDA will act upon the recommendation Kelly said, "I have no idea. The panel simply makes a recommendation, and the FDA does what they think is best."
Avoidance of long-term warfarin therapy, whose side effects include a risk of major bleeding, appeared to be one clear goal of the device, as suggested by the Boston Scientific representatives and nearly a dozen senior citizen patients and a handful of doctors who spoke during the public hearing. (Most of these speakers said their travel fees were paid for by Boston Scientific.) However, clinical trial protocols required subjects to be eligible for warfarin.
The Boston Scientific team also argued that the cardiovascular death rate was lower for Watchman than the control group, 1% compared with 2.3%. The FDA noted in background documents that none of the deaths were causally linked to the Watchman device, implantation, or anticoagulant therapy.
, chief medical officer at the division of Cardiac Rhythm Management, Boston Scientific, agreed with the panel that the device should not be seen as a first-line therapy. "Patients who are doing well on oral anticoagulation, patients who are anticipated to continue doing well on anticoagulants, ought not to be considered for this device."
Page, the committee chair, said that in spite of the complex nature of the debate, "I think the panel got it right." He summed up the decision saying the majority of the committee felt the device should be considered as a second-line therapy.
"I think this device has a home, and we just need the FDA and the sponsor to work together to develop a description for an indication that's appropriate," he said.
Also in background documents distributed prior to the panel the FDA concluded that the semifavorable results of the WATCHMAN group found in the earlier PROTECT AF, a randomized controlled trial that suggested device noninferiority on some measures, may have been due to a higher than expected hemorrhagic stroke rate in the warfarin group.
"[T]he PROTECT AF trial demonstrates a signal that the WATCHMAN device reduces the risk of hemorrhagic stroke compared with warfarin therapy. However, the robustness of this signal is diminished by several factors including non-use of warfarin in one subject, absence of imaging confirmation in one subject, concomitant use of antiplatelet agents in several subjects, and events associated with trauma in several subjects," the documents noted.
FDA staff also noted that when the panel voted last year in favor of approval, it did not have access to final data from PREVAIL. Only 28% of PREVAIL subjects had reached the 18-month follow-up window at that point.
In February 2014, the FDA received updated follow-up data from the PREVAIL trial showing six additional ischemic strokes in the WATCHMAN subjects and none in the control group. And as of June 2014, there were 24 primary endpoint events in the device group (including 13 ischemic strokes) and nine events in the control group. Even accounting for the 2:1 randomization in favor of Watchman subjects, FDA staff were still concerned by the disproportionate rate of strokes in the device group.
In addition, the PREVAIL study, which a representative of the FDA described as a "better executed trial," failed to show noninferiority and performed worse than the control group with regard to ischemic strokes. Boston Scientific attributed the higher ischemic stroke rate to the study's small sample size.
Boston Scientific said it has designed a "robust" post-approval trial, if the FDA responds positively, that would include 1,000 new patients who would be followed for 5 years.
From the American Heart Association: