1

DEPARTMENT OF HEALTH AND HUMAN SERVICES

FOOD AND DRUG ADMINISTRATION

CENTER FOR DRUG EVALUATION AND RESEARCH

 

 

 

 

 

 

ANTIVIRAL DRUGS ADVISORY COMMITTEE

IMMUNOSUPPRESSIVE DRUGS SUBCOMMITTEE MEETING

NDA 21-083 Rapamune (sirolimus) Oral Solution

Cyclosporine Withdrawal Maintenance Regimen

 

 

 

 

 

 

 

 

Thursday, January 24, 2002

8:30 a.m.

 

 

Holiday Inn Gaithersburg

Two Montgomery Village Avenue

Gaithersburg, Maryland

2

PARTICIPANTS

Janet A. Englund, M.D., Chairperson

Tara P. Turner, Pharm.D., Executive Secretary

AVAC MEMBER

Victor DeGruttola, Sc.D.

CONSULTANTS (VOTING)

Darrell Abernethy, M.D., Ph.D.

Hugh Auchincloss, Jr., M.D. (CBER

Consultant)

Steven Ebert, Pharm D. (Consumer

Representative)

Lynt B. Johnson, M.D.

William G. Lawrence, J.D. (CBER

Consultant)

Ron Shapiro, M.D.

Mannikam Suthanthiran, M.D.

CONSULTANT (NON-VOTING)

Lawrence Hunsicker, M.D.

GUEST (NON-VOTING)

Roslyn B. Mannon, M.D.

FDA

Renata Albrecht, M.D.

Marc Cavaille-Coll, M.D., Ph.D.

Mark Goldberger, M.D., M.P.H.

Rosemary Tiernan, M.D., M.P.H.

3

C O N T E N T S

PAGE

Call to Order and Opening Remarks

Janet Englund, M.D. 4

Conflict of Interest Statement

Tara P. Turner, Pharm.D. 6

FDA Introductory Remarks

Renata Albrecht, M.D. 9

Sponsor Presentation, Wyeth-Ayerst Research

Introduction: Randall B. Brenner, M.S. 12

Overview: John F. Neylan, M.D. 18

Design of Clinical Studies: John F. Neylan, M.D. 25

Efficacy Review: John F. Neylan, M.D. 33

Safety Review: John F. Neylan, M.D. 49

Pharmacokinetics: James Zimmerman, Ph.D. 70

Concentration-Controlled Trials:

James Zimmerman, Ph.D. --

Therapeutic Drug Monitoring:

James Zimmerman, Ph.D. --

Concluding Remarks: John F. Neylan, M.D. 87

FDA Presentation

Rosemary Tiernan, M.D., M.P.H. 114

Open Public Hearing

Alan Wilkinson, M.D., F.R.C.P. 167

Charge to the Subcommittee 185

Subcommittee Discussion and Vote 188

4

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

2 CALL TO ORDER

3 DR. ENGLUND: Good morning, everyone.

4 Welcome to the Subcommittee for Immunosuppressants

5 Meeting of the Antiviral Drugs Advisory Committee

6 group. I hope you are all in the right place here.

7 My name is Janet Englund. I am the Acting

8 Chairperson for this session. I am from the

9 University of Chicago and am a member of the

10 Antiviral Drugs Advisory Committee. We are very

11 grateful to have such knowledgeable guests and

12 voting members here to help us with the discussion

13 today.

14 At this point in time, I think what we can

15 do is ask everyone at the table to introduce

16 themselves, their name and their affiliation.

17 Perhaps, if we could start at the very back, to my

18 left.

19 DR. MANNON: I am Dr. Roslyn Mannon. I am

20 the transplant nephrologist at NIH and I am the

21 Medical Director of Transplantation at the NIDDK

22 Organ Transplant Program where we do kidney,

23 kidney-pancreas, pancreas transplants and, for the

24 past year and a half, have had extensive use in

25 rapamycin.

5

1 DR. HUNSICKER: Larry Hunsicker from the

2 University of Iowa. I am a transplant nephrologist

3 also. I am a clinical trialist. I think that

4 suffices.

5 MR. LAWRENCE: William Lawrence. I am an

6 attorney. I am Director of Patient Affairs for the

7 United Network for Organ Sharing. I am a liver

8 recipient of some fourteen years.

