P-R-O-C-E-E-D-I-N-G-S

(9:41 a.m.)

MR. DEMIAN: Good morning, everyone. I would first like to welcome you to this meeting. We're ready to begin this meeting of the Orthopedic and Rehabilitation Device Advisory Committee. My name is Hany Demian, and I'm the Executive Secretary of this Committee.

I'd first like to remind everyone that you're requested to sign in on the attendance sheets which are available outside the doors. You may also pick up an agenda and information about today's meeting, including how to find out about future meeting dates through the advisory panel phone line and how to obtain meeting minutes or transcripts

I will now read two statements that are required to be read into the record. The first one is the appointment to temporary voting member status and the conflict of interest statement.

"Appointment to Temporary Voting Status; pursuant to the authority granted under the Medical Device Advisory Committee Charter, dated October 27th, 1990 and as amended August 18th, 1999, I appoint the following individuals as voting members of the Orthopedic and Rehabilitation Device Panel for this meeting on January 10th, 2002; Kinley Larntz, Sanjiv Naidu, Leon Lenchik, Gene Siegal, John Kirkpatrick, Barbara Boyan, John Doull, Betty Diamond, and Hari Reddi. For the record, these individuals are special government employees and consultants to this panel or other panels under the Medical Device Advisory Committee.

They have undergone the customary conflict of interest review and have reviewed the material to be considered at this meeting. In addition I appoint Dr. Maureen Finnegan to serve as acting Chairperson for the duration of this meeting", and this is signed by David Feigal, Director of CDRH.

"Conflict of interest statement; The following announcement addresses conflict of interest issues associated with this meeting and is made part of the record to preclude even the appearance of any impropriety. To determine if any conflict existed the agency reviewed the submitted agenda for this meeting and all financial interests reported by the committee's participants. The Conflict of Interest Statute prohibits special government employees from participating in matters that could effect their or their employer's financial interests.

However, the agency has determined that the participation of certain members and consultants, the needs for whose services outweigh the potential conflict of interest involved is in the best interests of the government. Therefore, waivers have been granted for Doctors Stephen Li, Kinley Larntz, Edward Hanley and John Kirkpatrick for their interest in firms that could potentially be effected by the panel's recommendations.

The waivers permit them to participate in all matters before today's panel. Copies of these waivers may be obtained from the agency's Freedom of Information Office, Room 12A-15 of the Parklawn Building. We would like to note for the record that the agency also took into consideration other matters regarding Doctors Li, Larntz, Maureen Finnegan, Barbara Boyan and Gene Siegal.

Each of these panelists reported current or past interests in firms at issue but are not related to today's agenda. The agency has determined, therefore, that they may participate fully in all deliberations. Dr. Hanley has a past involvement with matters that are related to today's agenda. The agency has determined, however, that he may participate in the panel discussions.

We would like to also note that Doctors Rocky Tuan and John Kostuik are guests at this meeting and have reported interests in the firms at issue. In the event that the discussions involve any other product or firms not already on today's agenda, for which an FDA participant has a financial interest, the participant should excuse him or herself from such involvement and the exclusion will be noted for the record.

With respect to all other participants, we ask in the fairness -- in the interest of fairness that all persons making statements and presentations disclose any current or previous financial involvement with any firms whose products they may wish to comment upon".

Before turning this meeting over to Dr. Finnegan, I would like to introduce our distinguished panel members for generously giving their time and effort to help FDA in matters being discussed at today's meeting and other FDA staff seated at this table. So we'll go around the room and give your name and affiliation and your current areas of research. Dr. Finnegan?

CHAIRPERSON FINNEGAN: Maureen Finnegan, I'm an orthopedic surgeon at Southwestern Dallas and I do -- my research is mainly fracture repair.

DR. KIRKPATRICK: I'm John Kirkpatrick. I'm an orthopedic surgeon and spine surgeon from the University of Alabama at Birmingham.

DR. SIEGAL: I'm Gene Siegal, also from the University of Alabama at Birmingham and I'm an anatomic pathologist.

DR. HANLEY: Edward Hanley, orthopedic spine surgeon, Carolinas Medical Center, Charlotte, North Carolina.

DR. DIAMOND: Betty Diamond, Albert Einstein College of Medicine. I'm an immunologist and rheumatologist. I'm on sabbatical at NIH.

DR. DOULL: I'm John Doull. I'm a clinical toxicologist from the University of Kansas Medical School.

DR. LI: I'm Stephen Li. I'm interested in biomechanics and biomaterials. I'm current president of Medica Device Testing Innovations located in Florida.

DR. WITTEN: Celia Witten. I'm the Division Director of the Division of General Restorative and Neurological Devices which is the reviewing division for this product for FDA.

MS. MAHER: Sally Maher. I'm with Smith and Nephew Endoscopy and I'm the industry representative.

MS. RUE: I'm Karen Rue. I'm an R.N. I'm consumer representative.

DR. LARNTZ: Kinley Larntz, Professor Emeritus, Statistics, University of Minnesota and I'm a statistician interested in clinical trials.

DR. LENCHIK: Leon Lenchik, Musculoskeletal Radiologist from Wake Forest University in Winston-Salem, North Carolina.

DR. REDDI: I'm Hari Reddi. I'm a student of bone morphogenetic proteins.

DR. BOYAN: Barbara Boyan. I'm a professor at the University of Texas Health Science Center at San Antonio and my specialty is bone and cartilage cell biology.

DR. NAIDU: Sanjiv Naidu. I'm an orthopedic surgeon at Penn State College of Medicine in Hershey and my interest is in biomechanics and orthopedic surgery.

MR. DEMIAN: In addition, I'd like to introduce our three guests who are seated over here, Dr. Richard Miller, Rocky Tuan and John Kostuik.

CHAIRPERSON FINNEGAN: Thank you, Hany. As I previously stated, I'm Maureen Finnegan and I will be the chair for this meeting. Today the panel will be making recommendations to the Food and Drug Administration regarding a pre-market approval application for a spinal fusion cage with a growth factor soak in a collagen sponge use to treat lumbar degenerative disc disease.

I need to note for the record that the voting members present constitute a quorum as required by 21 CFR Part 14 and we will now proceed with the open public hearing session of this meeting I would like to ask at this time that all persons addressing the panel come forward and speak clearly into the microphone. The transcriptionist is dependent on this as a means of providing an accurate record of this meeting.

We would request that all persons making statements during the open public hearing of the meeting disclose whether they have financial interests in any medical device company. Before making your presentation, please state your name, affiliation and the nature of your financial interest if you have any. There's obviously someone who wishes to address the panel.

MS. TRISLER: Good morning, my name is Patsy Trisler and I'm a regulatory consultant at PharmaNet, Incorporated, a contract research organization. As an employee of PharmaNet, I have several clients who are orthopedic product manufacturers but I have no financial interest in any of them.

CHAIRPERSON FINNEGAN: Ms Trisler, you had made a request to make an oral presentation.

MS. TRISLER: Yes.

CHAIRPERSON FINNEGAN: What we would like to do is ask those people who had not made such a request, we do have two -- a request for oral presentations which we have put into the program and that's the next part of the program, so we would ask those people who had not made such a submission if they would like to make a presentation.

MS. TRISLER: All right, I apologize.

CHAIRPERSON FINNEGAN: So if you'd give us one second. Is there anyone else other than the two parties who had made a formal request to make an oral presentation?

(No response)

CHAIRPERSON FINNEGAN: All right, if not, then we have had two requests. One is from Osteotech and one is from Striker Biotech and we will start with Ms. Trisler. Go ahead.

MS. TRISLER: Thank you and I apologize for jumping the gun. As I indicated, I'm an employee of PharmaNet, Incorporated which is a CRO. I would like to thank the Chairperson and the FDA Executive Secretary for providing the opportunity to speak to you today.

The purpose of my brief presentation is to express some concerns in the form of potentially unanswered questions relating to the combination products of the type under review today by this committee. As you know, there have been several spinal fusion cages or systems approved by the FDA over the last five years. These products approved for treating degenerative disc disease are to be used with autogenous bone grafts.

Papers are being published reporting the successes observed with the use of the cages. It is clear also that there remains some problems or issues such as subsidence. The focus of my comments, though, is not on the cages but rather on the biologic component of the device, the bone morphogenetic protein or BMP.

As I'm sure you are aware even though BMPs have been under evaluation clinically now for about 15 years, only one has been allowed into the marketplace by the FDA. The approval is a limited one in the form of a humanitarian device exemption for treating long bone, non-unions when alternative treatments have failed. That product is human recombinant BMP-7 and bovine bone derived collagen.

This PMA before you today represents an important advance in medical device technology. BMPs and other growth factors are potent compounds that offer significant promise in many therapeutic areas. Further, the potential of BMPs or other growth factors combined with traditional medical devices is significant. However, before the first combined product of this type achieves market approval, it's very important to be certain all the appropriate questions have been addressed, have been both raised and addressed and as I indicated, this is the reason I'm speaking.

I do not know the full extent of Medtronic Sofamor Danek safety and effectiveness data. However, I do know this is a reputable company and my comments are by no means meant to challenge the capabilities, integrity of the data or quality of the studies performed by them. While we recognize that Center for Biologic staff participates in reviews of products of this type along with the Devices Center, the issues posed by the biologic component are quite different from the issues typically presented to this committee.

The standard of proof is different for drugs and biologics than for devices. Thus, the guidances provided by the FDA in those areas are different. The issues that prompt the following questions, in fact, are not covered in the devices guidance document for spinal systems. As a member of the public, I ask that you consider these questions in your deliberations.

My first point is related to cancer promotion. Cancer promotion by cytokine growth factors is well known, particularly when circulating blood levels are greater than normal or baseline as in the case of recombinant BMP. My question is, is enough known about the cancer promoting capability of BMPs. I question the status if this growth factor, BMP-2 as a cancer promoting compound.

In my review this morning of the panel briefing materials, my very quick review, I was surprised that FDA has agreed that certain non-clinical safety studies may be conducted post-approval. If transformed cells or other adverse events are seen after this implant has been released to the market, what is the surgeon to tell the patient?

I noted in the BMP-7 product that is approved, that patients with a cancer history are contra-indicated for the current -- for that product approval. Will it be necessary to similarly contra-indicate this BMP that is before you today or is the risk to benefit fully profiled?

My second point relates to the immunology area. Circulating antibodies to both Type I collagen and BMPs are reported. I know the FDA has dealt with this matter in the collagen area for many years. Are enough data available to demonstrate there is no correlation between these antibodies and health or medical events? Is the safety risk greater if a BMP is inappropriately and perhaps inadvertently applied directly to the spinal canal? Has the autoimmune reaction potential been fully evaluated?

My third point relates to cardiovascular. Cardiac adverse events and increased blood pressure and body temperature have been reported in animal studies. The effects reported are dose dependent. Is there sufficient assurance that benefits of the use of the BMP collagen mixture in a sensitive body area are outweighed by the potential risk to the cardiovascular system? The other BMP approved for orthopedic use and this one are provided with collagen as a carrier.

While I realize this may -- this question may not be particularly important, since there are potentially greater risks with the use of a growth factor in the spine than there are at a long bone non-union site, are data available to show that collagen alone is not effective in improving the rate of spinal fusion?

I believe this next point deserves particular attention. As is well-known, the early work of Dr. Urist demonstrated the ability of a bone inductive extract from adult bone to induce new bone formation at ectopic sites. Have the studies published since that time sufficiently looked at the quality of the bone produced and at the risk of uncontrolled growth of bone in the immediate and surrounding region of the implant? I have heard of one case in which bone grew into the spinal canal although I'm unaware of the extent of the problem.

Has a full enough evaluation been performed to be reasonably sure that if a large amount -- if a larger amount than indicated is applied in the spinal fusion area, the risk of in-growth won't occur. There is one final issue relating to a problem that is not limited to orthopedic devices, the expanded or off-label use in the medical community of a product approved for a very limited indication.

While I believe the medical community, not the government, should control the practice of medicine, in this case it seems the risk is significant for off-label use of the BMP component of this device system. Because of this, I feel that it is important for the panel to, perhaps, give more consideration than is usually done to this issue. As I noted earlier, many of the reported adverse events were dose dependent.

Since BMPs are potent compounds that have systemic effects and this particular product is the first of a kind for this use, I am concerned that the pharmacodynamics may not be fully understood. After this product type is out in the marketplace, if it is misused or misapplied, the potential for patient harm is great.

In closing, I ask the panel to give special attention to the potential for off-label use. I realize I have just scratched the surface of a number of areas and have not provided you with substantive information or data and that others, perhaps, will raise some more concerns. I am hoping, though, that none of these topics remain issues after today's review. This concludes my comments. Thank you for this opportunity.

CHAIRPERSON FINNEGAN: Thank you for having the interest. I did interrupt you when you were going through your financial interest. Would you mind reviewing those?

MS. TRISLER: Yes. No, I would not mind. I am employed by PharmaNet, which is a contract research organization. As such, I have clients in many areas. I have several clients in the orthopedic product area. I have no financial interest in any of them.

CHAIRPERSON FINNEGAN: So you're a consulting firm.

MS. TRISLER: Yes.

CHAIRPERSON FINNEGAN: Okay. And I believe our next presenter is from Striker Biotech.

DR. McCULLOUGH: My name is John McCullough and I'm an orthopedic spine surgeon from Denver, Colorado. I am not from Striker Biotech. I'm here to offer an opinion regarding today's discussion on the BMP-2 interfix threaded fusion cage PMA and I thank the panel for granting me permission to speak. My travel has been paid for by Striker Biotech. I have no financial interest in the company and I'm not a paid consultant.

I've participated in the Striker Biotech pilot study using BMP-7 in the human lumbar intertransverse interval for which my institution received research funds to cover the cost of the study. Contrary to questions by my colleagues prior to this meeting, I am here in a positive relationship with BMP, not a negative relationship.

Studies of BMP-2 and BMP-7 have shown great promise as potential osteo inductive replacements for iliac crest autograft for bone healing in appendicular and spinal fusion applications. In working with BMP-7 in the lumbar inter-transfers interval, I am impressed with its effectiveness, but I'm also impressed with the meticulous technique required to increase the likelihood of a solid fusion.

It is a much less forgiving milieu for fusion than the interbody interval mainly because it is a soft tissue bed on which the body never intended bone to form. A solid anterior lumbar interbody fusion is a relatively easy outcome to achieve but it's a technically demanding surgical approach fraught with serious complications. The lumbar intertransverse fusion is just the reverse. It's an easy, safe, posterior surgical technique but obtaining a solid fusion is much more difficult.

As an example, allograft bone will often successfully incorporate in an interbody fusion model but it is useless in the adult intertransverse interval. The success of fusion with BMP-2 in a collagen sponge with the interfixed threaded fusion cage, the subject of today's discussion, is well established by the research of Boden's Zdeblick, Sandu and Hine.

The researchers, the brave patients who submitted themselves to this largely successful pilot study and the company supporting the research are to be congratulated. It is not my purpose today to call into question the efficacy of BMP-2 and its use in an interbody fusion with the interfixed threaded fusion cage. My concern, as with the last speaker's last point, is the potential off-label use of BMP-2 soaked in a collagen sponge.

There is a potential for surgeons to take an off-label approach if the panel does not carefully consider the labeling and packaging considerations available with this product and provide the option that will prove to be the best direction and control of the product and its potential off-label use. One off-label use for this BMP-2 collagen sponge model is the lumbar intertransverse interval. My concern for such usage is found in the research work done by Martin Boden, et al in the posterior lateral intertransverse fusion non-human primate model.

The efficacy of the BMP soaked in a collagen sponge, in this particular intertransverse fusion application was negatively impacted by the soft tissue and muscle compressing the sponge and thereby, compressing the growth factor out of the sponge. This led to an unexpectedly high failed fusion rate. It is easy to conclude that this scenario would conclude in humans with BMP-2 and a collagen sponge carrier. In this setting the potential is for the muscle to compress the collagen sponge and leak the BMP-2 away from where the bone is intended to form.

To induce new bone formation, it is not currently feasible or practical to apply BMP directly into a bone void for the purpose of bone growth. A carrier is needed for a number of reasons. One of the main reasons is for the containment of the growth factor at a site where bone growth is needed. A carrier is also needed for stem cell attachment and to provide a structural matrix for bone growth.

Well, BMP-2 and BMP-7 have been shown to be two of the most effective BMPs in the bone healing cascade, the carriers used by these BMPs and ultimately the orthopedic application site in which they are placed can effect their efficacy regardless of their potency. BMP-2 and the studies being discussed today has been used with a fibular hemostatic collagen sponge carrier placed in the interfix titanium threaded interbody cage.

In this application, liquid BMP-2, a combination of BMP and sterile water, is applied into the collagen sponge inter-operatively and allowed to soak into the sponge. The sponge is then rolled and placed into the cage. In this application it is important to note that the BMP collagen sponge combination is protected by the structure of the titanium cage.

In the intertransverse interval, this same BMP-2 soaked collagen sponge would enjoin no cage protection. Rather as Boden et al suggested, its compression by muscle would possibly lead to extrusion and dissipation of the BMP-2 and a failed bone induction.

Bone morphogenic protein research represents an exciting and new opportunity for surgeons and patients alike that over time may revolutionize the way we treat our patients. Getting rid of the bond graft harvest is an exciting concept and could possibly overrun the relative lack of knowledge amongst my colleagues about this technology.

There may be a temptation to push the envelope when it comes to indications and applications. With this new opportunity, also comes the responsibility and challenge of not only appropriate patient selection but also appropriate product labeling.

My strong assertion and belief is that we need to make every effort possible to insure the health and benefit of our patients through appropriate labeling in regards to this PMA. The BMP-2 collagen sponge device has been tested in an IDE study for specific spine pathology with a specific type of branded cage in the interbody interval.

If approved, I hope the indication for use will be for this particular combination of BMP carrier and cage product. Since this product was tested as a combination product cage with BMP, the requirement that they be packaged together the way they are intended to be used is reasonable and logical. This provides an additional and important opportunity to further insure that the use of the growth factor will be used in an application where efficacy has been proven.

Off-label use in areas such as the posterolateral intertransverse fusion will not be eliminated as an option to my colleagues, but packaging the BMP and cage together will limit the product's use in an unproven and potentially flawed application.

In concluding, I defer to commenting on the final approvability of this product. I am an enthusiastic supporter of the BMPs but in the event it is voted to be approved, I would recommend that the combination of the interfixed cage and BMP-2 be specifically required to be ordered together and packaged together to insure the product is used as has been tested. I would also recommend that off-label use of BMP-2 and the collagen sponge in areas outside of the application such as the intertransverse interval be addressed in the product labeling.

Thank you.

CHAIRPERSON FINNEGAN: Thank you. Dr. Witten.

DR. WITTEN: We also need to ask the prior speaker who paid her way, whether she paid her way or whether her way was paid for.

CHAIRPERSON FINNEGAN: Okay, I'll have her come back. Dr. McCullough, thank you very much. Ms. Trisler, is she still with us? While she's coming up, are there any other persons who would wishy to make a comment? Go ahead.

MS. TRISLER: I'm sorry, what was the question?

CHAIRPERSON FINNEGAN: Who paid your way to the meeting?

MS. TRISLER: As a consultant Osteotech has.

CHAIRPERSON FINNEGAN: Has paid your way to the meeting?

MS. TRISLER: Well, yeah, I live here but they paid my time.

CHAIRPERSON FINNEGAN: Thank you. All right, if there are no other people wishing to make comments, Mr. Demian has received eight letters regarding this meeting and he will now read them into the record.

MR. DEMIAN: I've receive eight letters and seven of them are from spinal surgeons, all letters regarding the use of BMP. The first letter is from Dr. Regis Haid.

"I am currently the chief spine surgeon for the Department of Neurosurgery at Emory University in Atlanta, Georgia. I have no vested financial interest in the product being discussed before the panel. I've developed products for cervical spine for various companies, including Medtronic, Codman and Spinal Concepts. These may be considered by some to constitute an indirect conflict of interest. I've received no renumerations for my interest in BMP.

Our group has been involved in the use of BMP. We have been given presentation on fusion techniques at national and international meetings and have briefly discussed the experimental use of BMP. We have actually published a paper in the Neuroscience Focus on the use of BMP. From my knowledge of the studies and presentations I've heard presented by other spinal surgeons, I do believe that BMP offers a significant advantage in the practice of spine.

It is very clear from my experience that the literature in neurosurgery and orthopedics state that autograft sites do present a well array of complications. It is also commonly known that harvesting the autograft iliac crest adds time to surgery and expense in the operating room and pain to the patient is always part of the harvesting autograft iliac crest.

I would ask the panel to recommend to the FDA to expedite their approval of this product. This would prevent further suffering of patients that occur with every autograft bone harvest as well as potentially decreasing the time in the operating room and, thus, potentially decrease the total cost to the patient. Having reviewed the data from the academic perspective, it seems very clear to me its efficacy is clear-cut in the use in lumbar interbody anterior devices and that the product should be made available to the American public.

Although I'm not an expert on the FDA, it is my belief that FDA required a small pilot study for this device under review and this was done before a large pivotal study could begin. If this is indeed the case, I believe this was unnecessary and prompted a delay of the release of this product which definitely benefits patients. I would suggest that the Orthopedic Advisory Panel recommend to FDA not to require these types of pilot studies for similar issues in the future".

The second letter is from David Malone. "I would like to add some information to the pre-market approval application for a spinal fusion cage with growth factors soaked in a collagen sponge intended for treatment of lumbar degenerative disease. I took part as one of the investigators in the posterior lumbar interbody fusion BMP trial sponsored by Medtronic Sofamor Danek Corporation.

Dr. Frank Tomecek was the lead investigator for our small group. There were a number of patients that were treated with the PLIF. Two of the patients had significant posterior bony over-growth impinging on their nerve roots requiring additional surgery. One patient, who was my patient, required two surgeries to clear excessive bone growth from his spinal canal. He has had no new bone growth over the past year. I am unsure as to whether or not this data has been included in the application to the FDA.

