+ + + + +



+ + + + +


+ + + + +


JULY 27, 2001

+ + + + +

The Advisory Committee met in the Versailles Rooms, Holiday Inn Bethesda, 8120 Wisconsin Avenue, Bethesda, Maryland, at 8:00 a.m., Mark E. Molitch, M.D., Acting Chairman, presiding.


MARK MOLITCH, M.D., Acting Chairman

THOMAS A. AOKI, M.D., Member




PRESENT (Continued):


MARIE C. GELATO, M.D., Ph.D., Member

KATHLEEN REEDY, Executive Secretary


ERIC S. HOLMBOE, M.D., Ph.D., Rick Management


JODY L. PELOSI, F.N.P., Ph.D., Consumer










Introductions 4

Conflict of Interest Statement 7

Opening Statement, David G. Orloff, M.D. 10

Company Presentation, Eli Lilly:

Introduction, Jennifer L. Stotka, M.D. 12

History, Robert Lindsay, M.D., Ph.D.. 18

Nonclinical Overview, John L. Vahle,

D.V.M. 27

Clinical Efficacy, Bruce H. Mitlak, M.D. 39

Clinical Safety, Gregory A. Gaich, M.D. 58

Summary and Conclusion, Bruce H. Mitlak,

M.D. 80

FDA Presentation:

Preclinical Studies, Gemma Kuijpers,

Ph.D. 135

Efficacy, Bruce S. Schneider, M.D. 149

Safety, Bruce V. Stadel, M.D. 168

Public Comment:

Dr. Donald White 206

Deborah Zeldou 208

Dr. Peter Lurie 212

Charge to the Committee, David G. Orloff, M.D. 221

Advisory Committee Discussion 226

Closing Remarks, David G. Orloff, M.D. 329


(8:08 a.m.)

ACTING CHAIRMAN MOLITCH: Good morning. My name is Mark Molitch. I'm the acting chair this morning. This is the meeting of the Endocrinologic and Metabolic Drugs Advisory Committee.

Today we're going to be discussing NDA 21-13 -- I'm sorry -- 318, Forteo, teriparatide injection or recombinant DNA origin. The presenters will be Eli Lilly and Company and the FDA.

We'll begin by introducing members of the table up front.

May I remind everybody that the microphones are activated by pressing on the right to speak, and after you've spoken, please then turn off the microphone to decrease the ambient noise in the room.

And we'll start on the left with Dr. Holmboe.

DR. HOLMBOE: Hi. My name is Eric Holmboe. I'm a general intern from Yale University serving as a consultant today.

DR. PELOSI: I'm Jody Pelosi. I'm an oncology nurse practitioner at the Phoenix Indian Medical Center, and I'm here as the consumer rep.

DR. AOKI: I'm Tom Aoki from the University of California, Davis, in Sacramento, California.

DR. LEVITSKY: I'm Lynne Levitsky. I'm Chief of the Pediatric Endocrine Unit at Mass. General Hospital in Boston.

DR. TAMBORLANE: I'm Bill Tamborlane, Chief of pediatric endocrinology in the Pediatric Pharmacology Research Unit, Yale University.

DR. GELATO: I'm Marie Gelato. I'm a professor of medicine and an endocrinologist at SUNY, Stony Brook.

DR. KREISBERG: Bob Kreisberg from Mobile, Alabama.

MS. REEDY: Kathleen Reedy, Executive Secretary of the Endocrinologic and Metabolic Drugs Advisory Committee, CDER.

DR. GRADY: I'm Deborah Grady. I'm a professor of medicine and epidemiology from the University of California in San Francisco.

DR. SAMPSON: I'm Allan Sampson. I'm professor of statistics, University of Pittsburgh.

DR. BONE: I'm Henry Bone, Director of the Michigan Bone and Mineral Clinical in Detroit, Michigan.

DR. STADEL: Bruce Stadel, Medical Officer in the Division of Metabolism, Endocrine Drug Products.

DR. SCHNEIDER: Bruce Schneider, Medical Officer, Division of Metabolic and Endocrine Drug Products, CDER, FDA.

DR. KUIJPERS: Gemma Kuijpers, pharmacology reviewer at the Division of Metabolic and Endocrine Drug Products, FDA.

DR. ORLOFF: I'm Dr. David Orloff, Director of the Division of Metabolic and Endocrine Drug Products at CDER.

MR. JENKINS: And I'm John Jenkins. I'm the Director of the Office of Drug Evaluation II in CDER at FDA.

ACTING CHAIRMAN MOLITCH: Thank you, everybody.

Kathleen Reedy will now read the meeting statement.

MS. REEDY: The conflict of interest statement for Endocrinologic and Metabolic Drugs Advisory Committee, July 27th, 2001, considering Lilly's Forteo.

The following announcement addresses the issue of conflict of interest with regard to this meeting and is made a part of the record to preclude even the appearance of such at this meeting.

Based on the submitted agenda for the meeting and all financial interests reported by the committee participants, it has been determined that all interests in firms regulated by the Center for Drug Evaluation and Research present no potential for an appearance of a conflict of interest at this meeting with the following exceptions.

In accordance with 18 United States Code 208(b), a fully waiver has been granted to Drs. Mark Molitch, Barbara Lukert and William Tamborlane. A copy of the waiver statements may be obtained by submitting a written request to the agency's Freedom of Information Office, Room 12A-30 of the Parklawn Building.

In addition, we would like to disclose for the record that Drs. Deborah Grady, Robert Kreisberg, Barbara Lukert, Lynne Levitsky, and William Tamborlane have interests which do not constitute a financial interest within the meaning of 18 United States Code 208(a), but which could create the appearance of a conflict.

The agency has determined notwithstanding these interests, that the interest of the government in their participation outweighs the concern that the integrity of the agency's programs and operations may be questioned.

Therefore, Dr. Grady, Dr. Kreisberg, Dr. Lukert, Dr. Levitsky, and Dr. Tamborlane may participate fully in today's discussions.

With respect to the FDA's invited guests, there are reported interests which we believe should be made public to allow the participants to objectively evaluate their comments.

Dr. Henry Bone would like to disclose for the record that he was an investigator on a Phase 3 study of Raloxiphene Evista (phonetic), one of the competing products to Forteo, from 1994 to 1999. Dr. Bone has participated as an investigator in several clinical trials of Alendronate and other competing product to Forteo, some of which are still current. He's also acted as a consultant to Merck.

In addition, Dr. Bone's clinic has received an unrestricted educational grant from Novartis. He has given lectures sponsored by Merck and Novartis.

Lastly, Dr. Bone is an officer of the Michigan Consortium for Osteoporosis, which has received supplementary support from Merck and Procter and Gable.

Dr. Bone receives no salary from the Michigan Consortium for Osteoporosis, but is reimbursed for his expenses.

In the event that the discussions involve any other products or firms not already on the agenda for which an FDA participant has a financial interest, the participants are aware of the need to exclude themselves from such involvement, and their exclusion will be noted for the record.

With respect to all other participants, we ask in the interest of fairness that they address any current or previous financial involvement with any firm whose products they may wish to comment upon.

I mentioned Dr. Barbara Lukert, who was not able to be with us today.


We'll now have an opening statement from Dr. Orloff.

DR. ORLOFF: Good morning. I want to extend my own welcome to the committee and thank you in advance for the service to the agency and to the drug regulatory process.

I'm basically going to read a statement that I read yesterday since there's a new audience, a new sponsor, as well as additional members of the Advisory Committee.

The Advisory Committee process is an important aspect of FDA's review and regulatory decision making for new drugs, affording an opportunity for us to hear from experts in the field, from members of the public, as well as from the sponsor on the subject application.

At the outset, it should be understood by all in attendance that we, the agency, enter into this meeting without an established course of regulatory action. We are here to engage in a discussion between the committee and FDA and the sponsor on the scientific merits of the investigations, clinical and otherwise, of this drug and of the ramifications of the resultant data for a decision regarding marketing of the product for the proposed indications.

I want to remind everybody that the tone and outcomes of the deliberations today and the opinions expressed by the committee, as well as those expressed by the presenters for FDA, do not represent final agency stance on the application. Regulatory action will come only after further review, internal discussion, and clearly discussion with the sponsor.

So, again, as Director of the division that is responsible for review and regulatory action on this product, I want to thank you for being here, welcome you. I'll have further remarks later when I charge the committee after the sponsor's and FDA presentations.

And I'll turn it back over to Dr. Molitch. Thank you.


The company, Eli Lilly, will now give their presentation. They've requested that we hold questions for various speakers until the end of their presentation, and then at that point they'll be open for discussion amongst the members of the panel.

So we'll start with Dr. Stotka, who is the Executive Director of U.S. Regulatory Affairs of Lilly.

DR. STOTKA: Slides on, please.

Good morning. My name is Jen Stotka. I'm a physician and the Executive Director of U.S. Regulatory Affairs for Eli Lilly & Company.

On behalf of Lilly, I thank you for the opportunity to discuss teriparatide, which we will also refer to as recombinant human PTH 1 to 34.

The proposed trade name for teriparatide is Forteo.

The indication for which we are currently seeking approval is the treatment of osteoporosis in post menopausal women and in men.

The advantages and safety profile of this new therapy will be highlighted in subsequent presentation today. The extensive contents of this application meet or exceed all expectations contained in applicable FDA and ICH guidelines, and our clinical trials were conducted with advisement from and agreement with the FDA's Division of Metabolic and Endocrine Drug Products.

Today we will provide data that support the position that teriparatide is the first clinically useful agent in a new class of osteoporosis therapies. These new drugs are bone formation agents in contrast to the anti-resorptives currently on the market, and it will provide an important new choice for the treatment of osteoporosis in post menopausal women and in men.

Comprehensive information from clinical trials enrolling over 2,800 women and men in 20 countries was submitted to the FDA as a new drug application in November of 2000. Our clinical evaluation of teriparatide began shortly after our initial IND filing in August 1995. The clinical development plan was formulated following input from a number of external consultants and the FDA.

Key points of the FDA's draft guidelines on the clinical development of osteoporosis drugs published in April of 1994 were taken into consideration when we designed our clinical program.

The pivotal study in post menopausal women with osteoporosis began in December of 1996, while our pivotal study in men with osteoporosis began in July of '97.

In December 1998, Lilly reported to the FDA an unexpected finding of osteosarcoma in a two-year rat carcinogenicity study. We informed the FDA of our decision to voluntarily stop all ongoing trials with teriparatide while this nonclinical finding was evaluated further.

In April 1999, Lilly submitted the recommendations of an external oncology advisory board to the FDA. This advisory board was convened to assist in the evaluation of the nonclinical osteosarcoma finding.

Lilly and the FDA discussed the appropriate follow-up for patients.

Shortly thereafter an observational study was implemented to continue to collect safety information in all patients previously enrolled in our Phase 3 program of teriparatide.

In July 1999, Lilly, the FDA, and external experts from our oncology advisory board participated in a meeting held at the FDA's request to discuss this nonclinical osteosarcoma finding. In September 1999, we met with the FDA to discuss preliminary safety and efficacy results of our pivotal Phase 3 study and to propose the content for an NDA.

Agreement was obtained from the FDA that the NDA package was adequate to support submission of teriparatide as a new agent for the treatment of osteoporosis in post menopausal women.

In July of 2000, Lilly and the FDA held a pre-NDA meeting. Agreement was reached with the FDA that the data with teriparatide also appeared to be adequate to support submission of teriparatide as a new agent for the treatment of osteoporosis in men.

The NDA was submitted in November of 2000. The requisite four-month safety update was submitted in March of 2001, and today we will demonstrate that the data submitted in our NDA meet or exceed the burden of proof for efficacy and safety.

Our presentations today encompass a number of scientific and regulatory matters. In fact, we will address all questions that the FDA has asked you to consider regarding the mechanism of action of teriparatide, efficacy in women and men, bone quality, and overall safety.

We will also review the rationale for the selection of our 20 microgram dose, and we will provide you with an assessment of the overall benefit-risk profile.

We will follow this agenda. First, Dr. Robert Lindsay, Professor of Clinical Medicine at Columbia University, College of Physicians and Surgeons, Chief of Internal Medicine at Helen Hayes Hospital, and past president of the National Osteoporosis Foundation, will discuss history, mechanism of action, and the unmet medical need.

Following him will be presentations by Lilly scientist:

Dr. John Vahle, veterinary pathologist, will cover nonclinical pharmacology and toxicology.

He will be followed by Dr. Bruce Mitlak, Medical Director for the teriparatide team, who will review the clinical efficacy data.

Next Dr. Gregory Gaich, senior clinical research physician, will present an overview of the safety profile of teriparatide.

And finally, Dr. Mitlak will provide the overall benefit-risk summation in our conclusions.

We look forward to a full discussion of the issues raised. Dr. Mitlak will facilitate Lilly's response during the discussion period.

Additionally, we have a number of our key scientific staff and external experts available here today to help respond to your questions.

In fact, we wish to thank the following experts for working with us and for being here today to assist with your deliberation: Dr. Adamson, Bellizikan, Chabner, Lindsay, Neer, Potts, and Stewart.

We ask for your active consideration to approve teriparatide for the treatment of osteoporosis in post menopausal women and in men. We believe the documentation provided will support such action, and we look forward to a mutually productive session.

I now have the pleasure of introducing Dr. Robert Lindsay for the scientific overview.

DR. LINDSAY: Thank you very much, Dr. Stotka.

Mr. Chairman, ladies and gentlemen, members of the advisory panel, it is a considerable pleasure for me today to introduce to you the topic of parathyroid hormone, an agent that my group has had considerable interest in for the past 15 years.

To set the stage, I shall briefly review the history, mechanism of action, and clinical need for recombinant 1 to 34 human parathyroid hormone as a treatment of osteoporosis in both women and men. Much of the data I will use comes from our specialized center of research funded by the National Institutes of Health.

The parathyroid glands were originally identified by Sandstrom some 121 years ago, and for the next 25 years, their function was hotly debated.

In 1906, Erdheim produced evidence that the parathyroid glands were intimately linked in calcium homeostasis, and in 1925, Collip, working with Eli Lilly Company, prepared a purified, stable extract that was clinically active, and was subsequently marketed.

That parathyroid hormone can be anabolic. It's not new nor novel. In 1929, Orb (phonetic) working with Fuller Albright, first demonstrated the anabolic effect by injecting the extract prepared by Collip into rodents, a finding confirmed some three years later by Hans Selye.

These experiments were largely forgotten until the early 1970s when Nile, Jerry Aerbach, John Potts first sequenced and synthesized the 1 to 34 of minor terminals of parathyroid hormone and subsequently the complete peptide.

This allowed sufficient purified peptide to be synthesized to more fully evaluate its pharmacological profile. Today, of course, 1 to 34 human parathyroid hormone is reduced by recombinant technology rather than by protein synthesis.

Initial experiments confirm the anabolic action in rodents and subsequently in other species, including dogs and nonhuman primates.

The first human experiments were initiated in the 1970s by the late John Parsons in collaboration with John Potts, Bob Neer, Jonathan Reeve and Pierre Munier (phonetic) and others. These studies confirmed that parathyroid hormone could exert an anabolic effect on the human skeleton, first published in 1980.

During the 1990s, several relatively small, controlled clinical trials have been completed. These trials all showed that one to 1 to 34 human parathyroid hormone could produce marked increases in bone mass, particularly in the lumbar spine, but also in the total hip.

The doses that were used varied from 400 to 800 units, international units, in the original concept, roughly equivalent to the dosage used in the Phase 3 studies about which you will hear later.

These data are exemplified by data from our own group published by Felecia Cosman (phonetic) and colleagues in 2001 that demonstrate an increase in vertebral bone mass over a three-year period of approximately 13 percent in an experiment in which parathyroid hormone was delivered by daily subcutaneous injection on top of already coexisting hormone replacement therapy. These data show the increase in bone mass in the spine.

In addition to these changes in the spine, there was also a significant increase in bone mass in the hip, again, over a three-year period, somewhat less than in the spine, but amounting to slightly more than four percent.

Similar data have been published using parathyroid hormone by itself.

Although this study was not powered to detect reductions in fracture, we were able to demonstrate statistically significant reductions in vertebral fracture during the three years of the study primarily because we actually saw no fractures in the PTH treated group.

The effects of PTH on bone mass occur by mechanisms that differ markedly from currently available anti-resorptive agents. About a year or so ago, Tony Hodgeman (phonetic) published data on iliac crest bone biopsies obtained one month after starting parathyroid hormone. After only four weeks of therapy, Hodgeman demonstrated an increase in osteoid surface, an increase in the surface of bone covered by osteoblasts, and a dramatic threefold increase in bone formation rate.

Later this year at the American Society of Bone and Mineral Research, we will present further data from these biopsies that demonstrate that those increases in bone formation occur not only in sites of prior resorption, but also on inactive surfaces, and that they occur in both the trabecular bone and osteo bone and periostea bone.

Our biochemical data confirm these histomorphometric responses. This slide demonstrates the increase in osteocalcin, a marker of bone formation, and an NTL (phonetic) peptide, a marker of bone resorption during the early course of treatment with parathyroid hormone.

You can see that osteocalcin increases dramatically and quickly, such that by one month of treatment there is about a 55 percent increase. There is a slower lag in the increase in NTX (phonetic), but by six months the full pharmacological effects of parathyroid hormone are evident. Parathyroid hormone stimulates both bone formation and also bone remodeling.

The consequence of these phenomena is not only an increase in bone mass, but an improvement in the structure of the skeleton with normal amellor (phonetic) bone being laid down.

Data currently in press from the studies that we have conducted in collaboration with John Bellizikan demonstrate that in both men and in women there is improvement in the connections among trabeculari (phonetic) within a bone.

These trabecular connections are best seen in a single patient slide shown in the next slide in which we have compared a biopsy from a 64 year old woman before parathyroid hormone, with an iliac crest biopsy from the opposite side in the same woman approximately two and a half years after parathyroid treatment.

It is clear that not only is there more bone present in the slide on the right, but also there are increases in the numbers of trabeculari that are present.

In addition to the numbers of trabeculari and the proved connectivity shown here, there is also rather surprisingly to us initially an increase in cortical thickness shown here and shown here. These improvements in cortical thickness differentiate the use of parathyroid hormone as an anabolic agent when delivered by subcutaneous injection from the disease primary hyperthyroidism.

Currently available treatments for osteoporosis are clearly effective. These agents work by reducing bone remodeling and allay bone loss. However, many patients remain at significant fracture risk.

Osteoporosis -- I beg your pardon. Remain at significant fracture risk.

Next slide.

The reductions in fracture risk that one sees with anti-resorptive agents amount to some 35 to 55 percent over a three-year period in patients with vertebral fracture. In addition, these agents are unable to restore bone matrix or architecture in the way in which we have demonstrated with parathyroid hormone.

We believe, therefore, that an unmet medical need continues to persist. Osteoporosis is not a trivial disease. We are well accustomed to the concept that hip fracture is associated not only with increased morbidity, but also with increased mortality.

Data published from the fracture intervention trial by Jane Collie (phonetic) and colleagues just last year demonstrated one feature of the disease, and that is that not only is hip fracture associated with an age adjusted increase in the relative risk of mortality, but that spine fractures are also, and that there is almost a linear correlation between the number of spine fractures that present and also the increase in mortality.

Data that we published in the Journal of the American Medical Association earlier this year demonstrates that when a patient presents with a new vertebral fracture, he or she will have a 20 percent increase in the likelihood of yet another fracture within a single year.

Next slide.

Unlike current agents, parathyroid hormone stimulates new bone formation and remodeling, rapidly increases bone mass, and by this unique mechanism of action, restores skeletal architecture.

In conclusion, therefore, ladies and gentlemen, teriparatide or recombinant human parathyroid hormone 1 to 34, as will be shown in the following presentations, reduces fracture risk significantly, and as I have demonstrated, works by a unique mechanism of action that I believe changes the paradigm for the treatment of osteoporosis and offers benefits to patients with osteoporosis that cannot be seen with current therapeutic options.

It's now my pleasure to introduce Dr. Vahle from the Eli Lilly Company, who will review the preclinical data.

DR. VAHLE: Thank you, Dr. Lindsay.

My name is John Vahle, and I am a veterinary pathologist with the teriparatide team.

I will briefly review the key findings from the animal studies conducted with teriparatide. Overall our nonclinical pharmacology and safety studies meet or exceed all worldwide regulatory guidances.

First, I'll describe the skeletal effects of teriparatide in our most relevant animal model, the mature ovariectomized Cynomolgus monkey.

Then I'll review the nonclinical safety data by briefly reviewing key findings from the animal toxicity studies.

And I will conclude by presenting the results from the two-year rat study previously mentioned by Dr. Stotka in which osteosarcomas were observed.

In monkeys, teriparatide increases bone mass and improves skeletal microarchitecture. These high resolution CT scans of the fifth lumbar vertebra were obtained in an 18-month skeletal pharmacology study in which ovariectomized monkeys were given teriparatide for up to 18 months and illustrate increased trabecular bone from a monkey given five micrograms per kilogram per day as compared to that from an ovariectomized control monkey.

Histomorphometry of the vertebra show that teriparatide stimulated new bone formation on both cortical as well as trabecular surfaces, resulting in increases in trabecular number, in connectivity, as well as increases in cortical area.

These improvements in skeletal architecture are not achieved with anti-resorptives. Most importantly, these effects on bone mass and microarchitecture result in increases in bone strength at both the vertebra as well as the hip.

There have been concerns that the substantial increases in trabecular bone produced by parathyroid hormone might occur at the expense of cortical bone. However, in this long-term monkey study, there were no adverse effects on cortical bone based on the following data.

Cortical bone mass was maintained at the mid-shaft of long bones, such as the radius, humerus and femur. Histomorphometry at these predominantly cortical locations revealed the anticipated teriparatide mediated enhancement of cortical remodeling.

A natural manifestation of this process was an increase in endocortical porosity which was accompanied by enlargement of cortical area and thickness.

There were no deleterious effects on cortical bone strength, and in fact, the net effect was that there is increased resistance to fracture at the mid-shaft humerus and radius.

I will now briefly summarize some of the key findings from the nonclinical safety studies. In the rat and monkey toxicity studies which supported clinical development, the important effects were primarily related to the known pharmacology of parathyroid hormone on either bone or mineral ion metabolism.

The most important effect in the monkey was the histologic observation of interstitial basophilia in the renal medulla. This effect was closely related to the magnitude and duration of hypercalcemia and did not appear to have an impact on renal function.

In contrast, renal histologic changes did not occur in the 18-month pharmacology study I previously described. As will be shown on the following slide, difference in these two monkey models account for the differing effects on renal histology.

In the toxicity studies in which renal changes occurred, the animals were young, immature, intact male and female monkeys who received a dietary calcium intake approximately six times higher than that of a post menopausal woman receiving calcium supplementation.

In the pharmacology study, there were no renal alterations even at doses that cause similar changes in the toxicity studies. The monkeys in this model are mature, ovariectomized females with a daily calcium intake approximately two to three times higher than a supplemented post menopausal woman.

Therefore, the lack of renal effects in this more clinical relevant model in which monkeys were treated for up to 18 months at doses which provided exposures up to eightfold that of a human receiving a 20 microgram dose provide substantial evidence that the histologic alterations in the toxicity studies do not represent a substantial safety concern.

In addition to the effects in the chronic toxicity studies just described, other important findings included a lack of genotoxicity and a full battery of in vitro and in vivo assays that meat global regulatory standards, and the findings in the two-year rat study.

In the next few minutes I'll review the primary findings from this study, which include exaggerated increases in bone mass, bone proliferative lesions, including osteosarcoma.

Importantly, there was no increase in the incidence of tumors in any other tissue or organ. As is standard practice in these types of studies, treatment with teriparatide was initiated in skeletally immature rats six to eight weeks of age and was continued for two years, which constitutes near lifetime treatment.

These high resolution CT images of the proximal femur illustrate the dramatic effects on bone mass that occurred in this two-year study.

This image from a control rat shows a normal pattern of cortical bone, trabecular bone, and intervening marrow space. In all teriparatide treated groups, there is a marked increase in both cortical bone as well as trabecular bone. In fact, the effect is so profound that it leads to near obliteration of the marrow space.

In terms of serum concentrations of teriparatide, these doses provided exposures that were three, 20, and 58 times that patients given a 20 microgram dose. These images and the supporting quantitative data show that even the lowest dose in rats results in exaggerated effects on bone mass that do not occur in patients, as illustrate in the following slide.

These figures compare the effects on bone mass in the two-year rat study to those observed in osteoporotic women and in monkeys. In the left panel are data from the diaphysis, primarily cortical bone site. On the right, the vertebra, a primarily trabecular bone site. On the Y axis is bone mineral content, a measure of bone mass expressed in these figures as a percent above control values. On the X axis is systemic exposure to teriparatide expressed as area under the curve or AUC.

The data points are from women with osteoporosis given the high dose, 40 micrograms, in the Phase 3 trial; rats given the low dose, five micrograms per kilogram, in the two-year rat study; and monkeys given the high dose of five micrograms per kilogram in the 18-month pharmacology study.

These data sets were selected because they are the most closely comparable in terms of duration of treatment, ranging from 18 to 24 months, systemic exposures to teriparatide, and the skeletal locations examined, and they show that over a comparable range of exposures, osteoporotic women and monkeys have similar increases in bone mass.

In contrast, rats have marked increases in bone mass at both cortical as well as trabecular sites.

It is also important to note that this increase in the rat in above peak bone mass for a normal rat, while the value shown for women is the percent above a woman with osteoporosis. So that although women who received teriparatide treatment have increases in bone mass, their bone mass does not exceed peak values for normal, healthy women.

In addition to the exaggerated increases in bone mass, the other important finding in this study was the occurrence of bone proliferative lesions. The majority of these lesions were osteosarcomas that occurred with a dose dependent incidence in all dose groups in both males and females.

There were 60 rats per sex per group in this study, and at the high dose of 75 micrograms per kilogram, the incidence reached approximately 50 percent.

These lesions occurred at multiple sites in both the axial and appendicular skeleton, and metastasis to soft tissue occurred in approximately one third of the rats with osteosarcoma.

In addition, there was a low incidence of benign proliferative lesions of bone.

In addition to the profound increases in bone mass and the bone proliferative lesions, including osteosarcoma I've just described, there was no increase in the incidence of tumors and other tissues, including the mammary gland and the kidney, tissues with high concentrations of PTH receptors.

Based on the initial observation of bone tumors in rats, Lilly made the voluntary decision to stop treatment of patients in the Phase 3 trials while the findings in the rats could be studied. We extensively reviewed these findings with a variety of internal and external experts, including the formation of an external oncology advisory board composed of oncologists, epidemiologists, and pathologists.

After considering data from the rat study and the relevant scientific literature, the advisory panel reached the conclusion that in spite of not identifying a no effect level, the findings from the two-year rat study are not likely to be predictive of an increased risk of osteosarcoma in patients with osteoporosis who were treated with teriparatide.

A variety of factors have been considered in assessing the predictive potential of the findings from the rat model. First, there are important differences between the rat model and the intended clinical use which account for the extreme effects seen in the rodent skeleton.

First, rats are exposed for a relatively long proportion of their lifetime, which is in contrast to patients who would receive treatment for a relatively short proportion, approximately two to three percent.

In addition, there are distinct differences in skeletal biology between rats and humans. For example, rats continue to have longitudinal skeletal growth throughout life, and their growth plates remain open, which is in contrast to humans whose growth plates close at the time of adolescence.

Also, rats lack the mechanism to replace old cortical bone with new cortical bone, a process known as osteonal remodeling.

Importantly, teriparatide is not genotoxic, and it is known that rodent carcinogenicity assays are not always predictive for non-genotoxic agents. The exaggerated effects, skeletal responses observed in the study were mediated by the interaction of teriparatide with the PTH receptor on the osteoblast, and in two-year rat studies with a variety of agents, it has been shown that continual hormonal stimulation such as this can induce tumors in rats which are not relevant to humans.

For example, proton pump inhibitors, such as omeprazole cause gastric carcinoids in rats due to chronic increases in gastrin levels. However, similar neoplastic responses have not occurred in humans treated with omeprazole despite the fact that they also have chronic increases in gastrin levels.

Because of the differences in rats and humans, and there are questions about the predictivity of the rat findings, it is important to consider the data from other species. In terms of other animal data, the most relevant is a lack of bone lesions in an 18-month pharmacology study in which 80 skeletally mature ovariectomized animals were given teriparatide for up to 18 months at exposures up to eightfold greater than women receiving a 20 microgram dose.

We also carefully reviewed the literature on human hyperparathyroidism, and while we recognize important temporal differences between hyperparathyroidism and the daily administration of teriparatide, there is no evidence of an increased risk of bone cancer in patients with hyperparathyroidism, despite the fact that there is chronic stimulation of the osteoblast in new bone formation in both cases.

In summary, the nonclinical evaluation of teriparatide has been rigorous, and the following conclusions can be made. The pharmacology studies clearly show that teriparatide stimulates new bone formation resulting in increases in bone mass, improvements in skeletal microarchitecture, and increases in bone strength while maintaining cortical bone quality.

In particular, these improvements in skeletal microarchitecture are not achieved with anti-resorptive.

In animal toxicity studies, effects were primarily related to the known activity of PTH or related peptides on bone or mineral ion metabolism, and the findings do not represent important clinical safety concerns.

And, finally, a thorough review of the two-year rat study in the relevant scientific literature, we believe that the osteosarcoma findings are not predictive of an increased risk of bone tumors in osteoporosis patients treated with teriparatide.

This concludes the nonclinical data review. It's now my pleasure to introduce Dr. Bruce Mitlak, Medical Director, who will review the clinical efficacy data.

DR. MITLAK: Thank you, Dr. Vahle.

Good morning, Mr. Chairman, committee members. My name is Bruce Mitlak. I'm a physician and Medical Director on the teriparatide team.

I have the pleasure of reviewing the evidence with you that teriparatide treatment increases bone mineral density, improves bone architecture, and prevents fractures in patients with osteoporosis.

The clinical program included 25 trials that enrolled more than 2,800 women and men worldwide. The study codes and titles for the fully enrolled Phase 3 programs and our ongoing observational follow-up study are shown on this slide. I will use these codes to identify the studies in my presentation.

As I will describe this morning, the pivotal placebo controlled study in women was GHAC, and the pivotal study in men was GHAJ. Studies GHAF and GHAH are supportive studies which are included in your briefing document, but will not be included in my presentation this morning.

Study GHBJ is the ongoing observational follow-up study in which prior Phase 3 patients are currently being followed.