9 DR. AUCHINCLOSS: My name is Hugh

10 Auchincloss. I am a transplant surgeon at Harvard.

11 DR. ABERNETHY: Darrell Abernethy,

12 National Institute on Aging. I am a clinical

13 pharmacologist.

14 DR. DeGRUTTOLA: Victor DeGruttola,

15 statistician at Harvard School of Public Health.

16 DR. TURNER: Tara Turner, Executive

17 Secretary for the Committee.

18 DR. EBERT: Steven Ebert. I am an

19 infectious diseases pharmacist at Meriter Hospital

20 and Professor of Pharmacy at the University of

21 Wisconsin.

22 DR. SUTHANTHIRAN: Mannikam Suthanthiran.

23 I am Chief of Transplantation Medicine at New York

24 Hospital, Cornell Medical Center.

25 DR. SHAPIRO: I am Ron Shapiro. I am

6

1 Director of Renal Transplantation at the Thomas E.

2 Stassel Transplantation Institute at the

3 University of Pittsburgh.

4 DR. TIERNAN: Rosemary Tiernan, medical

5 reviewer, FDA.

6 DR. CAVAILLE-COLL: Marc Cavaille-Coll,

7 medical team leader, Division of Special Pathogen

8 and Immunologic Drug Products, FDA.

9 DR. ALBRECHT: I am Renata Albrecht,

10 Acting Director, Division of Special Pathogen and

11 Immunologic Drug Products.

12 DR. ENGLUND: Thank you. Welcome,

13 everyone. I would like now to have Tara Turner,

14 the Executive Secretary, read the conflict of

15 interest statement.

16 Conflict of Interest Statement

17 DR. TURNER: Thank you. The following

18 announcement addresses the issue of conflict of

19 interest with regard to this meeting and is made a

20 part of the record to preclude even the appearance

21 of such at this meeting.

22 Based on the submitted agenda for the

23 meeting and all financial interests reported by the

24 committee participants, it has been determined that

25 all interests in firms regulated by the Center for

7

1 Drug Evaluation and Research which have been

2 reported by the participants present no potential

3 for an appearance of a conflict of interest at this

4 meeting with the following exceptions.

5 Dr. Ron Shapiro has been granted waivers

6 under 18 USC 208(b)(3) and 21 USC 355(n)(4)

7 amendment of Section 505 of the Food and Drug

8 Administration Modernization Act for his lectures

9 supported by a competitor on unrelated matters. He

10 receives more than $10,000 a year.

11 Dr. Janet Englund has been granted a

12 waiver under 18 USC 208(b)(3) for her consulting

13 for a competitor on unrelated matters. She

14 receives less than $10,000 a year.

15 Dr. Lawrence Hunsicker has been granted

16 limited waivers allowing his participation without

17 voting privileges under 18 USC 208(b)(3) and 21 USC

18 355(n)(4) amendment of Section 505 of the Food and

19 Drug Modernization Act for three grants and

20 contracts to his employer. The first is a grant

21 from the federal government and a competitor

22 involving competing products funded for less than

23 $100,000 per year. The second is a contract from a

24 competitor involving competing products and the

25 product at issue. However, Dr. Hunsicker is

8

1 unaware of the details of this contract. The third

2 is a grant from the federal government involving

3 competing products which receives funding greater

4 than $300,000 per year.

5 A copy of these waiver statements may be

6 obtained by submitting a written request to the

7 agency's Freedom of Information Office, Room 12A30,

8 of the Parklawn Building. In the event that the

9 discussions involve any other products or firms not

10 already on the agenda for which an FDA participant

11 has a financial interest, the participants are

12 aware of the need to exclude themselves from such

13 involvement and their exclusion will be noted for

14 the record.

15 With respect to all other participants, we

16 ask, in the interest of fairness, that they

17 address any current or previous financial

18 involvement with any firm whose products they may

19 wish to comment upon.

20 Thank you.

21 DR. ENGLUND: Thank you. I think we have

22 Dr. Johnson here with us, if you want to introduce

23 yourself.

24 DR. JOHNSON: I apologize for the

25 tardiness. Sometimes, it is hard when you have a

9

1 meeting that is at home. I am Lynt Johnson. I am

2 the Director of Transplantation at Georgetown

3 University Medical Center here in Washington, D.C.

4 DR. ENGLUND: Thank you. Glad you're

5 here.