I've been told that the posterior lumbar in a body fusion cage trial was halted. I assume it was because of this bony overgrowth problem. With regard to the patients with bony overgrowth, I personally experience -- my personal experience in re-operation on both of these patients, the bone quality from the BMP is robust and excellent. The fusions are solid. I do feel that the BMP is a useful adjunct to bony spinal fusion.

However, BMP may lead to excessive bone growth and may cause significant neural impingement if placed in posterior lumbar interbody type of device. There does need to be at this point in time some type of barrier between the area where the bone can overgrow and the neural elements. I note that Dr. Frank Tomecek and Sofamor Danek did further experimental studies on the PLIF model but I do not have the data. I know the data does exist and may be helpful if you are considering approval of this material for a posterior lumbar interbody fusion type of approach.

If BMP is approved for spinal fusion, and I feel that it would be useful adjunct, the caveat is that it must be placed in such a manner that bony overgrowth cannot grow into the spinal canal as I think this would cause significant problems for a proportion of the patients whom it is used in".

The next letter is from Dr. Robert Banco. "As chief of the spine section of the New England Baptist Hospital, my colleagues and I are pleased to have participated in the rhBMP-2/ACS/LT open clinical trial. Serving as the principal investigator, two co-investigators and myself are members of the Boston Spine Group, four orthopedic surgeons and one physiatrist with a practice dedicated solely to spine.

As a group we perform over 400 spinal fusions annually many of which are accompanied by iliac crest harvesting. As you know, harvesting patients with iliac crest increases the risk of complications, including but not limited to infection, nerve damage and possible damage to the muscles and vessels. Donor site pain is by far the most common complication and patient complaint.

BMP-2 supplants the need for harvesting the iliac crest and therefore, negates the risk of these complications. We at the Boston Spine Group have heard many presentations and have read the literature regarding BMP-2. We are excited by the reported outcomes. We are anxious for this product to get out of the lab and into the clinical practice and are looking forward to the use of InFUSE^TM in the clinical setting".

The next letter is from Dr. Paul McCormick. "I'm a full time faculty member at Columbia University of Physicians and Surgeons. My practice is exclusively limited to the evaluation and surgical management of patients with spinal disorders. By the way of disclosure, I have no financial or other vested interest in the products that are being discussed before the panel.

As a full time spine surgeon at a major academic center, I'm well aware of active research that has been conducted for years regarding biological enhancement of spinal fusion. Like many other spine surgeons, I look forward with great anticipation when effective agents will be commercially available for the utilization in spinal fusion. Spinal fusion is an important technique for many patients who have lost their mechanical integrity of their spinal elements through trauma, degenerative changes, neoplasm, or disc herniations in prior surgery.

A major problem related to the spinal fusion is the harvesting of the autograft which is usually required for a vast majority of spinal fusions currently performed. The pain and morbidity associated with autograft harvest can be considerable. Often this pain persists over time and may be permanent.

Further, despite significant advances in fusion techniques and spinal instrumentations, a measurable number of patients continue to suffer from failed fusion or pseudoarthrosis. Therefore, any useful adjunct that can be utilized to facilitate and enhance spinal fusion would be of tremendous benefit to patients with spinal disorders requiring this type of surgery. In essence, there's a tremendous need for biological fusion enhancers such as BMP that diminish the reliance on autograft harvesting as well as enhancing the rate and the success of spinal fusion.

I'm also well aware of the research that is currently being conducted at numerous centers regarding BMP. To my critical review, BMP has shown exciting promise in enhancing spinal fusion and bone incorporation. I fully appreciate the responsibilities of the FDA in general and of your panel in particular in acting in the public interest through oversight on the approval and introduction of these devices and agents. I would respectfully request that such evaluation be carried on in an expedited fashion so that if BMP satisfies the FDA requirements for approval, we can utilize this substance in a timely manner.

Such an expedited approval would likely reduce the pain and suffering of future patients that are requiring spinal fusion".

The next letter is from J.J. Abitol. "I'm a practicing spinal surgeon, also a current board member of the North American Spine Society where I have been a past scientific program chairman. Although there is no current official position statement from the Society, I would like to express my opinion about bone morphogenetic proteins or BMPs. Being familiar with the research in this area, I can say with certainty that BMP has been one of the most heavily researched subject matters in all of orthopedics.

Since the late Dr. Marshall Urist first discovered these proteins over 30 years ago, and unprecedented amount of publications and research efforts have been dedicated to studying these proteins. For all practical purposes, all of these studies have demonstrated to the research and medical community that safe and new alternative to taking autograft is now at hand.

I strongly urge this panel to approve these desperately needed proteins and make Dr. Urist's dream of having bone graft in a bottle a reality. It is time to take these type of proteins out of research and make them available to surgeons to use in our clinical practice to treat patients".

Our next letter is from Dr. John Peloza. "I'm a nationally recognized spine expert with a tertiary specialty practice in Dallas, Texas. In my practice I perform many spinal fusion procedures on all levels of the spine from the skull to the sacrum. These fusions are done from an anterior, posterior and sometimes combined approach.

I'm often challenged by difficult spinal reconstruction problems secondary to disease processes including spinal deformity, degeneration, trauma, tumor and infection. My team and I have been and are presently involved in multi-center studies evaluating spinal surgical implants minimally invasive and non-surgical technologies as well as biological technologies for the treatment of spinal disorders.

I'm an authority on bone morphogenetic protein from my experience as a clinical investigator with rhBMP-2, professional presentations, knowledge of the scientific literature, national and specialty meetings and think tanks. I have direct experience with the impressive clinical results on my own patients utilizing this protein.

Presently we have a number of bone graft alternatives. The gold standard is the patient's own bone or autograft. It is osteogenic, contains viable bone cells at transplantation, osteo inductive, actively promotes or enhances bone formation and osteo conductive, acts as a structural framework or scaffold for bone formation. Unfortunately, it is in limited supply, e.g. the patient's iliac crest.

In many cases we have very little or no autograft at all. Additionally, the bone graft harvest surgery contributes significantly to post-operative pain that can be permanent and can lead to other complications. When autograft is not available or inadequate, surgeons use allograft, bone bank or cadaver bone. Allograft bone is mainly osteo conductive, weakly osteo inductive and has no osteogenetic properties.

Depending on the surgical construct, allograft fusion rates are lower than autograft and take much longer to heal. Due to the massive demand for bone graft worldwide, allograft bone is also in limited supply. Additionally, allograft bone has a risk of disease transmission. Modern bone processing is effective in eradication of bacteria and viruses. However, prions are very difficult to detect and no processing has been validated for their removal.

After autograft and allograft, surgeons can use bone graft extenders, demineralized bone matrix. These products are mainly osteo conductive, poorly osteo inductive if at all, and not osteogenic. They are the last line of bone graft material and informed surgeons have little confidence in their efficacy in obtaining a solid fusion.

Recombinant human bone morphogenetic protein is an attractive infusion surgery for many reasons. The fusion rates in animal models and in human trials is the same or better than the gold standard autograft. With a production facility there would be an unlimited supply of rhBMP. RhBMP will eliminate the need for bone graft harvesting surgery which will eliminate the associated pain, potential complications and cost.

There will be no chance of disease transmission. The major cost of sponsored surgery is when the surgery fails. This can occur secondary to a major complication such as infection but the most common reason for failed surgery is the failed fusion or pseudoarthrosis. This is a problem that vexes all spinal surgeons.

A product that markedly enhances our ability to heal bone with few documented side effects is an extremely powerful tool in the treatment of spinal disorders. RhBMP is a breakthrough product that represents the best of our research and advances technologies. It has been thoroughly tested in animal models and human trials. It has consistently proven better and safer than our present alternatives. It is time to get rhBMP into the clinical arena where it is desperately needed for the optimum care of people.

Finally, I would like to state I have no financial interest in the product rhBMP-2 nor do I have a financial interest in the company that is sponsoring rhBMP-2".

Second to the last letter from a trio, Stephen Papadopoulos, Curtis Dickman and Volker Sonntag. "We practice primarily spine surgery at the Barrow Neurological Institute in Phoenix, Arizona. Our practice consists of regional and national and international referrals. We have no vested financial interest in the specific product being discussed before the panel.

We have been aware of the field of BMP research for several years through peer review, publications and scientific presentations. We strongly believe the clinical availability of this product in the United States will significantly enhance patient care. Graft site complications from autograft harvest are well described and documented. The availability, quality and healing issues related to allograft is also well known.

Fusion failure may result in chronic pain, deformity and the need for additional spinal reconstructive procedures. We believe that the approval of BMP will provide a significant advance in the patient outcome and satisfaction".

The last letter is from Dr. Doug Morrow. "It is my understanding that you are about to discuss and vote on approval or rejection of rhBMP. I want my voice to be one that you may not otherwise get in the sense that I am both a physician and a patient waiting on the approval of this enzyme to fuse my lumbar spine. I have rather an unusual set of circumstances wherein I got an infected disc in my lower back because of an injection called a discogram.

The infection all but destroyed two lumbar vertebrae, leaving me in constant pain for the instability associated with the deformity. I've been keeping up with all the literature on the subject and especially this enzyme which speeds up the natural healing process of growing bone. I'm a perfect candidate for this material to be used in surgery on my back. I've been waiting for its use for some time delaying my surgery because of it.

I have back pain every day all day. I urge you prompt approval of this material so that my doctor can then use it on me as soon as possible. There are many people just like me who need help. Please help us. I beg you and thank you from the bottom of my heart".

That's it.

CHAIRPERSON FINNEGAN: Thank you, Mr. Demian. You may get an award for that. We will now proceed to the presentation of the pre-market approval application P000058, Medtronic Sofamor Danek InFUSE^TM bone graft/LT-cage lumbar tapered fusion device. I need to remind the public observers at this meeting that while this portion of the meeting is open to public observation, public attendees may not participate except at the specific request of the panel.

We will proceed first with the sponsor's presentation followed by the FDA presentation. I would like to ask each speaker to state his or her name, their affiliation and I would ask everyone to please speak into the microphone so that people in the back of the room can hear you but also most importantly so that transcriptionist can hear you. The sponsors, if they would like to come up, could start.

DR. LIPSCOMB: Members of the Orthopedic and Rehabilitation Devices Advisory Panel, my name is Bailey Lipscomb and I'm the Vice President of Clinical Affairs at Medtronic Sofamor Danek in Memphis, Tennessee. We have the pleasure and the long awaited privilege to present to you the results of decades of research, development, and clinical studies. At the outset, we would like to thank literally thousands of people who have worked over the years to make these presentations possible. For the next 90 minutes we will present for the first time to an FDA advisory panel the culmination of work arising from a discovery made by Dr. Marshall Urist in 1965.

Dr. Urist found that certain proteins which he later terms as bone morphogenetic proteins, stimulate the formation of bone and these proteins can literally make bone where bone did not exist before. In the early 1980's researchers of Wyeth-Genetics Institute in Cambridge, Massachusetts developed a method to synthesize several of these bone morphogenetic proteins using recombinant methods.

The BMP-2 yields from these methods are much greater in quantity and much purer in nature than can be obtained from natural sources. The bone morphogenetic protein that will be reviewed today is recombinant human bone morphogenetic protein 2 or more commonly known in its abbreviated form as rhBMP-2 and this is made by Wyeth-Genetics Institute.

The rhBMP-2 is supplied as a sterile freeze-dried powder that is reconstituted at the time of surgery with sterile water to a concentration of 1.5 milligrams per milliliter. The solution is then applied to absorbable collagen sponge. The sponge provides the matrix to retain the rhBMP-2 in the desired location sufficiently long to stimulate the formation of bone cells.

The absorbable collagen sponge is a commercially available product that is made by Integra Life Sciences of Plainsboro, New Jersey. FDA approved the PMA application for the absorbable collagen sponge back in 1981. Medtronic Sofamor Danek has named the combination of rhBMP-2 with the absorbable collagen sponge as InFUSE^TM bone graft. This PMA application for InFUSE^TM covers this use with Medtronic Sofamor Danek's LT-cage lumbar tapered fusion device, not the interfix device that Dr. McCullough mentioned but the LT-cage device.

The LT-cage device is a hollow fenestrated titanium alloy threaded interbody fusion device. FDA approved the PMA application for this device over a year ago. Typically two cages are inserted in parallel from an anterior surgical approach in lumbar spinal fusion procedures. In current medical practice bone graft is harvested from the iliac crest, is packed into the LT-cage devices. This is stipulated in the labeling of the device.

Autogenous bone graft augments the fusion of the treated segment and it is now considered a standard of care graft material. Today, however, we're seeking an approval recommendation from this panel to use infused bone graft instead of autogenous bone graft to pack the central cavities of the LT-cage devices. Let's focus more closely on what is at issue here today.

It is not the LT-cage device. This product is commercially available for the same medical indication, that is symptomatic degenerative disc disease, and for the same manner of use, anterior antibody lumbar fusion procedures. It is not the absorbable collagen sponge that has been FDA approved as an implantable hemostatic agent. It has a long history of safe and effective use dating back over 20 years. The real issue today is the safety and effectiveness of rhBMP-2 when used with the two previously approved products and whether it is a suitable replacement for autogenous bone graft, the gold standard, in antibody fusion procedures.

We believe the years of basic research and development of this product have yielded considerable evidence to support the safety and effectiveness of the product. Further, the infused bone graft LT-cage device is supported by clinical data arising from a large multi-centered prospective randomized clinical trial, a desirable scientifically valid study design, but one that is rarely used for orthopedic implants in the United States due to its difficulty in execution.

This study embodies an idyllic scientific research. Eliminate as much of the variation as possible except for the variable being studied. In this study that's exactly what occurred. Patients met the same study entrance criteria and received the same interbody fusion cage. The only variable was the 50/50 chance that a patient would receive either infused bond graft or would receive autogenous bone graft in their surgery.

It is our opinion that this study presented today overcomes virtually all of the objections to study trial designs that have been voiced over the years by this orthopedic advisory panel. These clinical data as well as the pre-clinical test results, manufacturing information, and labeling were submitted to FDA as a modular PMA application with the first module being submitted in April of 2000. The PMA application has been under review by FDA since then and presenting this information to this advisory panel is part of the review process.

As typical for these meetings, we plan to present overviews of the relevant information contained in the PMA application. Dr. Gerard Riedel, the senior project director of the rhBMP-2 program at Wyeth-Genetics Institute, will make the first presentation and he will cover the origin and biology of rhBMP-2 and the pre-clinical safety studies. Dr. Riedel will be followed by Dr. Scott Boden, an orthopedic surgeon from Emory University. Dr. Boden will discuss the results of pre-clinical testing of infused bone graft in animal studies as well as the results of the pilot trial involving infused bone graft with the LT-cage device and that study supported the initiation of the larger pivotal trials.

Dr. Hallett Matthews, an orthopedic spine surgeon from Richmond, Virginia, will review the results of the large scale pivotal IDE trial of the infused bone graft with the LT-cage device. Dr. Matthews was an investigator in the open surgical approach study. I will then return to the podium for concluding remarks.

In addition to these speakers, we have assembled here today a group of physicians and scientists who should be able to answer the questions you may have about the product under review. These experts include several clinical investigators, the inventor of the cage, radiologists and immunologists, an OB/gyn physician, a histologist, a statistician, basic scientist and the discoverer of the rhBMP-2 that has been used in the study.

So without further ado, I will now turn the podium over to Dr. Riedel.

DR. RIEDEL: Thank you, Bailey. Good morning. My name is Gerard Riedel. I'm employed by Wyeth-Genetics Institute, a pharmaceutical company, that collaborates with Medtronic Sofamor Danek in the development of BMPs in spine surgery. In my presentation I will briefly describe the origin and the biology of recombinant human bone morphogenetic protein 2. I will also summarize the pre-clinical studies we have conducted that compliment the pre-clinical studies conducted by Medtronic Sofamor Danek.

As a reminder, the letters rhBMP-2 represent recombinant human bone morphogenetic protein 2. Scientists at Genetics Institute use molecular biology techniques to isolate the human gene in coding BMP-2. This gene was inserted into a chromosome of an industry standard mammalian cell line and this cell line was subsequently engineered to enable it to produce high levels of rhBMP-2 protein. This cell line can grow in large vessels and synthesize the protein as it does so. RhBMP-2 is purified from the media, filled into sterile vials and subsequently freeze-dried.

RhBMP-2 is a member of a large protein family whose members all have activities associated with the growth and differentiation of tissues. Endogenous rhBMP-2 plays a key role in bone repair and embryonic development. Recombinant human BMP-2 is a homodimeric glycosylayted molecule with a molecular weight of approximately 30,000 Daltons. The protein is highly conserved and active across species. This conservation allows the use of recombinant human BMP-2 in all of the animal studies I will present rather than having to prepare specie specific versions of this protein.

Finally, the biological activity of rhBMP-2, that is the basis for its therapeutic development, is its ability to induce bone in both animals and humans.

This slide demonstrates that bone induction activity of rhBMP-2 in the classic in vivo assay known as the rat ectopic implant assay. This assay was originally developed in Dr. Reddi's laboratory. In this assay, rhBMP-2 is implanted at a non-bony site. Typically bone is induced at this site within seven to 14 days following implantation. The photograph on the left shows the gross appearance of an ossicle of bone induced by rhBMP-2 in the subcutaneous space of a rat's thorax. Histological analysis of this new bony tissue reveals extensive formation of trabecular bone corresponding to the dark pink regions in the photo micrograph on the right, a complete compliment of bone associated cells such as osteoblasts, osteoclasts and stromal cells and a highly vascularized structure with all bone marrow elements corresponding to the light pink regions in the photograph on the right.

This activity in this rat model has been labeled osteo induction. Only rhBMP-2 and several other bone inducing BMP proteins exhibit this biological activity. No other protein or drug has demonstrated this activity in this model. Some of the biological events comprising bone induction have been identified. Following the implantation of recombinant human BMP-2 cells initially migrate to the site and undergo several rounds of cell replication.

Subsequently, fibroblasts appearing mesenchymal cells differentiate into osteoblasts. Bone is formed, initially as woven trabecular bone and subsequently remodeled by the combined action of osteoclasts and osteoblasts into lamellar bone.

The newly induced tissue is highly vascularized as demonstrated by the numerous blood vessels in the photo micrograph on the left. The entire sequence of events induced by rhBMP-2 recapitulates the physiologic process of bone formation.

Considerable information has also been published about the mechanism of rhBMP-2 action. Responsive cell types and major cell surface receptors have been identified. Additionally, elements of the signal transduction pathway have been identified by which rhBMP-2 exercises its effects on cells. Finally, and very importantly, it has been shown that it is necessary to apply rhBMP-2 locally in order to obtain bone induction in vivo.

To facilitate the local application of rhBMP-2, the protein is combined with a biomaterial that is generally called a matrix. The use of a matrix with rhBMP-2 enables its surgical placement at the treatment site, facilitates retention of rhBMP-2 at that site and ideally provides an environment that is compatible with bone induction.

The matrix selected for clinical development in this specific program is an absorbable collagen sponge, abbreviated as ACS. This sponge was selected after screening dozens of matrix candidates. The sponge is a commercially available product marketed in the United States since 1981 as a surgically implanted hemostatic agent and has an extensive commercial experience of safe use. The sponge is composed of bovine tendon- derived type 1 collagen. Its manufacturer meets or exceeds all regulatory requirements.

This next slide shows an example of the dry absorbable collagen sponge in its original packaging prior to the addition of rhBMP-2. This diagram describes the preparation of rhBMP-2 ACS. The vial containing the freeze-dried powder of rhBMP-2 is reconstituted with an appropriate volume of sterile water, abbreviated WFI in the diagram. The resulting sterile solution of rhBMP-2 is subsequently applied uniformly to the dry ACS, generating a cohesive pliable implant that can be readily manipulated in the operating room as depicted in this slide which shows that the wetted sponge can be rolled and subsequently inserted into an LT-cage.

I mentioned before that one desirable attribute of a matrix is its facilitation of rhBMP-2 retention at the site of implantation. This slide describes one experimental system we used to assess this attribute. We generated rabbit ulnar osteotomies onto which we implanted rhBMP-2 ACS contained radioactively labeled rhBMP-2. Following implantation, we measured the amount of rhBMP-2 retained at the implantation site over time by a non-invasive technique of gamma camera scintigraphy. Basically, we measured the radioactivity remaining at the site over time and this method has been validated by several supplementary analyses including direct explant measurement and biochemical characterization of the radio labeled BMP-2 derived from the explants.

This slide shows the retention of rhBMP-2 in the rabbit ulnar osteotomy model. The Y axis represents the percent of the rhBMP-2 initial dose remaining at the implantation site and the X axis represents time in days. The graph shows that the use of ACS as a matrix facilitates the local retention of rhBMP-2 at the site, in contrast to the application of rhBMP-2 in buffer depicted by the line with black squares. Significantly more rhBMP-2 is retained at the implantation site when it is applied in combination with ACS depicted by the line with yellow diamonds. Radio-labeled rhBMP-2 can be detected at the implantation site for as long as 14 days in this model.

With this background information in mind, I will now discuss the non-clinical safety studies that have been conducted. The safety of rhBMP-2 alone or combined with ACS has been assessed in a variety of studies. Implantation of rhBMP-2 ACS to assess its implant safety has been conducted. The absorption, distribution, metabolism and excretion of rhBMP-2, abbreviated ADME, has been assessed. Finally, the safety of rhBMP-2 alone has been studied in a panel of assessments.

The safety of rhBMP-2 ACS implantation has been evaluated in three anatomic sites, including a spine safety study conducted in Dr. Hanley's laboratory. The results of this spine study have already been published. The two other implant safety studies used rat and canine models with follow-up extending through six or 12 months. In these two studies we used rhBMP-2 dosing that greatly exceeded the specie specific therapeutic range and I should explain this.

I have previously mentioned that recombinant human BMP-2 is biologically active in all mammalian species. However, different species require different concentrations of rhBMP-2 within ACS for optimal bone formation and specifically the optimal therapeutic concentration of rhBMP-2 is lowest in rodents, higher in canine and even higher in non-human primates and patients.

We took advantage of this phenomenon to deliberate exceed the species specific optimal concentrations of rhBMP-2 within ACS in order to assess any toxic effects. The safety results of all these studies were uniform. There were no systemic effects observed, no gross pathology or histopathology findings, no effects on blood chemistry, hematology or urinalysis and no incidents of bone formation distant from the site of implantation.

