Medical Device Daily Washington Editor
GAITHERSBURG, Maryland — This past Tuesday's meeting of the FDA's orthopedics and rehabilitative devices panel demonstrated that the agency's advisory groups are becoming increasingly serious about post-approval studies (PAS) — similar to the new emphasis on post-approval follow-up being espoused by the FDA. And despite the short life of most devices, this panel recommended follow-on tracking of essentially a decade.
The panel voted 5-2 for recommending agency PMA approval of the Scandinavian Total Ankle Replacement (STAR), with the condition that the follow-up study run to between eight and 10 years, and with heightened wear testing.
If approved, STAR will be offered for sale in the U.S. by Link Orthopaedics (Rockaway, New Jersey), a sister firm to Waldemar Link (Hamburg, Germany). Link said that the STAR would be the first three-piece ankle replacement on the U.S. market and would compete with two-piece ankle replacements such as the Agility, made by DePuy Orthopaedics (Warsaw, Indiana).
According to Andrew Greenberg, president of Link, the STAR is "the most widely used ankle outside the U.S.," available in Europe since 1990.
The sought-for indication is for replacement of "a painful arthritic and/or severely deformed ankle due to rheumatoid arthritis, primary arthrosis or post-traumatic arthrosis." Arthrosis is a degeneration of bone tissue, and the device was studied as an alternative to fusion.
The STAR is a non-cemented joint that the company tested for a compression force of about 3,000 newtons (N), thought to replicate the amount of wear induced by a patient weighing 163 lbs. However, despite having agreed to that metric in the initial trial protocol, FDA complained that this was not the worst-case load, expressing a preference for wear-testing of 250 lbs., or slightly more than 6,000 N.
One of the complaints that the agency expressed was that the study was observational in nature rather than driven by hypothesis. However, the agency did not explain why it allowed such a study to go forward, given its reservations.
The pivotal trial was for a total of 267 subjects, but a total of 320 subjects were included in the longer-term follow-up, the "continued access" trial. However, only 211 of those managed to stay with the study for 24 months for an attrition rate of about one third.
The original trial defined success as including no device failures, revisions or removals, and no radiographic evidence of loosening. However, these standards were changed to allow for up to four mm of radiolucency, or apparent space between the device components from a radiographic examination.
Another change was to allow patients who demonstrated more than the allowed radiolucency to be defined as a success if they showed clinical success at 48 months.
The firm had agreed to a post-approval study prior to the hearing, which would run to eight years, but FDA insisted it wanted the post-approval study to be driven by a hypothesis rather than a tracking study. The panel indicated that it would like to see the eight- to 10-year follow-up because of the potential for device wear. And they asked for wear testing at a 6,000-newton load.
During the discussion, Kathleen Propert, ScD, associate professor of biostatistics at the University of Pennsylvania School of Medicine (Philadelphia), said she is "becoming accustomed to non-randomized studies in this setting," but that "the baseline differences between the groups . . . made me worry that there are unmeasured co-variates there."
Propert said that she was concerned about the primary safety endpoints due to changes between the pivotal study and the longer-continued access study. Because of this, "when they change, we have to make sure that we change them uniformly across the board," she said.
Pamela Adams, the industry representative and acting president/CEO of Etex (Cambridge, Massachusetts), addressed the safety issue also, saying that many PMA studies get safety decisions with 250 subjects or less, but "if we have 400 and there was a safety signal, I think we'd see it."
Propert, who voted against recommending apporoval, expressed doubts about whether a study could retain a sufficient number of patients to make such a long follow-up meaningful, wishing the company "good luck" in the effort.
"I certainly struggled with this vote," she said.
Despite acknowledging that the sponsor established reasonable efficacy, she said she had doubts about safety. "I don't think it's unsafe — I just don't see valid scientific evidence that it is safe."
Regarding the follow-up of eight to 10 years, Danica Marinac-Dabic, MD, the chief of the epidemiology branch at CDRH, told MDD that she would prefer to see a retention rate of 80-85%, and that "sample sizes does correlate directly with retention."
However, she declined giving an estimate of how large a sample size would be needed to establish that kind of retention.
The hearing also served to suggest a correlation between the devices that go to a hearing and a change in the lead reviewer.
The first lead reviewer for the device under discussion at Tuesday's panel meeting was Hollis Rhodes, who recently took a job as the director of orthopedic regulatory affairs at Musculoskeletal Clinical Regulatory Advisers (New York) (Medical Device Daily, Feb. 12).
Likewise, the lead reviewer of the study for the Cormet 2000 hip resurfacing system, made by the Corin Group (Cirincester, UK), had also changed well before the firm and FDA took their differences to an advisory panel (MDD, Feb. 26).
Greenberg told MDD that Brian Pinder, the lead reviewer on this application, was not the original lead reviewer for the PMA. He said that the post-market requirements were not what the company expected when it first applied, but that "they are about where we would have expected last week."
Greenberg said that the difficulty of getting the STAR through FDA would force the company to rethink the kinds of products it might try to bring to the U.S. market in the future.
"It's become very tough" to get a device through the FDA process, he said — a remark that other device makers would probably agree with.