This diagram includes the two pivotal clinical studies that I will present this morning. Study GHAC enrolled 1,637 women with osteoporosis to evaluate the effect of teriparatide treatment on the risk of fracture.

Study GHAJ enrolled 437 men with osteoporosis to evaluate the effective of teriparatide on bone mineral density.

In December 1998, we voluntarily stopped these studies and asked patients to complete an early discontinuation visit. This action was taken to allow further evaluation of the finding of osteosarcoma in a concurrent long-term toxicology study as just described by Dr. Vahle.

Women participated in Study GHAC for a median of 21 months and men in GHAJ for a median of 12 months at the time of the respectively study closeouts.

We created an observational follow-up study called GHBJ. The primary purpose of this study was to collect safety information and to maintain contact between the study sites and our study patients.

All patients who had been enrolled in these studies, as well as our other Phase 3 studies were invited to participate. Now I will first focus on results from Study GHAC.

Study GHAC, the pivotal study in women, enrolled 1,637 women. It is a prospective, randomized double blind trial that was performed in 99 sites at 17 countries. Post menopausal women who were at least five years post menopausal and who had a radiographically confirmed vertebral fracture were eligible to participate.

The primary endpoint in this study was the proportion of women with one or more new vertebral fractures. All women self-administered a once daily subcutaneous injection that included either teriparatide, 20 micrograms, teriparatide, 40 micrograms, or placebo, and all women were provided a supplement that included 1,000 milligrams of calcium and 400 to 1,200 units of Vitamin D.

The baseline characteristics for women in the study are shown by treatment group, and in this presentation the placebo group will be shown in white, the teriparatide 20 group in yellow, and the teriparatide 40 group in blue.

The groups were balanced for the characteristics shown, as well as for other factors which could affect the risk of fracture. The mean age was 69 to 70. There was a slightly greater proportion of women greater than 70 years of age in the two teriparatide groups compared with the placebo group.

The mean number of years since menopause was 21 to 22 years. Prior treatment for osteoporosis was reported by 13 to 16 percent of the women, but no treatment was permitted for between six and 24 months prior to the beginning of the study, depending on the specific treatment.

Baseline spine bone mineral density expressed in standardized units was approximately 820 milligrams per centimeter squared, corresponding to a T-score of about minus 2.6, and as shown, approximately 60 percent of these women had two or more prevalent fractures at the beginning of the study.

Because of early closure, women completed different lengths of time in the study. This panel shows the number of women who completed the specified months on the X axis. Because women were asked to suspend study medication and then were scheduled for their final visit, exposure to study medication was on average eight weeks shorter than the duration shown on this slide.

You can see that the duration of observation was similar across treatment groups. Relatively few women in any group left the study before 18 months. The maximum duration between baseline and final radiograph for a patient was 27 months, and the median was 21 months.

Eighty-one percent of the women in this study had an adequate baseline and follow-up radiograph.

This figure shows the scale used to grade both baseline and incident vertebral fractures in this study. Vertebral bodies that are either normal or a fracture that is crushed in the anterior, mid or posterior part of the vertebral bodies are shown.

Radiologists who were blinded to treatment assignment called vertebrae either normal or reported to us the presence of a mild, moderate or severe fracture using this scale as specified in the protocol. While this is a semi-quantitative scale, these grades correspond to approximately a 20, 25, or 40 percent or greater loss of height of the vertebral body.

In this study, a fracture was reported if a vertebrae had a score of zero at baseline and was found to have a score of one, two, or three at follow-up. Over the 21 months of the study, 105 women were found to have one or more new vertebral fractures.

Results for the primary efficacy endpoint are summarized on this slide. Let me review the format which will be used also on the subsequent two slides.

The number of women with one or more new fractures in each group is shown on the respective treatment bar. The height of the bar corresponds to the proportion of women within each group with a fracture. The relative risk in 95 percent confidence intervals are shown for each comparison to placebo, and all p values refer to comparisons with placebo.

Teriparatide reduces the risk of vertebral fractures. In women assigned to treatment with teriparatide, the relative risk for fractures were .35 and .31, corresponding to a highly statistically significant 65 and 69 percent reduction in the likelihood of a fracture. The absolute risk of fracture was reduced from approximately 14 percent to five percent and four percent.

Additional analyses were performed to evaluate the effective of treatment on more severe fractures in this study. This figure shows the results for women who had one or more vertebral fractures that were at least of moderate severity.

While ten percent of women assigned to placebo had fractures that were moderate or severe in degree, only one and two percent of women assigned to treatment with teriparatide had such a fracture. The relative risk of .1 and .22 corresponds to a 90 and 78 percent reduction in the risk of having a moderate and severe fracture.

In this study, we found that regardless of treatment, women with more severe fractures were more likely to report back pain or to suffer height loss.

This panel shows results for women who had two or more new vertebral fractures during the study. The relative risk for multiple vertebral fractures was .23 and .14, corresponding to a 77 and 86 percent reduction in the risk of having two or more new vertebral fractures

Teriparatide treatment reduces the risk of overall nonvertebral fragility fractures. This figure shows the proportion of women who reported one or more nonvertebral fragility fractures both overall and by specific skeletal site.

As specified by the protocol, site investigators determined whether a fracture was associated with excess trauma, such an association with an automobile accident or fall greater than a standing height. Fifty-eight women had fractures that did not result from excess trauma, and these were considered fragility fractures.

Teriparatide treatment significantly reduced the risk of nonvertebral fragility fractures. The relative risk of .47 and .46 correspond to a 53 and 54 percent reduction in the risk of fracture in each group compared with placebo.

And while there were a small number of women with fractures at any specific skeletal site, the figure shows that there was a similar or lower proportion of teriparatide treated women with a fracture at each site compared with placebo, including the radius, which I will return to in a few minutes.

This analysis of the same data for the placebo group in white, the teriparatide 20 group in yellow and 40 group in blue now shows the data as time to first event, and it demonstrates that the effective treatment on the risk of nonvertebral fracture became progressively apparent after about nine months of treatment.

It also shows that at no time during the study was there evidence for an increase in risk for these fractures.

Teriparatide treatment increases lumbar spine bone mineral density. Lumbar spine bone density increased significantly with teriparatide treatment at each visit where it was assessed, including the first visit at three months, where nearly a four percent increment in bone density had already occurred.

At endpoint, the difference in bone mineral density between the 20 microgram group and placebo was nine percent and between the 40 microgram group and placebo was 13 percent.

Ninety-six percent of women in the study assigned to teriparatide 20 micrograms had an increase in bone mineral density. These increases in bone density were associated with rapid increases in biochemical markers of bone formation and secondarily bone resorption.

Teriparatide treatment increases hip bone mineral density. Total hip bone mineral density was measured in approximately one half of the women in the study at a subset of study sites, and femoral neck bone density was measured in all women.

At endpoint, total hip bone mineral density decreased by about one percent, and in contrast, increased in both of the teriparatide groups.

The mean difference between the teriparatide groups and placebo at endpoint was 3.6 percent and 4.6 percent. Each comparison was statistically significant.

At the femoral neck compared with placebo, the increase in bone mineral density at endpoint was four percent and six percent. Other hip sites were also significantly increased by teriparatide treatment.

Ultra distal and distal radius bone mineral density was measured in about 450 women. At the ultra distal radius, bone density declined slightly in the placebo group, but did not change significantly in any group, nor were there differences between groups.

At the radial shaft bone mineral density decreased about one percent in women assigned to placebo. The difference between the treatment group and placebo was one percent in the women assigned to treatment with 20 micrograms and two percent in women treated with 40 micrograms of teriparatide.

The 40 microgram group differed significantly from the placebo group. This early decrease in bone mineral density likely reflects increases in cortical bone remodeling and as demonstrated by PQCT in a subset of approximately 100 women was associated with preserved cortical thickness and evidence for periosteal new bone formation.

Importantly, it was also associated with a humerically lower number of wrist/forearm fractures in the teriparatide group, as I had previously highlighted for you.

Importantly also, teriparatide increases total body bone mineral. Total body bone mineral was measured in about 400 women at a subset of study sites. Compared with the placebo group which lost .7 percent, the increase in the 20 and 40 microgram groups were 2.6 and 3.5 percent, each comparison statistically significant.

This confirms that the increases in spine and hip bone density are associated with a net increase in total body bone mass.

Transiliac bone biopsies were obtained from 61 women at baseline and then again at either 12 months or study closeout. This slide shows the baseline and endpoint bone biopsy from one patient in the 20 microgram group and one patient in the 40 microgram group who had spine bone density responses similar to the mean for their respective treatment groups.

The green stain shows calcified elements, including both the inner and outer cortical shells, as well as trabecular bone.

Also apparent is marrow space and a small amount of extraosteo soft tissue. Trabecular bone volume, TBV, is indicated below each biopsy.

Dr. Eric Erickson, the reader for these biopsies, determined in blinded fashion that there was no evidence for woven bone, abnormal mineralization, cellular proliferation, or abnormal architecture in these biopsies.

Among the biopsies taken at 12 months, there was an increase in intra cortical remodeling in the 40 microgram group, but not the 20 microgram group. This was no longer observed in the biopsies taken at study closeout.

This remodeling transient is consistent with the results observed in the primate study and did not adversely affect biomechanical strength in the monkeys.

In addition to the favorable effects on trabecular bone volume just shown, there was significant increases or trends to increase in mineral apposition rate, wall thickness, trabecular thickness, and a measure of connectivity, connectivity of the trabeculae.

So to summarize the results from this study, teriparatide treatment was effective at preventing spine and non-spine fractures in women with osteoporosis. Treatment with teriparatide 20 and 40 micrograms reduced the risk of vertebral fractures by 65 and 69 percent; reduced the risk of nonvertebral fragility fractures by 53 and 54 percent; increased bone mineral density at the spine and hip but not the forearm; increased total body bone mineral and had favorable effects on bone architecture.

Now I will present the results from our study in men. Study GHAJ was a prospective, randomized double blind study in men with osteoporosis performed at 34 sites in 11 countries. Four hundred thirty-seven men with osteoporosis either associated with hypogonadism or with idiopathic osteoporosis were enrolled with low bone mineral density at either the spine or the hip.

The primary endpoint of the study was change in bone mineral density at the spine. All men self-administered a once daily subcutaneous injection, again containing either teriparatide 20 micrograms, 40 micrograms, or placebo, and all were provided supplements containing 1,000 milligrams of calcium and 400 to 1,200 units of Vitamin D.

The baseline characteristics for men in the study are shown again by treatment group. The groups were well balanced for the characteristics shown. On average men were 58 to 59 years of age. Twelve to 18 percent reported the use of other treatments for osteoporosis prior to the study, but, again, none were permitted for six to 24 months prior to randomization. Mean baseline T-scores for the spine, femoral neck, and total hip are shown by treatment group.

This figure shows the exposure in GHAJ from the time of randomization to the time of the last bone mineral density measurement. The median duration of follow-up in this study was 12 months.

For the same reason as the study in women, the actual time receiving study medication was in this case about four weeks on average less than the duration shown here.

Teriparatide treatment significantly and rapidly increased spine bone density in men. At endpoint spine bone density had increase 5.4 and 8.5 percent in the 20 and 40 microgram groups compared with placebo. The bone mineral density response was rapid, with a significant increase compared with placebo at the first measurement point in the study at three months.

Importantly also, response in bone density was similar in men with osteoporosis associated with hypogonadism and those with idiopathic osteoporosis.

Because most men were, in fact, completing an early discontinuation visit rather than a formal 12-month visit at the 12-month time point, the data will be shown as baseline to endpoint. At endpoint total hip bone mineral density had increased .63 percent in the 20 microgram group compared with placebo, which itself had increase .54 percent. This comparison reached a p value of .074.

The mean increase between the 40 microgram group and placebo was 1.6 percent. At endpoint femoral neck bone mineral density had increased 1.2 and 2.6 percent in the 20 and 40 microgram groups compared with placebo. Each of these comparisons was statistically significant.

However, at other hip sites the comparison for the 20 microgram group was not significant.

Importantly teriparatide treatment increased total body bone mineral in men. Total body bone mineral was measured in 254 men at a subset of study sites. At endpoint total body bone mineral had increase 1.1 and 1.3 percent in the two treatment groups compared with placebo. Each comparison was statistically significant.

So to summarize, treatment with teriparatide was effective at increasing bone mineral density in men. Treatment with teriparatide 20 micrograms and 40 micrograms increased bone mineral density at the spine and femoral neck. Total hip bone density was significantly increased only for the 40 microgram dose.

There was a significant increase in total body bone mineral for both doses.

To further evaluate the effect of gender on response to treatment, we compared the mean actual change in bone mineral density from women in Study GHAC, in men in Study GHAJ. We compared the actual change because we found, unlike percent change, the actual change was independent of baseline bone mineral density, and men in Study GHAJ started with a higher bone density than did women in Study GHAC.

As you can see, the actual change in bone mineral density for women and men for a comparable period of treatment are nearly identical.

Similarly, actual change in bone mineral density at the femoral neck for comparable period of time is identical for men and women. This is shown for men with a measurement up to the 12-month time point in the protocol.

These two panels support that gender was not an important factor in the expected response to treatment.

So to summarize, despite early study completion, both Studies GHAC and GHAJ clearly reached their specified primary endpoints.

Treatment with teriparatide 20 micrograms and 40 micrograms significantly reduced the risk of vertebral and nonvertebral fractures in both menopausal women. The reduction was similar for each dose.

Treatment rapidly and significantly increased bone density in post menopausal women and in men, and treatment improved bone microarchitecture.

That concludes this presentation. I would now like to introduce Dr. Gaich, who will review the clinical safety.

DR. GAICH: Thank you, Dr. Mitlak.

Good morning, Mr. Chairman, committee members. My name is Gregory Gaich. I'm a physician on the teriparatide team, and I am pleased to show you the data which establishes the safety and tolerability of teriparatide in the treatment of post menopausal women and men with osteoporosis.

Like the efficacy data just presented, the data that I will show you also supports the 20 microgram dose as the proposed marketed dose.

I'll review the overall safety experience, the results of the clinical and laboratory safety evaluations in our study in post menopausal women and in our follow-up study and in our study in men with osteoporosis.

I'll conclude with the results of the drug interaction and special population studies which were performed.

Our clinical investigations included 25 trials, which enrolled over 2,800 women and men, more than 1,900 of whom received teriparatide. Does of five to 100 micrograms were used in the clinical pharmacology studies, and doses of 20 and 40 micrograms were studied in our long-term Phase 3 studies. Total patient exposure to teriparatide was over 1,900 patient-years.

This slide shows the overall results of the clinical safety evaluations in the two placebo controlled Phase 3 studies combined. In this slide, the total number of patients in each dose group is shown at the top of the column, and each row shows the number and the percent of patients in each treatment group who had the listed event.

As shown in the table, the number of patients experiencing any adverse event was similar in all three treatment groups. There was a significant increase in the number of patients who discontinued due to adverse events in the 40 microgram group, but not the 20 microgram group.

The discontinuations in the 40 microgram group were primarily due to nausea.

The number of patients experienced in the teriparatide treated groups experiencing any serious adverse event, cancer, or death was similar or lower in the teriparatide treated groups compared with placebo. No osteosarcoma or other primary bone tumor occurred in any patient.

There were very few deaths in the studies, and the difference in the treatment groups was not statistically significant. None of the deaths were judged to be related to study drug by the investigator, and there were no patterns in the cause of death.

In addition, there was no difference in the morality among treatment groups in patients in older or younger age groups.

The evaluation of treatment related clinical and laboratory effects is based on the data from all of our studies. I'll focus on the data from the pivotal Phase 3 study in post menopausal women, GHAC, in which 1,637 patients were treated for up to two years.

I'll also show the data from the clinical pharmacology studies and/or other Phase 3 studies where it provides additional information.

In Study GHAC, the adverse events in the 20 microgram group were general mild and did not lead to discontinuation from the study. Leg cramps were reported by two percent more patients in the 20 microgram group than in the placebo group, and this was statistically significant.

In the 40 microgram group, headache and nausea were significantly increased compared with placebo, but this was not observed in the 20 microgram group.

There was a numerical, although not statistically significant, increase in the incidence of nausea in the 20 microgram group, and nausea may also be treatment related at the 20 microgram dose as well as 40 microgram dose.

There was also a treatment related reduction in the incidence of new or worsened back pain in both treatment groups, and this is consistent with the reductions in vertebral fractures which Dr. Mitlak presented.

Similar significant reductions or trends in back pain were also observed in the other three phase three studies.

Next I'd like to review the results of the pharmacokinetic and safety laboratory evaluations in Study GHAC. All of the laboratory effects observed in our studies were expected based on the known pharmacology and physiology of parathyroid hormone.

This is a best fit analysis of the serum teriparatide concentrations obtained from 360 patients in Study GHAC. The solid line shows the mean teriparatide concentration following a 20 microgram dose. The hatched area shows the 25th to 75th percentile range.

The upper limit of endogenous parathyroid hormone 1 to 84 is shown in the horizontal line. The serum concentrations of teriparatide peaked at approximately 30 minutes post dose and declined rapidly thereafter, with an apparently elimination half-life of approximately 60 minutes. By three to four hours post dose, very few patients had measurable teriparatide in the serum, and there was no accumulation of teriparatide with repeat dosing.

The average 24-hour exposure of teriparatide and endogenous PTH combined did not exceed the upper limit of normal for endogenous PTH.

Serum calcium was also measured at every visit, and we performed a similar best fit analysis on the serum calcium measurements.

This graph shows the serum calcium analysis overlaid on the pharmacokinetic analysis. The vertical axis on the left shows the teriparatide concentrations, and the vertical axis on the right shows the serum calcium concentrations. The upper limit of normal for serum calcium of 2.64 millimoles per liter or 10.5 milligrams per deciliter as shown by the horizontal line.

As expected, based on the known effects of parathyroid hormone and on the transient exposure to teriparatide following each dose, there was a brief, transient increase in the mean serum calcium concentrations following a 20 microgram dose. The mean baseline serum calcium concentration was 2.3 millimoles per liter or 9.2 milligrams per deciliter, and the mean peak serum calcium concentration occurred at 4.25 hours after the dose and was 2.4 millimoles per liter, or 9.6 milligrams per deciliter.

Very few patients even transiently exceeded the upper limit of normal.

Serum calcium returned to baseline by 16 to 24 hours after the dose, and the serum calcium at this endpoint was not increased in either the 20 microgram or the 40 microgram dose.

In the 20 microgram group, these transient changes in serum calcium were small. Median increase was 0.3 to 0.5 milligrams per deciliter at each study visit, and 97 percent of the patients never exceeded 11 milligrams per deciliter. The highest observed value was 11.6 milligrams per deciliter.

Eight percent of the patients had a single high serum calcium and exceeded the upper limit of normal, and three percent exceeded the upper limit of normal on two consecutive four to six-hour post dose measurements.

The transient changes in serum calcium were greater in the 40 microgram group, with a median increase ranging from 0.5 to 0.7 milligrams per deciliter and with more patients exceeding the upper limit of normal.

Transient increases in serum calcium which exceeded the upper limit of normal were not associated with clinical adverse events in either treatment group, however.

The pre-dose serum calcium was measured 16 to 24 hours after the preceding dose in a subgroup of approximately 450 patients. This graph shows the medians and the 25th to 75th percentile range for the pre-dose serum calcium at each visit during the study. The upper and lower limits for serum calcium are shown by the horizontal lines.

There was a small decrease in the serum calcium in the placebo group at three and six months, but the pre-dose serum calcium in the teriparatide treated groups remain similar to baseline throughout the entire course of the study.

We also observed expected effects on urinary calcium excretion, which were consistent with the known physiology and pharmacology of parathyroid hormone. The median urinary calcium excretion in the placebo group was 3.9 millimoles per day or 156 milligrams per day. There was a small increase in the 24-hour urinary calcium excretion for the first six months, and the median increase was 30 milligrams per day at the six month time point.

There was no difference among treatment groups in the number of patients with elevated urinary calcium excretion, and the highest observed 24-hour urinary calcium excretion was similar to placebo and the two teriparatide treated groups. The result showed no increase in the incidence of urolithiasis or related events.

We've shown a lot of data on the serum and urine calcium. Let me summarize those results before moving on to the remainder of the presentation.

The magnitude of the serum calcium effects were small, 0.3 to 0.5 milligrams per deciliter in the 20 microgram group, and the effects on serum calcium were brief, with the serum calcium returning to baseline after every dose.

There were small increases in the 24 hour urinary calcium excretion. The median was 30 milligrams a day, and there were no clinical adverse events associated with the increases in the serum or urine calcium.

These data indicate that monitoring of serum in urine calcium is not necessary in patients treated with 20 micrograms a day of teriparatide.

Parathyroid hormone has known effects of uric acid clearance and effects on uric acid were also observed in our studies with teriparatide. This slide shows a dose dependent increase in the serum uric acid which was observed at one month and remained at a similar level throughout 12 months.

The serum uric acid concentration in the placebo group was 270 micromoles per liter or 4.5 milligrams per deciliter. The median increase was 0.9 milligrams per deciliter in the 20 microgram group and 1.2 milligrams per deciliter in the 40 microgram group.

The increases in serum uric acid resulted in 2.8 percent of patients in the 20 microgram group and five percent of patients in the 40 microgram group, exceeding the upper limit of normal at least once during the study.

These increases in serum uric acid did not result in an increased incidence of gout or arthralgia, however.

There are a number of conditions that have been historically associated with hyper parathyriodism. We examined our clinical trial data to determine if these conditions were associated with teriparatide administration.

The incidence of cardiovascular disease, hypertension, peptic ulcer disease, renal insufficiency, and urolithiasis were not increased in the teriparatide treated patients.

The next few slides summarize the renal and hemodynamic evaluations in more detail. Clinical and laboratory data were examined in order to evaluate potential effects on the kidney. There was no significant effect on the incidence of urolithiasis or related terms, on serum creatinine concentrations, on measured creatinine clearance, or on routine urinalysis during the study.

Routine vital signs were obtained in the Phase 3 studies, and more extensive hemodynamic evaluations, including serial orthostatic blood pressure measurements were performed in the clinical pharmacology studies.

In the clinical pharmacology studies which enrolled health volunteers generally over age 50, we were able to detect small changes in the post dose heart rate, which were also detected as a shortening of the RR interval on the electrocardiogram. There was no QTC prolongation or other clinically significant effect on the electrocardiogram following a 20 microgram dose or any other dose study.

There were no significant effects on standing or supine blood pressure in the 20 microgram dose, although there have been occasional subjects who experience transient symptomatic postural hypotension following teriparatide administration. This was observed once following a 20 microgram dose and more frequently at higher doses. Symptoms were relieved by lying down, and they did not preclude further dosing.

A number of subject receive subsequent and sometimes higher doses of teriparatide without experiencing orthostatic hypotension.

In the Phase 3 studies in which there were no restrictions in activity. There was not an effect on sitting blood pressure or pulse or on the incidence of postural hypotension. Nevertheless, it is possible that a patient may experience transient, symptomatic, postural hypotension following a 20 microgram dose of teriparatide.

I'd now like to describe the clinical and laboratory effects after discontinuation of treatment. These are the interim results from the ongoing follow-up study, GHBJ. Patients who had participated in any of the previous Phase 3 studies were invited to participate in the follow-up study. Approximately 80 percent of the women and men who enrolled in the prior treatment studies enrolled into Study GHBJ.

The patients have completed the first two visits, which were approximately six and 18 months after the end of the prior treatment studies. This represents a total of 39 months of cumulative observation for the women previously enrolled in the previous Study GHAC and 30 months in the men previously enrolled in the pivotal study GHAJ.

When we first discussed the results of the patients previously enrolled in Study GHAC, at the first study visit approximately six months after the end of the treatment study, there is no longer a difference from placebo in nausea, headache, leg cramps or clinical laboratory endpoints, except for the serum uric acid.

The increase in serum uric acid concentration had declined to less than two percent, but it was still statistically significant.

The number of patients in the teriparatide treated groups with abnormal serum uric acid concentrations was no longer different from placebo. There was a small, a less than two percent, but statistically significant increase in the serum creatinine. There was no decrease in the measured serum or I'm sorry. There was no decrease in the measured creatinine clearance, and only one patient in the placebo group and one patient in the 40 microgram group had a clinically significant increase of greater than 0.4 milligrams per deciliter.

These effects were not observed in the other Phase 3 studies.

Through visit two of the follow-up study, approximately 18 months after the end of the treatment study, there were no new clinically significant safety issues identified. There continued to be no increase in the incidence of cancer, urolithiasis, gout or arthralgia, and there continued to be a reduction in the incidence of new or worsened back pain, which is consistent with the observed continued reduction in radiographic vertebral fractures.

We also recorded non-vertebral fractures in the follow-up study, and this analysis shows the time to first non-vertebral fragility fracture for the women in Study GHAC, who were then followed in Study GHBJ. This horizontal line represents the period of time during which treatment was discontinued.

The initial part of this curve is identical to the one previously shown by Dr. Mitlak. The risk of non-vertebral fracture following discontinuation of treatment did not increase in the teriparatide treated groups. The absolute risk reduction in teriparatide treated patients at the end of study GHAC was three percent, and the absolute risk reduction was approximately five percent at GHBJ visit two.

That concludes the presentation of the safety data in the pivotal study and the follow-up study in post menopausal women.

I'd now like to briefly review the safety evaluations in the men with osteoporosis. Study GHBJ was the pivotal study in 437 men with osteoporosis, and the results are similar to the study in post menopausal women.

This slide shows the results of the clinical and laboratory effects in the study in men. As was observed in the post menopausal women, there was a dose dependent increase in the number of patients with at least one serum calcium exceeding the upper limit of normal at four to six hours after the dose, but the number confirmed on repeat testing was only 1.3 percent in the 20 microgram group.

The magnitude and the time course of the serum calcium was also comparable to what was shown in the post menopausal women.

There was a significant increase in nausea and headache at the 40 microgram group, but not the 20 microgram group.

There was no trend towards increase in leg cramps in the men. However, there were too few events in this study to evaluate that effect adequately.

There was also a significant increase in the number of men, again, in the 40 microgram group, but not the 20 microgram group who discontinued due to adverse event, and just as was the case in the post menopausal women, the discontinuation in the 40 microgram group were largely attributable to nausea.

The other clinical and laboratory effects, such as effects on serum urine calcium and urinary calcium excretion were also comparable to the effects in post menopausal women.

We also performed pharmacokinetic measurements in the men. The time to peak concentration and the apparent elimination half-life were similar in men and women, but the serum concentrations of teriparatide were 20 to 30 percent lower in men than in women.

As Dr. Mitlak and I have described, the effects on spine and hip bone mineral density, clinical adverse effects, and laboratory tests were similar in men and women.

Well, not an endpoint in Study GHAJ, spine radiographs were obtained as a screening test and follow-up spine radiographs were obtained at visit two of the follow-up study of GHBJ in order to provide a more complete set of data with which to compare to the women.

This slide shows the vertebral fracture incidence in men and the time between the baseline and follow-up radiographs includes both the treatment and follow-up phase, a total of 30 months.

There were fewer fractures in this study than in the pivotal study in post menopausal women, and there were too few fractures to have adequate statistical power.

Nevertheless, the observed patterns in vertebral fractures in the men and in moderate and severe vertebral fractures in the men was similar to the patterns observed in the post menopausal women.

In addition, the number of men sustaining new vertebral fractures or new moderate to severe vertebral fractures was identical in the 20 and 40 microgram groups.

While this analysis is not a pre-specified analysis of the study, it does illustrate the similarity of the similarity of the response to treatment in men and in women, and it supports the 20 micrograms as the appropriate dose in men as well as in women.

In addition to examining potential gender differences, we also examined special populations based on age, renal function, cardiac function and blood pressure. There were no clinically significant pharmacokinetic or safety findings in these special populations, and restrictions or special monitoring of patients with these conditions are not necessary.

We also performed clinical pharmacology studies which evaluated potential pharmacodynamic and safety interactions with hydrochlorothiazide, furosemide, calcium channel blockers, Atenolol, Digoxin, hormone replacement therapy, and Raloxifene. There were no clinically significant interactions with teriparatide in these drug interaction studies, and restrictions or special monitoring of patients taking these medications was also not necessary.

Now, let me conclude by summarizing the results of the clinical and safety evaluations of teriparatide. In the Phase 3 studies, leg cramps and possibly nausea were treatment related at the 20 microgram dose.

Forty micrograms per day was more likely to cause nausea, headache and discontinuation due to adverse events.

The increased incidence of symptomatic postural hypotension observed in the clinical pharmacology studies was not observed in the Phase 3 studies.

Finally, there was a lower incidence of back pain in both the 20 and 40 microgram groups, which was consistent with the reduction in vertebral fractures.

The laboratory evaluations showed the expected transient effects on serum calcium, and the expected pharmacologic effects on serum uric acid and on urinary calcium excretion. These effects were small and were not associated with clinical adverse effects, and 40 micrograms a day was more likely to causae increased serum calcium and serum uric acid.

After discontinuation of treatment, nausea, headache, leg cramps, and the laboratory effects quickly resolved, except for the small increase in serum uric acid. Through 18 months of post treatment follow-up no new clinically significant adverse effects were identified, and there continued to be no increase in the incidence of cancer, urolithiasis, gout or arthralgia, and there was no increase in the rate of nonvertebral fractures.

There continued to be a continued significant reduction in the incidence of new or worsened back pain.

In conclusion, teriparatide 20 micrograms and 40 micrograms a day were safe and well tolerated in our studies of treatment of post menopausal women and men with osteoporosis. The effects on the clinical laboratory tests were small and consistent with the known physiology and pharmacology of parathyroid hormone, and routine laboratory monitoring in patients taking 20 micrograms a day is not necessary.

Likewise, restrictions or monitoring in the special population study are not necessary. There were no significant drug interactions identified, and finally, although both doses were safe, teriparatide 20 micrograms a day was associated with fewer adverse effects.

I thank you very much for you attention, and Dr. Mitlak will conclude this morning's presentations with the summary and conclusions.

DR. MITLAK: Mr. Chairman, members of the committee, I have the pleasure of concluding the formal presentation from Lilly this morning.

We've provided evidence for you that teriparatide is a bone forming agent that increases bone mineral density, improves bone microarchitecture, and prevents fractures in patients with osteoporosis. Teriparatide was safe and well tolerated by patients in the clinical trials.

To summarize the presentations, Dr. Lindsay outlined the pressing clinical need for such an agent and reviewed the breadth of prior experience with teriparatide.