6 At this point, I would like Dr. Renata

7 Albrecht, who is Acting Director of the Division of

8 Special Pathogens and Immunological Drug Products

9 at the FDA, to give us some opening remarks.

10 FDA Introductory Remarks

11 DR. ALBRECHT: Thank you, Dr. Englund. On

12 behalf of the Division, I would like to extend a

13 welcome to you, Dr. Englund, to the members of the

14 committee, our distinguished guests and

15 representatives from Wyeth-Ayerst. We very much

16 appreciate your being here today to discuss a new

17 Rapamune regimen in the management of patients with

18 renal transplants.

19 Specifically, this is the first time the

20 agency and the committee has been asked to consider

21 a regimen, a maintenance regimen, in which

22 cyclosporine is withdrawn as the Rapamune dose is

23 increased to target blood levels.

24 Many of you will recall the original

25 application for Rapamune was brought before this

10

1 subcommittee in the summer of 1999 and resulted in

2 the approval of Rapamune, the 2 milligram dose, in

3 combination with cyclosporine and steroids for

4 maintenance. Results were also presented for the 5

5 milligram dose which was interpreted as showing

6 similar efficacy and increased toxicity.

7 One of the noteworthy findings from those

8 original studies was the reduction in

9 glomerular-filtration rate noted in the Rapamune,

10 cyclosporine and corticosteroid arm relative to the

11 other arm. This raised questions about long-term

12 consequences of the regimen and also prompted the

13 agency to ask the sponsor to conduct some phase IV

14 studies.

15 Now the company has submitted to us a

16 supplemental application containing studies in

17 which many patients were randomized to the

18 cyclosporine-withdrawal arm and had the Rapamune

19 doses increased. Questions that arise are whether

20 the cyclosporine withdrawal may have affected

21 efficacy either favorably or unfavorably.

22 The other questions are regarding safety.

23 Are there changes in the safety profile. Has the

24 GFR been preserved? Are there other new toxicities

25 that may be introduced with this new regimen?

11

1 These are some of the questions that we will be

2 asking you to deliberate during the course of this

3 meeting.

4 Finally, I would like to express our

5 appreciation to Wyeth for putting forth a great

6 effort in planning in bringing forth this

7 application to the committee for discussion. I

8 would also like to recognize some of my colleagues,

9 Dr. Marc Cavaille-Coll, Rosemary Tiernan, Karen

10 Higgins and Cheryl Dixon for the intense effort

11 they have put forth into this project.

12 In the first part of the morning, Wyeth

13 will present a number of talks on the clinical and

14 pharmacokinetic findings from their studies. This

15 will be followed by a presentation by Dr. Rosemary

16 Tiernan. Finally, as I mentioned, we do have a

17 number of questions that we would like the

18 committee to deliberate and give us guidance on

19 this application and on issues relative to

20 clinical-study endpoints.

21 With that, thank you and I will return it

22 to you, Dr. Englund.

23 DR. ENGLUND: Thank you. At this point, I

24 think I would like to introduce Randall Brenner

25 from Wyeth-Ayerst Research to start your

12

1 presentation.

2 Sponsor Presentation--Wyeth-Ayerst Research

3 Introduction

4 DR. BRENNER: Good morning, everyone.

5 [Slide.]

6 I am Randy Brenner from the Regulatory

7 Affairs Department at Wyeth-Ayerst. On behalf of

8 our organization, we are pleased to have this

9 opportunity today to review the data supporting our

10 supplemental NDA for the cyclosporine elimination

11 indication for Rapamune for use in renal-transplant

12 patients.

13 [Slide.]

14 Our presentation today has the following

15 agenda. Upon completion of my brief introductory

16 remarks, Dr. John Neylan will discuss the need for

17 a calcineurin-inhibitor-free immunosuppressive

18 regimen in renal-transplant patients. He will

19 review in detail the designs of our pivotal-study

20 Protocol 310 and a supportive phase II study

21 Protocol 212 and provide a review of the collective

22 efficacy and safety data from these studies.

23 Following Dr. Neylan, Dr. James Zimmerman

24 will review the pharmacokinetics of Rapamune in

25 concentration-controlled trials and therapeutic

13

1 drug monitoring in this patient population.

