There are three key issues that will impact use of Watchman and future left atrial appendage (LAA) closure devices for prevention of stroke in atrial fibrillation patients -- who is covered, how patients will be followed, and with what comparator.
Who Is Covered
The Centers for Medicare and Medicaid Services (CMS) issued a draft National Coverage Determination (NCD) that Watchman is not reasonable or necessary, so it is not covered by Medicare. But CMS is proposing to cover Watchman under a Coverage with Evidence Development (CED) paradigm. However, CMS created some confusion about just which atrial fibrillation patients would be eligible for CED coverage of this device.
The FDA approval specified patients had to be able to take warfarin, and the NCD said it would be covered to patients for whom warfarin is contraindicated. CMS has since clarified that the FDA label is broader than the proposed Medicare coverage determination, but that Medicare proposes to cover what the FDA approved and not any off-label use.
Thus, it is likely that the confusion over eligible patients will be resolved or clarified in the final NCD. It appears that CMS simply wants to ensure that the device is "necessary," not just a convenient replacement for the need to take daily pills.
, an interventional cardiologist, a past president of the American College of Cardiology (ACC), and a member of the FDA advisory committee that recommended approval of Watchman, said, "There are some nuances between the patient population that CMS offers coverage for Watchman that is different from the FDA."
"I could argue as a clinician, and a conservative clinician, that I would look at the devices for patients who are poor candidates for anticoagulation, as opposed to as a replacement for a drug," he continued. "The argument against this is that the reality is that of patients without a contraindication for anticoagulation, more than 50% are on an anticoagulant. Why is that? Is it the failure of physicians being good doctors and following evidence-based medicine? Is it clinicians leery of bleeding? Or is it patients refusing to be on an anticoagulant? I suspect it is a little of all three."
Brindis -- who is the senior medical officer, external affairs, for the ACC-National Cardiovascular Data Registry (ACC-NCDR) -- said there are still a lot of unanswered questions about LAA closure that need answering:
- "There are concerns about long-term efficacy related to stroke.
- "Certainly, when one rolls out a new device into the community, you are concerned with safety and efficacy as more non-academic centers are starting to use it.
- "You are very interested in questions related to novel oral anticoagulants [NOACs], which were not studied in the Watchman trial but which might be utilized in the community.
- "You need to collect data to understand these things related to safety and efficacy...And we need to understand what patient profiles are getting the device."
How Patients Will Be Followed
Progress is already being made to establish an LAA closure registry. In fact, the infrastructure of the LAA closure registry is pretty much done and has been sent off for IT development. This could be ready to go by spring 2016.
Brindis said that an effort has been going on the last year, with input from interventional cardiologists, electrophysiologists, outcomes experts, professional societies, and others, to put together the data elements that would provide, from a clinical perspective, a lot of the answers needed to form an LAA registry.
"This is not just a professional society effort," he said. "These data elements have been vetted with the FDA and industry. In this case, Boston Scientific, from their studies and the data elements they had in their trials, could offer a lot of excellent comments and feedback to be sure we have correct data elements to answer the key questions...and be sure that it meets the postmarketing study requirement the FDA has appropriately mandated."
The LAA registry will be a separate registry from the , but a lot was learned from setting up the TVT registry that helped with the LAA registry. TVT is not run by industry at all, and neither will the LAA registry.
"In retrospect, the TVT registry looked simple, but it was an incredible challenge...and a true 50/50 partnership between ACC and the Society for Thoracic Surgery (STS)," Brindis said. "The fact that CMS and the FDA could see the two societies -- who you could argue might be not only collaborative but competitive -- hold hands and arm-in-arm work together, that alone led the governmental bodies to say this is the right answer."
"Then, FDA and CMS figured out they wanted to be with us, and we put them on the Steering Committee," he added. "The FDA was happy about the pre- and post safety data the registry would collect, and CMS was happy because we could start collecting good data to see if this is usual and necessary. And patients and clinicians were happy because we could study safety and ways to improve [TAVR]."