Dr. Vahle presented nonclinical data demonstrating that teriparatide is a bone forming agent that increases bone mass and strength in several species. He also described the finding of osteosarcoma in a long-term study in rats and outlined factors that are important in understanding the relevance of the findings to the proposed use in women and men with osteoporosis.

Dr. Gaich and I presented the favorable overall clinical profile for teriparatide.

Let me begin now by reviewing our considerations on the nonclinical findings. In 1999, the following experts were convened to review the findings in the nonclinical study and to provide advice on the follow-up of study participants. These include Drs. Chabner, Adamson, Antman, Henderson, Fletcher, Raymond, Kronenberg, and Doppelt. Drs. Chabner and Adamson are in attendance with us today.

This PTH oncology board reviewed the available nonclinical and clinical data and provided the following conclusions for us.

Based on current information, the findings in the rat study were unlikely to predict for the development of bone tumors in patients who had received teriparatide in the clinical trials. This conclusion was reached with considerations of the following:

The lifetime duration of treatment in the rats compared with a relatively brief exposure intended in humans;

The fact that treatment was initiated during the rapid growth phase of the animals;

The difference in rat and human bone biology and response to PTH;

And the lack of clinical association between hyperparathyroidism and osteosarcoma in humans.

Since then we have evaluated additional nonclinical and clinical information and have had ongoing discussions with our consultants and with the agency. In specific, as described by Dr. Vahle, no skeletal lesions were observed in an 18-month study in monkeys given four to eight times the exposure of subjects in the Phase 3 trial.

While we recognize that there are temporal differences in the profile of PTH exposure in patients with hyperparathyroidism and those who had received teriparatide as treatment for osteoporosis, osteoblast stimulation occurs in both, often to a greater extent in patients with hyperparathyroidism, and patients with hyperparathyroidism can have elevated levels of parathyroid hormone for years.

New bone formation also occurs in patients with hyperparathyroidism, but resorption usually occurs to a greater degree.

We identified a single case report of the co-occurrence of hyperparathyroidism and osteosarcoma in the literature. Dr. Olaf Unell (phonetic) then assisted us by performing a systematic search of the national cancer registry in Sweden which covers the entire population and 40 years of exposure. We were able to identify 12,644 patients who had been identified as either having a parathyroid adenoma or parathyroid hyperplasia and linked this to the cancer registry.

There was no case where the diagnosis of hyperparathyroidism and osteosarcoma occurred in the same patient.

As previously described also, Study GHBJ, the observational study, was designed with input from the oncology board and to date has provided approximately 2,000 additional patient-years of follow-up. No primary bone tumors have been reported in any patient.

We've concluded that it is unlikely that the findings in the long-term study in rats predict a risk for bone tumors in patients who had received teriparatide for treatment of osteoporosis.

We have promptly shared information about the animal findings with the scientific community and with the regulatory agencies. We reported the rodent findings in clinical presentations, the initial presentations given at the Endocrine Society by Dr. Neer, at the American Society of Bone and Mineral Research by Dr. Marcus, and the American College of Rheumatology by Dr. Gennant, and it included information about the animal findings in many subsequent presentations.

We have also included a description of the animal findings in the primary publication of the study data.

The GHBJ study was also designed to collect some additional safety information, but also to facilitate information sharing and, therefore, we had set the study up to maintain contact between the physicians and our prior study patients.

Now, looking forward, we would propose to exclude groups that increased risk for osteosarcoma, such as those with Paget's disease, unexplained elevations of alkaline phosphatase, adolescents or those with open epiphoces (phonetic), and those with a history of radiation to increase the certainty with which we can begin to exclude or further exclude a relationship with teriparatide treatment over time.

To insure the most favorable benefit-risk for this important potential therapy for patients with osteoporosis, we also proposed to limit the duration of treatment for up to two years in post menopausal women and men based on currently available data.

We continue to put patient safety first and provide a commitment to the following elements of a post approval safety surveillance program. Lilly has a worldwide system for assessing spontaneous adverse reports that is already in place to collect information on men and women who did not elect to participate in Study GHBJ. This system will be used to track safety in a post approval setting.

We will continue long-term follow-up of women and men in Study GHBJ, and by 2005, we'll have accrued approximately 7,000 patient-years of follow-up on these subjects.

We are working with the agency to create an active program with a goal of collecting and assessing information on a substantial proportion of cases of osteosarcoma that occur in the United States each year regardless of any treatment they may have received.

Because of the very low incidence of this disorder, we propose to utilize large, stable, population based databases, such as the NCI's SEER database, and also to work with sentinel sites, that is, specialty referral centers where such patients with the disorder receive care.

We will provide a periodic update on prescriptions by geographic region to the agency. We will work and review new information on a periodic basis with an external safety review board. This program will be ready to be implemented at launch.

Now, to summarize the clinical data. Teriparatide treatment improves skeletal architecture. These CT scans of baseline and follow-up iliac crest bone biopsy from a patient treated with teriparatide provides evidence for enhanced architecture, that is, improvement in the trabecular network of bone from the baseline state to the follow-up state after treatment. It is data similar to that which was shown earlier this morning by Dr. Lindsay.

This effect of teriparatide was associated with significant favorable effects on clinical outcomes on study patients, that is, treatment prevented fractures.

We have considered the following in dose selection. In the Phase 3 trial, vertebral and nonvertebral fracture risk was reduced to a similar extent in the 20 and 40 microgram groups in women. While there was a rapid and dose related increase in the surrogate outcome of bone density at the spine and hip in women and men, the actual increase in spine and femoral neck and total hip bone density was similar for women and men.

The 40 microgram dose was more likely to cause adverse events, transient elevations in serum calcium, and resulted in a higher rate of discontinuations from the trials in women and in men.

Teriparatide 20 micrograms is an appropriate dose for treatment of osteoporosis in post menopausal women and in men.

Pharmacokinetic and pharmacodynamic analyses supported that dose adjustment is not required for gender, weight or age.

To summarize the effect of teriparatide 20 micrograms, in women in Study GHAC, teriparatide 20 micrograms reduced the risk of vertebral fracture by 65 percent; reduced the risk of nonvertebral fractures by 53 percent; increased bone mineral density at the spine and hip without a significant effect at the forearm; and increased total body bone mineral. There was no increase in fracture risk for at least 18 months after cessation of treatment.

In the study in men, teriparatide significantly increased bone mineral density at the spine and femoral neck without significant effect at the total hip, and there was a significant increase in total body bone mineral.

The adverse effects associated with teriparatide treatment in the Phase 3 clinical trials in women were nausea and leg cramps. The overall pattern was similar in men, except for that leg cramps were not reported at an increased frequency.

In the clinical pharmacology studies, postural hypotension was observed, but almost always after doses of 40 micrograms or greater.

While the incidence of clinical apparent postural hypotension was not different among groups in the Phase 3 trials, we believe that this is a potential treatment related effect.

We observed increases in serum calcium between four to six hours post dose that had returned to baseline by 16 hours post dose. The levels transiently exceeded the normal range of repeat in only about three percent of women, and there was no difference from baseline in pre-dose serum calcium at any visit.

There was a median increase in serum uric acid of about 20 percent without effect on the incidence of gout or arthralgia.

There was no increase in the risk of cancer, no primary bone tumors were reported, and there was no effect on mortality.

Teriparatide treatment restores bone architecture and bone mass. No other osteoporosis treatment can do this. The now demonstrated ability to prevent fractures confirms that teriparatide can fulfill an important unmet medical need in women and in men with osteoporosis.

Clinical trials support that 20 micrograms per day is an effective and safe treatment for osteoporosis in post menopausal women and in men.

This now concludes the presentation from Lilly. Thank you very much for your attention.

ACTING CHAIRMAN MOLITCH: I'd like to thank the sponsor for a crisp presentation that came in on time.

We now have the opportunity for the panel to ask questions of the sponsor. At this point we'd like to try to ask questions that are specifically related to the presentation, the data presented, as far as factual questions regarding this.

I think additional philosophical questions and other types of things we'll have the opportunity to discuss later.

So if any members of the panel would like to start with questioning, please do.

Dr. Bone.

DR. BONE: Thank you.

I appreciate your very nice presentation. I have one or two -- actually I have several questions, but I'll try to ask them one or two at a time.

With regard to the osteosarcomas, when you investigated the animal tumors, what did you find out about their responsiveness to parathyroid hormone? Do they have receptors? Do they respond in vitro to parathyroid hormone? Are these tumors ones that may have been a result of an effect on early differentiation but no ongoing effect of the tumor by the hormone or is it something that's stimulated as we go along?

DR. MITLAK: Let me invite our toxicologist, Dr. Vahle, to response.

DR. VAHLE: We've not isolated the osteosarcoma cells in vitro to study PTH receptor density or responsiveness to teriparatide. So we don't have any direct evidence to address your question one way or the other.

DR. BONE: Was the receptor expressed in the tissue, in the slides?

DR. VAHLE: We've not done any receptor identification in those specific slides or have grown them in culture either.

DR. BONE: Why?

DR. VAHLE: Because there are technical difficulties in getting to that PTH receptor in those specific slides. Also, in investigating that, it was not clear whether that was going to give us clear information about their relevance to humans.

DR. BONE: I'm a little disappointed that you didn't look.

Okay. I have a couple more questions if nobody else has one right now. Okay.

Could you show us the nonvertebral fracture data in men, the actual data?

DR. MITLAK: Well, the actual data are that there were six nonvertebral fractures in the male study, three in placebo, two in the 20 microgram dose group, and one in the 40 microgram dose group. Is that sufficient?

DR. BONE: Okay. Where were the fractures? What sites? Were they hip fractures?

DR. MITLAK: No, they were not hip fractures.

DR. BONE: None of them?

DR. MITLAK: None of them.


DR. LEVITSKY: Do you have any data or can you summarize data on the serial or concomitant use of bisphosphonates with this agent?

DR. MITLAK: I'm sorry?

DR. LEVITSKY: Do you have any data on the serial or concomitant use of bisphosphonates with this agent?

DR. MITLAK: We have just limited data to share with you on this. Let me ask for slide 4261.

What this slide shows is information from the 58 patients who had reported prior use of bisphosphonate prior to enrollment in the study. Because the study began enrolling in 1995 and '96, the bisphosphonates that were more commonly used and were available included primarily atidronate. There were also a few patients who received alendronate or toludrinate, and in one patient who received abandronate.

These data show that compared with placebo, the overall change in bone density was similar to the larger population. I do not have a lot of information on precisely how long the patients used these, but they had stopped treatment for between six and 24 months prior to enrollment in the study.

DR. KREISBERG: I also have several questions. I'd like to ask whether you conducted any studies in orchiectomised (phonetic) male primates. I didn't understand from the presentation in your experimental models whether the male primates were androgen deficient or not

DR. MITLAK: Dr. Vahle, please.

DR. VAHLE: Consistent with the guidances, the 18-month pharmacology study I described was limited to ovariectomized females. So we have not studied the similar model in males.

DR. KREISBERG: The other question that is partially related to that is whether in the human studies, where you were treating hypergonadal men and men with idiopathic osteoporosis, whether the hypergonadal men also received androgen replacement.

DR. MITLAK: The study in men included approximately half of the men that had idiopathic osteoporosis and half were hypergonadal. Testosterone treatment, if it was being used by men, could be continued during the study, but was not permitted to be started de novo during the study.

A small proportion of men, in the range of ten percent or less, had been taking testosterone or an androgen replacement into the study, and as we said, overall the response in men with idiopathic and hypogonadal osteoporosis to teriparatide treatment was similar.


DR. AOKI: Do you have any data or are you planning any studies on monkeys older, for periods longer than 18 months, or on rats that are older than six to eight weeks to determine if the osteosarcoma is, in fact, somehow age related in the rats and to see the more relevant model, whether or not the osteosarcoma question can be laid to rest using longer term studies?

DR. MITLAK: Let me ask Dr. Vahle again to comment on the work that's ongoing.

DR. VAHLE: Since the initial observation, we've worked closely with our experts as well as the FDA in developing some ongoing research that I'd be happy to share with you.

If I could please have slide 4222, let me briefly highlight the two main components of this.

First, in response to the second portion of your question, yes, we are conducting a follow-up rat study which looks at two things: one, the effect of treatment duration and, two, the effect of age at treatment initiation.

In this respect it addresses the question. We have treatment arms which avoid the phase of rapid skeletal growth, and this is a study that was conducted or designed in collaboration with the agency as a Phase 4 commitment.

In terms of additional monkey work, what we are doing is an additional study which has an 18-month treatment period. This represents approximately eight percent of the monkey's lifetime at exposures up to eightfold human exposures, but it contrasts with our prior work in that it's followed by a minimum three year observation period to allow us to have some extended follow-up data in the primate model, and again, this is a study that we are in the early stages of and designed with the agency.

DR. BONE: Going back to the series of questions, could you discuss what studies you are conducting concerning the -- or have conducted -- concerning the mechanism by which these osteosarcomas were induced, biological mechanism?

DR. VAHLE: As part of that ongoing research program, another component of that was to convene a group to try to discern what type of mechanistic studies would be useful in trying to assess the relevance to humans, and again, this is something that we have discussed with the division.

It has not been clear that there are a direct set of experiments that will help us understand the mechanism in the rat and then clearly differentiate it from the humans at a cellular or molecular level. Rather, we have focused on these effects of treatment duration and age of initiation because it's clear these differences between the rat model and the human that we want to more clearly establish.

In those follow-up studies, we are continuing to evaluate new technologies, such as gene array or genetic characterization to see if they would provide any assistance or any additional insight.

DR. BONE: Have you completed any studies addressing this mechanism at all?

DR. VAHLE: No, there have been no studies completed to date. The studies and the concepts I've outlined are all in progress. What I can share though is interim results from the long-term rat study, and that following six months' treatment duration, both during the rapid phase of skeletal growth as well as after the rapid phase of skeletal growth, there are no bone proliferative lesions, and there are the anticipated exaggerated effects on the skeleton, but again, that study is still in progress.

DR. GRADY: I'd like to ask a little bit about nephrotoxicity. It seemed that in one of your monkey studies at least there was a fair percentage of the animals who had nephropathy, and one out of eight in that study with renal failure, and I don't think you talked about that at all.

DR. MITLAK: Please.

DR. VAHLE: I'd certainly be happy to address the renal findings.

If I could go back to the main slide 28, please, we've studied renal tissue and renal function in two different models. In the toxicity studies, and there are a group of three different toxicology studies represented here, we observed these subtle histologic observations in the kidneys of monkeys over a range of doses and over a range of duration of exposure, both three-month studies and up to one year.

We conducted -- because in those routine studies there was no clear evidence that these renal changes had an impact on renal function, we conducted a special study to determine if these changes had effects on renal function.

That study was conducted at a high dose of 40 micrograms per kilogram. That provides exposures that are in excess of 100-fold what a woman would receive in a 20 microgram dose would do. In that study, eight monkeys received this particular dose. Seven of eight of those monkeys we were able to reproduce the lesion, and one monkey did not develop the lesion.

Of those seven monkeys that had the lesions, six of these had no changes in renal function as far as creatinine clearance, urinary concentrating ability, urinary acidification ability.

One of those monkeys developed a sustained hypercalcemia. Serum calcium pre-dose, not post dose, but pre-dose serum calcium was up to 14 milligrams per deciliter. That monkey did develop renal failure in association with that hypercalcemia, and that monkey after removal of teriparatide treatment and after the hypercalcemia resolved, renal function returned and the lesions were at least partially reversible.

Does that address the question?

And in addition, I didn't highlight those are all findings from the toxicology model. Going back, again, to the pharmacology study, this is a study where monkeys were treated for up to 18 months, and there were 20 monkeys per group. So a more robustly powered study, and we did not see any renal alterations.

ACTING CHAIRPERSON MOLITCH: Just to pursue this particular area, in the human studies was urinary concentrating ability looked at?

DR. MITLAK: No. In the human studies, we measured creatinine and creatinine clearance. Concentrating ability was not measured.

ACTING CHAIRPERSON MOLITCH: I mean, certainly in even the hypercalcemic states and hyperparathyroidism, concentrating ability is probably the earliest thing that's noted. Why wasn't that looked for?

DR. MITLAK: What we found in the clinical studies was that urinary calcium changed to a very small degree. Urinary calcium, as was highlighted by Dr. Gaich, changed on average by about 30 milligrams per day.

We also saw no difference in the proportion of patients with hypercalceria across the treatment groups. So between those changes and the assessments that we made, we felt we had assessed renal function. We did not measure concentrating capacity.

ACTING CHAIRPERSON MOLITCH: And, again, continuing with this line, the patients who were treated with hydrochlorothiazide and furosemide at low dose, plus the PTH, there was no particular change in serum calcium that occurred in those patients; is that correct?

You said there was no drug interaction.

DR. MITLAK: Dr. Gaich.

DR. GAICH: Yes, that is correct. Among the patients treated with thiazide diuretics in our Phase 3 study, we looked at the serum calcium response, and it was similar.

In addition, we did a specific clinical pharmacology study to specifically look at the interaction on serum in urine calcium between teriparatide and thiazide diuretics, and likewise there is no interaction there.


DR. GELATO: Hi. This is just to follow through with this.

In going through your safety data, it was noted that there were a small number of patients who had calciums that exceeded 11, and so what I wasn't clear about was did that -- was that also a transient elevation or did it persist?

And were they the same patients who had increases in urinary calcium excretion?

And there was a subset, I think that continued with impairment or at least elevated serum creatinines, and I wondered if there was a link between those findings of the elevated calcium, urinary calcium in the creatinine to sort of get at some of these issues.

DR. MITLAK: Let me invite Dr. Gaich back to address those questions for you.

DR. GAICH: Thank you.

Let me start from the bottom and work my way up.

First, we did look for a relationship between the increase in serum calcium and effects on serum creatinine or creatinine clearance, and we did not find one.

Second, all of the calcemic effects that were observed were transient. So even the patients that had the highest serum calciums, the baseline serum or the pre-dose serum calcium is back down to normal.

And finally -- what was your third question?

DR. GELATO: Was there a relationship to those patients because --

DR. GAICH: Between serum calcium and urine calcium?

DR. GELATO: And the elevation of serum creatinine.

DR. GAICH: Okay. There was not a relationship between -- a strong relationship -- between the patients who had high serum calcium and high urine calcium, nor was there any relationship between the patients who had high serum calcium transiently and an increase in serum creatinine.

DR. TAMBORLANE: Again, on the same, just even from that individual animal experiment, the case, it seemed to me I was hearing the suggestion that serum calcium did not have to be monitored during therapy, and maybe under the normal circumstances, but it's likely that patients with hyperparathyroidism might be exposed to the drug, and there's very limited data.

Is that your continued suggestion that calcium not be monitored?

DR. MITLAK: Let me invite Dr. Gaich again to help address this question.

What we are suggesting and what we have observed in the clinical studies is that the incremental change in serum calcium in patients seemed to be independent of the baseline serum calcium, that is, whether somebody is in the low, mid, or upper part of the range, the increment in calcium was fairly consistent with the dosing.

Therefore, we recommend that high calcium, hypercalcemia be excluded before patients are considered for treatment, and once that has happened, we found based on the clinical trial results that that is a reasonable course of action.

Let me see if Dr. Gaich has --

DR. GAICH: Thank you.

Could we look at slide 4455?

We actually did look at a number of factors to determine if there were any particular characteristics of patients who would have higher responses of serum calcium, and I will show you what we evaluated.

I'm sorry. I need 4455. We need to go one back. There we go.

We looked at the relationship between the highest post dose serum calcium and baseline serum calcium, baseline serum, 25 hydroxy Vitamin D, the body mass index, the baseline intact parathyroid hormone 1 to 84, and age.

Now if we can go to 4456, please.

The only significant relationship or strongly significant relationship was the relationship between baseline serum calcium and the highest post baseline. The correlation coefficient was .45, which was highly statistically significant, and as you can see, based on the graph, the higher the baseline serum calcium, the higher your post baseline serum calcium.

And this was the only strong predictor of baseline -- of post baseline serum calcium.

As Dr. Mitlak also mentioned, we also looked at the relationship between baseline serum calcium and the change in serum calcium, and there was not a positive relationship.

So patients who started with high baseline serum calciums did not have an exaggerated response.

May I have the next slide, please, 4457?

Among the other things, there were some weak negative and weak positive correlations. There was a weak positive correlation with 25 hydroxy Vitamin D, the correlation coefficient of .13.

Weak negative correlations of particular interest to your question is there was a negative correlation between baseline intact parathyroid hormone and the highest post baseline serum calcium. So patients who started with higher intact PTH at baseline tended to have lower post baseline serum calciums.

Nevertheless, we do believe that patients with hypercalcemia should not be treated with teriparatide.

DR. KREISBERG: I have two questions. One is other than a reduction in pain, do you have any other quality of life indicators about these patients? Did they generally feel better, worse or the same?

The reason I ask is that in primary hyperparathyroidism, which I'm not suggesting this is comparable to, there are neuropathic and muscular types of symptoms that patients have other than just cramps.

DR. MITLAK: Based on assessment of adverse events, those sorts of symptoms were not seen.

DR. KREISBERG: Okay. The other question actually relates to the longest duration of therapy that patients have received teriparatide, and I believe in one of Dr. Lindsay's slides, it was up to 36 months.

Based upon the change in the markers of bone formation and bone resorption, one would predict eventually that that would come into balance and the bone density would plateau. So one of the question is: is that true?

And then a follow-up question is: how long would you intend to use teriparatide for the treatment of osteoporosis? Do you see that as an indefinite exposure to the hormone?

Because I think that gets to the issue that is troubling everybody, and that is longer term exposure might, in fact, bring out some side effects that haven't been brought out by short term exposure.

DR. MITLAK: Let me answer in part, and then invite Dr. Lindsay up to comment on part of your question.

As I laid out in my final comments, I think based on the available data and to maximize the benefit-risk for patients, we would propose to limit duration of treatment for two years until further information is available to help us.

DR. LINDSAY: We have treated people for up to three years with parathyroid hormone 1 to 34, and in those studies, the bone mass changes continue for the three years of the study.

We subsequently followed those patients still remaining on hormone replacement therapy, and their bone density plateau has remained stable.

During the third year of treatment, it's interesting that the biochemical markers of formation and resorption are returning back to baseline, despite continued treatment with parathyroid hormone, and we think that the increase in bone density that you see during the third year is the phase of secondary mineralization that would follow the synthesis of newborn matrix.

And I would agree with you that longer term use is probably going to be associated with a plateauing. We just don't have data out beyond that three years.

ACTING CHAIRPERSON MOLITCH: Dr. Lindsay, while you're still there, you cited the well known data that there's an increase in mortality associated with fracture. I don't think you meant to imply that there are any studies that show the intervention to increase bone mineral density with perhaps decreased fracture as might decrease mortality rates.

DR. LINDSAY: No, I did not show data about that.


Dr. Schneider, do you have one?

DR. SCHNEIDER: I had one small, beginning technical question. Could you just tell us the multiple comparisons procedure that you used to adjust your p values, given that you were looking at two active doses?

I couldn't find that in my briefing document.

DR. MITLAK: Dr. Wang, would you please come to the microphone?

DR. WANG: My name is Ouhong Wang. I'm the statistician on the teriparatide product team.

To answer your question, the study was designed to control for the primary efficacy variable at the .05 level. For the secondary comparisons, everything is reported at the nominal .05 level. It's not adjusted.

But, in essence, the protocol is designed in a way that we wouldn't report any secondary efficacy results if the primary efficacy result is not significant. So it is kind of a gatekeeper strategy.

DR. SCHNEIDER: I'm sorry. Could you say again something about the primary efficacy variable? How did you handle multiple comparisons on that?

DR. WANG: The primary efficacy actually is the combined -- well, when you look at the particle, it is the combined teriparatide doses, 20 and 40 microgram groups compared with placebo. So that's the primary, and to separate the doses we will also look at the separate doses versus placebo.

DR. SCHNEIDER: Thank you.

The second question I had is in GHAJ the primary efficacy variable was noted as a change in lumbar bone mass density, and you presented the percent change and later indicated that the change was independent of baseline.

Do you have data or analysis on just the change from baseline in lumbar BMD?

DR. MITLAK: Let me ask our group if we have the slide.

I can tell you that the analysis of change rather than percent change was identical. The statistical inferences were identical, except for that the change at the total hip, which was not significant by percent change was significant for actual change.

DR. SCHNEIDER: And the final question I had was in the AC study, did you look at BMI or weight, the effect of that as a covariant in either the adverse experiences or the efficacy variables? And what level of effect did it have?

DR. MITLAK: Let me invite our pharmacokineticist, Dr. Satterwhite, to come to the microphone to address your questions.

DR. SATTERWHITE: My name is Julie Satterwhite. I am a senior research scientist at Lilly, and I was responsible for the pharmacokinetic and pharmacodynamic analyses.

For the pharmacodynamics we looked at -- in terms of efficacy, we looked at the biochemical markers and BMD response. We did evaluate body mass index and weight and found that neither one of them was a significant covariant governing response.


DR. SATTERWHITE: Yes. We saw that in both AC and AJ, and it was not a significant covariant.

DR. KREISBERG: Do you remember what the slope was? Was it positive, negative? In fact, the correlation, even though it wasn't significant?

DR. SATTERWHITE: I can get that answer for you.

DR. KREISBERG: Thank you.

DR. GRADY: I'd like to ask about calcium intake. In this study it was recommended that women take, I think, a gram of calcium per day, and I think one of the things we've perhaps been fairly successful at is getting most post menopausal women to take calcium supplementation.

I wonder if you adjusted calcium supplementation during the study and also if you planned to recommend calcium supplementation in addition to the drug during treatment.

DR. MITLAK: Let me invite Dr. Gaich up also while I tell you that the mean intake at baseline in the women was in the range of seven to 800 milligrams per day so that a 1,000 milligram supplement resulted in a total of approximately 1,700 to 1,800 milligrams of calcium.

We expect that going forward, that patients who receive treatment would take calcium supplements. We would recommend that their calcium total intake be adjusted to that recommended for patients with post menopausal osteoporosis or osteoporosis in men.

I might ask Dr. Gaich to help comment on any dose adjustments that have occurred in the study.

DR. GAICH: Okay. Thank you.

First of all, a flat dose of 1,000 milligrams a day was prescribed for all of the patients, was recommended for all of the patients. So we did not adjust based on dietary intake to bring up to some level, and again, we think that's fairly more typical of the clinical practice than doing an extensive dietary survey and doing an adjustment.

The physicians were allowed to change calcium supplements to or to adjust calcium supplements based on side effects, especially GI side effects with some supplements, and also if the patients had transient increases in serum calcium or urine calcium, which was documented on repeated measurements.

And the number of patients who underwent adjustments in the calcium supplements was fairly small.

DR. GRADY: What does "fairly small" mean? And was it the same in the two groups?

DR. GAICH: I'm sorry. The question was what was fairly small and was it the same in the two groups?

Yeah, first of all, let's see. If I can have slide 3373.

This slide will show the incidence of the number of patients who had one and more than one increase in serum calcium, as well as the number of patients that had adjustments in calcium or study drug.

And this is the line that we're looking at. Among the patients that had an increase in serum calcium, 7.2 percent or 7.2 percent of the patients had a decrease in their calcium intake as a result of a transient increase in the serum calcium. So that's what I mean by "fairly small." It was lower in the placebo group, and because there were more patients with transient increases in serum calcium in the high dose group, there were more in the high dose group.

Thank you.

DR. GRADY: I'd also like to ask about uric acid. You know, I know you kind of sort of mentioned, but could you just tell me the percentage of participants who had elevated uric acid in the two groups? Because it does seem that that also is a persistent problem.

DR. GAICH: Yes, the increases in uric acid were similar to the order seen by other things, such as thiazides and aspirin therapy, things along those lines.

The number of patients with increased uric acid in the 20 microgram group was 2.8 percent, in the 40 microgram group was five percent.

By study endpoint and six months follow-up, the serum uric acid concentrations were very nearly back down to baseline. The difference between placebo and the treatment groups was less than two percent, even though that was still statistically significant.

And at that time, there was no difference in the number of patients with high uric acid concentrations.

DR. GRADY: What did you define as high?

DR. GAICH: The upper limit of normal was -- let's see. If we can have my main slide.

DR. GRADY: I really just want to know the percent or proportion above whatever you defined as high in the two groups.

DR. GAICH: Correct. My main slide 83.

It has that on there. I just want to make sure I give you the right number. Yes, it was 9.0 milligrams per deciliter, and the reference ranges are based on a large database, over 20,000 clinical trial patients and are adjusted for age and gender as well where appropriate.

DR. BONE: Thank you.

I have a series of questions as well, and I'll just continue, if I may, with Dr. Grady's line of questions about the uric acid.

There's two issues here. I think one is the number of patients who exceed the fairly high upper limit that you used, and the other is, you know, how the sort of overall curve shifts for "uric acidemia," if I can put it that way.

Did you get an idea of the interactive risk of hyper uric acidemia in patients taking other concomitant medications, such thiazides, or any other risk factors for the development of either an overtly elevated uric acid level or an increase in the uric acid level of, let's say, two milligrams per deciliter or so?

DR. GAICH: The data that we looked at is we looked at all of the data, including concomitant medications, adverse events, laboratory effects for all the patients that had an increase in the serum uric acid above the upper limit of normal, and in that group there were not patients who -- a lot of patients who were on thiazides. There weren't enough patients who had high uric acid and who had thiazides in our study for us to do a meaningful kind of analysis looking for the interaction.

DR. BONE: We may come back to that question later. Let me ask you some questions about the Vitamin D status of the patients. Obviously the Vitamin D status of patients who would potentially take this medication is of considerable concern because if we accelerate bone turnover at the same time as having insufficient Vitamin D, we may induce a mineralization defect that might not have been apparent in the clinical trials.

Can you tell us what the baseline 25 hydroxy Vitamin D status was for your patients and also what was the effect on 125 dihydroxy Vitamin D levels in the patients in the treatment groups?

DR. MITLAK: Let me ask Dr. Gaich to come to the microphone again.

Let me also first show you the slide to answer your second question, which is the 125 dihydroxy Vitamin D change during treatment for the three groups.

It's slide 4260.

What this panel shows is measurement of 125 Vitamin D over the first year of the study, and it shows that with treatment, 125 Vitamin D levels increase.

In combination with this, we actually see a slight decrease in 25 Vitamin D levels, which we presume is part of the conversion process.

And now if I could ask Dr. Gaich to answer the first part of your question.

DR. GAICH: Thank you.

