Medical Device Daily Washington Editor

GAITHERSBURG, Maryland — Stryker (Kalamazoo, Michigan) has had its share of regulatory headaches, including a December class I recall of cranial implants and a warning letter issued last year for distribution of the company's OP-1 Implant, a bone morphogenic protein product without study site approval. The company had to deal with more bad news Wednesday in the form of a decision by an FDA advisory committee to recommend that the agency not approve the PMA for its other bone morphogenic protein, the OP-1 Putty.

Stryker picked up a license for the OP-1 line of products from Curis (Cambridge, Massachusetts) in December 2007 (Medical Device Daily, Jan. 2, 2008). The agreement between Stryker and Curis entailed an initial payment of $1 million and as much as $41 million in royalty payments should the OP-1 products hit unspecified sales figures. Both OP-1 Putty and the OP-1 implant consist of the bone morphogenic protein (BMP) variant known as BMP-7 and a collagen derived from bovine sources. OP-1 Putty differentiates partly due to the addition of carboxymethylcellulose to the mix.

According to company documents, the putty product has been used on more than 15,000 patients in the U.S. and another 25,000 in other nations with "no specific safety trends observed." The pivotal trial compared OP-1 Putty to the use of bone graft harvested from the crest of the hipbone to fuse lumbar vertebrae for patients with degenerative spondylolisthesis who have not responded to six months of conservative therapy. The idea behind the product is to eliminate the need to harvest bone from the hip, a procedure described as painful and requiring longer procedures and more hemorrhage. The original plan was to follow the patients for 24 months using conventional X-ray scans to evaluate whether the vertebrae have fused.

The first analysis of the completed trial data showed that the OP-1 performed better than bone graft in terms of neurological success and the need for re-operation. However, patients in the control arm (bone graft) scored better on the presence of new bone by X-ray and on scores of the Owestry disability index, a now-commonly used index of patient mobility. Thus the trial failed to demonstrate non-inferiority of the OP-1 to graft.

One of the problems encountered in the first analysis of the pivotal trial was that the Putty tended to migrate from the original point of placement, and the location of bone growth was consequently not where clinical investigators anticipated, leading to questions about the amount of regrown bone. This was the impetus behind the use of CT scans to re-evaluate the patients at 36 months. However, the company also expanded the margin for non-inferiority from 10% to 14% on scans at the same time, generating a new set of conditions for FDA to consider and the panel to consider. The number of fatalities in the study arm was 11, compared to five in the control group, but the rate of death was lower for the study subjects due to enrollment of 208 in the study arm and only 87 in the control arm.

Eugene Poggio, PhD, a consulting biostatistician, noted the original protocol's requirement of a non-inferiority margin of 10% and that the first analysis of the trial showed an overall success rate at 24 months of 38.7% for OP-1, and 49.4% for autograft. He said, "our intent in designing the extension study was to use the same primary endpoint," and get all the original patients back in. "Approximately 80% of eligibles participated in the extension study," he said, adding that the extension analysis led the firm to conclude that "OP-1 is at worst 11.6% worse than autograft and at best, 12% better than autograft."

Poggio also argued that the sensitivity analysis indicated that "the two treatments are virtually identical in overall success rates," and that "the two treatments are very similar, regardless of how you handle the missing data," a reference to the fact that not all the original patients participated in the extension study.

Jianxiong Chu, PhD, a biostatistician with the agency, told the panel, "the sponsor's proposal to allow a larger non-inferiority margin [in the extension study] is not justified from my statistical point of view." Chu remarked that FDA has "issues with such a post-hoc analysis," adding that "the sponsor's [modified intent-to-treat] analysis ... still fails to support the non-inferiority claim."

Another issue that cropped up was whether the device would spur immunological reactions that would create adverse events, a concern that seemingly was unmollified by the extensive data behind the device. This concern revolved around the possibility that the sterilization of OP-1 with irradiation would carry a toxic effect.

Also presenting on the company's behalf, David Wong, a former president of the North American Spine Society (Burr Ridge, Illinois), said the use of conventional X-rays is "an unreliable way to determine whether there is a solid bridge" between the vertebrae. Another spokesman for the company said that the data demonstrated "no difference in terms of biomechanical stability."

When it came time for the vote, panelist Brent Blumenstein, PhD, of Trial Architecture Consulting (Washington) moved for non-approvability, stating "I am unconvinced that the data provide sufficient evidence of efficacy." Janine Mason, MD, of Jason and Jarvis Associates (Hilton Head, South Carolina) commented, "I think that the data is really not quite adequate" regarding immunological safety and opined that a new study would be needed to establish this.

Raj Rao, MD, of the Medical College of Wisconsin (Milwaukee), remarked that he voted for disapprovability because of a "lack of clear radiographic superiority." He said he had "no major concerns with regard to safety issues, but some concerns about ... fetal maldevelopment down the road."

Panelist John Kirkpatrick, MD, of the University of Florida College of Medicine (Jacksonville), commented that he was "concerned about the post-hoc analysis that had to be done to yield a positive result" and whether "fusion happens, even when there's bone there." Surgeons typically abrade the vertebrae in the fusion site to trigger bone growth, and some panelists indicated that they felt this could account for some of the effect seen in the study arm patients. Kirkpatrick also said "I still have the concern over the very rare instance" of a potentially "catastrophic event."