November 14, 2008

Testimony on Collagen Scaffold Medical Device

Testimony of Jonas Hines and Peter Lurie, M.D., M.P.H.
Before the Orthopedic and Rehabilitation Devices Panel
of the Medical Devices Advisory Committee
Center for Devices and Radiological Health, Food and Drug Administration
Meeting on Pre-market Notification [510(k)] Application
for ReGen Biologics, Inc.’s Collagen Scaffold Device

Thank you for the opportunity to address the committee today.  My name is Jonas Hines and I am a researcher with the Health Research Group at Public Citizen.  I have no conflicts of interest to disclose.

I would like to start out by addressing the proposed regulatory pathway for the device, a collagen scaffold (CS) that is secured with sutures into the space left by the damaged cartilage (meniscus) that has been removed from the knee during arthroscopic surgery.

As you know, the two basic paths to approval for a medical device are the Pre-Market Application (PMA), analogous to the New Drug Application (NDA) for drugs, and the pre-market notification process. The pre-market notification, or 510(k), process is used to gain Food and Drug Administration (FDA) clearance to market a medical device that has been determined to be “substantially equivalent” to another, previously cleared predicate device.  FDA guidances make clear that “devices which do not have the same intended use cannot be substantially equivalent.”[1]

In determining substantial equivalence for a new device, the FDA’s guidances state that the agency’s “scientific expertise enables it to exercise considerable discretion in construing intended uses.”[1] Different indications are permitted, as long as the modifications do not present questions of safety and effectiveness different from those posed by the use of the predicate device.[1]

ReGen Biologics, Inc. is seeking clearance under 510(k) for its CS device “for use in surgical procedures for the reinforcement and repair of chronic soft tissue injuries of the meniscus (one to three prior surgeries to the involved meniscus) where weakness exists”[2] by establishing substantial equivalence to 23 other surgical meshes.  However, according to the FDA, the agency “has not previously cleared a surgical mesh device for this specific indication.”[2]

The sponsor provides only laboratory data, not clinical data, comparing CS to the proposed predicate devices and these data are, themselves, unconvincing.  First, the FDA points out that, even though the suture retention strength of CS is comparable to the other suggested predicate devices, none of the predicate devices is intended “to replace or repair damaged soft-tissue in a weight-bearing, articulating joint”[3] and none is a substitute for cartilage. Furthermore, none is intended to dissipate or transfer loads in a weight-bearing joint.[4] The mechanical requirements for a knee mesh implant are thus likely to be more extensive than those for the predicate devices.

In order to ascertain whether the CS could “withstand the functional demands placed upon it over a multi-year period of time prior to complete resorption,”[3]  the FDA asked the company to compare the tensile and suture pull-out strength of CS to that of a human meniscus. The company instead compared its product to a dog meniscus; as it happens, the suture pull-out strength of CS is notably weaker than dog meniscus.[5]

The inappropriateness of considering clearance through 510(k) is illustrated through a simple thought experiment. Imagine a clinical trial comparing the CS to one of its predicate devices, e.g., comparing a mesh implant in the knee with a mesh implant surrounding the shoulder.  Because the implant sites are so different, no reasonable conclusions could be drawn from such a study. This brings us back to the fundamental problem: the CS is for a different intended use than any of its proposed predicate devices and thus is not suitable for clearance through 510(k). 

A more appropriate path to market would be through the PMA process. Indeed, the sponsor has conducted a study that would be ideally suited to a PMA. The only problem is that the study fails to demonstrate either the effectiveness or the safety of the device.

That study was a randomized, controlled trial comparing partial meniscectomy (meniscus removal) plus CS with partial meniscectomy alone. In its analyses seeking to establish clinical benefit, the sponsor would have the FDA and this committee ignore the control group by having the patients act as their own controls. Its submission states that “patients with chronic meniscus injuries who received the CS device showed improved clinical outcomes as demonstrated by statistically significant improvement from their pre-operative status in pain, function, self-assessment, satisfaction and activity level at a mean of 4.9 years.”[6]

However, comparing changes in the CS-treated patients to the control, there were no differences between groups for any of the primary clinical outcomes. 

Confining ourselves for the remainder of this testimony to the patients with “chronic” injuries (one to three prior meniscal surgeries), the group for which the sponsor is seeking approval,[7] changes for all three of the primary clinical endpoints (visual analog scale pain score, Lysholm knee score, and patient’s self-assessment) were not significantly different between the treated and control groups.  Thus, it can be reasonably concluded that CS offers no clinical benefit to patients with damaged menisci. 

The sponsor also identified other primary surrogate endpoints (arthroscopic, histologic, and radiographic improvement), but these suffered from major methodological problems. In particular, there was no control group for the arthroscopic or histologic outcomes, investigators were not blinded, and the radiographic outcomes were criticized even by the sponsor.  Any benefits claimed for the surrogate endpoints are thus extremely suspect. Regardless, these outcomes are less important than the clinical endpoints. 

Furthermore, the two positive findings reported by the sponsor have been refuted by the FDA’s analysis of the sponsor’s data. 

First, the sponsor reports a statistically significant difference in the Tegner index for patients in the CS-treatment arm even though, as the FDA notes, it “was not a pre-specified primary or secondary effectiveness endpoint.”[8] The sponsor acknowledges the “clinical significance of the Tegner index has not been reported in the literature.”[9] In contrast, the Tegner activity score was a prespecified secondary endpoint, but an analysis of this endpoint was not provided in the FDA’s material or the sponsor’s published article.[10] Furthermore, the FDA states that “the Tegner activity scale was designed as a score of activity level to complement other functional scores (e.g., the Lysholm knee score).”[8] But there was no difference in Lysholm scores between CS-implanted patients and controls.

Second, the sponsor claims that the reoperation rate was 9.5% for CS-implanted patients compared to 22.7% among controls.[11] By employing a more conservative definition for reoperation, including procedures conducted during the mandatory second-look arthroscopy for CS-implanted patients, the agency data indicate a reoperation rate of 21.4% and 16.6% in the CS-implanted and control groups, respectively.[10],[12]

In summary, we are left with a device that, based on a randomized, controlled trial, is no better than the standard of care (partial meniscectomy alone). The trial failed to achieve statistical significance for its primary clinical endpoint and it appears this is also true for the primary surrogate endpoint. The two benefits the sponsor claims have been convincingly refuted by the FDA. 

Furthermore, the sponsor reports 37 serious adverse events (43%) in the CS-implanted group compared to 23 in the control group (33%). Moreover, there were 14 (16%) serious adverse events ascribed to the CS implant compared to two (3%) among controls.

While we reject the use of the 510(k) clearance process in this instance, we believe that the device fails to meet the device approval standard (“reasonable assurance that the device is safe and effective”[13]) regardless of which path to marketing is followed. When a device has been shown to add nothing to conventional therapy, it is hard to see how the public health (and coffers) are served by its approval.

---