All the patients in our clinical trials required to have a 25 hydroxy Vitamin D above the upper limit of normal. Some of them at screening were below the upper limit of normal, but then came into the normal range with supplementation.

DR. BONE: I think you misspoke.

DR. GAICH: I'm sorry.

DR. BONE: You said that all of the patients had to be above the upper limit of normal?

DR. GAICH: I'm sorry.

DR. BONE: I'm sure you didn't mean that.

DR. GAICH: Had to be above the lower limit of normal. Thank you.

DR. BONE: Meaning what?

DR. GAICH: We used the standard laboratory reference range, and I'd have to look that up for you.

DR. BONE: Well, as you know, most of the standard laboratory reference ranges are considered to be -- in most of the standard laboratory reference ranges what is presented as the lower limit of the reference range is widely regarded by clinicians in this field as consistent with Vitamin D insufficiency.

So I think it's a specific question we'd like a specific answer to as to what the distribution of 25 hydroxy Vitamin D levels actually was in the trial, and we may want to give some further thought to whether we can really account -- I think in the briefing document you said there was about a 25 percent decrease in mean 25 hydroxy Vitamin D levels, and this was explained or supposedly explained by the conversion to 125 dihydroxy Vitamin D, but since the ratio of the actual mass of 25 hydroxy to 125 dihydroxy Vitamin D is a ration of nanograms to picograms, I think that we will have to invoke some additional explanation for that phenomenon, and perhaps you will be able to comment on that after lunch.

One or two additional questions. One of the striking findings in your results was the failure to protect height. The usual result in trials where there's a substantial reduction in the rate of vertebral fracture, such as you have very nicely described, is that there is also a measurable difference between the height loss in the treatment groups and the height loss in the placebo groups.

And I'm wondering what you've done to try to identify a basis for that phenomenon. For example, since you have the radiographs, was an attempt made to assess the effect on actual vertebral heights to determine whether the height loss in the patients in the different groups could be explained in that way? Did people look at disk spaces? What was done to try to figure out why there was a discrepancy between your very impressive reduction in fracture rate and the lack of any apparent effect on height?

DR. MITLAK: We actually don't think that there is a discrepancy. I think because height changes likely occur in patients with fractures, and most patients in the treatment groups did not have fractures, it was not surprising to us that we didn't see overall differences.

But to answer your question about what we did to try and address this, let me ask for slide 4246.

What we did in this analysis is to take all of the patients in the study regardless of treatment assignment and stratify them by the most severe fracture grade. In other words, we took patients who did not have a fracture, those who had a mild fracture, a moderate fracture, or a severe fracture, and based on these grades looked at change in height.

And just as you might expect, patients with more severe type of fractures actually did lose height. We believe, again, because most patients did not have fractures in this study that it was not possible to see this effect if we looked at all of the patients together.

DR. KREISBERG: Did you use a stadiometer to measure height? How did you measure height in this study?

DR. MITLAK: Yes, stadiometers were used.

DR. BONE: To continue, one of the questions that we're concerned with is the duration of treatment, and it's clear from your data that most of the increase in height occurs in the first year or -- excuse me -- most of the increase in bone density occurs in the first year on treatment with a smaller, much smaller increase in the second year, and Professor Lindsay has described the phenomenon in the third year of increased density despite declining turnover.

This suggests that somewhere between the end of the first year and the end of the second year you start having more of a phenomenon of filling holes than you do of actually laying down more matrix.

Some of your patients completed about two years on treatment, and many only completed about a year. Did you look at what happened to -- since we know that there was not much of an increase in the second year in bone density compared with that in the first year, an interesting question is what happened to the relative risk of fracture in the patients whose second year of observation was off drug compared with those whose second year of observation was on drug.

In other words, was there a protective effect of being on drug in the second year or was the protective effect against fracture mostly carried over from the main gain in bone mass in the first year?

DR. MITLAK: There's several parts to that question. Let me try and address them, and then I'm going to invite Dr. Neer up to make a comment also.

I think that as we look at the data from these studies, we certainly agree that the rate of change in bone density in the spine becomes less over time.

However, and I think importantly, if we look at the rate of change in bone density at the hip or the total body, it's more of a linear change, and that is that patients do have proportionate increases in those two important measurements over time.

I think to your question about looking at fracture risk at an earlier time point or for those who were treated for a shorter amount of time, we cannot do that in this study for spine fractures because spine fractures were only assessed by baseline and endpoint radiographs.

We can do it for nonvertebral fractures, and I think the data show that for the fractures that we track, that after nine months there was a progressive reduction in the risk of fractures, and that as we followed patients out off of treatment, the risk of fracture did not increase, and I think that's the answer that I have.

And let me ask if Dr. Neer would like to comment further.

DR. NEER: I'd like to make a comment. I'm Robert Neer. I was involves in helping to design and conduct the trial GHAC.

I'd like to make a comment in response to Dr. Bone's question about height. Approximately 14 or 15 percent of the women in the study GHAC had a fracture. That means that 85 percent did not, and as in prior trials of, for example, alendronate, it is very difficult to demonstrate effects on height if one dilutes the therapeutic effect by including large numbers of people who don't have an adverse endpoint.

So, for example, as in trials of alendronate, if one analyzes the entire patient population, there's no change in height as a consequence of treatment. That is, treatment doesn't protect against height loss.

But as with alendronate, if one restricts the analysis to people who had an incident fracture, then there's a very clear effect on protecting against height loss. The treatment in those patients is clearly associated with less height loss.

As we reported in the paper in the New England Journal of Medicine, there was a statistically significant height loss in women in GHAC in the placebo treatment group, but there was no statistically significant height loss in either of the PTH treatment groups, and the difference between the PTH treatment groups and the placebo group was also statistically significant.

So it depends upon trying -- if you want to see an effect on height loss in studies of such patient populations regardless of the drug being evaluated, you need to restrict the analysis to people who had had a new incident fracture.

DR. BONE: I expect you have the data to answer the question I posed a little more specifically, but I'm not sure you conducted the analysis, and that was to look at the patients who completed one year on therapy, and then you followed this out.

You said you had the nonvertebral fracture data because those are spontaneous reports of clinical fractures. You didn't do vertebral height measurement or didn't do spine films after the interruption of the trial?

DR. MITLAK: Yes, we did. We did.

DR. BONE: Well, if you have the films --


DR. BONE: -- for the spine films, then I'm not completely clear why you can't look at incident vertebral fractures in the group that got a year of treatment and then were followed compared with the group that got two years of treatment.

Maybe I'm missing something here.

DR. MITLAK: Let me try and answer again.

What we have done is to collect radiographs in the follow-up phase after all of the patients had discontinued treatment with drug. I do not have data to show you for patients who may have discontinued treatment during the study and then were followed in the study to the endpoint visit.

If you wish, I can show you the data that we had collected systematically after all of the patients had been asked to stop treatment, if that would address your question.

DR. BONE: Well, that's what I'm talking about.

DR. MITLAK: I'm sorry. then I misunderstood.

DR. BONE: Do you have the patients -- if I'm not mistaken, you have patients who completed about a year and then were stopped, right?

DR. MITLAK: In the -- I think the point of misunderstanding -- in the study in women, the median duration of treatment was 19 months.

DR. BONE: Right.

DR. MITLAK: Okay, and what we've done is to have follow-up radiographs done now about 18 months after the time that they stopped treatment.

DR. BONE: So some of those patients, the ones who were about a year, have about an 18-month follow-up after one year of therapy, and those who got closer to two years would have 18-month -- would have a period of observation of about two years.

Does this just mean that the analysis I asked for -- I'm not expecting you to have done every single conceivable analysis. I'm just asking if you have that information.

What I'm trying to find out is whether the fracture risk reduction is mainly the result of the first year treatment or whether there's an incremental effect on fracture risk that's due to the ongoing application of the drug.

DR. MITLAK: We don't have data to answer that question for you.

DR. BONE: You haven't analyzed the follow-up data for that purpose?

DR. MITLAK: That's correct. We have --

DR. BONE: Thank you.

DR. GAICH: We just have the data that Dr. Bone asked, and very few patients had a year or less of treatment prior to the study closeout. It was only between ten and 15 percent in each treatment group. so not really enough to do an adequate vertebral fracture analysis.

DR. BONE: How many had between 12 and 18 months, in other words, below the median?

DR. GAICH: I'm sorry. You found it?

DR. BONE: I guess it would be about half.


ACTING CHAIRPERSON MOLITCH: We'll take a last question before the break from Dr. Tamborlane.

DR. TAMBORLANE: I think you showed the post -- sort of the follow-up data after the study was stopped as far as fracture rate, but in regard to sort of follow-up of Dr. Kreisberg's question, do you have the -- because it relates to duration of treatment -- do you have the bone marrow density data post discontinuation of the trial?

DR. MITLAK: yes.

DR. TAMBORLANE: Over time?

DR. MITLAK: yes.

DR. TAMBORLANE: Because that would say whether the density then goes back. I think those data -- I don't believe we saw those data.

DR. MITLAK: Let me ask you to put slide 4304 up, please.

Let me also explain as a preface, as Dr. Gaich had highlighted, approximately 80 percent of the patients who had previously been enrolled in the prior study elected to continue into the follow-up study, the follow-up study was an observational study. After the primary study database was locked, patients were unblinded to treatment assignment, and in the follow-up phase, patients could take other treatments for osteoporosis.

About half of the patients by 18 months out had begun to take some other treatment for osteoporosis, but the use of these treatments, whether it was any specific treatment or the use of any treatment was statistically similar across groups. So we have these data shown with that piece of background information.

What this slide shows for the spine is that in the first -- the data shown are for the endpoint of the prior study, the first visit for the follow-up study, and the second visit for the follow-up study. This is six months and then an additional 12 months.

It shows that the bone density decreases from the endpoint visit, but remains statistically significant for the next 18 months and is different from placebo even 18 months after treatment.

Let me ask also for you to show the next slide 4305, which is the same type of analysis at the hip.

All right. Thank you.

ACTING CHAIRPERSON MOLITCH: I think at this juncture we will take a break. We will be able to ask the sponsor additional questions after the FDA presentation.

It is now 10:32. We'll resume at 10:47.

(Whereupon, the foregoing matter went off the record at 10:34 a.m. and went back on the record at 10:55 a.m.)

ACTING CHAIRPERSON MOLITCH: We will now continue with the FDA presentation. The first person to present will be Dr. Kuijpers, who will be discussing the preclinical studies.

DR. KUIJPERS: Thank you, Mr. Chairman, ladies and gentlemen.

My name is Gemma Kuijpers. I'm a pharmacology reviewer in the Division of Metabolic and Endocrine Drug Products.

I thank you for giving me the opportunity to talk today about the preclinical safety of teriparatide. After my presentation, Dr. Bruce Schneider will address clinical efficacy, and Dr. Bruce Stadel will talk clinical safety of teriparatide injection.

In this presentation, I will focus on the main preclinical safety issue that emerged during the development program of teriparatide, namely, that teriparatide injection causes bone neoplasms in the rat.

First, I will briefly describe the purpose and design of carcinogenicity studies. Then I will address the data obtained in the two-year study, and finally, I will discuss the clinical relevance of the tumor findings.

For most new drugs for long-term use, the FDA recommends testing for carcinogenic potential. The most elaborate and stringent test for carcinogenicity is the in vivo rodent bioassay. This bioassay is usually done in both the rat and the mouse. It's carried out over a large part of the animal's life span, usually one and a half to two years, and with multiple dose groups, including a maximum tolerated dose to maximize the potential for detecting tumorigenicity.

Animals are sacrificed at the end of the study. Old tissues are examined histologically, and the statistical analysis is carried out to determine the significance of the tumor findings.

Finally, an attempt is made to evaluate the clinical relevance of the findings using all the data that are available on the pharmacologic and toxicologic effects of the drug.

To assess the carcinogenic potential of teriparatide, the sponsor carried out a carcinogenicity study in one rodent species, the Fisher 344 rat. The animals were treated for two years by subcutaneous injection. There were four dose groups: control, low, mid, and high dose group. And the drug was given to 60 animals per sex per group.

All tissues were examined of all animals in the study. Histologic evaluation took place after the animal was sacrificed per protocol at the end of the study or after the animal had died prematurely due to any cause. No interim sacrifices were done.

The bone sites examined were the femur, tibia and sternum in all animals, the vertebrae in most animals, and all gross palpable lesions at other skeletal sites.

As mentioned by the sponsor, teriparatide caused a number of different types of bone neoplasms in the rat, the majority of which were malignant osteogenic sarcomas. This graph shows the incidence of animals with any bone neoplasm in the four different dose groups. The incidence is expressed as percent of animals affected.

There were no tumors in the controls, and in the treated groups, the incidence varied between about five percent and 60 percent in the males and between about seven percent and 40 percent in the females. The effect was clearly dose dependent.

The bone tumors that were observed originated from cells in the osteoblast lineage and are very rare tumors in the rat. They were often seen before the end of the study as grossly palpable bone lesions. Several of them were malignant osteosarcomas that were fatal and metasticized to soft tissue sites.

Teriparatide did not cause a significant increase in the incidence of any other type of tumor.

This slide shows the systemic exposure to teriparatide and the human exposure multiples associated with the three different doses used in the two-year study. In the low dose group, systemic exposure was equivalent to approximately three times the human exposure at a clinical dose of 20 micrograms per day, while in the higher dose groups the AEC multiples went up to about 60 times in the high dose group.

This graph shows the relationship between the systemic exposure to teriparatide and the osteosarcoma incidence. Note that the exposure on the X axis is expressed as multiple of human exposure, again, at the 20 microgram clinical dose. The graph shows a clear relationship between systemic exposure and tumor incidence.

Osteosarcomas were detected at several sites throughout the skeleton as summarized in this slide. In males, the most frequently affected site was the tibia and after that the femur, and in females the most frequently affected site was the vertebra.

This graph shows the time of death of all animals in the male groups that were diagnosed with osteosarcoma. Note here that death occurred either due to scheduled sacrifice at the end of the study, around 730 days, or prematurely at some point before the end of the study.

For most, but not all of the animals that died prematurely, death was due to the osteosarcoma being fatal. Overall there was no increase of mortality with dose.

The conclusion from this graph is that in addition to an increased incidence, the osteosarcomas were detected earlier in the higher dose groups. The earliest tumor that occurred in the high dose male group was a vertebral osteosarcoma that was detected microscopically in an animal that died as a result of the tumor being fatal after 13 months of treatment.

A similar graph depicting time of death of females with osteosarcoma is shown in this slide. Although less pronounced than in the males, the same pattern can be seen, namely, osteosarcomas being detected earlier in the higher dose groups. The earliest tumor in the female high dose group was a fatal tumor in the skull bone in an animal that died at approximately 20 months.

As the sponsor has shown with QCT scans, teriparatide has a marked effect on bone mass in the rat. In this graph the relationship between bone mineral content of the vertebra in female rats is plotted against duration of treatment on the X axis for the different dose groups included in control.

Although most of the osteosarcomas in the two-year study were detected in the later time period of the study, it is not known when the tumors were actually first present in the animals.

The following slide shows the incidence of osteosarcoma in control Fisher 344 rats. In the current study with teriparatide, there were no tumors in either male or female rats in the control groups, and the incidence was zero percent.

Historical control data on osteosarcoma incidence in Fisher rats are also shown. These data are from control experiments carried out previously in the sponsor's research lab or from an historical control database of the National Toxicology Program. The data show that the spontaneous incidence of osteosarcoma in Fisher rats is extremely low and amounts to approximately 0.2 to 0.4 percent.

Since there were no osteosarcomas in the current teriparatide study, in the control animals of the current study, we used the average historical control incidence of 0.2 percent to calculate the relative risk of osteosarcoma in the teriparatide treated rats. The relative risk is shown in this line.

And note that even though the incidence of osteosarcoma in the low dose teriparatide group appeared fairly small, was about six percent for males and females, average, this translates to a relative risk in this dose group of 30-fold. Obviously the relative risk was increased in a dose dependent manner.

As the sponsor has clearly demonstrated, teriparatide markedly and dose dependently increases bone mass in the rat and in other species at all bone sites examined. However, this positive effect of teriparatide must be balanced against the adverse effect observed in the carcinogenicity study.

Those results were that teriparatide causes osteoblast neoplasms. The tumor induction is dependent on the dose and on the treatment duration, and occurred earlier in the higher dose groups. Tumors were detected in all dose groups and a no effect dose or threshold dose level was not established.

The main question we are now confronted with is what's the relevance of these animal findings for humans or what can we conclude from these data regarding the risk of bone tumors in humans treated with teriparatide.

First, some remarks about hormonal carcinogenesis. The current thinking is that in the multi-stage process of carcinogenesis, hormones can act as tumor promoters or co-carcinogens through a nongenotoxic or epigenetic mechanism. Specifically, a hormone can stimulate target cell proliferation and in that way confer a selective growth advantage to precancerous or initiated cells.

Although the exact mechanism underlying the teriparatide induced formation of bone tumors has not been elucidated, it's a plausible hypothesis that in conjunction with its positive effect on osteogenesis, repeated hormonal stimulation of the osteoblast would cause an increase in cell proliferation which would drive the accumulation of genetic errors and increase the chance of neoplastic transformation.

One other factor that could contribute to an increased chance of survival of precancerous cells is the inhibition of apoptosis, or programmed cell death, which is thought to be one of the effects of intermittent activation of the osteoblast PTH receptor.

Having said all this, the clinical relevance of the rat tumor findings depends on whether the mechanism of tumor promotion is operative in humans. Since we don't know whether this is the case or not, the simple conclusion here will be that the relevance of the rat tumors is not clear.

A number of considerations have been put forward to suggest that the rat bone tumor findings are unlikely to have any clinical relevance. These are the validity of the rat model, the lack of bone tumors in an 18-month monkey pharmacology study, and the lack of an association between hyperparathyroidism in humans and osteosarcoma.

Dr. Schneider will expand on the last points in his presentation, and I will elaborate on the validity of the rat model.

The sponsor has argued that the tumors found at the two-year study are unlikely to be predictive of an increased risk of osteosarcoma in humans. This position is based on the notion that the rat model is different from the human.

In fact, there is an exaggerated bone response to teriparatide that may be related to a difference in skeletal biology between rats and humans.

Also, the animals were treated from a young age and for a relatively large part of their life span.

Although true, all of these arguments relate to quantitative aspects of treatment and quantitative aspects of the two-year study carried out in the rats. I'd like to emphasize at this point that these kind of quantitative differences between animal and human studies, such as regarding dose and treatment durations, are intentional differences that are put into place in any type of toxicity study in order to maximize the ability to pick up any signal for a possible adverse event.

Despite the possible differences, quantitative differences between the rat model and the human, the main point, however, here is that there's no evidence that the human osteoblast is in any qualitative way different from the rat osteoblast in its response to intermittent PTH receptor activation. In fact, there is very strong evidence that the osteoblast mediated bone response to teriparatide is similar in rats and in humans, namely, an increase in trabecular and periosteal bone formation.

In our opinion, this qualitative similarity of the skeletal response to teriparatide is a strong reason to believe that the rat is an appropriate test model for evaluating effects of teriparatide on osteoblast behavior, including cell proliferation or neoplastic transformation.

Therefore, we believe that the quantitative difference in bone response between rats and humans related to the difference in treatment duration is no convincing reason to dismiss the tumor findings as irrelevant.

It is also our opinion that the tumor findings are likely to be relevant for any species that response to intermittent PTH receptor activation with an increase in bone apposition. To illustrate this, osteosarcomas have been observed in both rats and mice employing intermittent dosing with another PTH receptor like an analogue of PTHRP, which is a compound that acts on bone in a similar manner as teriparatide.

This indicates that the current tumor findings are neither specific to the animal's strain or species, nor specific to teriparatide. Rather, it seems to be related to intermittent PTH receptor occubation (phonetic) and the cellular events that are mediated by this particular type of receptor stimulation.

From the available data from the rat study, it cannot be concluded at what age the animals are susceptible to the proliferative effects of teriparatide. It's also unclear what duration of exposure to teriparatide is necessary to give an initiated cell a chance for neoplastic transformation.

For that reason, the sponsor is currently carrying out, as was mentioned this morning, a follow-up rat carcinogenicity study in which animals are treated from either a young age or an older age, a young age of two months or an older age of six months, and for different periods of time, either six months or 24 months.

In this study the animals are followed up until an age of 26 months before they're sacrificed.

The sponsor is also carrying out a monkey carcinogenicity study in which ovariectomized females are treated for 18 months and then followed up for another three years.

The results of these studies are not yet available.

In conclusion, the clinical relevance of the rat bone neoplasms induced by teriparatide is, in our opinion, unclear, and it would not be justified to dismiss the tumor findings as irrelevant until further information is available.

Therefore, we cannot exclude that there is a potential increase in the risk of bone neoplasms in humans treated with teriparatide.

I thank you for your attention, and Dr. Bruce Schneider will now address the clinical safety of teriparatide.

DR. SCHNEIDER: It's still morning. So good morning, everyone. I'm Dr. Schneider. I'm the endocrine and metabolic -- Division of Endocrine and Metabolic Drug Products. I'm an endocrinologist.

I'm going to spend the next 20 minutes giving you a very brief overview of the agency's view and interpretation of the efficacy results, and then I'm going to speak a little bit about my concerns relating to the risk of osteosarcoma.

I think we're all in agreement, and as I've indicated in my briefing document, that there is currently need for an anabolic agent for the treatment of many individuals with osteoporosis. I think it's clear that we have taken the strategy of using anti-resorptive therapy, including combinations of anti-resorptive therapy, about as far as we can go. They've been effective. They're helpful to many people, but there clearly is an unmet medical need for an anabolic agent.

Our task now is to consider whether in the case of teriparatide the benefit to risk profile merits approval. This decision made by the agency, which will be made by the agency, depends on our estimates of clinical efficacy, and these estimates must be derived solely from randomized placebo controlled clinical trials.

Other data are interesting, but we can't really accept them as efficacy data, and these must be balanced against safety concerns, and the principal one is the concern of osteosarcoma.

In a few minutes you'll hear a more complete safety review by Dr. Stadel.

Now, let me state at the outset that the results of the pivotal controlled clinical trials GHAC and GHAJ clearly established efficacy in the case of GHAC in post menopausal osteoporosis, osteoporotic women. The trial clearly established efficacy in reducing fracture risk and increasing bone mineral density in this population.

And trial GHAJ, the other pivotal trial, also clearly established efficacy in increasing spinal BMD in men with osteoporosis.

Although we don't have head-to-head comparisons, it's clear at least to me or it seems to me that for both men and women the beneficial effects at the lumbar spine, including BMD effects and fracture prevention, appear to exceed those of any currently approved agent.

Accordingly, these results would certainly be sufficient to meet efficacy criteria for approval of osteoporotic drugs based on our current criteria in the absence of any safety concerns.

These outcomes were the result of an extensive and thorough preclinical and clinical development program. The preclinical program, as you've heard, included mechanistic studies which clearly established anabolic action on bone and positive effects on bone quality.

The clinical Phase 1 and 2 studies demonstrated rapid anabolic action of teriparatide in humans with pharmacodynamic effects which were dose dependent in the 15 to 40 microgram range. The no effect dose was established at six micrograms. At doses greater than 40 micrograms, there was a rapid increase in adverse events in these early studies, and the positive effects were variable.

Safety tolerability profile was built up during these early studies, which led to, in my opinion, the proper dose selection for the pivotal clinical trials, GHAC and GHAJ and the other trials.

My only comment here is that it would have been interesting to have studied the effects of less frequent dosing, for example, 20 micrograms given every other day in terms of safety and patient acceptability.

And then finally, not shown on this slide I should bring up the fact that the assay usually, immunoradiometric assay that was employed, had a lower limit of detectability of 50 picograms per mL of PTH 1 to 34, which translates on a molar basis to about 123 picograms per mL of PTH 1 to 84, which is clearly above the upper limit of normal for PTH 1 to 84, and therefore, comments about absence of PTH 1 to 34 in the range that is below the hyperparathyroid range should be taken with caution.

The pivotal trial GHAC was described in detail by the sponsor. I'll just review it very briefly. The primary efficacy objective of this trial which studied the effect of teriparatide in the treatment of post menopausal women with osteoporosis, the primary efficacy objective was a reduction in the proportion of patients with new morphometric vertebral fractures. This trial had eight secondary efficacy endpoints and it enrolled about 540 patients in each of three treatment arms, as shown here.

The primary endpoint results are shown here. They were clearly achieved. There was a 65 or 69 percent reduction in the proportion of patients with new morphometric vertebral fractures which translates to about a nine to ten percent absolute risk reduction. These results were very robust, and the p value was less than .001 for each comparison of PTH versus placebo.

There were other fracture results which were not prespecified as outcome results, but which were methodology derived, and these are shown here, and I'm mentioning them because I thought they were quite impressive. There was an 80 to 90 percent reduction in the proportion of patients with multiple new vertebral fractures and a similar reduction in fracture severity using the Gennant grading system that the sponsor employed.

The key secondary endpoint was the proportion of patients with new nonvertebral atraumatic fractures combined. The study lacked the power to detect site specific differences at nonvertebral locations, such as the hip or wrist, which are very important for osteoporotic patients, and as shown here, there was about a 53 or 54 percent relative risk reduction in the incidence of all such fractures, with an absolute risk reduction of about three percent.

The p value was less than .02 for each comparison versus placebo without adjustment for multiple comparisons, so that these results were not quite as robust as the results at the lumber spine. Nonetheless, they were statistically significant.

This slide summarizes the percent of patients with fractures at each of seven different extravertebral sites, and there were very few such fractures throughout the study. All in all, the 20 microgram dose of PTH, for example, prevented two or three hip fractures in 540-some odd women treated for the duration of the trial, which is about a median of 19 months' exposure, and there were a few risk fractures that were also prevented by treatment.

None of these comparisons were statistically significant. The changes were in the anticipated direction.

The other secondary efficacy endpoint results are shown in this slide. In GHAC there was a significant increase relative to placebo in bone mineral density at the lumbar spine, hip, and total body. There was no effect in bone marrow density at the forearm. There was no effect on height loss in the entire population as a whole.

I might add that the effects in other trials of other agents have shown very small and inconsistent treatment related decreases in the populations when taken as a whole, treatment related differences of about one to two millimeters, which have been statistically significant.

Subgroup analyses of patients who do fracture consistently show greater height loss in that subgroup, but they can't be considered as efficacy outcomes because they're trial derived subgroups unless they're prespecified.

In any case, there was no effect on height loss in this population despite the substantial BMD and fracture prevention efficacy at the spine.

The histomorphometry results have been described. I won't go into them. They were basically positive. There was a positive effect, anticipated effect on biochemical markers which demonstrated an anabolic action of teriparatide.

And a final secondary outcome was health related quality of life indicators. The sponsor used five different instruments to measure health related quality of life changes.

I might add that every one of these indicators had back pain as a specific domain, and two were osteoporosis specific, and there was no effect seen. Back pain, which we've heard about about three or four times during the discussion this morning, was reported or recorded as a safety outcome, as an adverse event outcome with a p value attached to it, which I cannot accept as an efficacy outcome.

The other pivotal trial was GHAJ, the effects of teriparatide in the treatment of men with primary of idiopathic osteoporosis, and with osteoporosis associated with primary hypogonadism, the primary efficacy objective here was an increase in spine BMD, and the secondary endpoints were essentially the same as with GHAC.

This trial was smaller and enrolled about 145 patients in each of three treatment arms. The exposure was fairly small because of the early termination of the trial, of all the clinical trials. Actually there were very few dropouts relative to osteoporosis trials. There were about 88 percent, 82 percent, 74 percent of the patients in study at end. About 87 to 90 percent of patients received six months of treatment, and about 25 to 30 some odd percent of patients received 12 months of placebo controlled treatment.

The results, the primary endpoint was clearly achieved in this trial. They were highly significant increases compared to placebo for both doses, and the increase of 40 micrograms was greater than that achieved with 20 micrograms, and the key secondary BMD endpoints at eight other skeletal sites at 20 micrograms. There was statistical significance relative to placebo at the femoral neck only using endpoint last observation data.

For 40 micrograms, there were greater effects at almost every skeletal site, with statistical significance achieved at the total hip, the femoral neck, intertrocanter (phonetic), Ward's triangle, and whole body.

And it's for this reason that I raise the question in my briefing document as to whether the dose should be adjusted in treatment of men with osteoporosis.

The other secondary endpoint results were similar to GHAC.

So our clinical efficacy summary is shown in this slide, and let me go back here. In post menopausal osteoporotic women, teriparatide 20 micrograms is highly effective in increasing lumbar spine bone marrow density and BMD at other sites and in reducing the risk of morphometric vertebral fractures.

The drug is effective in preventing nonvertebral fractures combined, but the data are not as robust as in the spine. The 20 microgram dose -- and this is very important -- is as effective as the 40 microgram dose in reducing the risk of fractures, and this would establish in my mind that 20 micrograms is the appropriate dose, and the drug did not prevent height loss in this population.

In men with idiopathic osteoporosis with or without hypogonadism, primary hypogonadism, teriparatide 20 micrograms is highly effective in increasing lumbar spine BMD, but is either ineffective or only marginally effective in increasing BMD at other skeletal sites.

The 40 microgram dose was substantially more effective than the 20 microgram dose at nearly all skeletal sites. The drug did not prevent height loss.

I must emphasize again there were no fracture efficacy data from GHAJ or from any other randomized controlled clinical trials in men.

I should also add that, of course, this trial was truncated. It was stopped after a median of 11 months of exposure, and we really don't know what would happen with two years of exposure to the drug.

Now, these efficacy outcomes which clearly would meet our approval criteria, must be balanced, of course, against the risks, and the major risk that I see is the risk of osteosarcoma, and in the next few minutes I want to let you know why, although I certainly don't have any answers to this question, why I'm still concerned about it.

The major reasons for concern, of course, as we've heard this morning, is that this is a robust, dose dependent occurrence in rats, and we also know now in mice. There was no threshold dose demonstrated.

Now, unlike other preclinical outcomes that we often see, this was a biologically plausible outcome, and it involves hormonal stimulation of known target tissue.

In this slide I've listed seven reasons why we're told that we shouldn't be so concerned about it and why it is unlikely that osteosarcoma will occur in humans treated with teriparatide. I've listed every reason that I've heard and every reason that I can think of, and they appear here.

High exposure in rat studies. The treatment of rats began at six or seven weeks of age and was virtually lifelong. There's a negative monkey study. Rat bone differs from human. There's no increase in other malignancies in treated rats. Our experience with hyperparathyroidism in humans, and the observations in patients post treatment with PTH.