Furthermore, the local effects observed in these studies were consistent with the bone inducing biological activity of rhBMP-2. We also looked at the biodistribution of rhBMP-2 following its implantation. We used two implantation models in two species. The results are similar. RhBMP-2 is slowly released from the implantation site with a maximum of 0.1 percent of the implanted rhBMP-2 dose detected in the systemic circulation.

This slide shows the retention of rhBMP-2 at the site of implantation but this time in a rat femur onlay model. RhBMP-2 can be detected at the implantation site for as long as 14 days following surgical implantation in this model. In this same study we measured rhBMP-2 levels in the blood and showed that the maximum amount detected was 0.1 percent of the total rhBMP-2 implanted. I've not graphed those levels on this slide because they would all cluster at zero on this scale.

Nevertheless, because some small amount of rhBMP-2 was detected in the systemic circulation following the product's implantation, we studied the fate of rhBMP-2 following systemic administration. We used standard animal models of pharmacokinetics and biodistribution in rats and non-human primates and applied rhBMP-2 protein dissolved in buffer via intravenous administration. In these models we observed that rhBMP-2 is rapidly cleared from the systemic circulation with a terminal half-life of 16 minutes in rats to seven minutes in monkeys.

The liver is the principal organ of clearance. Subsequently, the protein is rapidly degraded and excreted into -- completely degraded and then the remnants are excreted into the urine. This graph shows the clearance of rhBMP-2 from the systemic circulation following intravenous administration of the protein in rats. The Y axis represents the percent of the rhBMP-2 initial dose remaining in the blood and the X axis represents time but now in minutes rather than in days as used in the previous graphs.

This slide shows that rhBMP-2 is rapidly cleared from the circulation and contrasts dramatically with its relatively slow clearance from the site of implantation. This slide summarizes the net effect of the local and systemic clearance of rhBMP-2. Slow rhBMP-2 release from the site of implantation combined with rapid systemic clearance, results in very low systemic exposure. This low systemic exposure has implications for the safety results described in the following slides.

I switch now to a description of various safety assessments of rhBMP-2 alone beginning with studies relevant to tumor formation or proliferation. Published studies have screened many different tumors and identified several tumor types that express either BMP-2 or BMP receptors. These published data do not indicate any role of BMP-2 in the initiation or the promotion of tumor formation.

We have also assessed rhBMP-2 in standard assays and determined that the protein is neither cytotoxic nor mutagenic. Additionally, we performed a thorough histological assessment of the local implantation site in the implant toxicity studies I previously mentioned. We detected no abnormal cellular features at the site of implantation in any study at any time point. For example, in our rat implant study, we implanted concentrations of rhBMP-2 on ACS that were 40 times higher than the optimal therapeutic concentration for this species.

The histopathology assessment of the implant site revealed no abnormal cellular features at any time point through the one-year follow-up period. These combined data suggest that rhBMP-2 has no role in the initiation or the promotion of tumor formation. To investigate the effect of rhBMP-2 on tumor cells that already exist, we have conducted in vitro studies. We focused our efforts on in vitro assessments because it is possible to achieve relatively high exposure levels to rhBMP-2 in contrast to the very low systemic exposure levels that can be achieved in vivo. These studies are summarized on the following slide.

We have provided FDA with the results of series of studies including our own that investigated the effect of rhBMP-2 on the growth of human tumor cells in vitro. Most of these studies have been published. In aggregate, these studies have assessed 51 human tumor cell lines to date. Three lines have shown some growth promotion in the presence of rhBMP-2 as compared with growth in the absence of the protein. I will discuss these lines first.

Two of these lines were derived from pancreatic tumors. When these lines were cultured in the absence if serum, rhBMP-2 stimulated cell growth by 12 or 25 percent above that of the controls. Both lines carried a mutation and a key component of the intracellular BMP signal transduction pathway. Other pancreatic tumor cell lines were evaluated and showed either no effect or growth inhibition.

The third tumor cell line demonstrating increased growth was derived from a prostate carcinoma. This cell line only showed growth promotion in the absence of serum or the absence of androgen. When either of these components was added back to the medium, rhBMP-2 actually inhibited the growth of this cell line. In the remaining 48 cell lines, the addition of rhBMP-2 had either no effect or resulting in the inhibition of tumor cell growth in approximately 50 percent of the cell lines tested. These lines included many different tumor cell types including seven osteosarcoma lines as well as the three additional pancreatic tumor lines and four additional prostate tumor cell lines.

Besides testing rhBMP-2 on established tumor cell lines, the protein has also been tested on primary tumor isolates generally obtained following surgical debulking procedures. Seventy-one independent tumor isolates have been tested to date and all show either no effect or inhibition this time in approximately 25 percent of the isolates tested.

Finally although inhibition of tumor growth has been observed most dramatically in the cell lines and primary isolates of multiple myeloma cells, the degree of inhibition is not extensive enough to consider rhBMP-2 for therapeutic applications in patients with these tumors.

Over the course of the review of this PMA submission, we met with FDA on several occasions to discuss additional tumor biology studies that could be conducted. We mutually agreed to perform that additional studies outlined in this slide as a post-approval commitment. These studies are intended to compliment the scientific literature and to more systematically assess rhBMP-2 effects on tumor cells that express the known BMP receptors.

The first study is designed to screen tumor cell lines for the levels of messenger RNA in coding each known component of the BMP receptor complex. We are performing the screening activity by using a sensitive polymerase chain reaction assay for each known receptor component and comparing the individual component messenger RNA levels in tumor cells with messenger RNA levels in other cell types known to respond to rhBMP-2. We used this comparison to operationally classify tumor cell lines as positive or negative for BMP receptor RNA.

The second cell line evaluates representative tumor cell lines from the first experiment. If possible, cell lines will be selected that represent a variety of tumor types and different messenger RNA levels of BMP receptor components. The growth of these cell lines will be assessed in vitro in the presence and the absence of rhBMP-2.

In the third study, representative tumor cell lines from the second experiment will be assessed if relevant as xenografts in an appropriate mouse model system in the presence and absence of implanted rhBMP-2. As I stated earlier, these additional studies constitute a post-approval commitment in agreement with FDA.

I will now discuss additional safety assessments of rhBMP-2 alone that have been performed. We have also conducted formal studies of rhBMP-2 safety in well-characterized animal toxicological models. We studied the systemic safety of rhBMP-2 in two species using intravenous administration to apply a single dose or doses repeated daily for 28 days at systemic exposure levels that greatly exceeded anticipated human exposure.

Similarly, because indigenous BMP-2 is active during embryogenesis, we have studied the reproductive safety of rhBMP-2 using repeated intravenous administration in standard rat and rabbit models that assess fertility or teratological effects again at systemic exposure levels that greatly exceed anticipated human exposure. In these formal toxicity studies, we assessed rhBMP-2 effects on clinical signs, ophthalmic evaluations, electrocardiograms and blood pressure, bone marrow and hematology parameters, blood chemistry, urinalysis, growths pathology and histopathology of all major organs.

In the reproductive toxicity studies, we additionally assessed rhBMP-2 effects on maternal and paternal mating performance and reproductive parameters, maternal toxicity, embryo lethality, litter size and viability and fetal abnormality. The results of all of these studies were similar. There were no effects observed.

Studies were also conducted according to the tripartite biocompatibility guidelines for medical devices and a series of general safety pharmacology studies were conducted using systemically administered rhBMP-2. The results of these studies were also similar. There were no effects observed.

In the course of the review of this PMA submission, we have recently met with FDA to discuss issues related to an immune response to rhBMP-2. Following consultation with FDA, we have mutually agreed to perform the additional studies outlined in this slide as a post-approval commitment. These studies are intended to more thoroughly assess the overall immune response to rhBMP-2.

Our first commitment in this area is to develop a broader clinical antibody assay to detect human antibody isotopes in addition to the major IgG isotopes that we currently detect. Our second commitment is to develop a valid assay to assess the ability of antisera to block or neutralize the biological activity of rhBMP-2.

Finally, we have also recently begun a discussion with FDA concerning experimental approaches to appropriately assess the potential of maternal anti-rhBMP-2 antibodies to have adverse effects on fetal development during pregnancy. In summary, the safety of rhBMP-2 ACS has been comprehensively evaluated in a series of non-clinical safety studies that used either local implantation or systemic administration. The overall pre-clinical profile that we have observed in these studies can be characterized as follows.

There was no observed systemic adverse effect of rhBMP-2 whether it was administered as an intravenous solution or implanted in association with the absorbably collagen sponge. We attribute this lack of adverse effects to low systemic exposure caused by the gradual release of rhBMP-2 from its implantation site combined with a very rapid clearance of rhBMP-2 from the systemic circulation. Local effects were observed with consistent with the bone inducing activity of rhBMP-2. We have observed no dose limiting toxicity related to rhBMP-2 in our studies at amounts substantially exceeding anticipated human exposure.

In conclusion, our pre-clinical safety assessment supports the use of infused bone graft in patients. Thank you for your attention. I'll now turn this presentation over to Dr. Scott Boden, who will review the data from several pre-clinical spine fusion studies and the results of the infused bone graft pilot clinical study.

DR. BODEN: Thank you, Dr. Riedel. My name is Scott Boden and I'm a Board certified, practicing orthopedic spine surgeon in Atlanta, Georgia. I'm also a professor of orthopedic surgery at Emory University. I've written extensively on the subject of bone morphogenetic proteins and I am familiar with the literature in this area. I have no direct financial interest in the product being discussed today before this distinguished panel but I am a paid consultant for Medtronic Sofamor Danek.

I also participated in the pilot study for the device being presented today which began five years ago when recombinant human BMP-2 was first used inside the LT-fusion cage device in humans.

I'd like to focus my remarks on the pre-clinical studies that led to the design and rationale and evaluation tools for the pivotal clinical trial which you'll hear about shortly. This slide summarizes four of the key pre-clinical studies looking at the recombinant BMP-2 protein in an interbody spine fusion environment.

Before I go into them individually, I want to point out the theme for these studies is that it was done in an animal model where the gold standard of the control had a less than 50 percent success rate in three of those four studies. And the empty cage or without protein had a zero percent success rate in the a fourth of those studies. Similarly or in contrast, the recombinant BMP-2 had a 95 or 100 percent success for inducing bridging bone in each of these challenging animal models.

The first study looks at single level interbody fusion with a titanium cage in a sheep model. In this study the cage was either filled with autogenous bone graft or the recombinant BMP-2 absorbable collagen sponge device. This slide highlights some of the challenges in non-invasively evaluating the presence or absence of bone formation in the interbody fusion area.

In the autograft controls, plain radiographs measuring lucencies of the primary determinant of presence or absence of fusion suggested that there might be 100 percent successful fusion rate. However, histologic analysis which directly visualizes bridging trabecular bone which is the criteria for a solid fusion, show that only 37 percent of those animals actually had bridging trabecular bone. In the case of BMP-2, again, the plain x-ray showed a very -- or indicated a very high success rate just using the lucency criteria but the difference here was that the histologic analysis showed bridging trabecular bone in each and every animal receiving BMP-2.

These pictures illustrate that point. Here is a micro-radiograph of an autograft control where there is clearly bone inside the cage but there are some areas that are not filled with bone and histology demonstrates that this is fibrous tissue shown in the pink color as compared to bone shown in the blue. This can be contrasted with both a radiograph and histology from one of the other animals in that study that had BMP-2 absorbable collagen sponge inside the cage rather than autogenous bone graft.

One of the other relevant questions that has already been raised this morning is what is the mechanical quality of bone that is induced using BMP-2 as compared to bone that would be initially formed and remodeled using autogenous iliac crest bone graft? And this study shows that using a variety of mechanical testing modes that the mechanical properties of the bone induced by BMP-2 were comparable comparing the red bar and the green bar to those seen with bone formed by autogenous bone graft from the iliac crest.

The second study looks at single level interbody fusion with titanium cage this time in a goat model. The end point was again six months and the two groups were again the same. Cage was either filled with autogenous bone or recombinant human BMP-2 on the absorbable collagen sponge. Once again, we see the challenges of using just plain radiographs to assess the presence of bone inside a fusion cage. In this case in the autograft group clearly less than 100 percent but more than the just under 50 percent that had an actual fusion based on histology, were assessed to be fused again using lucency as the primary criteria.

In contrast in the case of the BMP animals, although 100 percent lacked lucency on the plain x-ray criteria, only 95 percent, which means all but one animal had continuous bridging bone as measured by histology, the ultimate assessment of bone formation inside the cage.

Once again, biomechanical testing show that there was no statistical difference in the stiffness between the fusions that were formed with the autograft or fusions that were induced by the recombinant BMP-2. So, again, bone induced by BMP-2 absorbable collagen sponge functionally, inside the cage functioned similar to that of autograft bone. Now, one could argue that the presence of the metal cage might interfere with the ability to truly assess the quality of the bone and for that reason I'll just briefly show some mechanical assessment of bone with recombinant BMP-2 on the same carrier matrix, that's the absorbable collagen sponge but from a posterolateral fusion model where there's no metal.

The advantage of this model, which incidentally the autograft once again fused less than 50 percent of the animals and BMP-2 absorbable collagen sponge fused 100 percent of the animals but we can do testing that looks at just the strength and quality of the bone formed in the fusion without any confounding information from metal fixation. Once again, in this case, the mechanical properties of the bone formed with BMP-2 shown in the green bars were essentially comparable to those in terms of relative strength and relative stiffness seen with autogenous bone.

Now moving onto the non-human primate studies which are extremely relevant as was mentioned earlier by Dr. Riedel, because of the close parallel of the required concentration of BMP-2 to get efficacy of bone formation in non-human primates and how that translates to human clinical trials. In this case an allografted bone dowel or bone cage was filled with either BMP-2 and absorbable collagen sponge or with autogenous bone graft and inserted in a rhesus monkey single level, interbody fusions.

These two x-rays show an example from a control and you can see that the intervertebral disc space, that black line, is still present and there was no bridging bone or fusion across the segment when the allograft bone dowel was filled with BMP-2. In contrast, in this case you can see that bone has bridged the two vertebral segments completely obliterating the disc space with the allograft dowel was filled with BMP-2 and in addition, the allograft dowel has been remodeled.

These microradiographs again highlight this point. Two controls shown on the to, in this case, the remnant of the allograft dowel can be seen but bone is growing into and through it. In this case the dowel has fallen out of the histologic section but you can see that bone has not grown completely through the specimen. In contrast, these are two examples from animals that had BMP-2 with absorbable collagen sponge inside that cage and you can see that there's bridging trabecular bone across the interspace and remodeling of the allograft bone cage.

The last pre-clinical study is the one that most closely simulates what was to take place in human -- in the fact that this time BMP-2 with absorbable collagen sponge was placed into a scaled down titanium fusion cage. This was inserted into rhesus monkey at the lumbosacral junction with a six-month end point and the cage was either filled with the absorbable collagen sponge with buffer only, in other words, no BMP-2 or one of two doses, .75 or 1.50 milligrams per milliliter of recombinant human BMP-2.

What we found was that the sponge alone did not result in any spontaneous bone formation through the cage in either of those control animals. However, all of the animals that received either dose of recombinant BMP-2 on the collagen sponge had histologic bone formation through the cage as can be seen on these two examples.

This slide is important because it leads to the reliance and the importance of using CT scans to assess the presence of bone inside an interbody fusion cage. Here you can see that histologically in this case there was no bone forming inside the cage and on the CAT scan you can see that this dark area rather than a bright white area suggests that there is not a density consistent with bone inside the metal fusion cage. In contrast, in animals where bone grew through the cage, a CT scan revealed homogeneous bridging trabecular bone through the center of the cage on the CAT scan.

The quality of the bone and the normalcy in the non-human primates was as described earlier by Dr. Riedel in rodents and other animal models being entirely normal bone with osteoblast-line trabeculae remodeling and bone marrow elements.

Now I'd like to briefly describe the pilot clinical study that was undertaken following that rhesus monkey pre-clinical study as an introduction to validating the evaluation tools. In this study which was done at four investigational sites, the LT-threaded tapered fusion cage was filled with either autogenous bone graft in a small number of patients or with infused bone graft, that is recombinant human BMP-2 in absorbable collagen sponge.

First I'd like to briefly review the surgical anatomy for those who may be less familiar with the spine. The normal spine would have two vertebra adjacent connected by the intervertebral disc cartilage and in the approach that's being put before the panel today which is an anterior surgical approach, two cages are inserted from the front of the spine. It's important to note that there is residual disc or annulus material that serves as a microscopic barrier creating that compartmentalization that Dr. McCullough was asking for earlier preventing gross leakage of BMP-2 and carrier matrix into the area where the neuro elements would be.

Looking at this in cross section, we can see here the normal or pre-operative and then post-surgical schematic of two cages inserted side by side again with residual annular tissue serving as a macroscopic barrier. I should point out that while this is a macroscopic barrier and there can be fissures, in none of the animal studies nor in the human studies have we seen formation of bone posterior to the cage outside the confines of the disc space when the anterior surgical approach has been used which is under consideration today.

So the device before it goes in, has the wetted collagen sponge inside the taped fusion cage and in this clinical pilot study, 11 out of 11 patients were deemed by independent review of plain x-rays and CT scans to have achieved bridging trabecular bone through and/or around the fusion cages. Two of three of the autograft were shown to have successful bridging bone.

Another issue is the impact on clinical outcome. In the Oswestry Disability Index is a disease specific patient derived outcomes measure, which is commonly used in patients with low back problems. In this case having a lower score is desirable or indicative of less symptoms or less disability. In the Oswestry scores from pre-op to 24 months, gradually decreased on both groups, in this case it seemed a little bit quicker in the infused group and it nearly reached statistical significance but in the end the clinical outcome was at least as good in patients that used infused and did not have to have a second site harvest with autogenous bone graft.

Lastly, I think showing some representative pictures are really important to bring home the fact that bone is forming in an area where there otherwise wasn't bone. The format of the next several slides will look at a slice through the right-hand cage, the left-hand cage and a coronal view of both cages or frontal view using reconstructed thin sliced CT scans. And the three columns represent different points in time; six months, 12 months and 24 months. And you can see from this patient that started with autograft bone, that in fact, there was incorporation of the autograft in this case and there is bridging trabecular bone through the cages both on the lateral and on the frontal view.

Here's an example of another autograft patient where you can see that there's less density inside the cage perhaps with resorption of the autograft bone and over time there were lucencies that formed around the cage and absence of ridging bone suggesting a failure of bone formation and fusion. Contrast that with an example of some of the infused patients keeping in mind that the cage starts out without any bone in it and so when we see white bone growing through the cage, we know that it was induced as a result of infuser BMP-2.

You can see here in both cages there's bone growing through the cages and as time progresses you can see secondary ossification which is a normal adjunctive finding in any solid fusion around the cages through the disc space. Another example of an infused patient, again showing an increase in bone density in the cage over time and also the secondary healing around the cage normal for a solid interbody fusion.

Another question is what happens to the bone inside these cages over a longer period of time and although this was a pilot study, these patients have continued to be followed and this is an example of one of the autograft patients at 48 months showing that when autogenous bone was put inside the cage, then in fact, it remains bridging through the cage and remodels similar to the density of bone in the adjacent vertebral bodies.

Looking at a patient that received infused we saw this same trend of preservation of bone in the cage, not disappearing of bone, and continuing to mature the adjunctive fusion throughout the interspace. So the bone that's induced with BMP behaves ostensibly the same as bone that was put by autograft bone in the case of inside the fusion cage.

And just one other example of a patient with BMP, again with longer term follow-up showing that the fusion is maintained even at four years with bone both through the cage and around the side of the cage. So I think it's important to summarize the goals of the recombinant BMP-2 absorbable collagen device. You've heard much about systemic safety and toxicity in the first talk and I think at this point it suffice it to say very simply that we did not see any bone formation at a distance from the cages or in any place outside the caged in matrix in any of the animals or any of the pilot patients that have been discussed so far.

Equally important to safety is effectiveness. And effectiveness for this device really should be considered the ability to eliminate bone grafting morbidity which is substantial in these patients and to obtain equal or better healing success rate defined as bridging trabecular bone across the interspace and through the cage. In other words, stated more simply the goal of recombinant BMP-2 absorbable collagen sponge device is to make bridging bone.

So in conclusion, based on a series of detailed pre-clinical and a clinical pilot trial, I believe that recombinant BMP-2 absorbable collagen sponge has shown success in 95 percent or better in four animal studies, substantially better than autograft controls in those models. The bone formed is normal and biomechanically equal to that formed with autogenous bone graft. CT scan analysis correlates the best with the histology of bone and therefore, is an important indicator of the presence of new bone formation by BMP-2 inside fusion cages.

And finally, the concentration of BMP-2 that was successful in the rhesus monkey pre-clinical study was also successful in the clinical pilot study with 100 percent success in humans at the same doses that were predicted by the rhesus monkeys. At this point, I'd like to turn the podium over to Dr. Hal Mathews, who will describe the results of the pivotal clinical trial in greater detail.

DR. MATHEWS: Distinguished panel members, ladies and gentlemen, good morning. My name is Hal Mathews. I'm a practicing spine surgeon from Richmond, Virginia. My entire clinical focus is spine care. I'm an associate clinical professor of orthopedics and neurosurgery at the Medical College of Virginia in Richmond. I have no direct financial interest in this product under review here today and I'm not being paid for my participation in this meeting. I participated in the open surgical approach study of this device as an investigator.

I'm here today to present the results of the InFUSE^TM Bone Graft/LT Cage Lumbar Tapered Fusion Cage Device clinical trial. Before I discuss the details, I want to report to this advisory panel and to the audience the top line findings from this open surgical approach study. First and foremost the primary objective of the clinical trial as stated in the protocol was met, thus establishing the safety and effectiveness of the InFUSE^TM bone graft in the treatment of degenerative disc disease.

Secondly, the InFUSE^TM bone graft stimulates the formation of bone which results in very high fusion rates. Thirdly, the InFUSE^TM bone graft patients experience shorter operative times and less blood loss than the control patients. And finally, patients who receive the InFUSE^TM bone graft avoided the complications and significant post-operative pain associated with bone graft harvesting in the control group.

