From the April 30, 2007, issue of "The Gray Sheet"
An FDA advisory panel recommended approval of Link Orthopedics' STAR ankle replacement system April 24 despite FDA reviewers' conclusion that results of the firm's pivotal trial are unclear.
The Orthopaedic & Rehabilitation Devices panel voted 4-2 in favor of Link's PMA for the Scandinavian Total Ankle Replacement (STAR) system at a meeting in Gaithersburg, Md.
STAR is the first nonconstrained mobile bearing total ankle system to be reviewed by the agency. It is intended as an alternative to fusion (arthrodesis) surgery to replace an ankle joint deformed by rheumatoid arthritis, primary arthritis or posttraumatic arthritis, according to the indication requested in Link's PMA.
Trial Sways Majority Of Panel Despite Limitations
Link's 237-patient, 16-center, nonrandomized pivotal trial was designed to show that STAR is statistically noninferior to arthrodesis. However, the data in the original PMA filing failed to meet the primary safety endpoint - a composite of device failures, revisions or removals, radiographic evaluation parameters and major complications.
Under this composite measurement, STAR was successful in 71% of cases versus 82.7% for the arthodesis control, according to FDA. Subsequently, the Hamburg, Germany-based company submitted supplemental data with two new safety analyses.
In the original analysis, seven STAR patients who did not meet the radiographic success criteria at six or 12 months but met the criteria at 24 months were counted as unsuccessful.
In the first reanalysis, Link counted these seven patients as successes. However, with this change alone, the study still failed to meet the primary safety endpoint.
An additional five patients that were clinically successful at 24 months had been counted as failures because early radiographic findings suggested the implant would eventually fail. Nevertheless, four-year follow-up on these patients confirmed the long-term success of the implant.
Link submitted a second reanalysis reclassifying the latter five patients as successes. With this change, the overall reported safety success rate for the STAR ankle group was 79.6% - statistically noninferior to the 82.7% for the control.
FDA reviewers expressed skepticism over the company's approach. For example, agency briefing documents prepared for the meeting note that "How the changes in radiographic analyses are directly related to the clinical outcomes are not clearly defined in the PMA amendment."
At the panel meeting, FDA statistician Jack Zhou also stated that "Depending on different radiographic determinations, it is not clear whether noninferiority was established for the primary safety endpoint."
In light of that uncertainty, FDA asked the panel to comment on the "appropriateness" of the post hoc reanalyses and their "impact" on the interpretation of the safety data.
"The purists in us would say it is inappropriate; however, it's a realistic way to look at the data," Acting Panel Chairman John Kirkpatrick, M.D., University of Florida College of Medicine, Jacksonville, said in his summation of the panel's recommendations. Kirkpatrick, an orthopedic surgeon, added that "as far as whether radiographic success correlates to clinical failure or clinical success, we don't know."
Panel consultant Michael Mayor, M.D., Dartmouth Medical School, Hanover, N.H., who was deputized to vote at the meeting, succinctly expressed the ambivalence toward the reanalyses shared by most of the panel members: "Appropriate? No. Adequate? Probably. Impact? No."
Kirkpatrick suggested that STAR's labeling include an explanation of the discrepancy between radiographic and clinical success rates in the trial "so that the surgeons and/or the patients can make their own decision as to what is a success or not."
Despite their reservations about the post hoc statistical analyses, most panelists agreed that Link data provided reasonable assurance that STAR is safe, especially when compared with currently available replacement ankle joints. STAR is "as safe as its peers which are already on the market. Whether it's as safe as ankle arthrodesis remains a bit of a question," stated panel member Stuart Goodman, M.D./Ph.D., Stanford University School of Medicine.
The panel's statistician, Kathleen Propert, Sc.D., University of Pennsylvania, was among the minority voting against recommending STAR approval. "I do not think there is sufficient evidence that the device is safe. I'm empathetic and applaud the attempts to work out a very difficult problem, but ... with the uncertainty of the results, I'm not willing to say that there is no reasonable doubt" about its safety, Propert said.
The primary effectiveness endpoint was based on patients' Buechel-Pappas score. The scoring system assigns a score of 1 to 100 based on ankle deformity, function, pain and range of motion.
The results showed that improvement in total BP score was significantly higher in STAR patients. In addition, significantly more STAR patients had a BP score improvement of greater than or equal to 40 points. On average, STAR patients' BP scores improved 40.5 points two years after surgery, compared with 26.3 points for the control group.
Since arthrodesis is, by definition, the attempt to immobilize the ankle joint, the company also computed results excluding the BP scoring system's range of motion component. In that calculation, the STAR group improved an average of 36.9 points versus 30 points for controls.
However, FDA's Zhou pointed out that the baseline characteristics for the control and STAR groups were different. For example, the STAR group included a higher proportion of patients with rheumatoid arthritis and primary arthritis, whereas the majority of control patients had posttraumatic arthritis.
"The pivotal study's nonrandomized design created confounding effects that are difficult to control for," he told the panel. "The comparability of the STAR and control population is questionable."
Some panelists said they would have preferred if the trial had compared STAR to another ankle joint replacement device instead of the fusion procedure. Nevertheless, all panel members agreed that they were reasonably assured of STAR's effectiveness.
Panel Recommends Five Conditions For Approval
The committee voted to recommend five conditions for FDA approval of STAR, including an eight-year, post-approval clinical study with independent radiographic assessment. The firm also should update the surgical manual to include recent improvements in implant technique.
The three remaining conditions included bench testing to establish the maximum weight STAR can support; development of patient education materials; and modification of the proposed indications for use to remove the term "severely deformed," and to substitute "degenerative arthritis" for "primary arthritis."
- Reed Miller
CLICK HERE to sign up for a free trial of "The Gray Sheet"