Let's look at each one of these. The argument has been made that rats received excessively high doses of teriparatide, and there was an excessive response in rat tissues. Let's follow this line of reasoning a bit and see where it goes.

The rats, according to my calculations, based on AUC, which is a unit of exposure, multiplied by a fraction of a lifetime, the rats received about 25 to 1,000 times the proposed human dose, again assuming that humans would be treated for two years, which is about two or three percent of a lifetime.

Now, if the background rate is 0.2 percent in rats, and that's a higher number; it may be a realistic number, but it's a higher number compared to the background rate in humans, which is about four or five per million per year. If the background rate is 0.2 percent in rats, then the study dose range led to about a 30 to 200-fold increase in tumors, and one can compose ratios of increased tumor occurrence divided by increased dose, and you get a number like a range of about 0.2 to 1.0 across the dose range, and this would yield a risk in humans of about 1.2 to, let's say, twofold.

If the risk is less than twofold, given the low background rate humans, we'll probably never see it. If it occurs, we won't know about it. I don't know how comforting that is, but it will be very difficult to measure.

And so these risk projections depend on the basal rates of tumor recurrence because if the background rate in rats, for example, was 0.2 percent, then you'd have a 300 to 2,000-fold increase in tumors, and you might have a four or five-fold increase in humans exposed, and of course, these are totally speculative extrapolations.

One make assumptions of linearity, and so forth, but this is about as far as I can take this argument, and so it doesn't really lay the issue to rest.

The next argument that's been made is that the treatment of rats began at a very early age, six to seven weeks, and the question is are young animals particularly or exclusively susceptible. That is, we have already heard further experiments are in progress now to determine whether the effect is age dependent in rats. The dose it's my understanding is going to be given in a staggered fashion to rats in a long-term study carried out by the sponsor, and I think this is really a critical experiment which will tell us a lot about the timing of tumor formation.

Of course, it's always likely -- not likely; it's always possible -- that the older rats will be more susceptible than the younger ones. You have to do the experiment to find out.

The negative monkey study is presented as an example, and again, this does not allay my concerns completely because I believe that the number of animals is far too small to detect even a large increase in tumor occurrence if the background rate is low, and I think what's been absent from a lot of the conversation and the discussion is consideration of the background rate.

For example, if the background rate in monkeys, let's say, is even ten times that in humans, and if the drug causes or the doses of the drug cause, let's say, a 100-fold increase in tumor formation, you'd still expect only four monkeys to get osteosarcoma in every 1,000 monkeys studied per year. So that studying 80 monkeys for 12 or 18 months might not be enough.

The next argument is that rat bone differs from humans, and certainly it does in terms of its architecture or growth and remodeling patterns. They all differ. The real question is not architectural as far as I'm concerned, but the following. Do the two species, rat and human, differ in the ability of the osteoblast precursor pools to replicate and expand clonally in response to intermittent hormonal stimulation?

This is the key question in terms of tumor promotion as far as I'm concerned, and we don't know the answer.

The next is that there's no increase in other malignancies in the treated rats. Clearly PTH is not a carcinogen. The concern here is not with that, but with the promotional effects of a hormone in a specific target tissue.

Next is our experience with hyperparathyroidism in humans, and frankly, as an endocrinologist, I can tell you this would be the most compelling reason for me not to worry. Hyperparathyroidism, particularly mild, primary hyperparathyroidism, as we all know, is not uncommon, and I'm sure there were tens, if not hundreds of thousands of people walking around with mild elevations of PTH.

In fact, our clinical practice guidelines afford us the opportunity of delaying or not doing parathyroidectomy at all and letting many, many people live out their lives with mild primary hyperparathyroidism, and osteosarcoma is, to my understanding, unknown in this group.

And I think this is really the best experiment of nature which tells us the most, but assuming that there aren't different cellular responses to intermittent versus sustained elevations in PTH, as there are with the overall bone pharmacodynamics, I don't know the answer to that question.

And finally, there are the observations in humans post treatment with PTH. We have about 1,450 patients treated for more than three months.

Again, given the low background rate, which is about four or five per million per year, this number and this period of observation, it would be unlikely that we would be able to detect an increase in tumor occurrence, given these background rates.

Also, what we're waiting for is the occurrence of a clinically obvious tumor, something which presents as pain or swelling, and that, I think, will take some time, perhaps 25 or 30 doubling times, let's say. So that I don't know that two or three years is enough time.

And my last slide here is, again, to weigh the benefits versus the risks, and they're the benefits of a new, very promising anabolic agent which really I think offers a lot of hope and is very exciting for treatment of osteoporosis. There are known benefits from the clinical trials, which show substantial bone mineral density increases in men and women and fracture efficacy in women, again, especially at the lumbar spine.

We don't know the long-term benefits of these architectural improvements from an anabolic agent. I suspect they'll be quite positive. We really don't know.

And these must be weighed against the unknown risk of osteosarcoma.

Thank you.

DR. STADEL: Good morning. I'd like to begin by expressing appreciation to Dr. Sunita Zalani and her colleagues at Lilly who have been very forthcoming in responding to rather detailed questions from me. I've tried to explore the database very thoroughly, and I can make a generally brief presentation on the clinical trial program, beginning by saying that in general, with a few exceptions, I agree with the presentation that has been made by the sponsor on the safety findings in the clinical trial program.

So I will briefly go over some highlight points about the trials, and then as others have done, I will talk about osteosarcoma.

This is something that came out of some discussions as this was going forward. Safety analyses differ somewhat from efficacy analyses, and I've put up here simply that the analyses of efficacy hypotheses are ordinarily specified in advance, and the use of p values is focused on testing the prespecified hypotheses. In analyses of safety, there usually are no prespecified hypotheses, but there is still a need to assess the data to identify potential areas of concern.

P values as a descriptive tool are useful for this, with the understanding that a p value associated with a new safety finding does not have the same meaning as a p value associated with either the testing of a prespecified efficacy hypothesis or a prespecified, a previously observed safety finding. new safety findings from one study should generally be tested in others before arising at conclusions.

This is important because I show p values on new associations, and I do not want the opportunity of them being misunderstood.

Now, in the preclinical studies, there were some key issues that arose that were on my list of things to understand as I did the safety review in the clinical trials, and these were post dose hypotension and tachycardia, decreases in RR and QTC intervals -- I just put QT -- and increase in serum and urine calcium.

I will say that in the clinical trial program, I found no clinical events in excess in the treated groups which would have been the types of events associated with these phenomena. These are dose related phenomena. The doses in the trial produced minimal tachycardia. I will show some information on that later.

I also looked for any other kind of cardiovascular even that might be an offshoot of a hypotensive episode, and I did not find excesses in the treated groups.

Now, with regard to electrocardiographs, no electrocardiographs were obtained in the Phase 2 or 3 clinical trials. So that I was not able to evaluate electrocardiographic findings under conditions of the kind of clinical setting in which the drug would be used. I did not see clinical events suggesting cardiac bad clinical outcomes, but I could not evaluate electrocardiographic information. I found this somewhat troubling.

In the preclinical studies, you've heard before about these issues. So I need not dwell on them.

Just a reminder of the size of the key studies. These are the two main studies of the enrollment criteria, the numbers of patients in the treatment arms.

Again, just a reminder of what size studies are we dealing with. The main studies I've shown, GHAC and AJ, AC the main study in women, AJ the study in men. Two other studies that were important supportive studies that had active controls I've listed. Just to give the denominators a sense, I will be showing numerators with percents and p values. This is your opportunity to know what the denominators are.

Now, this, I think is very important information. In terms of the issue of possible long-term effects of duration of use, and this is a sort of lead to the osteosarcoma discussion later, this is most of what we know in the two main trials about duration of use. That is, 85 percent of the women were in the 13 to 23 months exposure to study drug, and 87 percent of the men in the six to 14 month exposure.

This is a way of looking at it a little differently. In the total program, 1,452 patients were treated for at least three months. Now, that provides 95 percent confidence that you will detect an event if it occurs once in 484 or fewer patients. You may notice I have not put person-time here. One can make this function for any number of person-years. You could say that number of people studied for five years would give you that confidence of seeing it in 484 patients followed for five years.

The reason I have emphasized the number itself is that for rare outcomes, the question of individual susceptibility to an adverse effect is at least as important as the duration of follow-up. So that I wanted to put some emphasis on this is the n that we're dealing with.

I think for a clinical trial program I'm not criticizing the n. In terms of dealing with the potential for a comparatively rare, but extremely important adverse event, one needs to understand the limitations that are inherent to the follow-up of such a data set.

I look now at many things, but I'll just mention serious adverse events as defined by the Food and Drug Administration are listed here. In this analysis, co-genital anomalies and drug overdoses don't matter much. So the main things are on the top.

I looked at each of these separately. I will show you, as you've seen a little of this before, but here it is for the two main trials and the supported trials, that the aggregate rates of patients who had one or more serious adverse events by treatment arm were very good. There is no increase.

I looked at these by individual adverse event terms by study, and there is only one serious adverse event term which is statistically significant, and that was that actually in GHAC the rate of breast cancer was lower in the treated groups than it was in the placebo group.

I do not put great weight on that as a finding, but it was statistically significant.

No other analysis was statistically significant or even met the criteria of a trend, of a .1 screen. So quite a generous screen.

Looking at adverse events of any severity, you've seen some of this. So I'll just mention again briefly back pain was decreased at both doses. Nausea and headache, not increased at 20, but increased at 40. Leg cramps increased. Gout and arthralgia and urolithiasis, both potentially important, gout because of the uric acid elevation and urolithiasis because of the calcium elevations in the urine; both of these as clinical events were not present.

Dizziness, syncope and vertigo I analyzed very carefully because of the postural hypotension. There was nothing in dizziness or syncope -- excuse me -- in syncope or vertigo. There were a few cases of patients who had more severe dizziness in the treated groups, and I wanted to mention that. So there was a little bit of a difference, but not enough that I would generalize it as an important overall finding.

Now, in routine measurements, there were no differences between treatment groups in sitting blood pressure measurements. However, very little post dose data were obtained. In only one clinical trial involving a relatively small number of post menopausal women, there was post dose data, and I've showed you there.

Now, one hour after dosing with 40 micrograms was the maximal effect, and it was quite modest, a mean increase of five beats per minute, and an interesting thing. The range, it seemed to involve the bottom coming up rather than the top rising, which I thought was kind of unusual. I don't know if it would replicate in another data set, you know, but I do nonetheless feel somewhat uncomfortable that we don't know more about post dose heart rates and electrocardiographic findings under the general conditions of usage.

So since the electrocardiograms were not done in the studies, we have discussed that if the drug is approved, that there would be a Phase 4 commitment to obtain these data and sort of round out the data set in the absence of any clinical events to give greater concern. I'll leave it at that.

Now, a couple of things that have been discussed before, but I feel that I should show. One is the frequency of four-hour post dose hypercalcemia, and I've put this out by showing the number of patients with one episode and the number with two or more and then a group p value, and then the range of the hypercalcemias, 2.65 to 2.89 millimoles per liter.

So there are episodes. Most of the patients have one. Some have two or more. A difficulty one faces with what appear to be small numbers in a clinical trial like this is that three percent of a couple of 500 patients isn't that many, but when a drug goes into the marketplace and thousands are treated, the dimensions expand.

And I just want to bring that up now and then as a reminder because I lose track of it sometimes, and I think probably everyone does looking at these data.

Now, this, I think, is an important slide. This shows actions that were taken in close temporal proximity to the serum calcium measurements. It's not clear that they were taken, definitely caused by the elevations. The nature of the data don't allow one to be, I think, absolutely sure of that, but I think it's probably reasonably sure that these events were related to the detection of hypercalcemia in the patient.

And I think it's important because I think what it says is that the physicians involved in caring for these patients were watching this, and when they saw things go too high, they were making adjustments, and I think that bears on the question of whether there's ever any need to monitor.

You know, so these patients were monitored.

You see study drug adjustments. I pushed the wrong button somewhere. Study drug adjustments were also made significantly, but study discontinuation not.

I have managed to push a wrong button. Thank you, George. Thank you very much.

Okay. Now, this is the 24-hour urine calcium. You'll notice here that although the median has increased, there is not a meaningful increase in the frequency of episodes. Actually it was one percent higher for one episode in placebo, and then two percent higher for two episodes, two or more episodes in 20 microgram.

So it looks like that although there was an increase in load of calcium on the kidney, this was not manifesting itself as defined hypercalcurea (phonetic), and I think that's of some comfort, and you can see the range, again, at the bottom that I've put of where the hypercalcurea episodes fell from 7.6 to 20.2 millimoles per liter for 24 hours.

Now, I put this up. It's not significant, but I put it up because it's not significant. These patients do have an increase in alkaline phosphatase when they go on the drug, which is expected, but the fact that at the 20 microgram dose you have no increase in people above the upper limit of norma I think has some value with regard to if you are following the patient and they have a very high alkaline phosphatase. You don't write it off as due to the drug. You work it up.

And so I think that's a valuable finding actually with regard to the 20 microgram dose. It means that alk-phos can still be used in work-up.

Now, the post treatment follow-up study briefly. These are the number of patients, about 77 percent aggregate and quite uniform from the different trials actually enrolled in the study.

Now, this study, there was still blinded treatment at first, but then it became open label, and of course, with this number of enrollees, there's the potential for selection bias. So this gets into really an observational data set analysis and is much less reliable, I think, than the blinded randomized data.

I did want to show the number of serious adverse events simply to show that in this follow-up data, although it goes from 12 percent to 17 percent, then it goes down to 13 percent in the main trial in women, it does go up in men. It's not significant, but then in the two other trials in women it's actually fairly strongly in the other direction.

So I conclude that this is not meaningful, and I'm somewhat reassured by that. I don't see a lag phenomenon, you know, in follow-up of something emerging.

This I wanted to show. This is the survival curve. Where those bars are is when people finish the studies. So this is from the beginning of randomization to the end of the observational follow-up study to give you the death rates by treatment group. As you can see, they're very, very close.

Now, interestingly enough, they're even closer when you correct for a small problem. In the large study in women, purely by chance, the women in the two treatment arms were each on average one year older than the women in the placebo group, the randomization p value of .1, and in fact, when you correct for age, it brings the death rate slightly closer together.

And I was a little concerned when I first saw them because although there was no significance, there were more deaths in the treated arms. And so when I was able to get some balance out of that, I felt better about it.

I have two findings which I regard as tentative that we've been working on from the follow-up study. There is an entity in adverse event coding called cardiovascular disorder, which is a place where people put things they don't know where else to put, things that don't go under coronary heart disease, that don't go under congestive heart failure, that don't go under the specific entities; go under cardiovascular disorder, things in the cardiovascular system.

So this was quite a collection of things. It turned out that it was about 55 percent heart murmurs. The reason I show it, the reason I'm a little concerned about it is that the pattern was present in this subset during the trial, and when I looked at all patients randomized during the trial, the pattern is there. It's not statistically significant, but the pattern is there.

And then in follow-up it gets a little stronger, and when you take it into the aggregate, it gets a little stronger.

Incidentally, your handout has a slight numeric error on this one, just in case it's of concern to anyone. It says 39 percent where it should be 55 percent, and a couple other things.

So we've been still working that up. I bring back the caveat at the beginning about new safety findings and p values and so forth. In stratifying this by age and looking at the effective age, there's more of an association over 70 than under, and the possibility that tighter control of age may dissipate is still there. I haven't done that yet. I've looked at a lot of things about it.

The las thing I'd mention I do think needs to be mentioned, and again, it's another tentative finding. This was found at the first. This represents events found at the first visit in the follow-up study were there was an increase in the 20 microgram group that I've shown here, but there was also a similar increase in the 40 microgram group.

My slides are 20 microgram group because that's what's proposed for marketing, but for consistency scientifically, there was also a similar increase in the 40 microgram group, and there was a bit of an increase in this direction in the Mayo study.

So I've wanted to follow this up. I don't have any strong interpretation to place on it. The creatinine clearance distributions were not different between treatment groups, and follow-up has been done on 18 of these patients thurs far, 18 including the 40 microgram set, and that's a little reassuring. It looks like it may regress towards the mean.

So I will simply mention those are the things in progress. I don't see anything alarming in the data, and I will now turn to the topic of osteosarcoma.

I think from my standpoint as an epidemiologist, I think we have to know about when approaching this, one of the most important things to realize is in women and men 50 years of age or older, the approximate treatment population to this drug, that the annual incidence, the average annual incidence is four cases per million per year. That's from the SEER system data for recent years.

Of course, it's a little lower at the 50 year age and a little higher at the upper ages, and that means a total in the country of about 300 cases per year.

SEER covers about -- I just got that from using population figures, and the occurrence is generally similar by gender and race. So that's the dimension of what one's dealing with as a base rate.

And the question is: how do you detect an effect on something like this? It's not easy.

I should stop to mention the one really important risk factor involved. For anyone who's not familiar with it, Paget's disease is a resorptive disease of bone in which osteosarcoma -- in patients who have serious Paget's disease, clinically manifest and followed for long periods of time, osteosarcoma occurs with about a one to five percent frequency in the reported series. These are cumulated frequencies over varying durations of follow-up.

And most of the cases are in Paget's patients who were over 50. Most of the osteosarcomas that arise in Paget's patients. They have to have Paget's disease for a long time.

And so I wanted to mention that and to mention a little bit about Paget's disease in the U.S. population. Now, we were speaking previously about overt clinical Paget's disease. Now I'm speaking about subclinical, little foci that are found on X-rays. This was using the national health and nutrition examination survey data from the early 1970s.

There was a read done of the X-rays, and one comes out that the prevalence of Paget's disease in the over 50 age group is about one percent on average and increases with age, similar by gender and age. In other countries, it's a little higher in Britain, and a little lower in some other countries.

It may have gone down somewhat. There's some reason to believe that the prevalence of Paget's disease may be going down, but this is to give just some idea of a ballpark idea of what the underlying prevalence of a disorder that one is a little nervous about because would PTH potentially stimulate any of this.

I will mention that there was one case of Paget's disease diagnosed in the clinical trial population, and that was a man who was diagnosed a couple of months after he had finished a year of 40 microgram treatment, and the diagnosis seems to be quite confident.

The initial read at the time of diagnosis did not describe the presence of any pathologic uptake on the bone scan then. Subsequent reads apparently have been that maybe some disease was present. So I think it's -- I'd have to say it's a bit unclear to me.

I guess with regard to conclusions, I'd have to agree with both the investigator and Lilly. I think it's possibly drug related and possibly coincidental. I really wouldn't want to tie.

I would want to say one thing that's important here. From the previous slide I showed you with the one percent prevalence of occult Paget's disease in a clinical trial program involving a couple of thousand women, there must have -- patients, women and men -- there must have been a reasonable number of people, you would think, playing the odds, who had suboccult Paget's disease who were enrolled in the trial and who were treated with the drug. The only case we've seen is this one case.

So to some degree, I think it really cuts both ways. I think it provides a measure actually on the positive side, although I think most people would agree, and your proposed labeling would agree that if Paget's disease is known, you would try to avoid the drug.

Well, to get to the end of it, what can be done? Well, continuing to follow up the patients in the observational study is a good idea. I've tried to convey earlier what the limitations of numbers are, the realities.

One learns something, but it doesn't answer a lot of questions.

Mapping drug use data I think is extremely important to know if the drug goes into the marketplaces, to know where does it get used, where could it be studied, where are the potentials.

And of course, we have to deal with adverse event reports, and I'll talk a bit on the last slide about that again.

We talked about two kinds of surveillance, getting referral centers and doing case ascertainment, first off, to find out how many cases one can get hold of, and then there's the potential of using those cases for case control studies. I think you'd have to use controls from the residential areas of the cases or something along those kind of lines to get a reasonably unbiased assessment.

The sponsor has talked about the potential of getting quite a large percentage I think, up to about 40 percent of the cases diagnosed in the country, which if that were done, it would help.

And the other is what's called the SEER system. It's an excellent resource for doing cancer research. It's an NCI sponsored, National Cancer Institute sponsored program. The only limitation is for very rare tumors, it covers 14 percent of the country.

So I will close with this slide. This has a couple of interpretations. This is purely hypothetical. I want everyone here to understand that I am not talking about risks that are real. I'm talking about a scenario for the purpose of trying to convey an idea.

If the incidence is four per million up here, okay, that one is, I think, a fact. Let us suppose that the drug was marketed and we reach a state and there was a relative risk of three, large enough for most people to think it has some importance.

If you look at the numbers, then a tripling of risk would take four per million to 12 per million, and you subtract out the base rate, the attributable risk is eight per million per year.

Well, if early in marketing a quarter of a million people used the drug at that threefold risk level, that would give two attributable cases per year. No study would work that out. We would not be able to.

So I think one of the most important things to convey is that if any epidemiologic effort is made to assess, it's going to take years. The drug would have to be in the marketplace for quite a long time before it would be possible to get hold of an association. I think everyone who's looked at it agrees about that.

And so whatever your decisions are in weighing benefits and risks, one's talking about a substantial period of uncertainty, four, five years, something of that kind.

The last comment is that this kind of data can help us in one way, is that it gives us some idea of how many exposed cases to expect if there were no effect, knowing the four per million per year, knowing how much drug is used, and that provides a basis against which to judge spontaneous reports.

Thank you.

ACTING CHAIRPERSON MOLITCH: The FDA's presentation is now open for questions. I'll just start with the first question for you, Dr. Stadel.

If the risk of Paget's in this population is one in 100 and the risk of osteosarcoma in the Paget's population is probably one in 100, as you've said, or maybe even one in 1,000 if you wanted to go down to patients that don't have symptomatic disease that's known, then we're still talking about a one in 10,000 or even one in 100,000 risk of osteosarcoma in the general population, which is far less than what is actually clinically detected.

So how do we reconcile these two numbers?

And then the final thing is that if the sponsor who wishes to exclude everybody with Paget's, how many patients develop osteosarcoma who don't have preexisting Paget's disease? And are we talking about a --

DR. STADEL: The majority.


DR. STADEL: to the best of my knowledge, in older patient groups where the Paget's association is strongest, it still only accounts for less than half of the osteosarcomas, association in the reports I've read.

If anyone knows otherwise, please speak up, but I've really looked for that and I've only found a couple of reports.

I think I can address your question in two ways. One is that we don't know. This is Paget's disease. The people who did this know what they're doing, I believe, but we don't know if these very small foci of Paget's disease have the same meaning with regard to the osteosarcoma risk as the lesions that are large enough that represent the cases that were followed in the clinical series, and I can only assume that it doesn't because otherwise, as you're pointing out, the numbers would work out differently.

ACTING CHAIRPERSON MOLITCH: I don't know whether you know or anybody else can help us with this. In patients who develop osteosarcomas in the absence of Paget's disease, do they develop elevated alkaline phosphatase levels? They do?

DR. BONE: I can probably add a couple of points here. In a couple of studies where population based or at least reasonable efforts have been made to get a population based estimate of the risk of Paget's associated osteosarcoma, the risk for all patients who could be identified as having Paget's disease, in other words, for this kind of risk population, it's probably in the one to 1,000 to one in 10,000 case range rather than the one to 100, but this is confounded by the variable observation periods.

So it's probably something like one per 10,000 per year is my assessment from having reviewed this not too long ago. So I think if you have a ten-year observation period, you might see one out of 1,000 patients, and this is roughly what you see in treated Paget's disease with effective therapy. You get a big reduction in the risk. There are only two or three cases that I'm aware of in the world of effectively treated Paget's disease in which sarcoma emerged after that.

I think the two main time points at which osteogenesis sarcoma occurs is in kids and in older adults, and the inference is drawn that an important reason for the bump in the older adults is the Paget's disease, but I think Dr. Stadel is right. It certainly doesn't account for all of the cases. You can't get a very solid figure about exactly what proportion, but half is fair.

The elevation of the alkaline phosphatase is not uniform, but it's typical of both Pagetic and non-Pagetic osteosarcomas, but it's not something you can absolutely count on, but the majority of patients will do that.

DR. STADEL: One of the things I had hoped, and I talked to the fellow who did this, Roy Altman, who did this analysis of the NHANES data, as to whether anything was known about the alkaline phosphatase levels in these as to whether these small lesions were associated, but unfortunately it does not appear the information is available.

DR. BONE: Typically clinically though the smaller the amount of volume of bone involved, the lower the alkaline phosphatase levels. It's a function of both intensity of the Paget's disease and sort of activity at the site, and the extent of the involvement just as you would imagine.

DR. STADEL: Thank you.

ACTING CHAIRPERSON MOLITCH: Other questions for Dr. Stadel or the FDA? Yes.

DR. GRADY: Well, I'm really confused. So the first speaker suggested that the rat low dose was about threefold the human dose. Then Dr. Schneider suggested it was quite a lot lower than that. So do we have -- I mean, I really think this is important because if the rat low dose was the equivalent of about a three-fold higher human dose, you know, that -- do you know what I --

DR. BONE: Maybe I can ask a question here.

DR. GRADY: Yeah.

DR. BONE: Do I understand correctly that the first presentation, the animal safety data looked at the ratio of the administered doses in micrograms per kilogram? And Dr. Schneider's presentation further adjusted this according to the percentage of the live span of the exposure, not just years of exposure, but fraction of the life span, which would give about a, you know, 40-fold increase in the apparent dosage because it was estimating that the percent of life span for a human would be about two percent of the life span.

DR. GRADY: Right. I think that's what the difference is. But let me just understand this. So that in terms just straightforwardly of dose, the equivalent human dose, I mean, the dose that was given to the rats is about threefold the equivalent human dose. Is that your assessment?

DR. KUIJPERS: On the database, yes.

ACTING CHAIRPERSON MOLITCH: But that's based on AUC, not the actual --

DR. GRADY: Right, and I think it's somewhat of a leap to then divide that by the sort of percent of life span of use. There's no evidence that that's a reasonable thing to do, is there?

DR. SCHNEIDER: I don't know what's reasonable. The sponsor has claimed in this analysis that animals were given a lot of drug times a longer time. So all I did in this really hypothetical presentation was to multiply the amount of drug in terms of AUC times the amount of time in these sort of ARB units, that is, percent of life span.

Accordingly, what I got was a number like about at the lowest dose three times the AUC, and then I multiplied that by some number, let's say, like ten times the life span units, and that would go up to the highest dose where you have like a 1,000-fold thing where the AUC differences were about 60 and the life span differences may have been -- I don't know -- 25, 30, 40 times, something like that.

DR. GRADY: And one more question. Also in the original presentations, the estimated relative risk in the rats at the low dose was 30, around about 30. Where did you get three?

DR. STADEL: Made it up.

DR. GRADY: You made it up. Okay. Just for illustrative purposes.

DR. SCHNEIDER: The relative risk that I derived in those calculations were based on a background rate of 0.2 percent in the rat, which Dr. Kuijpers did a meta analysis of all the data, and so that gave me the risks, and then I could formulate a risk range of 1.2 to 1.0 based on that background rate.

But as I cautioned, if the background rate is lower, it can go up tenfold or more.

DR. STADEL: The Figure 3 was purely to illustrate the relationship of relatively and attributable risk in a low tumor setting. I picked three because I thought it was reasonable to work with. You could even pick a larger relative risk, and it still comes out as something you can't really well deal with.


DR. SAMPSON: Dr. Schneider, Dr. Stadel, there's this current, ongoing carcinogenicity study that the sponsor is doing that's got two different start dates and two different durations of treatment, as I understand it. Is there anything, is there any reasonable outcome that one could expect out of that that would increase either of your levels of comfort if you saw the results of that?

DR. SCHNEIDER: Perhaps you'll get two answers. Gemma first.

DR. KUIJPERS: I guess one possible outcome would be when one starts treating animals at a later age, starting at six months of age, it might be possible that long-term treatment of those animals would not lead to the development of osteosarcomas, which means that the treatment spent in the early age would be critically important, and it would reduce our level of concern because we're treating -- we're planning to treat humans at a later stage in life.

DR. SAMPSON: Do you know when that study is scheduled to be completed?

DR. KUIJPERS: I think the results will be available by the end of 2002.

ACTING CHAIRPERSON MOLITCH: Are there other questions from the panel?

DR. KREISBERG: I'd like to ask Dr. Schneider if you would go back over the statement that you made about the immunometric assay for the 134 molecule vis-a-vis native parathyroid hormone. Was the implication there that the level was sustained higher than would be expected, higher than what would be the normal range for a period of time that was longer than the apparent half-life?

DR. SCHNEIDER: All I'm suggesting is that in the terminal portions of that projected curve that the sponsor showed that there would be times in which the -- since the lower limit of detection was 50 picograms per mL, that there would be times in which an undetectable level would, in fact, be accompanied by a level of biologically active hormone that was twice the upper limit of normal on a molar basis, that is, that that would translate into about a 120 some odd picograms per mL of PTH 1 to 84.

And so that one can't really say specifically. One can project the trajectory of these curves, but one cannot do that with absolute precision during the terminal elimination phases.

Furthermore, the statement that the total elevation of PTH over 24 hours is less than what is normally seen I don't think could be substantiated on the basis of those data.

Now, what this means I don't know. I mean there's certainly -- if one just looks at calcium and so on, we've discussed that, but strictly speaking, my impression is that, that in the terminal elimination phases of the curve, an undetectable level can still exist with twice the upper limit of normal on the basis of bioactivity.

ACTING CHAIRPERSON MOLITCH: Other questions from the panel?

DR. GRADY: Can somebody clarify for me how good the data is that there's no increase in risk of osteosarcoma in primary hyperparathyroidism? I mean, you know, this has been mentioned a couple of times, but what kind of studies are these, and what are the denominators and so forth?

DR. STADEL: About all I can tell you is that I went through PubMed looking for everything that dealt with the issue and could not find any evidence of convergence. They were usually a series of patients with hyperparathyroidism.

There was one report of osteosarcoma in a patient with hyperparathyroidism, and the authors of that had done a lot of searches of the literature on hyperparathyroidism and had not been able to find any other cases, and that was about all I can -- I did not find anything like, for example -- I really didn't find any good studies of osteosarcoma in the literature. It's too rare.

DR. MITLAK: Dr. Grady, if I could, in my presentation I included some work that we had done using the national cancer registry in Sweden. We had searched the literature in the same way as Dr. Stadel and had found this one single case.

We then went in a systematic way through the records in that database covering 40 years and the entire population in Sweden. Dr. Unell (phonetic), who assisted us, both searched the hospital discharge database to look for patients who had been hospitalized with a diagnosis of hyperparathyroidism, and we also looked in the cancer database for patients who had been entered for reason of adenoma, which by law in Sweden needs to be entered.