Let me offer a few additional observations that I believe will bring into sharper focus the clinical trial results. At the end of the day it is our job as physicians to help our patients in the least invasive and least painful ways. For the patients from whom a bone graft was taken from their iliac crest, at the time of discharge, 80 percent of patients registered a score of at least 10 out of 20 and nearly 15 percent of these patients had a score of at least five at 24 months. In addition, six percent of these patients experienced graft site complications, including bone fractures, nerve injuries, infections and hematomas.

Given its equivalent performance in achieving fusion, the infused bone graft is clearly the most humane way to treat this painful condition. I will now elaborate on the clinical trial and the results and I will conclude with a brief review of the laparoscopic clinical trial that was also conducted showing equivalent rates of fusion as well as some other potential patient benefits.

Let us now discuss the open surgical approach for device implantation. This study had a prospective randomized control design. The investigational treatment patients received the LT cage device filled with the InFUSE^TM bone graft. Henceforth, I will refer to these patients as the InFUSE^TM group. The control patients were treated in a similar manner with the LT cage device filled with autogenous harvested bone from the iliac crest. These patients will be designated the autograft group.

The primary objective for the clinical trial was to determine if the overall success rate for the InFUSE^TM is at least as high statistically as the rate for the autograft group. Overall success is a derived variable encompassing primary safety and effectiveness considerations. Secondary objectives focusing on equivalency and superiority of specific end points were also developed.

Bayesian methods were used for statistical comparison of study outcomes. Patients admitted to the study had a single level symptomatic degenerative disc disease as noted by back pain of discogenic origin with or without leg pain with degeneration of the disc confirmed by patient history and radiographic studies. There are a number of additional inclusion and exclusion criteria such as age, weight, mental competency, medical history, and existing medical condition.

Patients involved in the clinical trial were evaluated pre-operatively, at surgery and post-operatively at six weeks, three, six, 12 and 24 months. A total of 143 patients received the InFUSE^TM bone graft. There were 136 patients who were treated with autogenous bone graft. Patient follow-up compliance at all post-operative periods exceeded 90 percent. Sixteen investigational centers contributed to these patients.

Patients in both treatment groups had very similar demographic characteristics and pre-operative medical conditions. This enhances one's ability to interpret the treatment effects since potentially confounding factors did not impact with the results. In terms of surgery results, the mean operative time for the InFUSE^TM group was approximately one-half hour less than that for the autograft group and this finding was statistically different.

The blood loss for the InFUSE^TM group was also statistically lower than that for the autograft group. The mean hospital stays of patients in both treatment groups were slightly more than three days and did not have statistical difference. The results of other surgical variables such as treated level, operative approach and type of orthosis were similar for both groups. The outpatient and inpatient classification and return to work times were also comparable.

For clinical outcomes I would like to emphasize that 24-month data are being used a primary supporting evidence of the safety and effectiveness of treatments. In order to satisfy the FDA's guidance for spinal implant studies, a composite variable termed overall success was created and this variable is the primary end point of the entire study for PMA approval purposes.

Overall success is comprised of the effectiveness parameters of fusion, Oswestry success, neurologic success. It is also influenced by two important safety considerations, the occurrence of any serious adverse events possibly associated with the device and the occurrence of a second surgical procedure classified as a failure. The overall success criteria is very demanding.

The primary objective of this study was to determine if the overall success rate for the InFUSE^TM group was at least as high statistically as for the autograft group. As evidenced from this slide, the overall success rates for the two treatment groups at 12 and 24 months following surgery are very similar and stable over time. These rates were statistically equivalent at 24 months. Therefore, the primary clinical trial objective was met, thus supporting approval of this product.

I will now discuss in detail the safety and the effectiveness parameters that were used in this clinical trial. Safety was assessed as a function of the nature and the frequency of adverse events and second surgery procedures and the formation of antibodies to rhBMP-2 and collagen. Based on these assessments, the infused group was found to be as safe as the autograft group.

Now for more details. Reported adverse events in each group were classified by their nature, their severity according to the World Health Organization criteria, and their duration. Also Medtronic Sofamor Danek instructed investigators to report all adverse events that occurred whether or not the event was related to the treatment or the device. This conservative approach led to the reporting of many unrelated events that were included in the analysis. For the InFUSE^TM group only 17 patients or 11.9 percent had an event that was possibly related to the device and only 11 of these patients or 7.7 percent were the events considered serious.

Overall, a total of 113 InFUSE^TM patients had at least one adverse event with a substantial majority not being related to the device. As you can see from this slide, these rates are very similar to those rates for the autograft group. Adverse events were also categorized according to their nature and comparisons were made between the two treatment groups. There were no statistical differences for all reported categories of adverse events except for two. These categories in which differences were noted were graft site events and urogenital. Nearly six percent of the autograft patients had graft site complications. These complications included bone fracture, nerve injuries, infection and hematoma.

Obviously, there were not graft site adverse events for the InFUSE^TM group. This fact clearly supports the use of InFUSE^TM bone graft since it eliminates the need to harvest bone graft. Urogenital complication rates favor the autograft group. The difference in rates was mainly due to urinary retention following surgery and these events resolved in all patients prior to discharge from the hospital.

Overall the occurrence of adverse events in the clinical trial were considered typical for a patient population having an anterior lumbar interbody fusion procedure and were not unanticipated. Another component of safety assessment is the number and nature of additional surgical procedures performed after the initial study surgery. This slide lists the classifications of additional surgical interventions.

According to the protocol, revisions, removals and supplemental fixations are considered significant procedures at the treated spinal level that effect the assessments of the treatment outcome. Therefore, a patient having one of these procedures is considered a treatment failure for study purposes. On the other hand, re-operations and other surgical procedures that are believed to have no effect on the treated level are therefore, not considered failures. The second surgery rates for both groups were comparable and there were no statistical differences for any of the additional surgery category comparisons.

Because of the proteinaceous nature of both rhBMP-2 and the absorbable collagen sponge the development of antibodies was assessed as part of the IDE clinical trial. Serum samples were taken from each patient pre-operatively to establish their baseline condition and at three months following surgery. These samples were analyzed for the presence of antibodies specific to rhBMP-2 and to bovine type I collagen. If a patient had a positive response to bovine type I collagen, the serum was also tested for antibodies to human type I collagen. Antibody levels were checked in both InFUSE^TM and autograft patients, even though the latter group was not exposed to the InFUSE^TM product.

The rates of antibody formation were not different for the two treatment groups. There was one InFUSE^TM patient and one autograft patient who had authentic positive responses to rhBMP-2. The incidents rates were very low, at less than one percent. There were no adverse events that appeared to be related to these findings.

Approximately 13 percent of patients in both treatment groups had authentic positive responses to bovine type I collagen. These antibody responses did not appear to result in any clinical manifestation nor impact the overall success rates of the study. None of the patients who tested positive for bovine type I collagen had a positive result for human type I collagen. These antibody findings are similar to those from other Medtronic Sofamor Danek clinical trials involving the InFUSE^TM bone graft.

Since I've presented a lot of information, I want to briefly review the impressive safety profile of the use of the InFUSE^TM bone graft with the LT-cage device before moving onto the effectiveness results. Adverse events and second surgery procedures for the InFUSE^TM treatment were very similar to the autograft treatment. The rates of antibody formation were not different for the two treatment groups. In addition, any positive antibody response appeared to be without clinical manifestation.

The use of the InFUSE^TM bone graft eliminated graft harvesting adverse events that occurred in approximately six percent of the autograft patients and significant graft site pain in approximately 80 percent of patients peri-operatively. This finding is significant since it supports a major reason for using the InFUSE^TM bone graft.

There were also no cardiovascular adverse events associated with the use of the InFUSE^TM bone graft. Therefore, based on the data, the InFUSE^TM bone graft LT-cage device is safe for its intended use in the anterior lumbar interbody fusion procedures to treat degenerative disc disease.

Now, we'll focus on device effectiveness. Briefly in summary, these patients received the InFUSE^TM bone graft experienced exceptionally high fusion rates, pain relief, maintenance or improvement in neurologic status. Let's review specific effectiveness results in more detail. We consider fusion to be the primary end point since the intended use of the infused bone graft is to induce bone formation in spinal fusion procedures.

For this clinical trial, CT scans and radiographs were used to assess fusion. These films were evaluated at the University of California San Francisco under the direction of Dr. Harry Genant, a board certified radiologist. There were two teams of reviewers who were masked to patient treatment. Each team worked independently of each other. If their overall fusion conclusions differed, a third reviewer would adjudicate the findings. However, this occurred in frequently since the percent agreement between the two primary review teams exceeded 98 percent at all time points.

Fusion was based on evidence of bone, spanning the two vertebral bodies of the treated segments, using CT scans and radiographs. In addition, segmental stability and lucent line criteria also had to be met to be considered fused. Patents having second surgical procedures reported by the investigator as due to pseudoarthrosis or non-union were also considered as fusion failures regardless of independent radiographic findings.

This later condition dramatically impacts fusion rates for both treatments. For example, at 24 months post-operatively, all non-unions in the InFUSE^TM group were due to second surgery criteria and not the radiographic criteria. The fusion rates for both treatment groups were high at 12 and 24 months following surgery. At 24 months following surgery the InFUSE^TM fusion rate was 94.5 percent and was statistically equivalent to the autograft rate of 88.7 percent.

Frankly, for the study the most important aspect of the fusion criteria is whether bone, spanning the two vertebral bodies at the treated level could be detected by the independent radiologist. This would be indicative of whether the InFUSE^TM bone graft was effective in stimulating de novo bone formation. It is noteworthy that in all patients in both treatment groups with CT scans available, such spanning bone was detected at 12 and 24 months. CT scans were particularly important in detecting the bone. These findings are considered of prime importance since fusion cannot exist unless bone is connecting the treated segment.

In addition, these findings agree with the data previously presented by Dr. Scott Boden. The Oswestry Low Back Pain Disability Questionnaire was used to measure the effects of back pain on a patient's ability to manage everyday life. The Oswestry Questionnaire has 10 questions and is self-administered. Oswestry scores are expressed as a scale ranging from zero to 100 points with a lower score indicating less pain and disability.

As seen with this slide, the mean Oswestry scores for the two treatment groups were very similar at all time periods. At 24 months following surgery, the mean improvements in Oswestry scores from pre-operatively were approximately 29 points for both treatment groups. These findings are quite gratifying and represent an approximate 55 percent improvement.

This slide illustrates the distribution of patients demonstrating pre-operative to post-operative improvements in Oswestry scores of at least 15 points, which is a very rigorous condition mandated by the FDA. This is termed Oswestry success. Like mean Oswestry scores, the Oswestry success rates were similar for both treatment groups. At 24 months following surgery the Oswestry success rates were found to be statistically equivalent with rates of 73 percent in both groups.

The neurologic status of patients was also assessed pre-operatively and post-operatively and at every follow-up visit and is considered an indicator of safety and effectiveness. The neurologic evaluations consisted of measurements of motor function, sensory, reflex, and degree of straight leg raise producing pain. A successful outcome for each parameter was based on the post-operative condition being no worse than the pre-operative condition.

Overall neurologic success for a patient in any given post-operative time period was based on having successful outcomes for all four neurologic parameters. This slide shows the overall neurologic success at 12 and 24 months following surgery for the two treatment groups. The rates are very similar across time and treatment. The 24-month neurologic success rates for the InFUSE^TM and autograft groups were determined to be statistically equivalent.

In addition to these end points that contribute to overall success, other effectiveness measurements were made during the course of this study. These measurements included back pain, leg pain, disc height maintenance and general health status via the SF-36 survey. The 24-month results for these parameters were comparable for the two treatment groups and statistically equivalent between treatments was demonstrated for all but two comparisons. They were back pain and mental component summary or MCS of the SF-36.

I will not focus on the MCS finding since the difference between the two treatment groups was less than four percentage points and this is not considered clinically significant. For back pain the success rate favored the autograft group and it is believed to be due to arbitrary cut-off assumptions of the analysis, since a mean improvement of back pain scores for the InFUSE^TM group was actually higher, showing more improvement than that for the autograft group.

Another very important effectiveness parameter that was assessed was graft site harvest pain. This was measured in autograft patients using two numerical rating scales, one for pain intensity and the other for duration. The composite pain score ranged from zero to 20 with a lower number signifying a better outcome. This slide shows a mean graft site pain for autograft patients from time of hospital discharge to 24 months post-operatively. At hospital discharge the mean score was 12.7 and approximately 80 percent of patients had scores of at least 10. As expected, the harvest site pain improved over time.

However, nearly 15 percent of patients had a score of at least five at 24 months post-operatively. Aside from the pain approximately 16 percent of patients indicated they were still bothered by the appearance of the graft site at one and two years following surgery. When these rates are coupled with the adverse events associated with harvesting the bone, a very compelling case can be made for using infused bone graft in spinal fusion procedures since it eliminates the negatives of graft site appearance, pain and morbidity.

There is additional good news about the InFUSE^TM bone graft. Another clinical trial was performed examining the laparoscopic implantation of the device and the results are just as compelling as for the open study. The data from the laparoscopic study augments the safety profile of the device and support approval of that surgical method of cage implantation. The laparoscopic study had one treatment group, those patients treated with the InFUSE^TM bone graft and the LT-cage device. Other than this, the protocol was identical to that of the open study.

A total of 134 patients received the investigational laparoscopic treatment. Fourteen centers contributed the patients. There was no overlap in surgeons between the open and laparoscopic studies. On average the hospital stay for laparoscopic patients was approximately two days shorter and statistically different than for patients of either treatment group of the open study.

Further, nearly 45 percent of laparoscopic patients were treated on an outpatient basis as compared to virtually none in the open study. The laparoscopic patients also returned to work some 20 days sooner than for the open study patients. These surgery, hospital stay and return to work findings for the laparoscopic patients may suggest that there is a synergistic effect of the use of the InFUSE^TM bone graft and the laparoscopic insertion of the LT-cage device.

The overall success rate at 24 months following surgery for laparoscopic patients was more than 68 percent and nearly 12 percentage points higher than for the autograft rate of 56 percent. This rate was not only statistically equivalent to the autograft, but statistically superior, a finding that more than satisfies a primary objective of the laparoscopic study.

The safety profile of the laparoscopic use of the device was also comparable to the open surgical treatment groups. As expected retrograde ejaculation rate was higher than with the open surgical treatment due to the transperitoneal approach for laparoscopic patients. However, the rate was lower than previously noted in other large studies using the laparoscopic implantation of the LT-cage device.

The effectiveness results for the laparoscopic investigational patients were also impressive. This slide shows statistical equivalence can be claimed for all comparisons to the autograft group from the open study. At 24 months the fusion rate was virtually identical to that for the open InFUSE^TM bone graft treatment at approximately 94 percent, these compared to an 88.7 percent value for the autograft group. Again, bridging bone was noted in all evaluated patients at 12 and 24 months radiographically.

Since seeing is believing, I want to spend the next few minutes showing a few slides of some study patients using CT. According to the protocol criteria, these patients had not responded to non-operative treatment for at least six months prior to being included in this study and had significant amounts of pain. The first case is an example of a successful radiographic fusion in an autograft patient.

The patient is a 37-year old female that had an L5/S1 fusion procedure. Since cortico cancellous autogenous bone chips are placed in this autograft patient's cages, it appears radio-opaque immediately after surgery. Over time the bone chips begin to bridge and consolidate to form bridging bone through the cages. The second patient was a 38-year old female who had an L5/S1 fusion procedure. This particular patient's cage were filled with autograft. The patient was not a successful fusion.

As a result of the failed fusion and lack of stabilization across the disc space, the surrounding bone undergoes some absorption that becomes evidenced by radiolucencies and black lines around the cages. The third patient is a 42-year old female who had an L4/5 fusion procedure with the InFUSE^TM bone graft placed inside the cage. In contrast to the autograft filled cages, when InFUSE^TM bone graft is placed into the cage, it is not initially radio-opaque. The InFUSE^TM bone graft starts out as a dark appearance within the cage, so as to increase whiteness, this is due to new bone formation.

The CT scan clearly shows evidence of new bone formation at six months, evidenced by the radio-opacity. And over time this bone becomes denser. In addition, anterior bridging bone can be seen in front of the cage and around the sides of the cage. This is further evidence of mechanical stabilization across the disc space.

One question you may be considering is do these impressive CT scans infusion results hold up over time and the answer is yes, and this is based on four-year post-operative CT scans from the same InFUSE^TM treatment that Dr. Boden previously presented. So to summarize, as demonstrated in both animal and human studies CT scans are the most practical and definitive method of detecting new bone formation within cages and determining fusion status.

The scientific data I have presented has been impressive and we believe the results certainly support approval of the product. Science aside, patients need to be satisfied with their results. So study patients were asked at their post-operative visits to respond to three questions related to satisfaction. This slide vouches for the high levels of satisfaction at 24 months following surgery for both InFUSE^TM bone graft LT-cage device treatments and for the autograft group.

Generally 75 to 82 percent of the patients offered positive responses which are very gratifying findings considering the complex nature of low back pain and degenerative disc disease. In conclusion, the primary objective of the prospective randomized study of the open surgical implantation of the InFUSE^TM device was met. The overall success rate of the InFUSE^TM bone graft LT-cage device was found to be statistically equivalent to the autograft treatment.

The InFUSE^TM treatment was associated with shorter operative times, less blood loss than their autograft control patients. Two of the primary benefits of InFUSE^TM bone graft are that it induces bone formation and that it eliminates the need to harvest autogenous bone graft in spinal fusion procedures. The autograft group results attest to the need for InFUSE^TM bone graft treatment since 80 percent of the patients had significant perioperative graft site pain and nearly six percent of these patients had an adverse event associated with graft harvesting.

Further, the laparoscopic implantation of the infused device produced very positive clinical results as well. The overall success rate was statistically higher than the autograft group. In addition, the patients had hospital stays that were two days shorter than the autograft group and they returned to work some 20 days sooner. Therefore, the results of this study of the open and laparoscopic implantation of the InFUSE^TM bone graft with the LT-cage lumbar taper fusion device showed the device to be safe and effective in the treatment of degenerative disc disease.

Thank you very much.

DR. LIPSCOMB: Members of the panel, in conclusion as clearly demonstrated in these presentations, and in the information that was submitted in the PMA application, more than a reasonable assurance of the safety and effectiveness of InFUSE^TM bone graft with the LT-cage device has been presented. We understand that following our presentations the FDA will pose several questions to this panel. We believe that our presentations have provided much information to address FDA's questions.

For the sake of clarity, let me summarize what you have just heard as it relates to some of these questions. One question pertains to a theoretical issue of rhBMP-2 stimulating cell proliferation from existing tumor. A comprehensive review of the literature provides a preponderance of evidence that rhBMP-2 has either no effect or an inhibitory effect on tumor cell proliferation.

We believe that ongoing laboratory testing at Wyeth-Genetic Institute as well as precautionary labeling statements will address any remaining theoretical concerns. Again, to emphasize, this issue pertains solely to the effects of rhBMP-2 on an existing tumor and there is no scientific evidence to suggest that rhBMP-2 transforms a normal cell into a tumor cell.

Another FDA question to the panel involves an immunology issue, specifically, what effects, if any, do antibodies to rhBMP-2 have on a developing fetus and the mother. Again, this is a theoretical issue that has not been manifested in either animal studies or in our human clinical trials. The rate of authentic rhBMP-2 antibody response was less than one percent and was similar to that in the control group in our InFUSE^TM clinical trials that you heard about this morning.

We do believe that this issue can be adequately addressed via precautionary labeling statements and instructions to females of child bearing age. Also we intend to discuss further with FDA the necessity for a pregnancy register.

Another line of questioning to the panel pertains to radiological issues. One aspect, is there functioning bone inside the cage? The answer is yes. Histological results from animal studies have verified that InFUSE^TM bone graph causes normal bone to form and that accompanying CT scans show that the appearance of the bone radiologically. Our clinical study CT scans similarly reveal the presence of bone where none existed before in the InFUSE^TM bone graft patients. This bone also remains intact and dense over time as evidenced from the CT scans that have been presented to you some of which are out as far as four years following surgery.

Finally, the major panel consideration, is the use of InFUSE^TM bone graft with the LT-cage safe and effective in the treatment of symptomatic degenerative disc disease? The valid scientific evidence presented here today unquestionably provides an affirmative response to that question. A multitude of pre-clinical in vivo and in vitro studies attest to the safety of InFUSE^TM bone graft. Functional animal model testing and clinical data from a pilot study as well as two large scale pivotal studies demonstrate InFUSE^TM bone graft safely stimulates the formation of bone.

The data from nine animal species and from humans are consistent. They are compelling and they are convincing. InFUSE^TM bone graft can safely form normal bone where none existed before and is an effective substitute for autograft bone. These data provide more than a reasonable assurance that the device is safe and effective for its intended use and this is the criterion for PMA approval. We believe that you will acknowledge the importance and the validity of this information and make this breakthrough technology available to surgeons and their patients by recommending approval of this PMA application.

This concludes Medtronic Sofamor Danek's presentations. We are available to respond to any of your questions. Thank you.

CHAIRPERSON FINNEGAN: Thank you. Actually, I think we'll do the questions later on today. We are going to take a 10-minute break. We will return at 12:00 o'clock for the FDA's presentation and then we will break for lunch.

(A brief recess was taken.)

CHAIRPERSON FINNEGAN: We are going to have the FDA presentation and this is in two parts. The first part is the FDA panel. Those are the members of the FDA staff who will give their presentations and the second portion will be three guest reviewers that the FDA has asked to look at particular portions of this PMA. And the FDA presentation will be started by Dr. Aric Kaiser.

DR. KAISER: Good morning. I'm Aric Kaiser, an expert reviewer in orthopedics and the lead reviewer for the PMA. I would like to first introduce the other members of the primary review team for this PMA who will be making the FDA presentations this morning. Peter Hudson was the lead pre-clinical reviewer, Barbara Buch, the clinical reviewer and Telba Irony, the statistical reviewer. I'd also like to acknowledge the expertise and efforts of a number of other people involved in this project both from the Center for Devices as well as valuable input from the Center for Biologics.