Like TVT, the LAA registry will be device agnostic. It is not a Watchman registry but an LAA registry. SentreHeart's Lariat may be included, though CMS has not mandated this with either a carrot or a stick. Perhaps St. Jude Medical will proceed with Amplatzer; and Johnson & Johnson/Biosense Webster is buying Coherex Medical, which has a CE-Marked left atrial appendage closure device, WaveCrest.
Asked who will run the LAA registry, Brindis said, "We thought a lot about this. Should the LAA module be housed in TVT, in the ICD registry, or even as a module that could be placed in any individual registry? What happened at the end of the day is the FDA and industry asked it to be set up as a stand-alone registry. So rather than piggybacking any of our current registries, it is now set up as a stand-alone registry."
"At the present time it is totally under the financial oversight and intellectual oversight of the ACC-NCDR," he noted. "We are in formal discussions with potential partners, which would include the Heart Rhythm Society [HRS] and the Society for Cardiovascular Angiography and Interventions [SCAI]...but those discussions are still happening."
Who will pay for the LAA registry? Like TVT, development is being funded by the professional societies, but operational expenses going forward would probably be covered by fees charged to participating sites.
A Control Arm?
The medical societies plan to make a case against a warfarin control arm in the LAA registry. Brindis said:
- "CMS is interested in collecting data on people who are not receiving a device. The interest is reasonable -- to compare the device to patients getting anticoagulation for afib -- but there are huge challenges related to that, and I think there will be some important back and forth on what the final answer should be.
- "It is a very unusual situation where a payor would argue you need to enter people in a registry that involves work, effort, and cost who are not getting the device. That is a potentially huge burden for the hospital.
- "You could argue it [a comparator arm] would be of value if you had a decent comparative group...but there are huge problems in making the comparators in terms of an observational study with a lot of confounders. That is a huge problem. Why didn't the patient get a device? What is different about that patient? And in our registry we may not be able to assure CMS, FDA, or clinicians that we can anticipate all these potential confounders.
- "And will this be so much burden on hospitals that I am worried about the data they will enter just to meet a requirement...We are working with CMS and trying to understand what their intent and desire is in this particular area."
While the STS surgical valve registry could be considered a comparator "arm" to TAVR, Brindis said that is not a perfect comparator either because of confounders: "Yes, we have data from a respected registry. Do we have the ability to do true comparisons? No. There are a lot of frailty and other issues that are not well captured."
Yet, Brindis said NCDR understands the interest in a comparator, and there are some possible alternative data sources. One is the ACC's ambulatory registry, PINNACLE, which has 1,200 to 1,400 cardiologists using it and more than 20 million patient encounter records, looking at things such as secondary measures for coronary disease, diabetes control, and hypertension control.
"Germane to this, we looked at afib patients who are candidates for anticoagulation, if they get anticoagulation or not, and what agents they get," he said. "So, we could argue... that PINNACLE is a potential comparator [for LAA closure] ...CMS may say you can use comparator data from billing data or administrative data. This is still liquid."
Will these efforts to eliminate a control arm from the LAA registry be successful? It seems likely that CMS will accept some other type of comparator.
The End Product
What changes are likely in the CMS final NCD? Brindis said the professional societies -- ACC, SCAI, and HRS -- are likely to submit a multi-society response, "I'm sure from experience that CMS will look strong and hard at these recommendations, and I would intuit they would find many helpful."
Other groups, including the American Hospital Association, also are expected to try to encourage CMS to drop this requirement, he said.
"This is all about the lessons learned and the huge success that the professional societies, CMS, and FDA have learned from the TVT experience," he concluded, "and that really, in my mind, has set a new paradigm going forward for the evaluation of disruptive, innovative, cardiovascular technology in harnessing the power of our registry structure."
A version of this article first appeared in Trends-in-Medicine's Quick Takes. offers its subscribers independent news and analysis on the latest data, regulatory issues, and trends of interest to the healthcare community.