We crossed both of those groups of patients, about 12,000 patients, 114,000 patient-years of exposure, with a set of terms that might include osteosarcoma, and as I had stated before, found in no case was there both diagnoses in the same patient.

DR. BONE: I had a couple of questions about the emergence of timing of some of these laboratory abnormalities. We had some episodes of hypercalcemia and hypercalcuria (phonetic), and the issue was, you know, didn't the adjustment of the patient's calcium intake bear on this question about need for monitoring.

Is there an identifiable time period in which these increases in serum reviewing calcium typically become apparent or can this be at any time during the exposure?

That might be a question for either Dr. Stadel or Dr. Schneider.

DR. SCHNEIDER: In the review of the clin.-pharm. data and actually the population based data, the hypercalcemia was the peak in the calcium, was about four to six hours after the dose. I guess anyone on the sponsor's side could --

DR. BONE: I meant in terms of weeks of exposure.

DR. SCHNEIDER: Oh. Oh, I'm sorry.

DR. BONE: For example, with patients who were treated with calcitriol, most of the patients who are going to develop hypercalcuria or hypercalcemia manifest this within about three months, which is when the peak calcitriol levels that we saw were also achieved.

My question was: for example, does this speak to monitoring patients at three months, just for an example?

DR. MITLAK: If I could, we have looked at this question, and again, while we have shown that the elevations in calcium are transient, there is no increase in calcium prior to the next dose.

We did look at the question that you have suggested from the Vitamin D literature. In our analysis of the data, if patients had a calcium measurement within the first three months that was not elevated, there was a very low likelihood that they would have an elevated calcium in any subsequent point during the study.

DR. BONE: Well, that's kind of qualitatively what I was getting at, but I'd be very interested in the actual numbers. I'm sure you actually have that, the time point at which the dose adjustments for the calcium are made and at which those elevations that result in intervention occurred.

And maybe after lunch you could give us those data.

The same question for the creatinine elevation. When did that become apparent?

DR. GRADY: This is a question for the sponsor, and I think, of course -- I'm sorry. I can wait.

ACTING CHAIRPERSON MOLITCH: We'll let that wait until this afternoon.

Any other questions for the FDA speakers?

(No response.)

ACTING CHAIRPERSON MOLITCH: Then I think we'll move to the final phase of this morning's session, which will be the open public hearing. We have three speakers who will present comments, Ronald White, Deborah Zeldou, and Dr. Peter Lurie.

And if they would come up to the front microphone and please speak your name, your sources from where you're coming, and any potential conflicts and financial conflicts that you may have with regard to your statement.

Dr. White.

DR. WHITE: Good afternoon. I'm Ronald White, Assistant Executive Director for Education, Research, and Community Affairs at the National Osteoporosis Foundation.

On behalf of our more than 350,000 members and donors, I want to thank you for the opportunity to testify before you today.

The National Osteoporosis Foundation is the nation's leading nonprofit voluntary health organization dedicated to reducing the widespread prevalence of osteoporosis through programs of research, education, and advocacy.

The NOF is proud of its broad base of funding support which comes from large and small individual contributions, memberships and memorials, foundations and corporations including Eli Lilly & Company, federated campaigns, special events, and federal and state agencies.

One of our most successful federally funded programs is the NIH osteoporosis and related bone diseases national resource center, which is located on our Washington, D.C. headquarters facility.

Osteoporosis is a widespread disease that affects the health of ten million Americans and is responsible for an estimated 1.5 million bone fractures each year. One third of American women over age 50 will eventually have the vertebral fracture, and fractures also occur in younger people, as well, due to secondary causes.

Approximately 12 to 24 percent of hip fracture patients will die in the year after fracture, usually from fracture related complications such as pneumonia or blood clots in the lung or from the surgery to repair the fracture.

Quality of life is greatly impaired in persons with severe osteoporosis not only because of pain and deformity, but also because of limited ability to move and be active, as well as the fear of future fractures.

In addition to the significant impacts on health, osteoporotic fractures result in medical, nursing home, and societal costs of approximately $14 billion each year.

The Foundation is very encouraged by the evidence from the research literature of fracture reduction in osteoporotic patients using Forteo. The availability of a treatment option for osteoporosis that builds bone mass and improves bone architecture would be an exciting addition to currently available anti-resorptive medications.

Thank you very much for your attention.


Let's hear, please, from Deborah Zeldou.

MS. ZELDOU: Good morning. My name is Deborah Zeldou, and I'm the Senior Director at the Alliance for Aging Research.

Thank you for the opportunity to come before this committee today to address the promising findings of PTH.

The Alliance for Aging Research works to stimulate academic, governmental and private sector research into the chronic diseases of human aging. We receive funding from a wide mix of foundations, private philanthropies, corporations and individuals.

For the last 12 months, contributions to the Alliance from Eli Lilly & Company have represented less than 3.5 percent of our total operating budget, income in the form of unrestricted educational grants.

As the Strategic Director of a not-for-profit group eager to find cures, preventions, and overall better health and vitality for the elderly, my views on osteoporosis reflect the medical needs of the growing population of older Americans. Our organization takes up the cause of the vast majority of Americans who fervently wish to benefit from scientific discoveries that improve the human experience with aging.

Survey research we conducted in June tell us that most Americans believe the federal government has a critical role to play to prepare the way for new medical breakthroughs and to hurry applications of science and health care in order to relieve human suffering and improve the quality of life for their family members and for themselves.

Osteoporosis is one of our most significant public health challenges. Experts predict that the number of hip fractures for both men and women will more than double in the next 50 years with the pending senior boom. Because this insidious disease can operate quietly and without recognition for decades, the silent thief steals more than bone mass. It takes an enormous toll on human life, often crippling its victims and causing them pain, grief, permanent disability, loss of independence, diminished quality of life, and sometimes death. It burdens our health system and care giving infrastructure.

Osteoporosis and the 1.5 million associated fractures it causes cost our nation 14 billion annually or 38 million a day in medical expenses alone. The graying of America is expected to quadruple annual medical costs more than 60 billion by the year 2030.

Better information and education about the disease and improving technologies are brightening the outlook for people with osteoporosis. Updated labeling by the FDA, for example, on foods and nutritional supplements, on calcium content in consumable products has helped guide consumers to purchase those items that help build and maintain strong bones.

Using diagnostic tools, physicians today can identify patients who already have osteoporosis, who are at risk for it before fractures occur.

New medications are also available to prevent or treat this disease, and advances in research are being made each day. Despite these advances, there is no cure, and new approaches to preventing, detecting, and treating osteoporosis are urgently needed.

Studies suggest that osteoporosis may be a quickly progressing disease once a fracture occurs, making prevention of future fractures critical for those patients who already have suffered from them.

Current treatments for osteoporosis only slow down or stop bone destruction. They do not have the ability to stimulate the formation of new bone. The suffering from osteoporosis need a treatment that can do more than slow or stop bone loss. PTH at this juncture shows promise for fulfilling this unmet need.

We are hopeful about the promise of PTH in improving the quality of life for millions of Americans as they age. We urge the FDA and its advisors to carefully consider the many benefits to patients and quickly move advanced therapies for the treatment of osteoporosis to the mainstream.

Thank you.


We'll now hear from Dr. Lurie.

DR. LURIE: Good afternoon. I wanted to spend my time just summarizing the comments that have been handed out and should be on your table, and in particular, those things that have been relatively underemphasized so far.

ACTING CHAIRPERSON MOLITCH: Please state your financial --

DR. LURIE: Oh, I'm sorry. I have no financial conflict of interest whatsoever. Our group takes no money from either government or industry.

The first point with regard to the efficacy study GHAC in women that has not been mentioned is that, in fact, many of the vertebral fractures, in particular, that were mentioned were, in fact, silent.

I quote from the Medical Officer review. "Because the majority of morphometric vertebral fractures are clinically silent, it is difficult to evaluate the overall direct clinical impact of these data taken alone."

Indeed, the Medical Officer continues, "The sponsor did not provide an analysis of clinical with symptomatic vertebral fractures in this application." I think that's something very important to consider.

Another thing we haven't hear much about in all of the laudatory comments about the efficacy of this drug is what the number needed to treat to prevent a nonvertebral fracture is, and we've done that little calculation. It turns out to be for the 20 microgram dose 28 people over the 19-month course of the disease. So it certainly is an effective drug, but I think we need to remember how many people will need to be treated and exposed to potential risks in order to benefit a single person.

And finally, Dr. Kreisberg did ask clearly about the question of quality of life, and the sponsor didn't make it very clear what the results of the quality of life studies in women are.

There was a quality of life study done, and there's no benefit whatever for the drug over placebo. This is true for both the studies in men, as well as the studies in women.

Turning now to the efficacy study GHAJ in men, obviously the most important point here is that the primary outcome measure was the lumbar spine BMD and not fracture.

Also, there's some lack of clarity. According to the medical officer review, the subjects in the end were only followed for approximately 300 days or ten months, not as long as sometimes advertised.

But most importantly, quoting again from the Medical Officer review, they called into question the importance of BMD data in men as opposed to those data in women, and a quote again from the Medical Officer. "The risk estimates for a given BMD T-score in men are not as well determined as in women. Whatever the cause of the uncertainty, the clinical impact changes in BMD will be more difficult to judge in men compared to women in the absence of fracture data. For that reason, we don't think that in the absence of fracture data this drug should be approved for men."

Moreover, the Medical Officer goes on to say, "Since we have no fracture efficacy data for either drug in men" -- this mean alendronate or the PTH drug -- "we have no fracture data for either drug in men, it is difficult to conclude that the 20 microgram per day offers any advantage over current therapy."

So having talked about efficacy, let me turn then to safety and make the following points that I think in my view make it rather clear that these rat data are absolutely relevant and make a compelling case for the carcinogenicity of PTH in rats and conceivably in humans as well.

Most of the landmarks of a positive and important rodent carcinogenicity study are present in this one. Firstly, the increases in tumors are substantial, and they are statistically significant. They are dose related. There is no no effect level identified. There could be sarcomas occurring in these rats at even lower doses than those tested.

The higher the exposure, the shorter the time to tumor initiation and death. The increases in tumors occur in both genders.

The exposure levels are, in fact, small multiples of human exposures. Dr. Grady asked about this. The area under the curve was measured at 24 months and was threefold the human exposure.

At 18 months, it was only 1.6-fold higher than the human exposure. So I think that's worrisome.

As it has been emphasized, osteosarcomas are very rare tumors in animals. So the appearance of this in these studies is very compelling.

Moreover, as has been noted, the tumors are mechanism based. Bone is where you would expect to see the tumors. Bone is where we see the tumors.

Moreover, because the formation of osteosarcomas is mechanism based, the fact that there are no positive mutagenicity of genotoxicity studies is basically irrelevant.

Let me also point out that the FDA has noted, and there was glancing mention, I think, of this in Dr. Kuijpers' presentation, that there are examples of other parathyroid hormone induced osteosarcomas in other related parathyroid hormone drugs, and so, again, it adds to the likelihood that this is no false positive, to be clear.

Let me just point one other thing out about the histology that was done of these animals. They did only look in four bones in a consistent fashion for tumor. So it's quite possible that there were other tumors that were hiding and simply not detected, and even more of the animals might, in fact, have had osteogenetic sarcoma than appears to be the case.

My presentation also includes in the written form mention of some of the renal, cardiovascular, and hypocalcemic concerns that have been raised by the committee. So I won't reiterate those.

To close then, in our view we do not believe that the data presented by the company provide an adequate rationale for approving this drug in men. There's no evidence that the drug reduces fractures. There's no evidence the drug is any benefit in quality of life.

The carcinogenicity studies in our view are very strong, and in this case, we think that this more than outweighs any theoretical benefit that might be gained for the drug in men.

Clearly, it's a more difficult case regarding the situation in women, but again, we should remember that the absolute fracture reductions themselves are not large, and many of the fractures presumably are asymptomatic, and there's no overall evidence of benefit on the patient's quality of life.

Moreover, there are already four drugs that are approved by the FDA for the treatment of osteoporosis, and so we believe much more narrowly that the risk-benefit assessment for women tips against approval as well.

However, should the committee choose to vote in favor of approval, there are at least four things that we think you need to do to minimize the risk to patients.

First, the drug should be restricted to use as a second line drug to minimize the extent of exposure to the overall population.

Second, there needs to be a black box warning, particularly on the osteogenic sarcoma findings.

Third, there is a need for the patients for a requirement for a med. guide for patients, and by this we don't mean handing out the doctor's patient package insert, which patients do not understand, nor do we mean the drug company funded documents that are handed out as patient information leaflets in pharmacies which are very often misleading. We mean an FDA mandated med. guide.

And finally, we agree with the idea of establishing registries and the like to identify those rare patients with osteogenic sarcoma who show up in order to do case control studies.

Thank you.

ACTING CHAIRPERSON MOLITCH: Thank you for your comments.

At this point we'll take a lunch break and we will resume at 1:45.

(Whereupon, at 12:37 p.m., the meeting was recessed for lunch, to reconvene at 1:45 p.m., the same day.)




(1:49 p.m.)

ACTING CHAIRPERSON MOLITCH: Before we start our general discussion this afternoon, Dr. Orloff is going to have some comments for us.

DR. ORLOFF: Thank you.

Good afternoon. The first thing I want to do is to thank the sponsor and representatives from that side and the FDA reviewers and their presentations, and the testimony in the open public hearing. Everything was clear, and I think we're ready to proceed with the discussion.

I have a few remarks to make before the discussion. This is nominally the charge to the committee. As I said yesterday, I'm not going to read the questions. I think they're fairly clear as written. If any clarifications or modifications are required as we go along, we'll be happy to add that as needed.

What I'd like to do is take a few minutes and summarize the FDA's concerns and conclusions after review of this application, most of which I think, as I said, were quite clear in the presentations that you heard before lunch.

With regard to efficacy, I think it's been clearly stated that we concur generally with the sponsor that efficacy has been demonstrated, and that the weight of evidence from the preclinical studies, from the clinical studies in both men and women, and women to show increases in BMD and reduction in the risk for morphometric fractures and in men to show increases in BMD, do support the efficacy of teriparatide.

The issue of the clinical import of the largely asymptomatic vertebral or the impact on largely asymptomatic vertebral fractures that was raised at the end of the last session, I think, is something that bears some comment.

As I think most people are aware, we do rely on increased BMD and a reduction in risk for morphometric fractures as valid surrogates, if you will, for an expectation of clinical benefit with regard to reduction in perhaps more clinically significant fractures.

And so as Dr. Schneider made clear in his presentation, the data that have been presented with regard to efficacy for this drug do or would generally support approval on the basis of efficacy.

What we have before us and what we're interested in hearing the committee comment on is the situation in which there is a significant safety concern with the drug, at least as far as we're concerned. I'll touch more on that in a second.

But in light of that significant concern, I think it is reasonable to at least be aware that an effort at a formal risk-benefit analysis may become more difficult in the absence of any evidence of hard clinical benefit. I hope that was clear.

As I think was understood from the FDA presentations, we do have lingering concerns, if you will, or even significant concerns over the findings of osteosarcoma in rats, and though we agree that rat bone differs from human bone, we also realize, and the other arguments for and against the sort of extrapolation from those studies to an expectation of human risk were discussed in the presentations.

We also realize that the size and duration of the exposures in the Forteo human studies was adequate only to exclude adverse events, and in this case the risk of osteosarcoma occurring at relatively high rates, and Dr. Stadel and others have touched on that problem.

So to us I think the conclusion is that the matter is unresolved. So for the committee, while we realize that like the sponsor and the FDA, you do not have a crystal ball to definitively refute or support a hypothesis of osteosarcoma risk, we are interested obviously in your thoughts and discussion on this issue on whether and what further investigations may be needed before or after approval and how this theoretical risk, albeit arguably biologically plausible, should be managed should the drug be approved for marketing.

I want to call the committee's attention and the audience's attention to Dr. Holmboe, who is present at the end of the table here across from me, who actually brings to the committee as a consultant an expertise in risk management, and I would encourage comments from him and questions to him from members of the committee.

With regard to the question we'll be asking specifically, we'll ask you in the event of an approval should there be restrictions on the use of this drug by risk category, that is, by fracture risk category; by response to other drugs, that is to say, for example, second line therapy in treatment failures on other established effective therapies or presumed effective therapies; and how the risk of osteosarcoma, should you feel it's significant, should be communicated; and, again, how it should be assessed over time across the populations exposed. You heard some discussion of plans in that regard. We would encourage further discussion or comments.

And I think with that I'll let the discussion proceed. So I'm going to turn it back over to Dr. Molitch.

Thank you very much.


And the floor is now open for discussion amongst members of the panel, who can address questions to each other, to the sponsor, to the FDA, and make comments in general.

Dr. Gelato.

DR. GELATO: I just wonder if we could get some comments from our consultant on the risk-benefit ratio and what his thoughts are in this regard. It might be helpful.


DR. HOLMBOE: I think when you consider the risk management, it's helpful to break that down into its component parts first. I think of three main elements.

The first is identification of the risk both from a population point of view, but also from a patient point of view. So starting at the population point of view, we know at this point that there appear to be three main categories.

The first is what I call pathologic, which has the greatest concern around the osteosarcoma risk, which at this point has been found only in an animal model, but at fairly high rates, as pointed out by the FDA. And so that certainly raises a lot of concern and will clearly raise a sense of dread and concern in patients any time you talk about a risk for cancer in taking a drug. So that's the first issue.

Second is in the metabolic things we heard about, and then finally the symptomatic, which are less certainly serious than the first that everybody is concerned about.

The second is assessment. You know, how are we going to assess these risks if this drug is approved? As we heard earlier, there's a problem with the signal. By that I mean that we're talking about a condition, osteosarcoma, that occurs at a fairly low rate, somewhat rare.

So, therefore, how are we going to monitor that down the road?

We also have to be concerned as we think about assessing risk, if approved, about what's going to happen as it's used in expanded populations. Most of these trials are really designed to look at efficacy, as we've heard.

The issue will then become is this effective from an epidemiologic point of view when we put it out in the general population, and that patients who would not have been enrolled in the original trials will be exposed to this drug with other co-morbidities, that may enhance their risk in unknown ways.

Finally, this drug is likely to be used in combination therapy, even if not approved for such. How are we going to monitor that risk? How are we going to assess that?

And then finally, as we heard, there are some issues in methodology regarding assessment, case control, population databases, things like the SEER database.

The one thing we haven't talked a lot about yet today is communication, and that communication has to go across several levels.

The two most important, I believe, are going to be communication to the physicians who would use this drug, and the second is going to be how that communication then occurs with the patient, and there are a number of challenges, I think, that confront.

When you consider informed decision making, there are a number of elements that need to go into that, and I think it's very important to place that context with regard to Forteo and how that might look between a patient and physician contact.

Clarence Braddock and Wendy Levinson have developed a very nice model, University of Chicago, with the elements that need to go into that. Three of those elements are, one, to discuss the risk and benefits of the therapy with the patient.

Another element is to discuss the uncertainty surrounding the therapy, and I think, again, that's one of the big issues for this drug.

And then finally, discuss the alternatives.

Part of the difficulty here is that we don't have a lot of head to head comparisons with this drug, and so that's going to be a real challenge for the physician.

The other thing is what should the physician tell the patient in how should that baseline assessment look like. I'd be curious to hear from the sponsor about what they think should be part of the baseline assessment for all patients: calcium, X-rays, et cetera, and how they feel that should be communicated to the patient.

From a personal point of view, I think that it is important to disclose the potential risk of osteosarcoma, again, if this drug should be approved, recognizing that it may be very rare.

I think that we do have some history to look back that may help us. It was mentioned earlier by the sponsor this morning regarding omeprazole and carcinoids. There's a tremendous amount of concerns about gastronomas that was not realized. However, the fact that it was not realized did not reduce the burden or need to inform patients of this risk.

And as a general internist using this drug almost 15 years ago, I can tell you that was part of the discussion and I think an important part of the discussion. So I think that's something else we need to consider.

So as you think about risk management, it's really those elements, identification, assessment and communication, that really need to be considered, and I think we do need to spend a little bit of time thinking this afternoon if this drug is approved that that patient-physician communication needs to be part of the dynamic because that's most likely where adverse reactions and problems are going to occur.

We have seen that with other drugs, for example, Cisapride. Despite multiple attempts by that sponsor to inform physicians of the risk of that drug, the drug continued to be used inappropriately, and so I think, again, those are other things that we have to think about as we look at the risk issues surrounding Forteo.


Other comments?

(No response.)

ACTING CHAIRPERSON MOLITCH: I'll start then if nobody has any yet at this point. I'd like to ask the sponsor about one of the concerns that you raised with the osteosarcoma was that this was unique to the rat model because of the differences in the remodeling or lack of remodeling, if you will, in the rat model.

So what has been done in other species that do have remodeling to start drug very early in the weanling stage and then continue it lifelong?

I presume that there are other long-term studies going on in different species that can shed light on this. Can the sponsor answer that, please?

DR. VAHLE: Certainly. Let me do that in two ways. First, let me discuss the differences in remodeling and some of the differences between primates and rats. Would that be useful as a part of the response?

If I could have slide 4233, please.

It is true that rats differ in skeletal biology from humans, including primate, and then I'll discuss what our follow-up studies in primates are.

With respect to the remodeling that you mentioned, two things to consider. One is rats lack the ability to break down cortical bone prior to forming new cortical bone. So they have really little or no cortical osteonal remodeling while that particular process is present in humans, as we mentioned during our presentation.

They also continue to grow throughout life, as opposed to humans or primates where growth ceases at adolescence.

Another, as I understood it, portion of your question was around bone turnover, and this really combined a physiologic difference with some differences in duration kinds of comparisons that may be useful in your deliberations.

If you evaluate rats for a given period of time, say, two years, they will have undergone approximately 25 to 30 bone turnover cycles in that particular time. This is in contrast to humans who during that time would have one to two bone turnover cycles or the Cynomolgus monkey, two to four bone turnover cycles.

So the second part of the question: what have we done to address that? Briefly mentioned in the response this morning, and I could just bring back up slide 4222, additional studies in primates are limited to the 18 month treatment duration followed by a three-year observation period.

So in respect to a species that has similar bone physiology remodeling types of phenomena, this study which we mentioned earlier is the extent of our evaluations.

ACTING CHAIRPERSON MOLITCH: And what is the background osteosarcoma rate in the monkey?

DR. VAHLE: Unfortunately the spontaneous background rate for osteosarcomas has not been defined. We are not able to find in the literature any background incidence rate. There are sporadic occurrences of osteosarcoma reported in the literature for monkeys. These are individual case reports, but not population databases.

Part of the difficulty with that is monkeys come from many different sources. The demographics, if you will, are very different. So we do not have a firm estimate.

If we were pushed to speculate, we would say it's somewhere between the four in a million that was quoted for humans in that particular population. Again, these are mature ovariectomized monkeys, and the rate in rats, which is higher, about .2 percent.

ACTING CHAIRPERSON MOLITCH: And you haven't studied other species?

I mean if you're trying to say that this is unique to the rat, I don't know that that's true yet. I'd like to see some other data in other species to show that it's unique to the rat.

It would be nice to look at another species that has a certain background rate and do enough of a population of long-term studies to show that it doesn't exist in those animals.

DR. VAHLE: The reason we chose the Cynomolgus monkey as the appropriate species, and this was in agreement and consultation with the agency, is because it has the most similar skeletal biology. Many of the other species do not have significant osteonal remodeling, and likewise, it is difficult to find other animal species where the known rate of osteosarcoma is precisely defined.

We're able to define it in the rat simply because we have large, two-year studies from which to determine a database.


DR. LEVITSKY: If this --

DR. GRADY: Just before we leave that, can you tell us the sample size in those two studies?

DR. VAHLE: The sample size in the follow-up monkey study, which is 18-month duration, is 30 monkeys per group.

DR. LEVITSKY: If this were to be approved and used as a second line drug, which one would assume would be its use because of the injection nature of the treatment, it would be important to have some idea of or at least an informed physiologic guess about what would happen to people who had been receiving long acting bisphosphonates for five years and then were given this drug.

Is there anyone in this room who feels that they could comment on what they think would happen since I gather there aren't any hard data?

ACTING CHAIRPERSON MOLITCH: I presume the sponsor has some data in animals showing the combined use.

DR. LINDSAY: I can comment from the point of view of clinical -- short-term clinical trial data which we published in the Journal of Clinical Endocrinology about two years ago, in which we looked at people who were already on alendronate ten milligrams a day.

And we looked for biochemical responses similar to the ones that I showed this morning and demonstrated an almost identical response in terms of osteocalcin increases and later increases in antilopeptide (phonetic) in the presence of alendronate as we had seen in the presence of HRT.

DR. LEVITSKY: Are there any data related to bone mineralization? They're all short term?

DR. LINDSAY: The human data are short term. There are animal data in rodents that are mixed. There are animal data in aged ewes that are also mixed. There are some positive studies and some neutral studies.

Part of the problem is that in the animal data relatively large doses of bisphosphonates were used, in excess of what you'd normally use in a human situation.

So the meaning of those studies in terms of human responses is far from clear.

Dr. Potts is reminding me that similar studies in rodents with HRT and in humans with HRT have shown essentially no difference in response, and there is a cyclical study in which parathyroid hormone was used with a calcitonin, and again, there was no essentially negative outcome.

DR. LEVITSKY: The problem though with the bisphosphonates is they're not like HRT. They're there and they're there and they're there, and that's what I'm wondering about.

DR. LINDSAY: Yes, and in a human we only have short-term biochemical data.

ACTING CHAIRPERSON MOLITCH: If we can continue just with this, I understood that perhaps some of the protective effect in the human against the osteosarcoma is, in fact, the remodeling that occurs against a constant stimulation.

If we do combine therapy with an anti-resorptive drug that's quite potent like alendronate or residrinate and then add the PTH, does that affect this protective effect at all for the development of osteosarcoma at least in theory?

I realize there are probably no data, but would that alter our assessment of the risk or should that alter it?

Maybe one of the bone biologists can help us with this.

DR. VAHLE: First, let me clarify a statement that may have been taken in error. We do not suggest the fact that humans or monkeys have cortical remodeling as being protective. We're simply highlighting that as one of the differences. So I can clarify on that.

Then I'd ask if there are any of the consultants who'd like to address the concept of the combination therapy any further than Dr. Lindsay already did.

So we are simply pointing out that it is one of the differences between the two species. We're not suggesting that it's causal or protective.

ACTING CHAIRPERSON MOLITCH: Dr. Bone, we've got about 20 bone biologists over there. Would any of you like to comment on this?

My question is: would the risk of osteosarcoma in relationship to PTH be affected in any way by the concomitant administration of a bisphosphonate in theory at least?

DR. BONE: Well, I think if we had a theory, a specific theory about how -- if parathyroid hormone does increase the risk of osteosarcoma, how it might do that, then we would be able to better answer the question. We know that like C-fas (phonetic) is induced and all kinds of things are.

There's a very complex cascade across two signaling pathways downstream of PTH, and we don't know if there is an effect, and if so where in all of that it could be.

Bisphosphonate therapy appears to dramatically reduce the risk in Paget's disease, but of course, the presumed mechanism is completely different. I think the only thing we can say is that there's nothing whatsoever to suggest that this phosphonate therapy would increase the risk in any independent way or probably modify the risk very much.

ACTING CHAIRPERSON MOLITCH: It sounds like if anything, it might have a protective effect and probably not an additive effect.

DR. BONE: I wouldn't want to go that far to say that there would be a protective effect, but I don't think there's any reason to think it would -- that bisphosphonate therapy would increase the risk here.

To the extent that osteoblast activity might be indirectly stimulated by osteoclast (phonetic activity, which does appear to be the case in spontaneous remodeling without parathyroid hormone stimulation, modulation of that bone resorption and decreased release of growth factors from the matrix might conceivably have a moderating effect here.

But I think the main point is I think it's hard to imagine a mechanism by which the bisphosphonate would add to the risk.

DR. LEVITSKY: Henry, do you think that, say, five years of bisphosphonate treatment would alter the response to PTH in terms of its ability to enhance bone remodeling and increase bone mineral and reduce fracture?

DR. BONE: Well, I don't even know who's going to win to the World Series. So --


DR. BONE: -- I think you could reasonably expect that patients who had prior or continuing bisphosphonate therapy would be responsive to parathyroid hormone. Whether their response would be similar to or a little bit less or a little bit greater than that that we see with parathyroid hormone alone, I think that's an empirical question and we could make up stories either way.

I think it would be unlikely that the patients would fail altogether to respond. Some people think that you might see a better net effect in cortical bone with a combination, but that's, again, a speculation.

The idea behind that would be that controlling bone resorption at the same time that you enhance bone formation might give you a positive focal remodeling balance and perform wonders, but I think that probably most people here in the bone field would expect patients who have had extended treatment with, for example, alendronate, which is where there's the greatest relevance because of its availability for the longest period of time, would probably respond, you know, but to predict whether there would be a modulation of the response would be, I think, guessing.

ACTING CHAIRPERSON MOLITCH: Dr.Neer, did you have a comment?

DR. NEER: I just wanted to make a point of information that the committee might want to be aware of with respect to Dr. Levitsky's question, and that is that the National Institutes of Health is currently funding several studies, including one at our institution to try to answer exactly that question because nobody knows what the answer is.


DR. GELATO: Dr. Bone, I'll ask you a question, too. Is there any information that you could think of if the tissues and things were available from the animals who developed the osteosarcoma, anything that could be gotten retrospectively that would help in understanding mechanism or shed light on the issue?

I mean, I know sometimes retrospective studies, you know, are almost impossible, but if tissues could be looked at, I mean, is there something?

DR. BONE: Well, I'm certainly not an expert on the molecular pathogenesis of osteogenic sarcoma. I would be very interested in whether the consulting committee that advised the sponsor was asked to address that question, and if so, what their specific recommendations were.

I asked a couple of rather naive endocrinologist type questions about, well, were they receptor positive and that kind of thing. I wouldn't regard those as very sophisticated questions, and the sponsor apparently felt that they were not worth pursuing. I don't know exactly how they were advised. One could imagine.

DR. CHABNER: I'm Bruce Chabner. I'm an oncologist, and I chaired the committee that considered the question. I think this is, from an oncologist point of view, it's a very interesting animal model of osteosarcoma, and we did suggest that the company support studies that would look at the biology because I think there's something to learn about the disease, if not about the risk.

And they will do that. They're planning to do that in terms of looking at gene arrays and the molecular defects in these tumors.

We don't know a lot about osteosarcoma in people. So it's, I think, a stretch to think that we can solve this problem very quickly by studying these animal tumors.