The sponsor has gone into detail describing the product, their pre-clinical data and the clinical results. And I'd like to remind the panel that the device that we're seeking your recommendations on today is the InFUSE^TM bone graft LT-cage lumbar tapered fusion device which is a three-component spinal fusion device that consists of a spinal implant, a growth factor and a carrier.

The first component, the cage component, is a titanium alloy tapered spinal fusion cage and as the sponsor has already mentioned, it has received PMA approval for use in the treatment of degenerative disc disease when filled with autograft. The other two components, the InFUSE^TM bone graft consists of rhBMP-2, the growth factor which is soaked into the ACS collagen sponge carrier.

From a pre-clinical standpoint, there were two areas that we looked at, those having to do with the cage itself and those having to do with the BMP and carrier. Since the fusion cage has already received PMA approval and has not been changed since that approval, there was no additional review necessary and we will not be presenting any information that the sponsor has not already presented.

The BMP and carrier will be the focus of the FDA presentations and our presentation will focus primarily on the issues having to do specifically with our questions and not repeat the full review that the sponsor has already given. The same thing will occur for our analysis of the clinical and statistical data. The sponsor has given a detailed presentation and we'll focus our comments on those issues having to do with the questions.

After the last FDA presentation, I'll get up and go through an overview of what our questions are, but beforehand, I'd like to just give you an idea of the general areas of concern so that when you're listening to our presentations, you have some idea of where to focus. We will be asking you for your input on issues having to do with reproduction and teratogenicity with tumorigenicity, radiographic effectiveness, end point interpretation, issues having to do with instructions for use and we'll also be looking for some input on potential post-market studies.

And with that, I'd like to introduce Peter Hudson, who will be giving the pre-clinical presentation.

DR. HUDSON: Hello, I'm Peter Hudson and I'm the lead pre-clinical reviewer for FDA regarding this product. I'd like to acknowledge the collaborative effort of colleagues in the Center for Drugs, Biologics and Devices for review of this application. They have provided critical input into the review of the pre-clinical and manufacturing information. It's important to note that review of the manufacturing information of this application has met the full standards of review that the Center for Biologics uses for review of recombinant reproduced growth factors. I'm going to briefly go over the pre-clinical evaluations and identify the issues that FDA believes need further evaluation.

The sponsor has been informed of FDA's concerns and in part has either begun to address some of these concerns or is committed to addressing some of the concerns as post-market commitments. As you know, we would greatly appreciate your input and guidance regarding the issues that remain as concerns.

I'll first go over the extensive toxicology and biocompatibility testing conducted on the produce although not as detailed as Dr. Riedel has done. Then I will discuss the experiments used to demonstrate bone inductive ability of the product. Finally, I will present questions that have arisen in the course of review of the pre-clinical test information. I would like to stress that these questions did not arise specifically due to the experimental observations from the sponsor's pre-clinical or clinical studies. We have posed the questions in consideration of relevant research literature.

As has already been described, part of the device consists of recombinant human bone morphogenic protein 2 in an absorbable collagen sponge or matrix. The rhBMP-2 ACS is placed within the lumbar tapered cage or fusion device. RhBMP-2 alone was evaluated in acute single and multiple dose general toxicology experiments. The results of those studies indicated that the cytokine did not cause toxicity except for the occurrence of injection-site related tissue thickening.

To assess the chronic toxicity of the cytokine in sponge, a six-month canine mandibular maxiofacial defect study and a one-year rat femur onlay study were conducted. No toxic effects were observed. The cytokine sponge combination as well as the fusion cage itself was tested in accordance with internationally recognized standards of biocompatibility testing. The cytokine sponge product passed all the tests shown here.

Studies of the ability of rhBMP-2 on a collagen carrier to induce bone included critical size defect repair models and fracture repair models on various entopic sites in rabbits, rats, dogs and monkeys. Histologic analysis of a monkey ulnar defect model and other studies suggest that bone formation in response to rhBMP-2 ACS occurs through a process of spindle or mesenchymal cell infiltration, vascular invasion and a combination of endochondral and direct bone formation.

Histologic analysis indicated that the bone formation process temporally extended from the outside of the implant towards the center until the implant was replaced by trabecular bone. Many of the animal bone induction studies included dose ranging studies and from these results a broad therapeutic dose range was identified. The effective dose range is bordered on one side by inadequate bone formation and on the other by excessive bone formation.

The therapeutic rhBMP-2 concentration range shifts with the animal species tested. Higher concentrations are required in canines than in rats and higher concentrations are required in non-human primates. The ability of rhBMP-2 ACS contained within the fusion cage to cause interbody fusion was evaluated in non-human primates, sheep and goat studies. The cytokine collagen and fusion cage combination device cause more fusion in comparison to autograft control and the fuse bone was not significantly different mechanically than autograft fused bone. These results indicate that the bone induced by rhBMP-2 in combination with ACS and/or the fusion cage is comparable to autograft induced bone and mechanically is not significantly different.

As I have summarized the sponsor has conducted a number of studies to establish the biocompatibility and safety of the product and has used various animal, non-human models to demonstrate bond inducing capability of the product. However, the FDA has two questions related to the safety. Again, these questions don't arise due to pre-clinical and clinical observations of adverse effects due to the product but due to consideration of the potential for adverse effects that might occur.

The questions regard the potential for rhBMP-2 to stimulate transformed cells bearing BMP receptors to proliferate and the potential for an immune response to rhBMP-2 to cause adverse effects in developing fetuses in pregnant women. I'll first go over the question for the potential for rhBMP-2 to stimulate transferring cells in a patient's body.

Bone morphogenetic proteins form a sub-family within a transforming growth factor a super family of cytokines. Cytokines within the TGF beta family and BMP specifically have been shown to play crucial roles in embryogenesis. In addition members of the BMP sub-family have been shown to influence growth, differentiation and apoptosis of various cell types including osteoblasts, condroblasts, neuro cells and epithelial cells. BMP is type 1 and type 2, serum kinase receptors in order to induce cellular signal transduction.

Like other members of the TGF beta family, BMPs may elicit various types of responses in cells due to the cell type and/or receptor type expression. It is reasonable to attempt to investigate the potential for BMP-2 to stimulate transformed cells. Some pre-clinical testing was conducted to address this issue previously by the sponsor. Now I will review the information contained within the sponsor's application that is relevant to the topic. Then I'll go over the additional studies or actually, I'll kind of briefly go over that since Dr. Riedel has pretty adequately discussed that already, the studies that were recommended by us.

First of all, I'll go over the pharmakinetic information pretty quickly. The experimental observations indicated that the systemic availability of rhBMP-2 is low. The prediction is based upon pre-clinical evaluations and assuming a one milligram per kilogram dose suggested systemic exposure to rhBMP-2 would be in the low nanogram per NL range. The experiments indicate that the clearance of rhBMP-2 from the systemic circulation is rapid and that the residents time and tissues involved and clearance is brief. However, individuals implanted with the device will likely have some low exposure to rhBMP-2 outside the implant site.

To address -- to more directly address concerns regarding carcinogenicity or for lack of a better term, tumorigenicity or promotion or stimulation of transformed cells, the sponsor conducted the Ames mutagenicity assay in which they found that the results with that were negative. In addition, they evaluated the product in a one-year chronic toxicity study in the rat and they have evaluated the product's ability to influence the proliferation or growth of a limited number of tumor cell lines and primary tumor cell isolates.

No carcinogenic effects were observed in the one-year rat femoral onlay study. In in vitro tumor cell growth experiments, BMP-2 was observed to inhibit two prostate carcinoma tumor cell lines, one breast tumor cell line, one tongue cell line and one lung tumor cell line and not to effect the growth of four osteocarcinoma lines. In assessing BMPs activity against primary tumor cell isolates, Soda, et al, found that of 65 available specimens, 16 were inhibited. No tumors were observed to be stimulated. In neither in vitro cell study did the investigators evaluate the cells for BMP receptor expression.

We cannot state that these studies demonstrate a lack of stimulatory effects of BMP-2 on tumor cells or tumor cell lines expressing BMP-2 receptors. Traditionally, the two-year rats carcinogenicity study isn't recommended for evaluation of implanted devices. We don't believe this assay would adequately assess for the potential of rhBMP-2 to stimulate transformed cells. In consultation with the sponsor we devised a series of experiments that Dr. Riedel has already gone over.

FDA believes that the studies can be done as a post-market commitment. You will be asked to comment on the concern of the ability of rhBMP-2 to stimulate transformed cells and whether you believe additional studies are necessary.

Now I'll go over FDA's question regarding the potential for an immune response to rhBMP-2 to cause adverse effects. I'll briefly discuss the information contained in the application regarding anti-rhBMP-2 immune response findings and then I'll present research literature regarding BMP-2 knockout mice.

I'll mention what post-market commitments the sponsor and FDA have discussed with respect to revisions of the Elisa used to detect anti-rhBMP-2 antibodies. Finally, I'll tell you what we'd like you to think about in preparation for our questions at the end.

Enzyme link immune absorbants and assays were established to measure anti-rhBMP-2 anti-collagen type 1 antibodies in animals and patients implanted with the device. Anti-rhBMP-2 and collagen type I antibodies were screened on rats, dogs and rhesus monkeys pre-clinically. In the femur onlay rat models, serum samples were obtained pre-operatively at four weeks, 26 weeks and 52 weeks. No rats demonstrated a positive anti-rhBMP-2 or anti-collagen type I response.

In at 28-day daily IV injection beagle dogs' toxicity study, three of eight animals receiving a high does of rhBMP-2 were determined to have a positive immune response. No animals exhibited an anti-collagen immune response. In non-human primate studies the antibody responses to rhBMP-2 were evaluated pre-operatively at four, eight, 12 and 16 weeks post-operatively. Antibodies to rhBMP-2 were detected in 35 percent, 7 of 20 of the animals treated with the device. The antibody responses were transient of a low titer.

No control animals exhibited an anti-BMP-2 response. Eight percent of the animals exhibited an anti-bovine collagen type I response. These studies suggest that immune responses to implanted rhBMP-2 can be expected. The type of responding antibody was not determined in these studies. In addition, we don't know if the antibodies cross react with endogenous BMP.

In addition, the antibody response was undefined as to whether it was of a neutralizing character. For clarification, a neutralizing antibody would effectively prevent the BMP-2 from inducing signal transduction in responding cells. The sponsor plans to revise the Elisa to better characterize the immune response elicited by implantation of rhBMP-2 as a post-market commitment.

In the clinical study the immunologic findings of which Dr. Buch will further discuss, two of 277 patients implanted with the cage contained rhBMP-2 exhibited a positive immune response. One control patient exhibited a positive immune response to rhBMP-2. The incidents of antibody formation observed in this limited clinical study was very low and did not correlate with adverse clinical findings. In other clinical applications of rhBMP-2 and in the pre-clinical evaluations done in non-human primate models, the incidents of the immune response to rhBMP-2 was higher.

FDA's concern about immune responses to rhBMP-2 regard to two issues; throatigenecity (ph) and restimulation of -- and the potential for restimulation of an immune response in the women during pregnancy. This concern is driven chiefly by experimental observations obtained from research literature regarding BMP-2 deficient mice. Again, I want to stress that these concerns were not raised by observations of adverse effects in the pre-clinical or clinical evaluations done by the sponsor.

Also it should be stated that pre-clinical evaluations were not specifically designed to evaluate these issues. The sponsor conducted teratology (ph) and fertility pre-clinical evaluations of rhBMP-2 but these studies were designed to assess if exogenously added rhBMP-2 itself would have deleterious effects on the development of fetuses or if it adversely effected performance parameters of reproduction.

These studies did not investigate whether the deletion of rhBMP-2 due to antigen-specific antibodies would cause embryonic morbidity. In addition experiments were not done to investigation a fetal expression of BMP-2 in pregnant females immune responsive to rhBMP-2 could cause toxicity in the mother. It is important to be cognizant of research and literature regarding the role of that cytokine plays in normal bone physiology as well as in embryogenesis in order to anticipate potential safety issues.

The reason why we pose these questions is based upon experiments conducted in mice deficient for BMP-2 or BMP-2 knockout mice. In these mice investigators noted that a deficiency of BMP-2 was embryonically lethal. The embryos were noted to have failed to close the pro-amniotic canal which cause the malformation of the amniotic cavity and chorionic tissue. BMP-2 deficient embryos also exhibited a defect in cardiac development manifested by the abnormal development of the heart and the exocelimic (ph) cavity. Homozygous deletion of BMP family members has resulted in other embryonic lethal events as well. For example, as shown in this slide, BMP-7 and BMP-2 deficient mouse embryos, BMP-7 deficient mice had defects in the development of eyes and kidneys.

BMPs obviously, play significant critical roles during embryonic development. If antibodies rhBMP-2 were to cross the placental barrier, they theoretically could adverse effect embryogenesis. This diagram captures the essence of the issue. A woman of child bearing potential is treated with rhBMP-2 to fuse vertebrae. The implantation of the cytokine elicits an immune response. During a pregnancy, fetal expression of BMP-2 restimulates the anti-rhBMP-2 immune response would have potentially adverse effects for the embryo as well as the mother.

We would like you to discuss this issue and look forward to your recommendations. Specifically, what type of animal models do you believe would sufficiently address the question of whether maternal antibodies to rhBMP-2 can cross the placental barrier and cause deleterious effects on the developing fetus, also what type of animal models would you recommend to answer the question regarding fetal expression of BMP-2 and its potential for adversely effecting maternal or embryonic development in women who have anti-rhBMP-2 antibodies.

We would like you also to consider the use of a registry for women of child bearing potential in order to monitor for these potential effects. Thank you.

DR. BUCH: Good afternoon. My name is Barbara Buch and I'm the FDA's clinical reviewer for this PMA. I'd like to thank my colleagues for their assistance with this review, especially Dr. Martin Yahiro. As the sponsor has already presented a detailed account of the results regarding safety and effectiveness of the clinical trial, I will not repeat that but I would like to start by highlighting some of the key points relating to the effectiveness and the safety of this device based on the data that was presented by the sponsor in PMA. Then I will briefly review some additional considerations in this and supporting studies in the PMA regarding the interpretation of radiographic data specifically.

Given all this information, I will then as you, as the panel, to focus on the radiographic interpretation issues which will lead you into a discussion of the radiographic panel question. Overall, this clinical trial was well conducted. I'd like to point out that Bayesian statistical analysis were used and there was a high patient follow-up and data accountability. In addition, there was meticulous adverse event reporting. As previously been discussed many times, there were two arms of the randomized portion and one non-randomized laparoscopic clinical trial. I'd like to also point out that the follow-up rates, again, are very high and that in the laparoscopic group the follow-up rate does not take into account those patients who are not yet evaluated for that 24-month evaluation.

Dr. Mathews has already explained in detail these clinical end points that were evaluated to determine overall patient success for the determination of safety and effectiveness of this combination device. For the randomized clinical trial, as I've said, the accountability of patients and the data at 24 months was greater than 85 -- greater than 87 percent. This included the antibody testing for anti-rhBMP antibodies and for anti-bovine collagen antibodies.

In the randomized group there was very little difference in the co-variates between groups pre-operatively. This included a strong correlation of the pre-operative SF-36 obtained in the Oswestry evaluation scores. I'd like to briefly highlight some of the clinical results. In the randomized portion of the trial, the investigational group had less blood loss and less overall operative time than the control group. This is in part attributable to the lack of bone graft harvest as has been mentioned.

The laparoscopic group, as expected, had a shorter hospital stay when compared to the randomized treatment groups and also had a shorter operative time. I'm sorry, the operative time was equivalent. Regarding antibody testing, as been explained previously, the patients were tested for the presence of antibodies to rhBMP bovine type I collagen and then to human type I collagen. In each of the three treatment groups, including the laparoscopic, there was only one patient that had an authentic positive response to rhBMP antibodies for a total of three in the entire clinical trial.

The overall study outcome was a success for the control patient that was positive and failures for the patients in the investigational and laparoscopic treatment groups. However, because of the low rate of occurrence, the significance of this finding cannot be determined. There is also a relative low rate of authentic positive elevated antibody responses to bovine collagen in each of the three treatment groups. Of these greater than 60 percent of the patients in the randomized groups were overall success. No patient in any of the treatment groups has a positive response to type I human collagen.

What's important to know is that analysis was completed comparing clinical outcomes with antibody responses. There were no correlations of any of the rhBMP and bovine antibody results with the overall outcome individual end point success or failure or the occurrence of adverse events. One other interesting final result is that in the randomized treatment groups both sets of patients returned to work in an average of approximately 64 days following surgery and not unexpectedly the patients in the laparoscopic arm returned to work faster than those in the open procedure groups.

This clinical trial has demonstrated that the outcomes for patients treated with the investigational device were as effective as those in the control treatment group. As this table shows, for the results of the primary effectiveness end points, the investigational treatment group were equivalent to the control. The same can be said about the majority of the secondary end points.

When looking at the adverse events in general, the incident of any adverse event in either of the randomized treatment arms was high. This is, in part, due to the detailed reporting of the adverse events and the nature of the surgical procedures. Specifically, there was one death in a patient in the control group who had a history of cardiac disease. The most significant finding was the incidence of graft site related adverse events aside from pain that occurred only in the control treatment group and were absent in the investigational treatment group.

These included fractures, nerve injuries, infection and hematoma. Donor site pain was high immediately post-operatively but as we've seen significantly resolved by six months to a year. Finally, important to note is that there were six pregnancies in women in this clinical trial. Of the five pregnancy in the two investigational device treatment groups, that is the laparoscopic and the open investigational group, there were two early trimester miscarriages, both in the laparoscopic group and three healthy births.

In the overall analysis of adverse events in randomized treatment groups, there were 12 categories which -- in which both groups had a greater than five percent occurrence rate. Of those the investigational treatment group had a slightly numerically higher rate of non-device related events including back and leg pain, GI symptoms, retrograde ejaculation, spinal events, incidents of trauma, one vertebral fracture and urogenital events.

Of these only the urogenital event rate was statistically significant as a difference compared to the control. Approximately half of those urogenital events involved post-operative urinary retention. This is not an unexpected event following spinal surgery. All these events resolved prior to discharge from the hospital.

Other events in this category included kidney stones, bladder and rectal symptoms and erectile dysfunction. These events occurred at least six weeks post-operatively, a period unrelated to surgical procedure. Retrograde ejaculation was documented in five investigational and one control patient but this difference was not considered statistically significant. And as would be expected and has been mentioned many times, the incidents of graft related adverse events were statistically and numerically worse in the control group.

In reviewing all of three treatment groups, retrograde ejaculation was higher in both investigational groups but only statistically different in the laparoscopic treatment group. This is, again, attributed to the surgical approach. There were no directly linked immune related adverse events. There were five possible or potential events that may be considered immune related and none of these patients had authentic positive responses to either anti-rhBMP antibodies or anti-bovine collagen type I antibodies.

There were two cases of cancer diagnosed during the clinical trial. One case of pancreatic cancer was diagnosed in a patient in the investigational treatment group and a case of breast cancer was found in a control group patient. There were no cases of osteogenic cancers reported. The overall occurrence of device related events, as we have seen, was similar between the investigational control and laparoscopic groups. What I'd like to point out is that this includes malpositioning of the device, migration, loosening and subsidence in addition to non-unions. When non-unions are removed from this scenario the incidents is low, falls to less than one percent which is expected for the caged devices.

When compared to the control group, the laparoscopic group had a higher occurrence rate of migration, malpositioning and related anatomic difficulties as well as has been discussed retrograde ejaculation. These occurrence rates probably relate to the procedural approach and are consistent with other spinal literature. The occurrence of second surgeries is similar in both randomized groups. Although the rate of removals was higher in the investigational group, the rate of supplemental fixations was slightly higher in the control groups. The relative rates of occurrence, however, are very similar with eight percent in the investigational group, 10 percent in the control group, and seven and a half percent in the laparoscopic group.

Based on the clinical data provided in this trial, patients receiving the investigational device achieved equivalent fusion and clinical scores compared to the patients receiving autograft control while eliminating the possibility and necessity of bone graft donor site and its attendant morbidity. Again, they were mostly equivalent adverse event profiles and occurrences.

Now, I'd like to turn your attention to the question of radiographic interpretation. This question arises because this clinical trial was the first to use both x-rays and thin sliced CT scans with reconstructions to determine fusion which is an important primary end point. The sponsor has provided the results of an x-ray versus CT validation study as part of the IDE and Dr. Irony will discuss this in her presentation. All the radiographic results of this clinical trial, both plain radiographs and CT scans were presented in order to assess fusion.

One other of our concerns is whether or not we can interpret the radiographs of patients treated with this combination device in the same way as we do the radiographs of those treated with autograft given that the fusion sites may calcify at different rates and the progressive rate diffusion may be different. To this end, I will review the current definition of fusion in this trial, show you some examples of radiographs from this trial and discuss some issues and interpretation within this trial and other studies presented in the PMA.

Plain films were reviewed for the presence or absence of translational motion and angulation. They were then reviewed for the presences of bridging trabecular bone. If there was trabecular bone present and no motion on fluction extension films, the patient was considered to be fused. If there was no bridging bone apparent on the plain films, the CT scans were assessed for bridging bone. If there was bridging bone on CT, no motion on plain films, no lucencies, the patient was determined to be fused. The sponsor utilized both plain radiographs and CT scans to determine the presence of bridging trabecular bone in the assessment of fusion.

There were no instances where there was bridging bone on plain films that was not seen on CT scan. However, there were many instances of false negative plain films, that is cases where the CT scan showed no bridging bone when the plain film did not. I'm sorry, there were no cases -- there were many instances of false negative plain films, that is cases where the CT scan showed bridging bone and the plain radiograph did not.

This phenomena could potentially inflate the success rate in the open investigational group. As seen on this slide, the table compares the determination of bridging bone by x-ray and CT scan at various time points. At six and 12 months, the proportion of disagreement between the x-rays and the CT scans was high. This problem was minimized at 24 months which was the end point of the clinical trial. You will also notice that as the study progressed from six months to 24 in all three groups, there's an actual decrease in the number and rate of patients with bridging bone detected.

At 24 months, however, only 8.3 percent of the patients who were considered failures by plain x-ray became successes by CT. In the control group, approximately five percent of the patients who were failure by plain radiographs, became successes by CT. I'd like to direct your attention now to the next series of CT scans which are examples of patients in the clinical trial.

The slices are taken through the center of the fusion cage and while I realize the determination of bridging bone and fusion are naturally determined by multiple serial axial to sagittal and reconstruction views, these views are an attempt to provide you with some representative examples of the patients in the trial which can supplement those presented by the sponsor. When reviewing these scans, consider, if possible, the progression of fusion and differing densities of the material within and around the cages at different time points.

The first series are patients considered successful fusions in the trial. The second set represent patients who were considered failures. I'm going to ask you if you were able to determine which side represented the investigational device and which represented the control. On the left side of the slide the cuts represent patients who were in the investigational group and on the right represent the control.

Now let's look at what information we've learned from prior animal studies and human studies that looked at radiographic results compared to surgical findings. These studies that looked at animal and human subjects implanted with spinal devices using autografts -- autograft only and then were taken back to surgery. Fusion was determined at the time of surgery by manipulation and histologic analysis and then compared to pre-operative x-ray and CT fusion status analysis.

A summary of these studies showed the CT scan correlated in most cases to the findings at surgery and that the CT scans specifically had higher sensitivity and specificity for determining fusion status compared to the plain radiographs. In the case where BMP was used in surgical fusion in animal subjects, a similar analysis was done at second look surgery to determine fusion and then compared to pre-operative x-ray and CT radiographic fusion analysis.

In summary the CT scans again highly correlated with the surgical findings and histological analysis. In addition, it appeared that the density and rate of progression of repair or remodeling differed somewhat in comparison to what's known about autograft and allograft. In this clinical trial there was an extremely high fusion rate in all treatment groups when using both x-ray and CT to make the determination of fusion.

When considering the data from other studies regarding the radiographic interpretation of fusion, you should bear in mind a few key points. First, in the x-ray and CT validation study that was done by the sponsor, autograft was the basis for the conclusion used to consider this method for determining fusion in this PMA.

Second, we may not be able to extrapolate the information from animal trials to potential human responses. And finally, we need to recognize that potentially the rate and extent of radiographic changes between autograft and rhBMP may differ. With all of this in mind, you'll be asked to comment on the interpretation of radiographic data in this clinical trial. Please keep in mind the following additional issues when commenting on the determination of successful fusions in patients implanted with this combination device.

This includes the presence and absorption rate of the collagen sponge, the identification of progression of bone repair processes in the presence of rhBMP and the ability of bone formed at various time points to accommodate applied loads. And finally, I'd like you to consider the implications of all of these factors on the interpretation of radiographic fusion and physician training in the future. Thank you for your attention.

DR. IRONY: My name is Telba Irony and I'm going to comment on the statistical issues relevant to your consideration for the questions presented by the FDA. I have discussed two statistical issues. I will briefly report about the analysis of safety and effectiveness in this submission and these analysis were made through Bayesian methods. And second, I will talk about the statistical comparison of the use of x-rays and CT scans in assessing spinal fusions.

First, with respect to the Bayesian methods that were used here, as was said before, we have two main studies. One was an open study -- open surgery study which was a multi-center study prospective and randomized. There were like 143 investigational devices and 136 control devices. The second study was a laparoscopic study which was non-randomized and in this statistical comparison we made a comparison with the same control as we did in the open surgery study.

Well, Bayesian methods were used and I will very briefly explain the methodology that was used in this submission. First, non-informative prior distributions were used and I'm just stressing that because usually when Bayesian methods are used, in many cases they use prior information meaning information from other studies. That was not the case here. We computed posterior probabilities instead of p-values and predictions of results for 24 months were made from some cases in which the patients had only some data; in the cases of the patients for which the 24 month values of some end points were missing or some patients that were lost to follow-up and patients that were not yet due for their 24-month visit.

Such conditions improved the accountability at 24 months was already high especially for the open study and as a consequence it enhanced the accountability and improved the precision of the estimates at 24 months. Just to give you a brief idea of what was done here, let's think that for each endpoint, P₀ will be the chance of success for that point, endpoint in the control group and P₁ will be the chance of success for the endpoint in the treatment group. So if you think about the difference between zero minus P₁, you will conclude that if P₀ minus P₁ is large, that will mean that the control is better than the treatment and if P₀ minus P₁ is negative, meaning small, the control will be considered worse than the treatment.

So we are going to look for a large probability that the difference is small enough and that will provide us evidence that the treatment was not inferior than the control and then will declare equivalent. So we are going to look at this probabilities. Small enough will depend on the endpoint. For each endpoint, we had a minimal clinically significant difference and high probability or large probability was in this case 95 percent. So here are the results for effectiveness endpoints for the open surgery control group -- open surgery compared to the control group.

That's the table. For instance, for fusion, the chance of success in the treatment group and that chance is already corrected for the loss to follow up and for patients that for instance didn't come at 24 months was 92.8 percent and for the control group was 88.1 percent. The probability of equivalence, in other words, the probability that this difference was small enough was basically 100 percent. And you can see for all of the effected endpoints, the probability of equivalence was considerably high.

There were two endpoints which was back pain and the MCS. I put a little red star there to say that that probability was high but was not 95 percent. Now, for the laparoscopic group we have a similar table. The values are different and basically for all endpoints the probability of equivalence was larger than 95 percent. And for this group there was a higher -- the group started late, so there were more patients that had not yet reached 24 months by the end of the study, so predictions was -- were made for more or less 25 percent of the patients.

Now the second issue I'm going to assess which is the statistical comparison for the use of x-ray and CT scans in determining fusion. First I'm going to talk about a validation study that was done independently on the submission and second, I'm going to talk about the scenario in the current submission and how it compares to the validation study.

Our problem was the false positive rates. We don't want to have false positive rates because that will inflate the results on fusion and consequently on overall success which was the primary endpoint of this submission. So in the validation study, it was done, I will stress, independently on the study in this PMA. There was a surgical exploration of 53 spinal fusion methods in humans in order to assess sensitivity and specificity of both x-rays and CT scans for determining fusion.

So before the surgical exploration, which was the gold standard; they opened the patient, they could see if the patient was effectively fused or not, the fusion status was determined by both x-rays and CT scans and the relevant parameters evaluated in the study with respect to this PMA are the sensitivity which is the probability of testing positive, in other words, determining fusion when in fact there was fusion, the specificity which was the probability of testing negative, determine there was no fusion when there was no fusion and a false positive rate, the probability of saying that there was fusion when in fact, there was no fusion.

These are the results of the study. These are point estimates. And for instance, for x-rays, the sensitivity was 79, about 79 percent, specificity 86 percent, so the false positive rate was about 14 percent. For the CT scans both sensitivity and specificity were higher resulting in the lower false positive rates. In that study there was a third method to determine fusion, which was a combination of x-ray and CT scan and in that case, the patient was determined fusion -- fused only if both x-ray and CT scan determined fusion. That's a very conservative method.

The conclusion for this validation study is that sensitivity and specificity is higher for CT scans and for x-rays. False positive rate is lower for CT scans and the smallest false positive rate is from the combined x-ray/CT scan method. That's very conservative and was not used in this submission.

Now, the validation study characteristics were different than the ones in the current PMA and that's an important point for consideration for the people that are trying to answer the questions that FDA has posed. First, the patients in this study did not have cages, they did not have spinal fusion cages. The inclusion criteria in this study was different. The patients in this study were patients with continued or worsening pain following instrumented lumbar fusion for instability or degenerative disc disease requiring surgery.

Given that, we will expect that we'll have a higher prevalence of known fusions in this study but however, the distribution was even, where 24 patients that were fused and 29 patients that were not fused. Second, the time period of the exams were approximately a year after surgery. The x-rays examined were flexion extension. There was no presence of BMP is the study and the method of performing CT scans was different than the one in the present PMA.

Now, what happened in the current submission? I will briefly explain how fusion was determined. It was actually based in several evaluations. First, it was based on evidence of bridging bone and the determination was first made by x-ray. If bridging bone was not detected, then CT scan was used and if bridging bone was detected by at least one method, either x-ray or CT scan, then the evidence of bridging bone was considered present. After evaluating bridging bone, these other evaluations were made based on x-ray; segmental stability, and lucent line criteria and in addition if there was a second surgery due to pseudoarthrosis, that will be a failure and the patient was not considered fused.

Now, as a consequence, the actual comparison that we are making is in the methods of detecting bridging bone. The other factors are the same in both methods of evaluation. The adopted way of detecting bridging bone is not conservative because it's sufficient to have evidence of bridging bone with one of the methods. And finally, in the submission there was no case in which the presence of bridging bone was detected by x-ray and was not detected by a CT scan.

So I'm going to present a table in which I will show the disagreement between the x-rays and CT scans in the examination of bridging bone. But before I do that, I will note some important considerations. First, again, in all disagreement cases the CT scans indicate that fusion and x-rays did not agree. There is much less disagreement at 24 months than at 12 months and the relevant point for this submission is actually 24 months. So this is the table that shows the disagreement. For instance, at 12 months there was disagreement in 52 patients out of 130. In other words, 40 percent of the patients there was disagreement between the determination through CT scans and x-ray.

In the control group, that disagreement was also high and the same with the laparoscopic group. When we go down to 24 months, the disagreement is much smaller. It's about seven percent for the open group, five percent for the control group and there was no disagreement for the laparoscopic group. Now, what is the effect of this disagreement in the success rates of fusion? At 12 months, for instance, an open group that will be determined by x-ray to have only 57.3 percent of success whereas if it was determined by CT scans the success rate will increase to 96.9 percent.

The same behavior was seen in the control group; by x-ray the success rate will be about 30 percent and by CT scan will be about 92 percent and the same with the laparoscopic group. However, at 24 months which, again, I will insist is our final and primary endpoint, this disagreement is much smaller. You cannot -- I put in red just to say that in a few cases actually the success rates decreased but they decreases slightly.

Now, what's the impact of that in the overall success. Of course, there will be a large impact as well. The impact will be more pronounced at 12 months with x-rays. For instance in the open group, the success rate will be about 32 percent and with CT scans will raise to 59.7 percent. Again, at 24 months, primary endpoint this difference basically disappears, it was just reduced to a much smaller difference.

As a conclusion, the determination of bridging bone has impact on the determination of overall success and the impact is much more pronounced at 12 months than at 24 months. The validation study was performed in patients at approximately 12 months after surgery and in the study both the sensitivity and specificity of CT scans were higher than for the x-rays, but the characteristics of the study were different than the ones in the PMA and should be taken into account by the panel members. And that concludes my presentation.

CHAIRPERSON FINNEGAN: Thank you. Aric, we're going to actually ask you to do yours after lunch so that we can have the panel looking at questions right away. We're going to have three short, five to 10 minute presentations from our three guest presenters and the first one is going to be Dr. Rocky Tuan who is a -- who has expertise in the BMPs and their biological effect.

DR. TUAN: So the purpose of my presentation is to give you an overview of the biology both in vitro and in vivo, of BMP as a group of molecules. My talk will be divided into four topics. The first one has to do with the protein itself. The second part has to do with its molecular mechanism action. The third part has to do with its biological activity both developmental and also in post-natal pharmacological applications and then finally some discussion on issues related to potential biological complications related to the usage of BMP.

As already has been described and detailed by previous speakers, the BMP or bone morphogenic protein of bone morphogenic activity was actually first discovered quite awhile ago. In fact, more than 100 years ago the work of Nicholas Sen at Rush Medical College who used bone grafts to treat osteomyelitis was in fact, the first indication ever of bone inductive activity coming out from bone itself. The pioneering study of Dr. Marshall Urist, of course, was instrumental in identifying, discovering activity from the demineralized bone matrix. Later on the work of Hari Reddi and Dr. Huggins in the early '70's with the subcutaneous implantation of demineralized bone matrix substantiated all of these earlier findings.

Later on, Urist, Sampath, Reddi and several other groups isolated in at least a certain degree of purity, these bone morphogenic molecules. Partial sequences were obtained and then these were then used to generate nucleotide probes to screen libraries and subsequently these BMPs were cloned. And at this point there are at least 20 BMPs and they belong to the larger family of transforming growth factor beta super family which was originally isolated from tumor cell extracts and called transforming growth factor beta for that reason.

The only exception is BMP-1 which is actually an enzyme. It's a procollagen C peptidase actually, so it doesn't belong in the same group although it's called BMP-1. There are at least four sub-families of the BMPs. There's a BMP-2,4, BMP-3, the OP1 BMP-7 and then finally the cartilage derived morphogenic protein which are also called growth and differentiation factors. These are -- and there are others that have not been neatly fallen into specific sub-categories.

The structure of BMP as already has been described by other speakers. It's a dimer, some time homodimer, sometimes heterodimer. They are synthesized originally as larger precursor forms which are then photolytically processed into carboxyl and into yield a mature product. It contains canonical 7 cysteine residues, one of which is involved in a very critical disulfide knot (ph) structure important for its activity.

Okay, the second part now, the molecular mechanism action; BMPs because they are in the family of TGF beta super-family, they are actually signalling molecules, that's the best way to describe them. They interact with cell surface receptors. There are at least two receptors, each of which, of course, also has cousins and relatives. These are the BMP receptor 1 and BMP receptor 2, both of which are also enzymes. They are kineses.

Upon binding of BMP to the BMP receptor 2, BMP receptor 1 is then phosphorylated. This then activates a whole series of signaling events. This receptor complexed and interacts with a family of molecules known as SMAD or S-M-A-D. MAD stands for Mothers Against Decapentoplegic, which is actually a fly protein and which is also a member of the BMP super-family. Again, I just want to also reiterate that the BMPs have been found across all species so far.

So at any rate, the SMAD, these are the signaling partners, SMADs 1, 5 and 8 interact in a sequential manner. Later on they then interact with a common partner called SMAD 4, which then removes the entire complex into the nucleus of the cell to activate specific genes, so that's basically how this whole thing works. There are also anti-SMADs. They're also called SMADs. They's called SMAD 6 and 7. So these are the inhibitory SMADs. Now, I want to emphasize one point in this description of the molecular mechanism of action is that this signaling pathway does not work in isolation.

Nature being the way it is, there is tremendous cross-talk between one signaling pathway and other signaling pathways. Recent data from many laboratories have shown that this particular -- from the outside of the cells to the surface receptor and then to the nucleus, this particular pathway, this access, actually interacts with other axes, such as the extracellular matrix and receptor complex signaling mechanism via intergrins and fibronectin and collagen, et cetera, as well as most recently the wind-signaling pathway. Again, that's another distant relative of some fly protein. Again, the function there is to stimulate cell growth, proliferation, differentiation, morphogenesis, et cetera. So when we consider BMP family as a family of signaling molecule, we have to consider what they also do with other signaling pathways.

All right, the third topic is activity, developmental as well as post-natal. One thing I want to emphasize again is that BMP action is highly cell specific. There we should get it out of our mind that there is one cell that is -- all cells react to BMP the same way. That is absolutely not true. BMP -- the same BMP can do different things to different cells depending on its own repertoire of receptors and also all the signaling molecules inside the cell.

Now, during development as Dr. Hudson already pointed out earlier, BMP is absolutely crucial for survival and development of the fetus. The evidence is ample at this point. If you have a knockout, i.e. deletion of BMP genes, you will get embryo lethality. If you also are missing, again by transgenic methodology, the receptors for BMP, you also get embryo lethality. So now, remember these things happen way before there is any bone or cartilage or anything like that.

So where BMP works is actually depending on the developmental stage of the animal. For example, during very early development, during gastrulation, BMP works be defining the polarity of the animal, particularly the dorsal-ventral polarity, who's on top, who's on the bottom, and again, nothing to do with bone or cartilage at that point.

And a very crucial example would be the specification of the formation -- of the development, the differentiations of the paraxial mesoderm. In fact, those are the cells that ultimately give rise to the spine. So before you even have a spine, the cells that are precursors to the spine already are responsive to BMP. Whether it's in the same manner or not, we do not know at this point. Nevertheless, if there is a perturbation in the distribution of BMP, the dorsal aspect and the ventral aspect would get all mixed up.

In the development of a limb, which of course has bones and cartilage, et cetera, again BMP is very, very crucial. It is important for the initial differentiation event of forming cartilage as well as the death of cells in the interdigital area. So it can be a positive factor in the sense that it makes cells, becomes something else or it can make cells undergo apoptosis and die so that there will be a space between the fingers.

And also member of the BMP family, TGF-5, is absolutely crucial for joint development. Without that, there would be a decrease in the number of joints. There would be a fusion of joints, mainly the joint never will actually form in the first place. So that's developmental.

In terms of post-natal, we have already seen many pieces of data on the results of pharamocological application or therapeutic application of BMP to a post-natal animal, the historical findings of osteochondroinductive activity, of course, is obvious. Fracture repair, critical size defect, feeling, spine fusion, we saw many pieces of data and also recently in terms of osteointegration and that is how bone cells interact with an implant bio-material. In this case it may be relevant to the topic at hand, namely that there is a piece of alloy that is part of this device.

Now, how the presence of BMP may influence the interaction between the neighboring cells and the metal alloy is something for us to think about.

Finally, potential biological complications; one thing that needs to be considered is the distribution of the BMP that has been placed into a particular site and also retention of this material once it has distributed to other places. We notice the low systemic level. However, exactly where the BMP finally ends up is something that perhaps to think about.

Tumor induction in terms of either inhibitory or stimulatory in terms of cell proliferation for transformed or not yet transformed cells, hematological perturbation has been shown for the grandfather of the TFG beta super-family, TGF beta-1. The presence of TGF beta-1 can cause hematological perturbations.

Teralogical effects, which will be discussed later, for example, whether the BMP can cross the placenta, the window of time in terms of its action, whether the effect can be transgenerational is also something worthy of consideration and finally immunoreaction which has also been discussed, so I think I'll stop here.

CHAIRPERSON FINNEGAN: Thank you very much, Dr. Tuan. Now, Dr. Miller is going to talk to us generically about animal models and the use of registries for teratogenicity. While he's setting up, the really bad news is we do not need to have a closed session, so we will go to lunch once Dr. Kostuik has finished his presentation. We do need to start back immediately at 2:00 o'clock, so I would ask you to make it a short lunch. Consider it a way to get over your Christmas indulgences.

DR. MILLER: Good afternoon. I'll try to make it brief. You have a handout so some of the slides have been eliminated in the interim to shorten it a bit. I've been asked to look at some principles and concepts. The developmental toxicology animal testing is how it is done. The evaluation of biotech products which really introduces a whole new question into the reproductive and developmental area and then post-marketing pregnancy registries and I'll try to do that all within 10 minutes.

I am offering a course at the FDA next month that takes four hours to do this, so I'm putting it in perspective. Different molecules will produce a different spectrum of malformations depending upon the mechanisms of action as one sees here with Accutane, with valporic acid in terms of spina bifida and methyl mercury down below. We know that the pregnant female is a pharmacologic orphan. We don't study products in them by choice necessarily and that we really are treating two different organisms and that's actually one of our questions.

Dose is the problem, threshold concept is critical to the area of developmental toxicology and usually we say where is that threshold there is a toxic effect? Well, in many instances when we're looking at it there are doses that won't produce that toxicity and maybe that's what we should be looking at as well. Along those lines, dose, length of exposure and time during gestation are our critical areas of evaluation and this will certainly become more important when one is talking about different surgeries and what time they may be occurring in a woman if she is pregnant.

There are certainly sensitive windows along these lines and I've borrowed this from Keith Moore, where you see in red here many different organ systems that are sensitive from the brain down to limbs and other organs. If we move just a little earlier, the whole pre-implantation, implantation area is one also of equal concern. I added this just at the last moment. I had removed it. However, when we are thinking of mechanisms of action, the National Academy of Sciences just completed their work in the past year and published a volume called Risk Assessment of Developmental Toxicology, which I recommend to you and have been using the 17 different signal transduction pathways to look across species to see the commonality for agents and I do have an asterisk up there; one for transforming growth factor, because obviously, that is the family cluster in which BMPs are found.

So if these are some of our concepts, what are the animal testing that we must undertake and follow guidelines on? And there are three areas, one being fertility and early embryonic development, usually performed in the rodent, the rat; embryonic and fetal development, sort of the teratogenic period, that is done in two species; one the rat, the other the rabbit or some other species that is a non-rodent.

And then some pre/post-natal developmental studies to look at behavioral and functional changes. Along those lines, studies are conducted according to these guidelines and these are sort of the minimum requirements that we really need to think about. Treatment via the likely human route with obviously some exceptions but we would like to perhaps more closely mimic that.

Results should permit identification of a NOAEL and a LOAEL and also look at a maternally toxic dose. Maybe the compound one is looking at doesn't get to a toxic dose that one can demonstrate in the mother but at least you've tried to go there. Extrapolation from the most sensitive species unless there is evidence that the species is inappropriate might, in fact, be reason to look at another species based upon drug metabolism, formation of reactive metabolites, and then obviously, we need to have clarity of a thought about detection versus characterization.

If these are our primary roles, then we're looking at the identification of hazard and we're looking at the identification of risk. And along these lines, hazard really is an animal studies agenda, especially with a new agent because we don't have human case reports to rely upon in many instances and if we generated that, the dose response model from animal studies is what we would need to utilize in trying to identify risk. And probably the most important phenomena we need to consider here is biological plausibility.

And biological plausibility really encompasses all of what we know about pharmacology and about actions in pregnancy and is that, in fact, reasonable what you are seeing. So here in terms of the risk assessment model we're talking toxicokinetics, toxicodynamics and outcome and the toxicokinetics becomes an extremely important component here to try to extrapolate among species.

Range finding studies are important as well as doing the definitive study as listed below here and in that one would certainly stop at delivery, actually before around 21 days to do a Caesarean section to evaluate the animals and in other studies doing the follow-up post-natally to see survival and what may, in fact, be happening in terms of behavior. In the rabbit, one usually starts with the non-pregnant dose ranging study and then does a pregnant one to see if there is enhances sensitivity and then going on to embryo fetal developmental studies, exposures between seven and 25 days to look for those types of problems that may be associated with the agent under study.