You know, one of the interesting questions is how does this tumor relate to what we see in people. So parallel studies would have to be done in human tumors as well.

We do know something about the molecular basis of osteosarcoma in people. It occurs in people that have a defect in the RB pathway, in retinal blastoma deficient patients and retinal blastoma gene deficient patients.

It also occurs in certain families associated with P53 abnormalities. But those are very isolated cases, and the other risk factors that we know about are exposure to radiation therapy, thoratrast, osteomyelitis, a history of osteomyelitis, all of them not very well understood in terms of how that leads to osteosarcoma.

I think the company is going to undertake studies to look at that. The plan isn't entirely clear, and one of the reasons is that we just have so little information about what causes human osteosarcoma.

DR. POTTS: I'm John Potts.

I did want to add something particularly to Dr. Bone's comment, following up on what Dr. Chabner said. We do know a fair bit about the state of receptor in osteosarcoma cells, as some of you may know. One of the classic cells that's used is called an ROS cell. It's a rat osteosarcoma cell line, and the important point for the committee to appreciate is that these are receptor positive, and they respond to PTH. The receptor doesn't have anything to do with the transformed nature of the cell. In fact, it's used as a model of a normal osteoblast.

So the pathorward (phonetic) hormone is not really playing at the time you look at the cell anything particularly about it. In fact, if anything, it has an anti-proliferative effect.

So it's because something else has happened in the genetic make-up of the cell at the beginning which has caused it to develop its oncogenic potential, and then the pathorward hormone receptor is there, and it responds the same way a normal osteoblast cell line does.

It doesn't help very much, but I think Dr. Chabner has really touched on the reasons why it's hard for anybody to say exactly how these studies will go forward, but they are planning to do them.

There's something about the genetic make-up of these inbred rat strains that clearly makes them susceptible to tumors of various types, which is why they're used, and the PTH, when you take the cell out, responds as it does in a normal cell.

DR. BONE: John, thank you for your comment. Are you speaking specifically of the tumor cells that were isolated from these tumors?

DR. POTTS: No. What I'm saying is very analogous cells of the same type have been derived, and as they brought out, I believe, for you this morning, that the company has not done studies of that type specifically with these.

We're all struggling with this, and so in terms of making a prediction, this is a pretty reliable one, what you might expect, but there is no such data.

DR. BONE: Well, I thought that might be one of the early steps in attempting to characterize these cells.

DR. POTTS: And perhaps the company can respond to that.

DR. BONE: Simply looking for uniformity. For example, if these cells are -- the common features from these tumors from one animal to another would be, for example, one thing to look at if they're very heterogeneous or homogeneous in some of these kind of biological characteristics, that would be a starting point.

DR. VAHLE: Just to clarify, I think there was a request that sponsor clarify. We've not done that with any tumor cells from the original study. It is one of many things that have been considered not only in consultation with the consultants we have here. It has included consultations with Kevin Raymond, who is a molecular pathologist with expertise in osteosarcoma.

DR. GRADY: Just to get oriented here, could somebody review for me what is the exact indication we're considering? And is the use of this drug proposed to be restricted to any risk group, to duration of treatment?

I think you say two years, or to prior use of other drugs, and are you proposing any kind of work-up or follow-up?

DR. MITLAK: The indication that we have requested is for the treatment of osteoporosis in post menopausal women an din men. As I included in my presentation this morning, the indication would also reflect that the duration of treatment should be for up to two years and that patients who are otherwise at increased risk for osteosarcoma should not receive treatment.

The type of evaluation that we think would be appropriate is consistent with the standard evaluation of a patient who is being considered for treatment or prevention of osteoporosis, and there are standard practice guidelines that are in place for this.

We think that these would be appropriate to exclude secondary causes of osteoporosis, such as hyperparathyroidism, and also to exclude Paget's disease.

DR. GRADY: So you have no proposal that it would be restricted to any -- for example, these studies were conducted in women with prior fractures.

DR. MITLAK: We think that women and men at increased risk for fracture would be candidates for this, and those would include, for example, women who have had fractures or women with low bone density who are at high risk for fracture.


DR. PELOSI: I have three questions that basically hopefully tie together when we really look at if this drug is approved, things that we as clinicians need to look at.

And if the sponsors could tell me in terms of compliance, since we're looking at daily injections and oral supplements for two years, what was your compliance rate in terms of this actually occurring, and did you see any dose intensity? In other words, how much did they truly have to take in that period of time so that we knew that the results you get really can be seen in the patient population?

DR. MITLAK: In the clinical trials, compliance was assessed by measuring return study medication. Compliance was very good in the clinical trials. I believe that roughly 80 percent of the doses that had been distributed to patients were taken.

DR. PELOSI: The reason that I ask that, I'm in oncology, but in oncology many times we see if we don't get a certain percentage of the dose, we obviously see a difference in the outcome. And so is there any plans for long-term follow-up in those who may be under your 80 percent to see if there was a difference in those?

DR. MITLAK: No, we don't have plans now.

DR. PELOSI: The second question that I have is in terms of your claim to reduction of pain. Could you just give us a brief overview in terms of how that was assessed and at what points, and if the pain -- a decrease was seen after people went off medication?

And I ask that because I guess my thought is, again, with certain medications that we have seen a reduction in pain. Patients are very reluctant to go off of those medicines, and if we're having a risk or a concern that there may be a risk, we need to plan for that.

DR. MITLAK: The information that were reported on back pain included results from patients reports, spontaneous reports at visits of new and worsening back pain. There were instructions in the protocol to the physicians to alert them for how they should consider reports of back pain with respect to this likely being or potentially being part of the syndrome of vertebral fractures.

So these questions -- the reports of back pain essentially were elicited by the sites when they discussed how the patient had been doing since their last visit at the clinic.

The data we showed you showed a lower proportion of patients reporting back pain, and we saw that pattern continue beyond the time the treatment had stopped.

DR. PELOSI: And the very last thing, in terms of quality of life data that you said really you didn't see an effect, was there or is there any way to look at those patients who actually went off study? Because I didn't see the quality of life data on those patients who self-selected to go off study actually was gathered because that may be valuable information, again, to say why is it that they truly went off.

And if we look at it post treatment, as well, a year later, has that quality of life changed and how did they view that experience while they were on?

DR. MITLAK: We do not have data for you in follow-up to the patients who had discontinued from the study in a general way. We have offered patients the opportunity even after discontinuing from the Phase 3 studies to come back from the follow-up study so that we do track them, but I do not have a precise answer for you.

DR. PELOSI: Okay. My only comment was I was a little disappointed not to see more minorities represented in the studies.

Thank you.

DR. GRADY: Could I ask you one more question about quality of life? I guess I found it odd that you didn't find any improvement. Those are fairly commonly used measures, and with continuous outcomes usually.

You did suggest there's an improvement in back pain. Did you look at the various elements of the quality of life? Was there improvement, for example, in pain and a decrement in some other of the factors?

DR. MITLAK: We saw little significant change in the quality of life instruments, but I think there are several things that need to be considered with that.

One is that we needed to use several different types of instruments because this study was performed in different countries, and we needed to use instruments that were validated in the patient's native language. This may have affected the power of particular instruments to detect a signal.

Two, the studies were stopped early, and I think, frankly, the difference from placebo or actually the patients who had not received active treatment had perhaps not been followed long enough to see as much of a signal as might have been present toward a longer period of observation.

And finally, we are looking forward for instruments that may be a little more specific for specifically the back pain that we detected as an adverse event signal to follow this up prospectively with patients.

DR. GRADY: Isn't it true that in your own studies of Raloxiphene that within, you know, up to two years of treatment with less of a reduction in risk of vertebral -- and these were also morphometric vertebral fractures -- there was an improvement in quality of life, I think, using these very same instruments?

DR. MITLAK: What we showed in Raloxiphene, and I think what we also show here, is that regardless of treatment, patients who suffer fractures have an impairment in quality of life. I think our data support that also, but what we did not show was a specific treatment effect.

ACTING CHAIRPERSON MOLITCH: Dr. Aoki, did you have a comment?

DR. AOKI: I have two questions primarily for the sponsor, but for anybody who can answer this question. It seems that we're not going to be able to resolve at least at this meeting and probably not in the near future that the mechanism for the osteosarcoma issue. So it seems to me that the post market surveillance is going to be key, and that's basically, I think, how we're going to get the adequate power for this and any analysis, and so I'd like to address this primarily to the sponsor because I'm sure they have thought of the same problem.

How are you going to design a post market surveillance program that is designed to pick up cases of osteosarcoma to see if, one, this is a problem or, two, it is not a problem?

The second question I had was: if the therapy is only going to be offered for two years, 24 months, does this mean that the patient then goes off the drug, never to go on it the rest of his or her lifetime, or is there a rest period and then they restart the medication?

DR. MITLAK: With respect to the design of the follow-up study, I highlighted in my presentation some of the elements that we think are important and appreciate the tremendous assistance and collaboration we've had in discussing this with our reviewing officers at the agency.

The elements of the program, obviously, first are to be able to identify cases regardless of what sort of treatment the patients may have had before, and I think we have identified two approaches for this.

One is to use stable population based databases, and the second is to proactively go to sites where patients are cared for. It turns out that this, because it is a rare disorder and because there are specialized treatments, that most patients in the country are cared for at a fairly small number of sites.

We have already begun a discussion with one of the molecular pathologists at the M.D. Anderson and have begun discussions on how we might actually be able to link between sites so that we would know in a way with a sense of immediacy when cases are brought to the attention of the site, whether it is because the patient has come to the site or because the site is reviewing pathology slides in the consultation.

And in that way we begin to establish an ongoing case series, a database. We would then have to use epidemiologic techniques, such as those suggested by Dr. Stadel, to create case control studies to follow up on any signals that might occur.

And, again, just from the standpoint of where we are on this, we do not expect to see a patient develop an osteosarcoma as a result of Forteo treatment, but we are going to do this diligently to confirm that this is the case.

With respect to the overall duration of treatment, I think that for now two years is two years, until we have further information on the drug.


DR. HOLMBOE: I have a couple of questions regarding your communication program, if this drug would be approved. The first would be since it is a time limited drug, how are you going to educate physicians in that regard, particularly given the patients often change physicians? I think you hear earlier that patients may be reluctant to come off of it if they're getting actually some benefit, and there may be some confusion about when they started it.

So have you thought about how you would manage that, to make sure that they truly only get the drug for two years?

The second thing is how are you going to educate physicians. I gather that you plan for this drug to be used or not be restricted to certain groups' physicians such as endocrinologists, but be open for primary care practitioners. So it really raises an important question of educating the primary care practitioners and those who use this drug with regard to some of the risk communication issues with patients.

So I just wondered if you could address what sort of plans you have for those issues.

DR. MITLAK: In considering your questions, we look to the physician as really the person who is going to have to work with their patients to communicate information about this. It is a theoretic risk, and there are many things that need to be considered.

We have already highlighted that from the outset we have tried to be transparent with respect to the findings. We have included information about the animal findings at the scientific presentations that have taken place. We have included a discussion of the findings in the manuscript that has recently been published on the results.

We would propose to be sure that our sales force and the individuals in the company who interact with the physicians are well prepared to be able to communicate this information and would expect that the physicians will have to help communicate this to their patients.

DR. HOLMBOE: Have you designed any educational materials to help physicians in this regard?

DR. MITLAK: We have not as yet.


DR. KREISBERG: I've been trying to think how I would use this drug as a physician, and it's my understanding that anything that changes the balance between bone formation and bone resorption in a positive way is likely to be effective, and that in some of the studies with anti-resorptive agents, the relative risk reduction has been of the same order of magnitude even though the change in the bone density has been strikingly different among different drugs.

So the question that I have is do you see this as a drug to be used right from the very beginning in the management of a patient with osteoporosis, or do you view it as a drug to be used when other therapies for osteoporosis fail?

And if it is to be sued in the beginning, how do you decide which patient to use a drug that increases bone formation over a drug that interferes with resorption?

DR. MITLAK: What I'd like to do is ask some of our consultants to provide their comments for you. If I could ask Dr. Lindsay if he'd be willing to come up.

DR. LINDSAY: I wrestled with the same questions over the last several years that we've been interested in parathyroid hormone, and I draw a number of conclusions.

The first is that patients who present to me with fracture, especially if the fracture is relatively recent, are at a dramatically increased risk of future fracture and deserve something that will reduce that risk fairly rapidly.

An agent like teriparatide can increase bone density far more rapidly and far more greatly than other agents and, therefore, might be considered to be the treatment of choice for those individuals.

There are also individuals who present whose bone density is sufficiently reduced that the change in bone density that would occur with an anti-resorptive agent would not bring them back into the normal range, sometimes even for the age, and certainly not into the normal range for young adults.

Again, here this agent would have the clear advantage and be more likely to be able to achieve that.

The more difficult issue, I think, that you raise is what do you do with people who are already on treatments because we've already been into the discussion about what the response is, and I think that the theoretical conclusion is that these people will response, based on our biochemistry and very little other data in humans, but that the response may be greater or lesser.

And I would see that there certainly is a cohort of patients who fracture on current therapies, who may then be amenable to this sort of agent as in that case a second line therapy rather than a first line therapy in the first two cases.

We all realize when we give anti-resorptive agents that we're reducing risk, but of course, when a patient fractures, the patient considers that to be a treatment failure, and I think that that would drive that particular prescription.


DR. JENKINS: I'd like to ask the question of the sponsor, and you may have answered this this morning. I had to step out for part of the presentation. It's a follow-up of Dr. Grady's question and Dr. Aoki's question that goes to the proposed indication.

Can you articulate for me what's the rationale behind your decision to recommend limiting duration of therapy to two years? And could you address that from an efficacy and a safety perspective?

DR. MITLAK: I think that the most straightforward answer is this is the data. We believe that this is the duration of treatment that the data that we have accumulated support. We have a high degree of confidence in the effect of treatment over this period of time and, therefore, are comfortable going forward.

We think that it is an important piece when considering the overall risk-benefit for this drug, which we feel is an important potential new treatment to be sure that as its use is begun that, again, we do this within the context of the data that we have in hand.

DR. JENKINS: Is there any particular efficacy reason that you would go for two years versus one year versus 18 months versus three years? I'm just asking.

And also it sounds like you're suggesting limiting duration based on some safety concern. Because we often for drugs like this, we have two or three-year data for drugs for treatment of osteoporosis, and those drugs don't have duration limitations in their labeling.

DR. MITLAK: What we have is the data that established that 18 to 24 months of treatment is a very effective regimen for reducing the risk of spine and non-spine fractures and are very comfortable with that.

We do not see any specific safety concerns. We think that given the uncertainty that this panel is dealing with with respect to the animal findings, that it is important from balancing risk-benefit to have a set duration of treatment, and we think that the studies support two years.


DR. ORLOFF: Can I just follow up with one more question related to efficacy?

Could you make a comment on whether there's bene consideration and whether you believe there would be any rationale for perhaps even limiting the duration not as part of the overall directions for use, but let's say limiting duration based upon BMD response. So that you can imagine individuals who might have a robust response in a fairly short time frame such that let's just say for the sake of the discussion that they reach an incremental BMD that is in line with the mean seen in the clinical trials that demonstrated efficacy and safety.

DR. MITLAK: Let me answer this in part and perhaps ask one of our consultants also to comment. I think that we are in the position with this drug where there is not a very close correlation between change in bone density and reduction in fracture risk. So I think to gauge change in bone density as an adequate surrogate for duration of treatment is not supported by the data that we have.

I think what we do have is the study results which showed that 18 to 24 months is an effective regimen for reducing the risk of fractures.

PARTICIPANT: I'd like to make a comment. I've struggled with this thought also about how long to administer therapy, and my initial impression before Eli Lilly discovered this osteosarcoma finding was that this therapy should be administered until bone mineral density reached a normal level or until bone mineral density stopped increasing, whichever occurred first.

I think that it's important to recognize, again, that there's never been an osteosarcoma occurring in a patient treated with this agent, and so one approach might be to just adopt the position I just articulated.

A more cautious and conservative approach would be to limit the therapy to some duration until more information was available from studies of a larger number of humans, admitting that the risk is unclear in humans. It would obviously be desirable to have more information before one used it without limit.

Two years is a compromise position, and I think that it can be defended on a couple of grounds. One, as you heard today, the beneficial effects on bone mineral density are time dependent, and bone mineral density increases most rapidly in the first year, somewhat more slowly in the second year, and then as Dr. Lindsay pointed out, there's still some increased bone density in the third year, but during that third year indices of bone formation and resorption in his studies have returned to or toward normal.

In fact, they start returning to or toward normal after 18 months in some studies. So while it seems to me unreasonable to give it without limit, it also seems to me unreasonable to stop therapy after only 12 months because what we're trying to do is help patients, and it's clear to me that they're helped more by 24 months of therapy than by 12.

I don't see any absolute way to answer the question because there's no empirical basis on which to answer.

DR. BONE: I have a couple of questions that came up in the morning's discussion in which the sponsor was asked to come up with some data, and since they've done all of this work now, I think we're anxious to see it.

Three specific questions had to do with the time course of developing hypercalcemia and hypercalcuria; time course of seeing the increase in the serum creatinine level; and the spectrum of 25 hydroxy Vitamin D levels at baseline and how they predicted the response to treatment.

DR. MITLAK: I'm going to answer your third question first. At baseline the mean 25 hydroxy Vitamin D level was 79 across the board. It was even in all three treatment groups. The reference range is 25 to 153.


So that is in nanomoles per liter, and the reference range is, again, 25 to 153.

DR. BONE: And the mean was how much?

DR. MITLAK: Was 79.

DR. BONE: And what was the distribution?

DR. MITLAK: The standard deviation was 24. So if you assume a normal distribution and go down to minus two standard deviations, that takes us down to 34. So you have about two and a half percent of the patients between 25 and 34, at the low end of the spectrum.


Any other questions?

DR. BONE: Oh, excuse me. I meant to ask one more.

And what relationship was there, if any, between -- or did you look at the relationship between the baseline 25 hydroxy Vitamin D level and either fracture risk or BMD response?

DR. MITLAK: We've not specifically done that analysis, but it is certainly one of interest to us and that we hope to get to perhaps starting next week.

Serum calcium -- sorry.

I'm sorry. I stand corrected. We don't have a statistical analysis, but we did do the pharmacokinetic analysis, and there was no relationship between baseline 25 hydroxy D and either fractures or bone mineral density response.

Let's go on to the serum calcium question. The question was what was the time to onset of the transient increases in serum calcium.

If we could start with slide 4415, please.

I'll show you two slides in this respect. The first is the time course, the by visit analysis of the four to six-hour post dose serum calcium in the pivotal study in post menopausal women, and this, again, shows the median and 25th to 75th percentile range for the serum calcium, again, measured at its peak four to six hours after each dose at each visit during the study.

And as the graph shows, there was a significant increase as early as one month, and after three months, the medians were very similar throughout the remainder of the study.

So this data would suggest that all of the transient calcemic effects should be apparent by approximately three months.

If we could see slide 452, please.

And this next is actually a time to first even curve of the time to the first post dose increase in serum calcium. While it's getting up, let me just remind you that these changes are transient, and even in the patients who have increased post dose serum calcium, it's back down to baseline by 16 to 24 hours after the dose.

DR. BONE: Yeah, but as Dr. Grady pointed out, you adjusted therapy in seven percent of the patients. So that's where we were particularly interested in at what time point those therapeutic adjustments were going to be.

DR. MITLAK: Okay. I'll show you this, and then I will provide that data.

Could you zoom into this part of the graph, please, just the box?

This is the time to first event in the placebo, 20 microgram, and 40 microgram groups, and this is the time the first patient had a four to six-hour post dose serum calcium which exceeded the upper limit of normal.

And as you can see, there was a very small number of patients throughout the study in the placebo group who occasionally exceeded the upper limit of normal, and that's what's expected based on our lab reference ranges.

You can also see that especially in the 24 microgram group, but even also in the 40 microgram group, the patients who exceeded the upper limit of normal even transiently were by and large identified within the first three to six months of the study.

Now, there were some dose adjustments allowed, in fact, required by the study, and let me back up just a little bit.

You can turn that slide off now. Thank you.

Let me just back up a little bit and describe the reasons why we monitored serum calcium and did dose adjustments in the study.

We did that so that we could describe the effects on the serum calcium in this patient population, and we put in the requirements for dose adjustments for two reasons.

One is because we were not certain how big the effects would be and wanted to make sure that there was protection for the patient.

And, two, we much preferred from an intention to treat analysis and provide as much data as possible on the patients to keep a patient in the study on a lower dose rather than forcing them to discontinue due to a laboratory abnormality if, in fact, that could be handled by a dose adjustment.

In the 20 microgram dose, there were -- there are a small number of dose adjustments and a very few in the first six months. In fact, only 2.4 percent of the patients in the first six months of the study had a reduction or discontinuation of study drug, and so basically what you see, the data through six months is data for 97.6 percent of the patients.

DR. BONE: Are you speaking only of the PTH or are you also speaking of calcium?

DR. MITLAK: That was the injectable study drug reductions.

We haven't done oral calcium supplement analysis the same way that we've just done the injectable study drug analysis, but in general, oral calcium supplements were adjusted prior to injectable study drug. Even though that was not the case, the physicians were free to adjust either downwards as they felt fit.

I'd also remind you that, you know, again, even the number of patients having adjustments in oral calcium supplementation was fairly small. It was less than ten percent.

DR. GRADY: Could I ask you a quick question? Was this fancy 28-day injectable injection device used in the trial, the same one that you're going to market?

DR. MITLAK: We did use I wouldn't call it a fancy injection device. It actually does represent our expertise in delivering injectable drugs to diabetes patients in a convenient way, and it was used in the trials, and the patients accepted it very, very nicely. There were very few patients who withdrew from the study due to problems taking the injection.

And so, yes, we would hope to bring those same benefits to the patient with a marketed product if it's approved.

ACTING CHAIRPERSON MOLITCH: I have a question about one of the covariants that you talked about this morning and that you said there was no effect of renal insufficiency. I'd like to know how many patients had renal insufficiency and what degree of renal insufficiency it was, and would you really, in fact, want to treat patients who had renal insufficiency with PTH considering the fact that they already have some secondary hyperparathyroidism?

So it may be just a question of degree.

DR. MITLAK: Okay. Again, I'll start from the bottom and work my way up. First, regarding hyperparathyroidism, the patients in this study were not permitted to be in this study if they had a calcium or a parathyroid hormone level above the upper limit of normal. So patients did not have secondary hyperparathyroidism in the study.

With regard to renal insufficiency, this being an older population, based on the measured creatinine clearance, we actually had quite a few patients with mild renal insufficiency, creatinine clearances between 50 and 80. And, in fact, about 40 percent of our patient population had a creatinine clearance below 80, most of those being between 50 and 80.

We had approximately 25 to 30 in the moderate category, between 30 and 50 milliliters per minute. So our study population does represent patients with certainly mild and to a lesser extent moderate renal insufficiency.

We also looked at patients with renal insufficiency compared with patients with normal renal function and did not find that there was any significant difference in effects on renal function or on serum calcium or on efficacy. So we were very comfortable that within this age population that range of renal function is well represented.

DR. BONE: Speaking of which, you were going to give us the figures on the emergence of the rise in creatinine.

DR. MITLAK: Okay. If I could have slide 4422.

I'm going to try to show you a lot of data from both the treatment studies and the follow-up study because the difference in the serum creatinines, which was described, only occurred at visit one of the follow-up study, which is about six months after the end of the treatment study.

This is the serum creatinine during the pivotal treatment study, GHAC, again by visit. These are the means and the standard deviations, with the upper limits and lower limits of normal by the horizontal lines.

As you can see, there was no difference among the treatment groups in the mean serum creatinine during the study or at endpoint.

In addition, there was no difference in the number of patients with an elevated serum creatinine or with renal insufficiency based on creatinine clearance during the study or at endpoint.

Could we have 4417, please?

This is just the same data with the measured serum creatinine clearance in the same study population, showing, again, no difference among treatment groups.

Could I have 4430?

Now, let me move on and describe the findings in the first visit of the follow-up study. First of all, there was no significant change in the measured creatinine clearance, and there was no significant difference in the median serum creatinine concentration at endpoint.

There was a difference in the median change from baseline to endpoint, and that difference was about one micromole per liter or 0.01 milligrams per deciliter, which was statistically significant.

There was also a significant or a trend at least towards a difference in the number of patients with a serum creatinine above the upper limit of normal six months after stopping study drug, and that was two percent in the placebo group, four percent in the 20 microgram group, and four percent in the 40 microgram group.

We also looked at patients with individual increases, and our predefined lab limits of a significant increase are 0.4 milligrams per deciliter. So we looked at that, and there was one patient in placebo and one in the 40 microgram group with an increase of at least 0.4 milligrams per deciliter.

There was no one with an extremely high serum creatinine. The highest observed serum creatinine at this visit of the study was 1.5 milligrams per deciliter.

I think the important point is that we also looked across the studies, and we did not see similar trends, and let me just show you the data across the studies, and that is slide 4502, please.

And here you can see the change in serum creatinine from baseline to endpoint in the treatment study and post menopausal women, in men, in the study which compared HRT alone to teriparatide 40 micrograms a day plus HRT and the study which compared alendronate ten milligrams a day to 40 micrograms a day of teriparatide.

And as you can see, the changes from baseline, you know, all are very small, and studies to study, they go in different directions and have different inferences.

So we think that overall, taken as a whole, the data shows that there isn't any adverse effect on renal function.

Thank you.

Let me just also add a comment on what Dr. Stadel had mentioned. The patients in the follow-up study are in the midst of another study visit, and we do have follow-up on approximately a third of the patients that had serum creatinines above the upper limit of normal, and half of those are now back within the normal range.

And so, again, this finding in visit one may just represent some normal variability from visit to visit.

We certainly did not see evidence of progressive decline in renal function in any of these patients.

DR. BONE: Yes.

DR. MITLAK: Any other questions pending from this morning that you'd like me to answer?

Thank you.

DR. HOLMBOE: I guess this raises the question that we've been talking about: who should receive the drug? But from a risk communication standpoint, who should not receive the drug in your opinion?

Most of these people, again, had creatinines that were so relatively normal, which is the usual way of primary incurrence of measure. They wouldn't do a creatinine clearance. They may, you know, calculate and estimate one using the equation, but I guess I'd like to hear who should not get this drug and how, again, will you help primary care practitioners identify these individuals?

DR. MITLAK: We think that individuals that have other secondary causes for osteoporosis should probably not receive treatment, and this would include patients with abnormal renal function, disorders of Vitamin D metabolism, and other identified causes.

DR. TAMBORLANE: One of the issues that came up was a suggestion with the juvenile rats and stuff. I certainly think once this was approved, if it were approved, that there would be interest in using this in children with osteoporosis.

What are your proposals for labeling instructions about use in children?

DR. MITLAK: As we had highlighted before, we intend to include a statement that says that individuals at increased risk for osteosarcoma should not be treated, and these will include patients with Paget's disease, adolescents, or those with open growth plates, for example, or patients who had received radiation therapy.

DR. TAMBORLANE: The agency, I know, has a concern about the orthostatic hypotension that you saw in the early studies. Was this a first dose effect or was it persistent with multiple doses?

DR. MITLAK: When it was observed, it was most commonly with a first or first few doses. As Dr. Gaich had highlighted before, in several patients who did have symptoms, when they were given a subsequent dose and sometimes a greater dose, the symptoms did not recur.

DR. TAMBORLANE: Is this something that you might think about in the labeling, especially in our older patients, that the first dose they be monitored for several hours?

DR. MITLAK: We have included instructions to that effect. We have included an alert to this and instructions that if symptoms occur, that the patient should be allowed to sit or lie down until their symptoms resolve.

ACTING CHAIRPERSON MOLITCH: I have a question, again, about the proposed limit of duration of treatment, and I was wondering why you chose not to use a differential duration for men and women, given that the males, I think the median time was nine or ten months, and at 20 micrograms, you have, as Dr. Schneider noted in his write-up, less impressive efficacy.

DR. MITLAK: As I included in my presentation this morning and as we have found quite clearly, gender was not an important baseline factor in either response to treatment, that is, actual change in bone density, nor in the safety profile as assessed by a comparison of the adverse event profile in men or women.

Therefore, we think that the database reflects or would support the use of this for two years in menopausal women or in men.

DR. SCHNEIDER: If I might make a comment on the gender comparison that you made, those BMD curves, basically the number of men in that study, you were comparing 11 or 12 months' treatment in men to whatever, 12 months of treatment in women, and the number of men who had been exposed to 12 months of treatment was what, 25 percent of the men? And it's an extremely small number, and I felt that the comparison really was unreliable.

Furthermore, the critical issue to me -- and this came out in my review -- is not so much comparing across genders and two different trials and so on and so forth, but really what happened in the placebo controlled trial in men.

I mean, clearly efficacy was reached at the lumbar spine. I won't quibble if it was 5.2 percent or 5.3. The really issue whether you want to achieve efficacy within 11 months or a year or whatever at other anatomic sites, and although there were numeric changes in the right direction, it didn't make it anywhere else.

DR. MITLAK: Let me make one comment, and then I'd ask Dr. Bellizikan to comment also.

With respect to the figures that I showed in my presentation, the data comparing spine was, I believe, an observed case analysis. So all of the data for the spine was included. For the hip where there's a single point at 12 months, what that represents is essentially the 12-month visit, visit six in the protocol.

So for patients who had had a measurement before that time point even if it was an early discontinuation visit, it was carried to that visit and included in that analysis.

Let me now ask Dr. Bellizikan to make a comment, please.

DR. BELLIZIKAN: My name is John Bellizikan. I'm from Columbia.

And I'd just like to comment on study that we concluded and was published in the JC&M in September. This work was done in collaboration with Bob Lindsay and his group.

This was a study of men with idiopathic osteoporosis, a small group, placebo controlled, blinded with a dosage of PTH, not this particular form of PTH, but analogous with a similar dosage. This study was carried out for 18 months.

With regard to the lumbar spine bone density, it was exactly the same in terms of the slope of increase as was shown for this study, but with the 18 month data, we saw a clear divergence after 12 months such that the PTH treated men showed a clear departure from the placebo, and by 18 months, there was an approximately three and a half percent difference in both density, which was significant from placebo.

So carrying out the study as we did to 18 months, we were able to show significantly different total hip density and femoral neck density as compared to placebo.

DR. SCHNEIDER: That's encouraging.