So selective reproductive toxicity can be detected in range finding studies and if there aren't any, usually one would perform according to the ICH guidelines and do sort of the minimum requirements. However, if you have found toxicity, you would certainly want to characterize it better by doing the dose response relationships looking at the critical periods and looking at the adverse effects in terms of both structural and functional but also emphasizing again, this toxicokinetics in terms of the critical periods and determining whether the agents are formed and delivered in fact, to the conceptus.

In vitro models can be used to better understand the toxicity itself. So in an identification of risk, we have a decision tree and these you need to keep in mind. Is the species appropriate for risk assessment? Are there differences in pharmokinetics and dynamics among species especially trying to extrapolate to human risk? Are there differences in reproductive physiology or timing of development which should be taken into account? What is the likelihood humans will be exposed and treated under the conditions one has used in these animal studies for critical periods and characteristics of the patient population?

And finally, conclusions about the potential of the agent to produce reproductive toxicity, is that similar to other agents and in this case you sort of have a unique agent and you may not be able to look across therapeutic classes. One of the main questions here is though early in pregnancy and I show here a Doppler where you can see the heart in red here in the early embryo and you can also see the blood flow in the mother.

Along these lines, this is the time we really want to know what is going on inside. Are we, in fact, having our compound arrive at the site? Well, here in a more schematic diagram we are looking at the early embryo and we can see the exocoelomic cavity and the amniotic fluid cavity. Now, these, in fact, can be sampled very early on and Eric Jauniaux over in London has done such. And I want to share just a bit of data with you to help put some of your deliberations in perspective.

Now, neither you nor I can read this from this distance but what it indicates is that mother villous tissue, decidua and embryo all are producing different products. And, in fact, if we sample them, can we find some of those products there? And Eric has done a very nice job of looking at this and I have circled two in particular, IgG and IgA. And if you look at the maternal serum, you will find that there is substantial IgG there and also certainly substantial IgA.

If we though, look at the celomic fluid, we find very little IgA and non-detectible in amniotic fluid but still the IgG tends to get across. And one of the reasons for this is, in fact, that the placenta does have receptors that will allow for that to occur. In this particular study you can see also that HCG is produced and tremendous amounts are ending up in the conceptus.

So one can say do IgGs get to the early embryo and the answer is, yes. In terms of our evaluation of biotech products, we're really on the horizon here of unique agents to try to evaluate and how do we try to assess that? Well, we have many new proteins. Now cytokines are not new but Fab fragments which are used clinically such as RheoPro have, in fact, been unique agents. Do they cross into the conceptus, anti-sense (ph) compounds?

The issue here is the difficulty in testing in rodents and lagamorphs because if it's a human protein you might get a heterologous response. Also extrapolations from human are often difficult to do and the biological plausibility issue comes forward with SARs and pharmacokinetics and pharmacologic action. Well, what can we do to look at this and I'd like to leave two thoughts with you along those lines. One is again, returning to the non-human primate model, and in the non-human primate model there are some studies that have been undertaken with recombinant cytokines, G-CSF and stem cell factor and they have looked at these in terms of whether they transit the placenta and enter into the fetus.

And to abbreviate which was a very extensive study, is that when they administered it to the mother and these were recombinant cytokines, they found insignificant transfer even though they had huge concentrations of materials in the maternal circulation. They also did not find any fetal effect. However, when they directly administered it to the fetus, they found rapid rises in fetal neutrophil counts. So you can see that in terms of some agents, they may or may not be transiting and having fetal impact.

So the non-human primate model may be a good one to explore in the reproductive area as well. Another possibility is looking at human placental profusions where you isolate a lobule and you have a maternal circulation and a fetal circulation and the maternal circulation you can add your agents to and see if they cross into the fetal side. If they do, you can quantitate how much one has seen. I just give you a list of examples here of a few that might be of interest to you and here is G-CSF again and you find very little really transferring across in four or five hours in the Gregor study published a couple of years ago.

Interleukin-8 was not detected when it was given to the maternal side. Epidermal growth factor had a very, very small amount of the maternal dose appear in the fetal side, as also with recombinant erythropoietin where they really did not find any significant levels on the fetal side. And the same with the RheoPro which is an FAB fragment. So there are examples. Now, how do you explain this particular issue and one is the placenta is a wonderful machine for breaking down proteins and if you aren't bound to a receptor mediated process, such as IgG being bound, the rest of them may, in fact, be catabolized and reutilized as basic amino acids to support the placenta as well as the embryo fetus.

So many of these are being broken down and recycles. So if these are going on, then if we make a decision that a compound has survived all of the pre-clinical analysis, what do we do in terms of post-marketing evaluations. And in fact, this has been reasonably undertaken in about the last 10 years mainly because of Accutane. And that has been the most rigorous registry by the Boston group headed by Alan Mitchell but along those lines, these are some of the issues that one needs to think of whether you're doing an in-house one using the university center or using the Organization of Teratology Information Services.

But along these lines, pregnancy registries are epidemiologic studies and therefore, you have to think about what is an appropriate control group when you are initiating those? Prospective identification of exposed pregnancies based upon voluntary contact from patients or health care providers, i.e., the referral bias that can often come in; follow-up of exposed pregnancies to obtain complete information because often times, unfortunately, at the FDA as well as at many pharmaceutical companies, that information is not completely collected.

When should pregnancy registries be conducted? And perhaps the best time is when a new molecule entity is being introduced such as live virus vaccines or an entity like this one, BMP, expectations of high use in pregnant women or of women of child bearing age, agents necessary for conditions associated with high morbidity or mortality that cannot be discontinued as a recognition of pregnancy and an agent's suspected adverse effects during pregnancy based on SAR, pharmacology and laboratory findings and agents known to be harmful during pregnancy which obviously was an issue with Accutane when it was first marketed.

So what design of the pregnancy registries; one, you need a preparation of written protocol, you need to explore consent issues, you need to have incentives to promote voluntary reporting and have a preparation of information documents which is a means to communicate with the public and also have a major advertising campaign. And with these, the initiation at the time of marketing for new products is probably the optimal time.

Window of opportunity for collecting reports is really within the first five years; quickly identifying any potential risks and not likely to be used for agents on the market for an extended period of time due to diminution of voluntary reporting. So with that sort of background, the next part that obviously needs to be done is to provide annual reports, publish occasions of interim results and really get the care providers involved.

So I hope we've reviewed that in as quick an amount of time as I can possibly do.

CHAIRPERSON FINNEGAN: We thank you both for the quality of information and the time. Thank you. Dr. Kostuik? Dr. Kostuik is going to give us generics on different ways to image the spine.

DR. KOSTUIK: Thank you, Dr. Finnegan. My name is John Kostuik. I'm the director of spinal surgery at Johns Hopkins and as I was preparing this talk in the last few hours, since I was only notified that I would be asked to do this yesterday, I realize that I have been a spinal surgeon for 35 years. That, in itself, is frightening since it constitutes approximately 40 percent of the time since spinal fusion was first introduced to mankind so I am getting old.

What is the problem in assessing a fusion? It's a difficult proposition. It is far more difficult today than it was when I was an embryonic spinal surgeon because of the advent of rigid internal fixation. This is particularly true in posterior approaches to the spine. There is no doubt today that the implants used are much better than they were even seven to 10 years ago. It has been relatively easy to assess fusion from an anterior interbody approach up until the development of metallic interbody cages.

It has been stated that expiration of the fusion mass is the gold standard. One, of course, would be very reluctant to explore a fusion mass on the anterior side of the spine because of scarring and potential risk to vascular structures could result in serious problems. But even on the posterior side of the spine, it is not a very valid technique because a bone graft regardless of how it may be stimulated, may fuse to itself but not fuse to the underlying host and therefore, you have a pseudoarthrosis which is very, very difficult to ascertain in any form of radiological investigation, including three dimensional CT scanning.

When should one do plain x-rays following a spinal fusion? I think that they should be done, of course, immediately post-operatively, probably at about 10 days after that. Those are for medical-legal reasons and not pertinent, I think to this group; then I think probably at about eight weeks, four months, six months, nine months, one year, 18 months, two years, after that depending upon the problem, for instance deformity cases, on should follow those for life.

Now, it has been stated also that flexion/ extension x-rays are valid. They are not valid since the introduction of rigid forms or internal fixation. If you have gross motion on flexion/extension, that is fine but generally that is not seen particularly with modern implants. Therefore, I rarely ever recommend their use. I do recommend, however, at least four views, including oblique views, which are frequently not done but which can be of great value in looking at the spine from other angles.

There are many, many false positives with flexion/extension x-rays which brings us next to the question of CT scanning. This has been controversial as well. There is no doubt that I think that the CT scans as presented here today are of more value with the sponsor's product than it would be with autograft since there is no intercage density at the beginning. However, it isn't the beginning that we are interested in. It is the end point that we are interested in and we do not know or at least I don't think there is sufficient data to tell us when we see significant enough ossication or it has been called bridging within the cage. Moreover, the bridging we see is within the cage only.

There has been studies done, including some of our own work, to show that if you take a cage and fill it with bone and take a CT scan, the first day post-operative, it looks like it's fused. So I think CT scanning has a lot of potential error and although I do admit that it may be more valuable with this product.

What about the type of CT scans? There's no doubt that the modern new thin scans are much better. There are -- there is software available that allows you to subtract the metal artifact. However, it is still inaccurate and recently has been thrown out in a court of law where it was presented as a means of assessment in a particular patient where a legal suit arose. Probably one of the most difficult radiologically to tell whether or not you have a fusion is the question of radiolucency or loosening. Particularly loosening I think is obvious. Radiolucency around a device screw cage can be a bit more subtle. It is my opinion that any lucency means non-union if there is still the presence of pain which may be different than the original pain the patient presented with.

If there is a lucency and pain then I think investigation probably should lean towards injections of dye to see if it flows freely or semi-freely and, perhaps, the addition of local anesthetic for pain relief. Bone scans are very little value, radioactive bone scans are very little value in assessing fusion mass since most fusions will remain warm or hot until two years approximately two years from the time of fusion. Hot after that probably indicates a pseudoarthrosis at least unless proven otherwise.

The real way, I think with anterior to body fusions with metallic devices to tell whether you have a solid fusion is the presence of an anterior sentinel graft, that is anterior to the cage. Now, the big question comes up is how thick should that anterior bridge be and that is not known scientifically. I have certainly had the personal experience on many occasions to remove a few millimeters of solid looking anterior bone definitely fused to the vertebral bodies to find a significant underlying pseudoarthrosis.

So I think we do need to know the answer to how thick should an anterior bridge be? Should it be eight millimeters, five millimeters, 10 centimeters -- 10 millimeters rather. I don't know. A question also which has not been announced and I'm perhaps unclear here is, what about more than one level of fusion, if the devices are to be placed at more than one level because it is far more common to do a two-level fusion than it is to do a one level fusion and then when do you decide when you have a pseudoarthrosis.

That is about all I have to say. Thank you very much.

CHAIRPERSON FINNEGAN: Thank you so much. All right, we will reconvene at 2:00 o'clock.

(Whereupon, at 1:38 p.m., a luncheon recess was taken.)

A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N

(2:13 p.m.)

CHAIRPERSON FINNEGAN: Ladies and gentlemen, I do think we are ready to begin if you would take your seats. Dr. Reddi, we'd be waiting for you. We are now going to have the panel reviewers give their presentations. The panel reviewers are three. Dr. Hari Reddi is going to give his review of the pre-clinical. Dr. Kirkpatrick is going to give his review of the clinical and Dr. Larntz is going to give his review of the statistics.

DR. REDDI: Madam Chairman, I would like to do you a favor and finish my comments in five minutes. I'll give you an extra five minutes.

CHAIRPERSON FINNEGAN: Well, your comments are very valuable, so you actually get as much time as you'd like.

DR. REDDI: I just want to help out the Chairperson. First, as far as the biological problem is concerned, we are going into a stage where how can we improve on nature, that is bone itself. So the autograph has three main ingredients; the cells, the matrix, and the signals. It turns out these are the same three ingredients which are needed in manmade approaches for tissue replacement, what is generally referred to as tissue engineering.

Now, I want to make some comments concerning this isolation of BMPs and some of the questions which were posed by FDA and what comments I can make in relation in general, not specifically. First, the BMPs during evolution they have been found not just in mammals but in other organisms. Let me tell you what I mean by that. First, there are several ways to get out the signals during development. One can isolate it from a tissue like bone. One can look at other organisms and find out if they are found in other tissues.

In the case of BMPs, it happens to be a rather unconventional method in which signals were isolated from bone which is really one of the ingredients in the graft itself, so bone morphogenic proteins are natural substance. Dr. Tuan mentioned the history for about 100 years Nicholas Sen. I want to take it back further that the idea that normal substances present in human body could be used as therapeutic agents was first propounded by none other than our old friend Hippocrates. He said 23 centuries ago the best therapeutics are therapeutics mined from the human body.

Next. These signals, although isolated from bone, it turns out in this case was later described in other organisms. This addresses the issue of possible immunity. Of course, I can see one should keep in mind but I want to make a point that because it has been conserved in evolution, the insect BMPs could be equally effective and safe in humans. The other point which I want to make upon which considerable effort has been directed mainly because of some misunderstanding about the transforming growth factor, it terrifies people because it is something called transforming growth factor and if one reads the literature, it's only the name and the name is a misnomer.

Although it's considered as involved in cancer research, it is really due to an in vitro artifact that this particular factor was called transforming growth factor. In effect, all these proteins are very important to normal physiology. So the point which I want to make is, one needs to be careful but not to worry too much about the teratogenicity which, of course, will come up. I'm making general comments, not specifically for a particular BMP. I'm talking about signaling molecules. These are normal constituents to the human body and the word here is moderation.

And the point which I want to make in response to certain tests is again, one needs to really look at the physiology. This is normally present. I want to add one other point. BMPs are not just in bone. It turns out this particular family of molecules are involved long before bone formed during the pattern formation. That is, for example, your right hand and left hand are the same, yet the DNA is the same but the right hand is a mirror image of the left hand which is involved in a subject called pattern formation.

Simply put for a lay person, if you want to build a building, you need to have architecture. This is usually referred to in the parlance of developmental biology as pattern formation. The point which I want to make is long before bone and cartilage appeared, BMPs have arose in defining where particular tissue forms including that your heart is placed on the left side so that is very important, the pattern formation having that it is involved in the initial differentiation of tissues such as cartilage in the lumbar. For example in the developing spine it is involved in the early stages.

So in effect, what we are looking at in terms of repairing is really a recapitulation. Finally, these same molecules play a very important role in the maintenance of tissues and finally in the regeneration and repair. So in summary then, in conclusion, I want to point out that BMPs are normal natural substances, normally found in the human body and it is not surprising they may have therapeutic implications.

CHAIRPERSON FINNEGAN: Thank you very much. Dr. Kirkpatrick.

DR. KIRKPATRICK: Thank you. I was beginning to fear that not only did I need to have prompting for what I was going to say but I was going to prove the orthopedic surgeon still is not technology proficient but fortunately with the aid of our gentleman there.

I'm going to provide an overview from a clinician's standpoint of the clinical results that we heard about today. The device basically we know pretty well by this time. It's a combination product with a titanium fusion cage, recombinant human BMP-2 and an absorbable collagen sponge. The study groups, from my standpoint, appear to be reasonably similar. There was an open arm with a surgical placement of the device for the control, fairly evenly divided and randomized.

The laparoscopic arm was basically a non-randomized group but compared again to the controls that were done. There was -- it was multi-center, which is one of the things that we always like to hear and there were roughly 135 patients in each group meaning the numbers were reasonable from my standpoint but, of course, the statistician will add to that and blinding wasn't possible, obviously, because of the graft site implications.

The indications for the surgeries were degenerative disc disease at L4 to S1, single level disease, Oswestry scores of greater than 135. Spondylolisthesis had to be grade 1 or less. For those of you that don't recall, that means just a little bit of subluxation but not more than 25 percent of the size of the vertebral body and then failure of non-operative treatment for four months, although I heard it was six months. So maybe they could make sure that the paperwork -- just confirm whether the paperwork is right or what you said was right would be appreciated.

I do want to remind the panel that many of us have been away from spine surgery for a little while and maybe don't do spine surgery. Degenerative disc disease as an indication in itself, is a fairly controversial subject. Degenerative disc disease is thought to be a source of low back pain. Unfortunately we cannot attribute it specifically as the source of low back pain. And so when we look at clinical results in here, 70 percent range, we wonder if that means that they're really doing better for back pain. So from a panel consideration issue I think we need to have the clinical understanding that most of the patients that have degenerative disc disease and are having fusions for degenerative disc disease, it is for back pain and not for the disease that they're getting fused.

Many people have degenerative disc disease without having evidence of discomfort or clinical limitations or functional limitations. And in fact, one of our panel -- I mean, not our panel experts but one of the applicant's experts, in fact, has published well on asymptomatic degenerative disc disease. So we can't equate fusing degenerative disc disease with eliminating pain which may be a consideration that we have to raise later with our discussions.

At any rate, the other indications are all appropriate and for their device it may have been the only reasonable one that they could get us a defined population for. Sorry, I went the wrong way. They excluded appropriate things, prior fusion at the level, significant co-morbidities, inorganic behavior, that's people that are trying to fake out the doctor about their pain, for those of you that aren't familiar with what inorganic behavior would be, substance abuse and then a few others that made a lot of sense as well.

The patient populations were very similar in the three groups and so beyond just stating that once again, I don't think we need to belabor that point. The primary outcomes, radiographic fusion, Dr. Kostuik pointed out exactly why I put it in quotes. I don't think we need to go on with that. The Oswestry scales, the neurologic function and the overall success were the primary outcomes. Secondary outcomes, of course, were the back pain, leg pain, the PCS and MCS of the SF-36 and then they also had another issue of this type early on to see about subsidence, I believe was the main target of that.

Basically, they found either equivalent or superior results. I would attribute the Oswestry superior rating there as being simply because they didn't have to take the bone graft is in my mind one of the bigger things as well as the limitation of the surgical exposure. Surgical exposures do lead to some patient symptoms and some problems down the road and I think that's the main reason that we're seeing a difference there. I don't think it's attributable to the device itself.

As far as the secondary outcomes that they were looking at, again, we didn't find a great deal of benefit there or difference there between all the groups, but you do notice that there was really in the open procedure for back pain we're getting back into that clinical problem, I think, where even the morbidity of the exposure may be enough of a problem to give them a bad result from a back pain standpoint.

And then, of course, as you see the PCS on the SF-36 was superior. Antibodies, I really didn't find anything in the data to attribute a problem with the antibodies either from the BMP or the bovine collagen antibodies. I will be asking the applicants if they wouldn't mind taking the question now and writing it down so that I can have an answer later. I'm just curious as to why -- how would you explain the fact that you would get bovine collagen antibodies in the control group when they did not apparently get exposed to the bovine collagen? And I'm sure you have an answer prepared for that. I'd just like to hear your explanation.

Let's see, so in summary, the device in my review of the data appears as safe as a cage with autograft for a single level fusion in degenerative disc disease. It appears to be as efficacious as well again with the quantification of the fusion data and I really have to reserve the comment on some of the embryological effects and that sort of thing for our other experts on the panel.

Discussion issues again, will be a radiographic evaluation of the cage fusion from my standpoint, the immune response, the tumorigenicity and the time course of the BMP-2 presence and just to reiterate again, the quality of the films that we were provided on our CD roms, failure or success, I have no idea how to tell this as a clinician, which one is the failure and which one is the success.

I'd be interested to know if the audience could pick those out. Similarly with the CT scans, I'm not convinced that they could tell me which of these were study or control and so from a radiographic standpoint, we have those questions and I thank you for your time.

CHAIRPERSON FINNEGAN: Thank you very much. Dr. Larntz.

DR. LARNTZ: I'll just sit here if that's all right. I've got a few comments, actually not very many. First, let me say I don't know if anyone said it but compliments to the company and the FDA for their presentations. I thought they were very good presentations and very clear and well, I'll just say that, that's true.

Compliments to the company for their randomized open study. The device clearly meets the criteria set up in the protocol. As a statistician I should just stop there and maybe you should say please do. Not only in inferiority or equivalence or whatever you call it, the device clearly establishes that with respect to the primary endpoints and actually comes close to this from the secondary endpoints. I mean, that's pretty good. Some of us have been involved in studies where that's not always true. So I think it's a very clear case that as far as the protocol criteria is concerned, the device meets the specifications.

Again, compliments. The documentation, has anyone seen what we have up here? The documentation is actually quite extensive, but it actually gives me good confidence in what was done and even down to providing some interesting programs which gave me great confidence in how the calculations were carried out. I appreciate that.

Okay, enough of that. Now, one issue that seems to be coming up are CT scans and x-ray. I'm a statistician. I don't know anything about those things. I don't know anything about lots of things but I do know it's important to take measurements in a blinded fashion and to have adjudication and remeasurement and a process that allows multiple looks at these x-rays or CT scans. It appears to me that the company did that. It appears to me that with respect to some issues, with respect to that, it doesn't matter much in the sense that the -- all of the conclusions wouldn't change whether you use the x-rays or the CT scans.

So I feel very comfortable that with respect to the clinical aspect, we can talk about what fusion is or is not, I'm not going to deal with that. That's not a statistical issue but I don't think there's an issue with respect to the data or the conclusions with respect to that issue. So I would like to say that the study conclusions don't depend on that issue. How is that, don't depend on whether CT scans or plain films are used.

A terribly small point and this is only because the documentation, there are some missing data. It's not all perfect. The world sometimes doesn't collect all the data and there are some methods given in the documentation which no one talked about today which is fine, called intent to treat and I'm just going to say for the record, that particular method of dealing with missing data is not appropriate. We need to do some kind of sensitivity analysis, that is some kind of analysis that says, if the control group did better than the treatment group by a certain amount, would it change our conclusions. There wasn't, as far as I could tell, and it's a question, if any sensitivity analysis of that sort was done, I didn't see that and I think we saw a lot of what was done.

But I don't think it would change the conclusions much in this case because again, I think, as I said, the criteria set up and protocol, were met and met fairly convincingly. Another question which is I don't know how much analysis was done of covariates to find out if there are sub-groups of the patients that do better on one -- on the device, not comparatively but are there sub-groups of patients for whom this procedure is better indicated, "the overall success rate", quote, unquote, I realize it's a very stringent criteria, but they're not very high, so are there ways to pick out sub-groups that have higher overall success rates for instance, and I don't know the answer to that.

Again, a small minor point, this is not necessarily for the company, maybe for the FDA, a number of these measures are scales. The Oswestry, is that how you say it, is a scale, back pain are scales and they're analyzed as success/failures because of some arbitrary criteria and actually the company mentioned -- I was actually pleased. The company mentioned that they might have some slight difference on back pain because they actually have a higher average but their success proportion was lower.

And I think that when we have scales like that we lose information when we go to arbitrary cut points. That's a comment, it's actually a comment I made before and it probably -- if you're on this panel again with another study, it will be a comment I'll make again because it seems there seems to be this drive to call it success or failure rather than trying to measure the size of the effect and the amount of change, and I'd prefer to see the continuous variables analyzed as continuous variables and there was, of course, some of that analysis done in the reports. Okay. So that's the open randomized study.

Now, the laparoscopic study is not a randomized study. It's an extra arm, right? It was done as an extra arm and when you have -- but the comparison was done to the control group in the open study. When you do that, you've got to be very careful. And here I think we haven't been so careful because I do think there are differences between the patients. It was minimized, I think, at various points, but there are differences if I read it right and, of course, again, I'm just a statistician so I might not have read it right.

But if I understand right, the laparoscopic patients had less previous back surgery. What does that mean to me? Well, they haven't been in to the doctor so much. I mean, that's good. They're maybe at an earlier stage because they're on more non-narcotic as opposed to strong narcotic medications. Okay. They have better pre-op neurological scores, better SF-36 MCS scores, better leg -- they look like they're a healthier population, just from my perspective as an unbiased -- well, I think I'm unbiased, as a statistician looking at it, how is that? I'll say that.

Now, if you have differences in a group and you want to make a comparison to control and you have differences of some sort, you really should think about making some adjustment for that and I didn't see any adjustment. So that's a question I have. The results look very good. In fact, in some cases, they're called superior but I don't think any of those analysis I know of made adjustments for perhaps that the laparoscopic patients were better to begin with. I don't think that happened.

So with respect to laparoscopic, I think we're in a situation where there is a possibility that an adjusted analysis would change the conclusions somewhat but I don't feel, given they had reasonably good results, I don't feel it would change it so much that we should discard the good results we have. You do have some what I would say advantage there because at least the laparoscopic patients that were better turned out to do better, that's at least comforting. But I do think an adjusted analysis might give us a slightly different story there. And so on that note, I think I'll stop.

Thank you.

CHAIRPERSON FINNEGAN: Thank you. That was, as usual, very enlightening. What we're going to do now is I'm going to ask Aric to put up the questions and then we're going to go around the panel and I'm going to ask each panel member with their expertise and their comfort level, to discuss their concerns, questions and perhaps, some thoughts they have on what's been presented so far, and just to warn you, Stephen, we're going to start with you.

DR. KAISER: Okay, as I mentioned earlier, we've got some general topic areas where the questions are coming from and as we move into the first area, reproduction, teratogenicity, we have several questions that we would like you to discuss in this area and would you like me to go through all the questions in the topic and then come back to the beginning? Okay.

In this area, the first thing we'd like you to do is discuss the potential for an immune response in the mother to effectively block BMP-2 expression in the developing fetus. We'd also like you to discuss the potential that the fetal expression of BMP-2 could restimulate a maternal immune response and cause adverse effects in the mother.

I'll go back to the first question.

CHAIRPERSON FINNEGAN: Actually, I'm going to have you run through them all, all the questions.

DR. KAISER: Okay, all right, the next category, tumorigenicity, we'd like you to discuss the potential for rhBMP-2 to stimulate growth of transformed cells. Now, I want to mention and it's been mentioned previously that this category and the previous category are based on potential issues, hypothetical issues. These were not things that were seen in the clinical data set presented by the sponsor but these are things based on information from the literature that could happen in the presence of the growth factors.

Okay, next radiographic effectiveness and this is a question that comes actually from the data presented by the sponsor. Given what you've seen from the sponsor data and from our presentation, we'd like you to comment on interpretation of the radiographic findings at various time points in view of the following factors; the presence and resorption rate of the collagen sponge, the carrier for the BMP, the progression of bone repair in the presence of rhBMP-2 in the case of the investigational patients and the absence of rhBMP-2 in the control patients and the relative ability of bone formed at various time points to withstand the applied loads.

Now, we move onto some things that based on the data we'd like to get some input on some of the labeling issues with this product. The first thing, with respect to instructions for use, we'd like you to provide some suggestions for adequate instructions with respect to the radiographic interpretates, so based on the previous question if there's anything that we should be putting in the labeling.

In addition, we'd like you to discuss any other specific training that should be implemented with respect to this product. We have a number of questions that are related to post-market studies. The first has to do with reproduction in teratogenicity. FDA believes that additional animal studies may be useful for assessing an immune response effect on fetal growth and development and so we'd like you to comment on the need for these studies. If you decide that these studies are necessary, the types of studies that should be performed as well as appropriate animal models.

In the area of tumorigenicity, we've described and the sponsor has described that there's been an agreement to conduct some additional studies -- additional non-clinical studies to evaluate potential for rhBMP-2 to stimulate transformed cells. And what we would like you to do is comment on whether there are any additional studies beyond those ones that we've already agreed to, to address this issue and if you believe that there are, we'd like some comment on the type of studies to be performed as well as the appropriate animal models.

And then finally we'd like to have your comments on the use of ongoing post-market registry data bases to further assess potential for congenital abnormalities. And as with the previous two questions, if you believe that registries are recommended, we'd like to have some input on the types of data to be captured.

CHAIRPERSON FINNEGAN: Okay, thank you. Just to clarify for the panel, I would like you to give your opinions. If you have questions for the sponsor, you might give them a heads up but we're going to go around a second time with specific questions for the sponsor, so this is mainly a generic discussion of your concerns and any thoughts you have on the questions. Dr. Li.

DR. KAISER: I'll leave this first question up and then let me know when you want me to pop to the next one.

CHAIRPERSON FINNEGAN: I think we've got a copy.

DR. LI: Yeah, my materials and engineering background doesn't exactly equip me to answer this question directly. A heads up maybe to the FDA or experts or the sponsor, I guess my question would be, are there any examples of any agent, pharmaceutical or otherwise, that actually passed all the in vitro tests that you've done on tumorigenicity or teratogenicity or the other things that you've tested that actually in vitro did not cause any ill effects but actually turned out to actually have a clinical effect? Because if the answer is yes, then I'm really stuck with this question.

CHAIRPERSON FINNEGAN: Dr. Doull.

DR. DOULL: Yeah, my question is peripheral also, and it's a heads up. When Dr. Hudson was talking about BMPs he mentioned the fact that there's a lot of variability in specie sensitivity to these agents. That also was brought up a couple other times and it's in our book and my concern is usually when you have intraspecies variability like that, you also -- or interspecies variability, you usually also have intraspecies variability. Yet, as I understand it, you're talking about taking a vial of this material, a standard dose, diluting it up, putting that on the sponge and putting it in the cage and it's the same dose for everybody; old, young, male, female, whether they are immuno compromised.

And if that's true, it makes it a little hard to look at the worst case kind of assumption that one would like to make to evaluate systemic toxicity as opposed to local toxicity.

CHAIRPERSON FINNEGAN: Dr. Diamond?

DR. DIAMOND: I guess my concerns have to do with the antibody assays because I'm concerned that they represent arbitrary numbers with a definition of authentic response but no definition of a biologically significant response and I think that, you know, one doesn't know what a neutralizing antibody titer is and -- unless there are studies that haven't been done and we certainly do know that antibodies, maternal antibodies can cause problems in a developing fetus and the IgA antibodies may not get across the placenta but they certainly get into milk and get into neonate.

So it's about the antibody assay.

CHAIRPERSON FINNEGAN: Dr. Hanley, sir.

DR. HANLEY: Yes, I'd like to preface my comments by saying that I've been to many of these meetings before and served as the chair of many of these. I'm a non-voting member at this meeting because of my previous involvement in studies on the spine and particularly in some of the initial studies with BMP use on the spine prior to what is going on here, some with Genetics Institute but not with the current sponsor Medtronic Sofamor Danek.

Several years ago we had a meeting of the Food and Drug Administration Orthopedic and Rehabilitation Advisory Panel with regard to spinal conditions in an attempt to set some criteria such that sponsors would have a good idea of what was needed for us to pass scientific judgment on what they did. Heretofore, I have personally not seen studies put together in a fashion that we could do it well.

I would compliment the sponsors on meeting all the criteria which were set down several years ago and I don't know if anyone's here who participated in that but what they have done is exactly what we've asked people to do so that we could make our job easier and not spend all afternoon saying, "What did you mean by that", or trying to make up for things that weren't there.

So I applaud them on their issues. I'm a spine surgeon, a clinician and I view myself as reasonably knowledgeable with regard to the issues; spine surgery, the selection of patients, the performance of the procedures and the use of the implants and materials under discussion here.

We will not and cannot solve the enigma of back pain in the selection of patients for surgery for it here. It is not part of this discussion. I do not believe that radiographic issues brought up have great pertinence to this presentation. They are what they are and our opinions on what's better if any of CT, regular radiograph, bears not on -- in my opinion, on decisions that should be made here. Those are part of the clinical practice.

It's nice to see that they included CT. It just means they're trying to give us everything that could be meaningful but I don't think it matters in the long run if the device is deemed to be approved and is approved, that the criteria be set up for what a practicing clinician should do.

That's a study issue. I'm sure plain radiographs are just as satisfactory and the patients, ultimately if this were approved, operated on with a device that should not be -- need not be subjected to CT unless for specific instances such as a clinical failure.

Some of the other things that might not be apparent to non-clinicians are issues like blood loss and blood loss is so small in all the groups that it makes no difference. The length of stay, however, has some import and this is one time where not taking a bone graft probably does dramatically improve the length of stay issue, particularly in that other arm that we criticized a little bit, laparoscopic arm.

I'm not an expert on teratogenicity and tumorigenicity and that sort of thing. I think we'll let others who have more expertise work that out. I think the issue here at the table today is mainly one of labeling, indications for use, trying to put in the proper perspective for -- if approved for people utilizing a device, how best it can be controlled and doing some appropriate follow-ups on it. This is -- it's been a good experience, one of the easier ones I've seen. Thank you.

CHAIRPERSON FINNEGAN: Gene.

DR. SIEGAL: Well, if there's a good cop, I guess there has to be a bad cop, too.

CHAIRPERSON FINNEGAN: You said that with a smile.

DR. SIEGAL: I have a number of issues. I do think that the sponsor did everything that reasonably could be done as far as the radiology, especially by acquiring the expertise of Dr. Genant and his associates, who is a world renowned authority and I feel very confident that everything that could be done radiologically has been done. However, histology/pathology has been eluded to multiple times, both in the pre-clinical and the clinical.

And it was used as the gold standard, if you will, to validate the radiology. I have a multi-part question depending on the answer and the question goes something like this. Were those veterinary or human pathologists that did those studies, neither or both? Did they work for the company or was there an independent vetting of the pathology results? Were the pathologists recognized as experts in bone diseases?

Changing subjects, as I understand the carcinogenicity issues, two pancreatic cell lines showed increased proliferation in the presence of BMP-2 and one patient developed pancreatic carcinoma while receiving the therapy. I would like to hear that coincidence or a potential problem. I wonder too, way off the topic perhaps, that at the time of surgery, the rhBMP-2 must be rehydrated, if I could use that term, with sterile water and then must be, quote, "applied evenly" end quote to the ACS which is loaded into the cage. Why was it not discussed pre-loading the BMP-2 sponge to maximize even distribution, either requiring hydration by perhaps emersion in water or pre-hydrating it and packaging it to keep it intact?

And to come full circle back to the pathology question, I wonder would it not be of value to do a carefully controlled radiology histopathology study with pathologists to see, in fact, if there is a gold standard one against the other?

DR. KIRKPATRICK: I'd like to echo some of the other panel comments, that that was quite a well-prepared presentation and a substantial data set to review. A few questions that I'd like to see addressed are, you mentioned that you have data beyond 24 months. How complete is it and did you see deterioration in the clinical results which is something that we often see with other fusion techniques?

I would like to see if you could provide me with some general insight with the expression of BMP-2 normally in the time course of the fusion. In other words, do we have any information on when BMP is normally produced in the fusion healing process and whether the application of the BMP-2 at the onset is coincident with what it would be in the autograft group, for example? I imagine you already have that data.

You mentioned that the metabolic pathway of BMP was through the liver but I did not hear a specific of what that pathway was in the liver and the package insert, as I recall, indicates that no liver studies were done. I'm wondering since one of your explanations about the toxicity was the fact that it was rapidly metabolized, what liver impairments would prevent it from being rapidly metabolized and as such, what liver enzymes should be checked prior to giving the device or using the device.

And the next question is perhaps one, if we find it approvable, in light of the history of the pedicle screw off label use, how would you recommend guarding against off-label use of this product, especially the rhBMP-2? Thank you.

CHAIRPERSON FINNEGAN: I basically have two areas that I would sort of like to see discussed. One is there were some very nice elution studies of the BMP but you didn't look at elution or I didn't see any data for elution from -- of the BMP inside the cage and I would suggest that that's probably a different pattern than just BMP in a sponge.

And the other thing that fascinated me that I couldn't find anywhere is why did the cases that failed fail. You picked pretty straightforward pretty simple spine problems and I was wondering if you have any feeling for why the ones that failed failed.

DR. NAIDU: I'd like to comment -- congratulate the sponsors for doing an excellent outstanding study. I thought it was very well presented. The data showed that the device is effective and with regards to safety, I will hold on. I'm an orthopedic surgeon with a biomechanics background. I will defer that question to the biologists on the panel, but in general, from what I've heard at least, the antibody response was detected in only three of the patients and from Dr. Miller's comments, it appeared as if hardly any cross the placenta barrier nor the amniotic cavity and our respected panel member, Dr. Reddi, goes on to comment that this is a normal substance. We should not be too concerned about it. And so I will defer that thought to Dr. Reddi and the rest of the biologists on the panel.

But as far as the radiographic findings, I think that it appears as if at least from the CD Roms, the CDs that I got, the disks, the fusion mass started to show up at six months as the sponsor stated, and the thing is I don't have any time zero CT scans to judge as to what it would look like. I can only imagine the collagen sponge would be hollow and it would be black. It would only be logical to assume that. So I think that there is bone forming, but I don't know the mechanical integrity of this bone.

And Dr. Kostuik stated that at least eight millimeters of bone must be needed to -- a thick -- eight millimeter thickness of bone must be needed to stabilize an intersegmental fusion. And the other thing is in light of Dr. Kostuik's comments where he stated that it is hard to judge on flexion/extension views mainly because of the superior instrumentation that we have developed today, such as segmental spinal fusion devices such as pedicle screws, it's hard to depend on flexion/extension views. Those are the words that I recall from Dr. Kostuik.

These are not -- you don't have pedicle screws here. You just have two cages. And so I would assume that the flexion/extension criteria that you guys used would be credible at least. That's what common sense would dictate to me at least. But I think that's all to an issue to discuss further is just the stability of these devices. Are they similar to these pedicle screw constructs that flexion/extension views and angular distortion is not considered credible as a radiographic criteria.

I think that you guys have shown a pretty reasonable product here that seems to be safe and efficacious but obviously, there are some issues as far as packaging and I'm also assuming that -- this is actually a question to the sponsor, that you are seeking approval for this for degenerative disc disease with less than grade one spinal disc thesis (ph) for single level fusion. That's what I'm assuming that this device is up for. If I'm wrong as far as that goes, I would appreciate clarification but thank you very much.

CHAIRPERSON FINNEGAN: Dr. Boyan.

DR. BOYAN: I have just two issues that have come up in my reading of the documentation and the discussion today and overall I, too, want to compliment the applicant. It truly was a beautiful package to read, but the two comments I want to make have to do -- one with mineralization and the other one has to do with antibodies.

And the mineralization has to do with how the use of CT and x-ray. I would like to echo the comments from down at the end of the table that it's very difficult at least early on to determine whether or not something is bone or if it's just remineralized collagen and given that you're using a collagen sponge, even x-ray or CT isn't able to discern whether or not that's bone, bone or if it's a graft that was fortuitously structurally remineralized.

And I say that only as informational because the only way you could ferret that out is with histology and you're not likely to take a nicely fused human and do histology but to bear that in mind in interpreting the data. The other comment has to do with antibodies and while I may not be as concerned as some people are about this future consequence to a pregnant person and her fetus, I am somewhat concerned about elderly individuals and people who are likely to have more than one experience with this device in their lifetime.

And if there has been any consideration given to people that might have multiple surgeries at different times and whether or not we're sensitizing them to be BMPs and sensitizing them to type 1 collagen.

CHAIRPERSON FINNEGAN: Thank you. Dr. Reddi.

DR. REDDI: Yes, thank you for giving me this opportunity, Madam Chairman. First, I'd like to also compliment the sponsors for giving us a good package but however, I have some questions which I was not sure whether I should outlay while I made a brief presentation or not because I'm a novice at this but I will very soon learn.

But I would like to ask as far as the tumorigenicity is concerned, whether the sponsor or some of their contractors have done studies because we are really interested in induction of tumors as opposed to stimulating growth of transformed cells. I found copious amounts of data on about 60 cell lines. A lot of cancer research today in the United States has shied away from cancer research because it doesn't mean anything for the human patient, because you can get whatever you want in a cell line, it might please some FDA regulators, but we are really -- it's a very important issue from the point of your patient.

If you really want to study tumorigenicity, it needs to be done in a living animal and I wanted to find out if such attempts are being made or being thought about by the sponsor, so I'd like to find out whenever the time is opportune for that.

The other question concerning the antibodies, I wanted to find out if the sponsor in their volumes of study have developed antibodies to the native BMP-2 as opposed to anti-peptide antibodies or monoclonal antibodies because you might make an antibody to a peptide by one of the scientists in Wyeth-Genetics Institute but I would like to see if there is such data and if such data is available, I would like to strongly recommend that the transplacental passage of these antibodies to native recombinant BMP-2, does it cross and does it have any adverse effects on the fetus.

That's a very important thing because we have been dancing around this issue. I think we need to do definite studies in order for both the -- to allay the fears of both the patients as well as the surgeon.

CHAIRPERSON FINNEGAN: Thank you. Now we're going to get a definitive answer about x-rays and CT scans, right?

DR. LENCHIK: I thought we were still talking about teratogenicity. I don't have much to say about that but to your preview, the sponsor, I have a couple of questions relating to CT. The CDs that we were given, the quality of the CT really varied widely from having real quality CTs with beautiful coronal and sagittal recon to other CT scans that were virtually uninterpretable.

So my question to Harry Genant in particular is what was your experience in the study in terms of CTs that were potential equivocal because you couldn't -- because of metal artifact perhaps or due to reconstruction artifacts. And the second question, again to the sponsor, what do you think the explanation is why there were fewer patients fused by CT at 24 months compared to 12 months?

CHAIRPERSON FINNEGAN: Dr. Larntz?

DR. LARNTZ: I don't have any more to add than I already did.

CHAIRPERSON FINNEGAN: Ms. Rue?

MS. RUE: I have a couple questions. They're points of discussion I guess. One was there was an agreement made with the women in the group for them not to get pregnant and they talked that there were six pregnancies anyway, not to get pregnant for 16 weeks and it doesn't say at what time these women got pregnant. So I'm wondering what effect that agreement held.

And also, if there's going to be a pregnancy registry board, that it include miscarriages of the fetus or embryo at any stage and also some pathology on that to see if there's any effects on the fetus. And also, the fact that the majority of pregnancies are not planned and most women don't know that they are pregnant for the first at least five to six weeks, a lot longer than that, what is going to be done as far as that goes prior to surgery.

CHAIRPERSON FINNEGAN: Thank you. Ms. Maher?

MS. MAHER: I don't have much to add above and beyond what everybody else has said and I thought what I heard was very well put together. I would ask the panel to be cautious about trying to mandate the practice of medicine as you're going forward and talking about labeling that would determine when fusion has occurred. I think most surgeons know when fusion has occurred and will be making that determination on their own no matter what's in the labeling.

So I would ask us to all be cautious and think about that. I would go the same way towards the concerns about off-label use. I think labeling can go into the labeling but there's -- mandating packaging or something like that will increase the cost of the product to the consumer without probably stopping much of what you're probably trying to stop. Thank you.

CHAIRPERSON FINNEGAN: All right, we'll now start back around the table and you can ask your questions and we'll get them to answer them one at a time. Dr. Li.

DR. LI: Do I ask the same question?

CHAIRPERSON FINNEGAN: Yeah, the same question or a new one if you have another one.

DR. LI: Yeah, I guess my original question was, if you've done a variety of in vitro tests to test tumorigenicity, teratogenicity and other possible complications. My question actually was either the sponsor or the FDA or panel members, are there examples of any agent that actually would pass all these tests, yet turn out to be clinically something you'd want to avoid?

In other words, how -- the fact that you passed all these tests, is that actual assurance that these things will not happen clinically?

CHAIRPERSON FINNEGAN: I don't know who the most appropriate -- probably, Dr. Riedel, did you want to answer that or did you --

DR. RIEDEL: To the best of my knowledge and the knowledge of my colleagues who are professional toxicologists, there is no example of such an agent. It would be just hypothesize.

DR. LI: Okay, my question did ask earlier because I was limiting myself to teratogenicity, was on the x-rays versus CT. I guess my question on that is, was the determination of whether or not there was a bone bridge, is that just a yes or no determination? And if it was just a yes or no determination, there's like one spicule or one trabecula that goes from side to side, does that count as a bone bridge or was there some threshold amount of bone that had to be in there to be qualified as a bone bridge.

And a follow-up question to that is, no matter how you determined wh