I have a question actually which may be helpful. In dealing with an earlier question about prior use of alendronate, as I recall in GHAC, obviously concomitant use of bisphosphonates was not allowed, but there was a subset of patients there who had been on bisphosphonates, and then of course, they had to be interrupted.

Have you done a separate analysis? I mean, perhaps some of the answers are in your own database.

You showed that. Okay. All right.


DR. BONE: Yeah, it seems to me clear that there are two diverging approaches that we can take to obtaining some of the incremental information that everybody is sort of asking about in various ways, and these have to do with cancer risk and the long-term effects of the drug and concomitant use and a lot of other things.

And these are basically observational approaches, trying to do the best job we can with essentially passively acquiring data that's being generated by the use of the drug, and the other is conducting systematic trials, which tend to be more circumscribed in number, but have much better defined denominators and ascertainment.

And in our recent experience with drug in the diabetes area, for example, some of these issues were really highlighted about how well you can make these calculations.

I just have a couple of thoughts about this. One is that when we're talking about the risk of osteogenic sarcoma, the question has been posed in a sense that could there be an increase of some amount in the risk of osteogenic sarcoma, and it's going to be very difficult, as we've all heard, to tell the answer to that unless the increase is very large over the background rate, particularly if we subtract the Paget's patients from the population.

Another way to look at that is to say what level of risk can we live with with this horrible disease. I mean it's a really bad thing to have an osteogenic sarcoma. So we could make some calculation about, you know, what level of risk can we live with. Can we live with one in 1,000? Probably not. Can we live with one in 10,000? Maybe. Could we live with one in 100,000? We'll never know the difference between that and the background rate even if it's two and a half times the background rate.

So one of the things people could think about is what level of risk can be accepted. Now, generally speaking, people don't like to take any risk of having something really bad happen, and when a new drug is on the market, you have the problem always of having had a sample size which is, you know, in some way achievable, and we always have the problem that an event that's going to occur at a rate of one in 5,000 or one in 10,000 individuals probably won't be detected except by sort of a fluke.

One of the things we may want to think about is in addition to registry type reporting, which might catch an increase in the background, could several questions be answered by doing a larger scale clinical trial over an extended period?

In other words, a couple thousand patients per arm for three to five years, that's the kind of range where you would not eliminate the risk of osteogenic sarcoma, but you could say it's very likely to be below one in several thousand, and I would certainly want the advice of Dr. Stadel and Dr. Grady on this point and others because this is not my area of expertise, but my sort of back-of-the-envelope calculation is that we could probably improve our confidence by about an order of magnitude if you had a study with three arms in it of about that size in duration. I might be wrong.

Another thing that could be obtained from that kind of study is you certainly wouldn't do a placebo controlled trial in patients of this risk level over that period of time, but you might consider an active control trial against the best available therapy as an alternative, and an interesting opportunity would then arise of having a combination arm, which should answer for good and ever the question about whether there's a combination effect.

We only really answered this question by doing that kind of study when we were talking about bisphosphonates or at least alendronate and estrogen. In that study a bone density endpoint was used rather than a fracture endpoint, which may be more appropriate here.

But that seems to me to be complementary. There may be resource issues and a lot of other things, and I wouldn't want to necessarily be considered the author of the Osteoporosis Investigators Full Employment Act of 2001, but that might be complementary information to what would be obtained in the trial that -- in the sort of passive observations that's been proposed for looking strictly at the osteogenic sarcoma. It leaves a lot of the other questions unanswered that people have been coming back to, and it's quite apparent that absent some large scale experience and extended time period experience, we're simply going -- we're going to be asking ourselves the same questions in a year or two or three or five.

One of the advantages that we've had in estrogen therapy and in use of particularly alendronate is that there were very long-term studies with estrogen, including particularly Dr. Lindsay's landmark study from Glasgow, and we've had a very long running extension of the pivotal trials for alendronate which have now been just about concluded after ten years.

So that there were always a cohort of patients who were being observed systematically who had been treated for a longer period of time than anyone on clinical therapy.

Just a couple of thoughts of the committee to kind of chew on.

DR. GRADY: Well, you know, it's fun to ask questions, and I think we've learned some things, and maybe even it was helpful. I think we really need to kind of, in the interest of catching my plane, start cutting about some major issues here because I think there are actually quite a few of them.

And the key one I think that Dr. Bone has brought up is, you know, we're talking here about a preventive therapy. So we're talking about treating women and men who are at risk for disease, but don't have a symptomatic disease, and so we'd really like for that treatment to be safe and, if possible, completely safe.

So I think we're all worried about the incidence of sarcoma. I think if you look at the data the company has provided us and you say a simple thing, that is, there were zero sarcomas out of 2,000 people followed for an average of about 18 months, one thing you can say is that the rate of sarcoma is, you know, with about 95 percent confidence unlikely to be higher than 1.5 in 1,000.

Now, that's still probably too high for this terrible disease, and perhaps larger trials would answer that.

I think perhaps the other way to go at it would be to say, all right, let's take maybe a kind of worst case scenario, which in my mind is that perhaps the underlying rate of osteosarcoma in patients who might get treated with this drug is maybe one in 100,000.

And even if the relative risk is something like 30, we're now talking about 30 in 100,000 or three in 10,000, which would be the excess risk of this disease, although quite devastating.

And I think if you compare that to the number needed to treat to prevent any clinical fracture, which is around about 30, and even the sort of estimated number needed to treat to prevent one hip fracture, which is around about one in maybe 200, it's a low risk.

The problem, again, in my mind is that it's a devastating illness, number one.

Number two, I'm still a little worried about some of the metabolic findings, although they didn't seem to translate into clinical problems, you know, the hypercalcemia, hyperuricemia, and increased creatinine clearance, serum creatinine.

And then finally, there are options. So I think what we really need to spend some time talking about, the labeling for this drug and whether or not it ought to in some way be restricted to women and perhaps to men at much higher risk than the average person who gets treated for osteoporosis.

ACTING CHAIRPERSON MOLITCH: Any other comments? Because otherwise I think we ought to start to go down our questions that have been addressed to us.

(No response.)

ACTING CHAIRPERSON MOLITCH: Hearing none, I think we will start with the first question, which is a question based on efficacy, and the question, there will be an A and B part to this, and I think we'll go around the table. Each person will need to answer yes or no to these questions as we go around.

So question one on efficacy is: based on the information presented by the sponsor in the NDA, are the data adequate to establish that teriparatide, 20 micrograms per day, is an effective dose?

And then (a) for the treatment of post menopausal osteoporosis to prevent fracture risk, and (b) to increase bone mineral density in men with osteoporosis.

And so I think what we'll do is go around the table to answer both A and B at this go-round, and then we'll go around to the next question after that. Perhaps we could start with Dr. Holmboe.

DR. SAMPSON: Can I ask for just one clarification, please?


DR. SAMPSON: On BMD, is that bone marrow density in lumbar spine or to be construed in general.

ACTING CHAIRPERSON MOLITCH: My guess is lumbar spine.

DR. SAMPSON: Thank you.

DR. SCHNEIDER: The lumbar spine was the primary endpoint. We had meant generally BMD in general, that is, given the aggregate BMD responses to 20 micrograms.


DR. HOLMBOE: I'm not sure I'm a voting member.



DR. SAMPSON: I wasn't quite paying full attention. Would you repeat that one more time, please?

DR. SCHNEIDER: The primary endpoint was BMD at the lumbar spine. What I meant in the question was given the aggregate BMD increases across the body to 20 micrograms.

DR. HOLMBOE: As the questions are written, I would say yes to both.

DR. PELOSI: I would answer yes to both

DR. AOKI: Same.

DR. LEVITSKY: Same. Yes to both.

DR. TAMBORLANE: Yes to both.

DR. GELATO: Yes to both.

DR. KREISBERG: Yes to A, no to B.

DR. GRADY: Yes to both.

DR. SAMPSON: Yes to A, no to B.

ACTING CHAIRPERSON MOLITCH: And I will say yes to both as well.

We'll then go on to Question 2. Actually the -- we're supposed to have some -- yeah, can you give us a tally?

MS. REEDY: Question 1, fracture risk in treatment of post menopausal osteoporosis, yes ten, no zero.

In fracture -- increasing bone mineral density in males, yes eight, no two.

ACTING CHAIRPERSON MOLITCH: And I think perhaps, Dr. Kreisberg, maybe you can also give us a reason why you voted no.

DR. KREISBERG: Yes, I'll be glad to do that.

I believe that the number of men treated is small, that the results are confounded by the fact that a percentage of them had androgen deficiency that was not corrected. It's a heterogeneous group.


DR. SAMPSON: I just refer to the company's data, and they certainly show significance in lumber spine, but in a number of the other secondary measures the results don't reach statistical significance.

ACTING CHAIRPERSON MOLITCH: Okay. We'll then move on to Question 2 with regard to safety, and the question posed is: based on the information presented by the sponsor in the NDA, are the data adequate to define the safety profile of teriparatide (a) for the treatment of post menopausal osteoporosis and (b) for the use to increase bone mineral density in men with osteoporosis?

And we'll start with the opposite side, and we'll start with Dr. Sampson.

DR. SAMPSON: I don't think that's quite so fair to switch and ask a statistician to do the lead on that.


DR. SAMPSON: I would say no and no.


DR. GRADY: Could I just as for clarification here? So if what we're interested in is making sure that there's some sort of strict registry follow-up, assuming that I would feel comfortable given that, then am I supposed to vote yes?

ACTING CHAIRPERSON MOLITCH: Dr. Orloff, do you want to comment? Dr. Orloff?

DR. ORLOFF: The question is intended to elicit your response with regard to whether you think the database as it stands adequately defines the full safety profile of the drug. In other words, are you left confident of what the risks of the drug are or do you feel that more information is needed?

DR. GRADY: In that case --

DR. ORLOFF: Let me just say that this is separate from the question of approvability.

DR. GRADY: In that case I vote no for both.


DR. KREISBERG: No for both.

DR. GELATO: No for both.

DR. TAMBORLANE: No for both.

DR. LEVITSKY: No for both.

DR. AOKI: Same.

DR. PELOSI: No for both.

DR. HOLMBOE: No for both.

ACTING CHAIRPERSON MOLITCH: And I agree no for both, and I guess all of our concern is with respect to the still unknown risk of osteosarcoma.

DR. LEVITSKY: Actually the further concern as to how this is going to work in patients who have been treated with other drugs. So the osteosarcoma is the major risk. I still am not entirely sure that.


Did other people who voted no have other concern other than the two that were mentioned?

DR. HOLMBOE: I just want to mention the issue of the combination therapy. I think we know very little with regard to safety.


DR. GELATO: I also think that patients need to be monitored. I'm not sure I feel comfortable with putting them on this and calcium supplements and just say, "You don't need to check calciums," and some of the metabolic issues, although I don't think they're major, but I think they at least need to be considered.

ACTING CHAIRPERSON MOLITCH: I think that this may well come up as we pursue along with our discussions about what might be appropriate to say what patients might be safe to take the drug to begin with and then appropriate monitoring of that patient once they're on drug.

We'll move on to number three, and I guess we'll come back. I don't know. Dr. Orloff, do you want us to discuss that aspect now or later?

DR. ORLOFF: You can move on to Question 3.

ACTING CHAIRPERSON MOLITCH: Okay. We'll then move on to approvability, and based on the data presented by the sponsor and the NDA, do you recommend approval of teriparatide (a) for the treatment of post menopausal osteoporosis and (b) to increase bone mineral density in men with osteoporosis?

And now we'll start back on the other side.

DR. HOLMBOE: I'd say yes to A with limitations, no to B.

DR. PELOSI: I would say yes to A but no to B.

DR. AOKI: I'd say yes to both.

DR. LEVITSKY: Yes to both.

DR. TAMBORLANE: Yes to both.

DR. GELATO: Yes to A, no to B.

DR. KREISBERG: Yes to A, no to B.

DR. GRADY: Yes to both.

DR. SAMPSON: Yes to A and no to B.

DR. MITLAK: And I will say yes to both

MS. REEDY: And that tally for Number 3, approvability for the treatment of post menopausal osteoporosis, yes ten, no zero.

For the increase of bone mineral density in males, yes five, no five.

ACTING CHAIRPERSON MOLITCH: I think we're going to make a C question here as well, which comes up on the next page about whether this will be appropriate to use as first line versus second line therapy, and perhaps with this approvability that we've just talked about, perhaps we can go back -- or maybe even a C and a D -- as a proof for first line drug versus second line drug, and maybe we can just do this by comments around the table, and again, I think we will just go around the table and talk about whether people think this would be appropriate for first versus second line therapy.

Dr. Holmboe.

DR. HOLMBOE: I think given the safety concerns that have been raised, I think it should be second line therapy at this time. I think with regard to men, I don't think we have enough clinical outcomes yet to conclude that it should be approved for that indication at this time.

DR. PELOSI: I would concur with that.

DR. AOKI: I agree.

DR. LEVITSKY: I concur that it should be approved for second line therapy, but I'm still a little uneasy as to what it should be second line therapy as second to because I don't think we have enough information yet to be sure.

DR. TAMBORLANE: I would approve it for first line therapy also.

ACTING CHAIRPERSON MOLITCH: Everybody else said second line. So you are first line or second line?

DR. TAMBORLANE: I said first line.


DR. TAMBORLANE: Sorry. The "also" was confusing.


DR. GELATO: I would approve it for first line for women with some, I guess, specifications of the group that gets it first line. For men if it were going to be approved, I would approve it as second line.

DR. KREISBERG: I would approve it for first line therapy in women, but I think the indications ought to be clearly defined.

DR. GRADY: I'd be in favor of restricting it to as a second line drug because I think the efficacy and safety of the other first line drugs is much better.

DR. SAMPSON: I'm going to pass on that as a nonclinician.

ACTING CHAIRPERSON MOLITCH: I like Dr. Lindsay's description of the patients who might be considered as first line therapy, the patients with very severe disease or the patient who has a fracture who's at high risk for a second fracture in the near future. So I'm not sure that I would want to actually restrict it to being a second line therapy, although I think that's probably how I would use it for most patients.

So I'm not sure how to answer my own question.


ACTING CHAIRPERSON MOLITCH: I guess I would not restrict it.

MS. REEDY: The tally on that is as first line therapy, four; as second line, five.

ACTING CHAIRPERSON MOLITCH: We'll then go on to Question 4. If the answer to either question in Number 3 is yes, which I think it was, given the theoretical risk for the development of osteosarcoma in humans treated with teriparatide, (a) should the treatment be limited, and if yes, how much or how long;

(b) Should the use of teriparatide be recommended only for certain subgroups of patients? If yes, please comment on the recommended target populations.

(c) Should teriparatide be limited to use as a second line therapy? If yes, please comment on what criteria should be used to define second line therapy.

And (d) comment on how it should be labeled in the labeling for the bolded warning or black box, and I think we can -- this is more of a general discussion at this point, and why don't se start with Dr. Holmboe again?

DR. HOLMBOE: I'll go through each of these. With regard to A, I actually agree with the sponsor that given the amount of data we have at this time, I would agree with the two-year limitation on the therapy with close monitoring.

With regard to B, it really relates to my previous answer. I, again, would restrict to women at this time, given the clinical efficacy data, and again I would recommend it as a second line agent, although I was somewhat persuaded by your argument with regard to those who are at higher risk as first line therapy, i think which gets to C.

D, please comment on how the findings. I think there are several things that I would recommend. I actually would recommend a bold warning with regard to the osteosarcoma risk. I think that would be very important.

Number two, I do think some patient education materials and provider education materials should be developed.

Although I appreciate the sponsor's response with regard to being transparent about the risk and certainly reporting in conference proceeding, I can tell you as a general internist a lot of the things I don't see, and it often takes time to disseminate that sort of information through that particular mechanism.

So I think that there's going to be a need for other mechanisms to get this warning out, and I also think that it would be helpful for them to provide some guidelines with regard to monitoring even though metabolically overall things look fairly safe. There is still some concerns with regard to some of the metabolic changes.

DR. PELOSI: I agree that it should be two years as well until we get more data.

In terms of B, I think I agree with the sponsor in terms of those people who should not receive it, in terms of those with Paget's disease, the adolescents.

For certain subgroups, I would see women who have already experienced fractures.

Number C, should it be limited as a second line use, I think we've had that discussion, and I did vote yes on that, but I think we do need to have some more criteria. I'm not sure at this point what that would be.

My concern would be, again, making sure that we have a registry of all patients, whether it's first or second line, to follow. I also would caution us in terms of looking at the SEER database. Again, it doesn't cover the entire United States. I think working with M.D. Anderson and some of the other places that have sarcoma centers is wise, but tumor registries really don't have that information and that detail. The question is: do we register physicians like they're currently doing with thalidomide so that we can actually go to that particular physician to get that information?

I think it would be hard to pull out of the SEER database, and I agree in terms of education, but I would also add education to the nurses because nurses are usually the ones that play a significant role in following, as well as educating the patient.

DR. AOKI: To A it would be yes, 24 months.

(b) With the proviso of the Paget's disease, I would recommend that those obviously not receive the drug.

(c) I would limit it to second line therapy at the present time.

(d) I'd have bolded warning.

DR. LEVITSKY: I think a two year limit is as good as any right now. So I would agree with that.

I would agree with the sponsor's recommendations about eliminating certain subgroups of patients as being reasonable ones to be treated.

I had said that I think it should be limited as a second line therapy, which I think should be the FDA recommendation, although I recognize that there might be certain very sophisticated people with special patients who might choose to use it as first line therapy.

And I think that there needs to be some sort of guideline drawn up so that a few calciums are checked rather than just ignoring that issue.

And I think a bolded warning would be sufficient in regard to the osteosarcoma.

DR. TAMBORLANE: Well, consistent with my neighbors, I would say that two years is fine for A.

(b) the stipulations as indicated by the sponsor as far as people should not use the drug is okay.

I voted for first line therapy for the same reasons that Mark had articulated.

And bolded warning. I'm not actually sure I know the difference between a bolded and black box warning, but whatever says "beware of this finding in rats" is important.

DR. ORLOFF: Maybe we should clarify that because the question mentioned black box warning, and a couple of people have talked about bolded warnings, and those are maybe strangely, but they are different in regulatory terminology.

A black box warning is just that. The information is usually set off at the very beginning of the labeling. It's enclosed in a black box, and it's designed to bring your attention to that, the first thing you see. It also has implications for promotion of the drug. When a product has a black box warning in the labeling, it limits the ability of sponsors to hand out promotional materials without handing out the full prescribing information.

So handing out, you know, the trinkets, et cetera, is restricted because you have to hand out the full prescribing information.

That's opposed to a bolded warning, which simply means that somewhere in the labeling -- it could be at the beginning; it could be in the warning section; it could be anywhere -- you put it in bolded type, but doesn't have all of the same promotional implications.

DR. TAMBORLANE: Well, I'm glad I asked. So the black box sounds the most appropriate for me.

DR. GELATO: I think two years is reasonable based on what the sponsor said and also the information that Dr. Neer gave us.

I think that, again, I agree with the sponsor that it should be very clearly laid out the patients who should not be given this drug in terms of a variety of things, any of the risk for osteosarcoma, renal insufficiency, and so on.

And I did say that I thought it could be used as first line therapy in women, but again, I felt that it should be limited to those women as suggested by Dr. Lindsay who had fractures and were at increased risk to fracture again because there I think you might want to go with something that's going to benefit them relatively quickly.

And I would keep it as second line for men who have maybe failed other therapies, and I agree with the black box warning.

DR. KREISBERG: I think that two years is reasonable. I think there should be specific indications and contraindications for the use of the drug as first line or second line therapy.

I agree that it should not be used in patients with Paget's and children and a variety of other things. I'd like to add the contraindication right now that it should not be used in combination whether they are anti-osteoporosis drugs because we don't have any information on that that obviously could be removed in the future.

The indications, I think, ought to be not only for recent fractures, but for failure of previous therapy, and it can be used in patients with very severe osteoporosis even in the absence of fractures, and I think it needs a black box warning.

DR. GRADY: I more or less agree with all of that. I think two years is reasonable. I think the contraindications suggested by the sponsor were reasonable. I think it should be a second line therapy both for women and for men, and perhaps used as first line therapy in women at very high risk if they have had prior fractures or very low bone density.

It's actually the final issue that worries me the most. I think that our two options for how to kind of get more data on this are to -- and perhaps we ought to discuss a little bit the idea of recommending that the sponsor wait to market this drug until the findings in the new animal studies are completed.

Actually I personally don't think that's going to help us a whole lot. I think the sample sizes are too small and we'll be still stuck with this prior rat study where 50 or 60 percent of them got osteosarcoma, and I believe that that leaves us only with some sort of registry, and I'd like to hear a little more discussion about that.

You know, we keep saying registry, but in my mind registry means that the people who take the drug sign up at the time they begin to take it and that you get identifiers for those people, that typically that would be, you know, full name, full birth date and Social Security number so that you can subsequently link with the tumor registry and perhaps also link with the national death index because I think anybody who gets osteosarcoma dies in relatively short order, and that's usually the cause of death, and that would clearly be, I think, the strictest way to go, is to sort of require prospective follow-up on persons who take this drug.

I'm not quite sure how registries usually work.

DR. KREISBERG: Why don't we just come back to this whole issue? That's actually Question 5 about the post marketing surveillance also. I think this whole issue of how to do this should merit full discussion. So why don't we come back to that before Dr. Stadel speaks?

DR. GRADY: Yes, and I like the black box.

DR. SAMPSON: I think I agree with the sponsor's recommendation of two years in duration.

Abstaining on B and C, and bolded warning on D.

ACTING CHAIRPERSON MOLITCH: I agree about the 24 months. I think the certain subgroups of patients I would like to have very clearly defined clearly the Paget's and the adolescents, but then other secondary causes of osteoporosis, which it should not be used in. It should be outlined in the package insert and certainly in educational, promotional materials, and maybe even specific tests recommended such as alkaline phosphatase and PTH before we start any therapy in patients that are essentially mandatory, and I would think also serum calcium level is mandatory prior to starting, and perhaps at designated intervals after starting therapy at three months and then perhaps calcium and alkaline phosphatase every six months for the duration of therapy would seem to me to be a reasonable thing to do in these patients.

As far as the warning, I think it should be a black box warning pending any additional information, and then I think we do need to come back to this whole issue of how are we going to track patients. What kind of either post marketing or perhaps as Dr. Bone suggested even pre-marketing studies or continuation studies to follow patients like this, and additional ideas, I think, are welcome.

Then we can start perhaps with Dr. Stadel and then come back through Dr. Sampson.

DR. STADEL: I don't think I have a hard and fast view about registries, although with such a rare disease, the orientation and discussion has been towards an active mode of case finding with the idea that exposure would be determined through the case if they're still alive or through the case's physician if they're not.

Then if you use that series for a case control study, you would have to mount a parallel measure of ascertainment.

You know, a registry could be used, and if one wanted to sweep the registry, register people through the national death index, that would make sense doing it that way. I think the reason that I haven't thought in that direction is simply that 99.9 percent of the information isn't useful. The outcome is so rare that going in the other direction is much more efficient.

DR. GRADY: Well, I guess I wonder have you done that because it seems to me it takes a long time to do a case control study. You have to accumulate a lot of cases, and then you have to go back and get their exposure data, and I just worry a bit that by the time you're able to accomplish a case control study, there might be a lot of osteosarcomas.

DR. STADEL: What we've been talking about doing would be a system of ascertaining cases in referral centers fairly rapidly, getting basic data on them. Whether they're used to mount a case control study, we haven't worked out all of the details, but the first step would be to ascertain the cases while they're still alive, ascertain them fairly rapidly.

I think we should think more about the issue. As I say, the biggest reservation I have about a registry is simply that it's a lot of work for the patients, a lot of work for the doctor, and most of the registry information would not get used for anything.


DR. BONE: Since we're not really voting and just commenting, a couple of thoughts based on our experience with something a little bit analogous, which was the triglitozone (phonetic) issue, and we have major issues there about ascertainment of cases, and there were estimates that the number of cases could be as high as ten times those counted or not, and there was a lot of discussion.

And the other thing is what's the denominator. So I think that whatever needs to be done from that public health sort of approach to ascertainment, as rigorous as possible identification of both the denominator and the numerator would be extremely important.

And I'm inclined to think that Dr. Grady's suggestion would be very useful if it could be implemented.

I'll just mention that complementary to that would be something along the lines I mentioned during the earlier discussion. It won't get you all the way to discovering whether there's a modest increase in the risk, but it will help with being assured there's not a big increase and also address a number of the issues that we've been wrestling with in a way that would be definitive, I would think.


DR. SAMPSON: I think I concur with Dr. Stadel. You're looking for such a rare occurrence, and if this is something that would be noted, it seems to me that just the identification of cases and then proper plan study after that would be a pretty good way to go.

ACTING CHAIRPERSON MOLITCH: Do you have any further comments, Dr. Grady?

DR. GRADY: The term "registry" has been used. I was just wondering how you would usually go about that. What do you mean when you say there's a "registry"?

DR. STADEL: Sorry. I don't understand your question.

DR. GRADY: Well, I think one of the people have been saying, "Oh, well, we should have a registry to get additional post marketing data on."

DR. STADEL: The term "registry" usually means that you register exposure and do follow-up. With the rarity of this outcome, I'd be much more inclined to go after the outcome, and it's a relatively easy exposure also. It's really -- at the first cut it's a binary exposure. Have you taken this drug or not?

And so I think actually it would be better to put the effort into as wide a case finding as possible, the ability to ascertain exposure in cases, and to use the case series, depending if it reaches a point where it looks like one would learn anything by doing a case control study.

You know, if the intercept is nothing, if you ascertain a large series of cases and none of them are exposed, then you're relatively sure of what the odds ratio will be. So I think this is a case where a lead edge on case ascertainment with the ability to mount a case control study when one has some idea of how the case series is panning out would make sense.


DR. KREISBERG: Well, my answer is brief. I think we ought to get some people who know how to do this and have them tell us what to do.



DR. GELATO: I would vote with Dr. Kreisberg. I think that's a good idea.

DR. TAMBORLANE: I would, too, except I would try to make sure it was proactive, you know, that there was a plan. Going after the cancer registry sounds like the way to go to me.

DR. LEVITSKY: I have a particular concern, which is that unlike the triglitozone issue, it's not going to be that you take the drug and you turn yellow. It's going to be that you take the drug; you forgot you took the drug; it's 15 years later, and something good happened to you, and I think really you need to have people who know exactly how to do that and to retrieve that information.

DR. TAMBORLANE: Especially since they're 100 years old by then and can't remember.



DR. AOKI: I agree with Dr. Kreisberg.

DR. STADEL: Something I tried to emphasize, and I'll take the opportunity to say again, this could take a very long time, and we need to be clear that we're talking about a large element of uncertainty that will go on for a protracted period of time even if we do the very best we can.

DR. PELOSI: I would just say with the registry if you register the patients at the beginning, run it through tumor registry, but all of the tumor registries, but not just SEER because SEER isn't comprehensive, but I really like the idea of the death index. It would be actually a quicker way to go to do case findings.

And it's my understanding we're answering number five right now. Is that -- well, I would just want to make sure that we did see future studies looking at more mature rats than the immature rats. I think that might be of help, but with a number in the sample that might be meaningful to us.

And I would also hope that we could look at quality of life data post treatment to see if, indeed, what the impact was on their life through this treatment as well.

DR. HOLMBOE: I think from a risk management standpoint I agree with Dr. Stadel that the registry has a lot of logistical difficulties, particularly at the point of care, and these filled out whether it be in the office or in the pharmacy, and I think you'd have to collect a huge amount of data to recognize a signal quickly.

So I think if you're going to do that, you still need to do the kind of surveillance you're suggesting with regard to looking for cases at referral centers using SEER data, you know, from a case control standpoint as well. I don't think you can get around not doing that when you have such a rare outcome.

I think that's where actually case control studies help. It's true, as Dr. Grady points out, you have to wait until the cases occur, and that's the down side to that. So it's always going to be retrospective, and so prospectively other possibilities would be to take advantages of some of the large pharmacy databases that now do exist around the country, including through the DOD and VA and some of the others. I don't know if you have access to those, but sometimes those can be helpful.

DR. STADEL: Only to say that with an expected rate of 400 million per year, I respectfully submit that most of those databases are too small.

DR. HOLMBOE: Agreed, but again, if we're looking for any signal, I wouldn't rule them out because we don't know what the rate is. I mean, granted, if it turns out to be that low, you'll never see it, but we don't know that, and we really don't. So I think taking advantage of what you do have, particularly with regards to some logistical digital registry, I'd still consider taking a look at those.

ACTING CHAIRPERSON MOLITCH: Well, I would agree. I think some post marketing registry does need to be done. I think if we rely upon physicians filling out forms in their office, it's not going to happen, speaking from personal experience.

However, one possible way might be to get patients to send in a $10 rebate slip to the company to register their name and their Social Security number, et cetera, when they get their first dose of drug and perhaps every six months thereafter for the duration of therapy to send in that slip to get their rebate, ten or $15 or $1,000, whatever Dr. Kreisberg thinks is appropriate.


ACTING CHAIRPERSON MOLITCH: But that may be one way to get some compliance in that regard.

I think right now we'll take any additional comments that any members of the panel might have that might be useful. Dr. Kreisberg.

DR. KREISBERG: Well, I'd only reiterate I'm sure the company -- I'm not telling the company anything that they haven't already considered -- is head to head as well as combination studies in a prospective randomized fashion would be very helpful to those of us who practice and take care of patients.

I may be naive, but I've always thought of osteoporosis being either increased bone resorption or decreased bone formation or a combination of that, which means there must be some pathophysiologic approach to identification of patients that might benefit from one form of therapy or another, and while that may be pretty unsophisticated, I would hope that the company would look at possible ways of identifying patients right from the very outset without having to resort to fractures or resort to extensive osteoporosis as being able to finger a patient who really should be a candidate for a bone forming agent rather than for an anti-resorptive agent.


Dr. Orloff.

DR. ORLOFF: If there are no further comments, I'd just like to make a couple of closing remarks.

I want to thank everyone again for participating in this process, the members of the Advisory Committee, the consultants sitting around the table, as well as the consultants who are sitting with the sponsor today.

I also want to recognize and applaud the collegial and successful collaboration that's gone on between the division and the sponsor to date and to say that I hope it continues.

So, again, thank you all and good afternoon.


MS. REEDY: People at the table, if you would like us to ship to you again your documents, please put your name on top and indicate so. If you would like us to shred them for you, simply leave them there on your chair.

(Whereupon, at 3:51 p.m., the Advisory Committee meeting was concluded.)