2 For a conclusion, Dr. Neylan will return

3 and summarize the results presented today and

4 address any questions you may have.

5 [Slide.]

6 The oral solution formulation of Rapamune

7 was first approved in the United States in

8 September of 1999. This application received a

9 priority review from FDA and was presented to this

10 advisory committee in July of 1999.

11 The approved package insert recommends

12 fixed dosing of this product in combination with

13 cyclosporine. Specifically, a 6 milligram loading

14 dose followed by a 2 milligram fixed daily dose is

15 recommended for most patients. A 5 milligram dose

16 has also been approved.

17 Immediately following approval of the oral

18 solution formulation, an application requesting

19 approval of a tablet formulation was submitted to

20 FDA. The 1 milligram tablet, which was approved in

21 August of 2000, provided significant advantages

22 over the oral solution in terms of patient

23 convenience while not compromising safety or

24 efficacy.

25 [Slide.]

14

1 The original advisory committee

2 presentation was supported by two phase II pivotal

3 studies, Protocols 301 and 302. These studies

4 demonstrated that, when used in combination with

5 cyclosporine, patients receiving fixed doses of

6 Rapamune had significantly lower rates of acute

7 rejection at less than 18 percent while maintaining

8 excellent patient and graft survival at greater

9 than 95 and 90 percent respectively.

10 As such, this committee voted unanimously

11 that this product was safe and efficacious. One of

12 the more important issues discussed in detail was

13 the unexpected impact of the Rapamune-cyclosporine

14 combination on renal function. As a result, this

15 committee and the FDA recommended that Wyeth

16 further evaluate this finding.

17 We were optimistic that we could

18 demonstrate that the observed renal effects in

19 Protocols 301 and 302 were due to the exacerbation

20 of cyclosporine toxicity and were not directly

21 related to Rapamune.

22 [Slide.]

23 To demonstrate this, we looked at the

24 information we knew from our phase III pivotal

25 studies, Protocols 310 and 302, which used fixed

15

1 dosing of Rapamune in combination with

2 cyclosporine. We also looked at information we

3 knew from additional phase II studies which used

4 Rapamune as base therapy demonstrating a favorably

5 safety profile with significant improvements in

6 renal function.

7 This was further supported by animal data

8 demonstrating Rapamune to be nonnephrotoxic and an

9 effective immunosuppressive agent when evaluated

10 alone. Rapamune's inherent absence of

11 nephrotoxicity is what makes a

12 calcineurin-inhibitor-free regimen with this

13 product potentially so beneficial to

14 renal-transplant patients.

15 As a result, we designed the current

16 registration studies, Protocols 212 and 310. These

17 studies evaluated the currently approved

18 combination of Rapamune plus cyclosporine versus a

19 group of patients that had cyclosporine eliminated

20 from the immunosuppressive regimen two or three

21 months after transplantation.

22 Additional details regarding the designs

23 of these studies will be presented by Dr. Neylan in

24 the design portion of this presentation.

25 [Slide.]

16

1 Protocols 212 and 310, the studies in the

2 current application, demonstrate equivalent

3 efficacy with excellent patient and graft survival

4 with an improvement in safety specifically in

5 regard to renal function and blood pressure.

6 Importantly, despite a difference in the number of

7 acute-rejection episodes immediately following

8 cyclosporine elimination, by month 12, there were

9 similar rates of acute-rejection episodes in both

10 arms.

11 Dr. Neylan will relate the impact of acute

12 rejection immediately following cyclosporine

13 elimination as it relates to severity, long-term

14 patient and graft survival and the impact on renal

15 function.

16 [Slide.]

17 The application currently under review and

18 in front of this committee today seeks approval of

19 an indication that will allow for the elimination

20 of cyclosporine from the immunosuppressive regimen.

21 The Rapamune dosing for this new indication

22 recommends fixed dosing for the initial

23 post-transplant period.

24 At the time of cyclosporine withdrawal, at

25 two to four months post-transplantation, Rapamune

17

1 dosing will be based on trough concentration levels

2 within a recommended range. As this new dosing

3 will require patient dosing utilizing trough

4 concentration levels, therapeutic drug monitoring

5 will now be required.

6 Dr. Zimmerman will discuss therapeutic

7 drug monitoring in detail during his presentation.

8 [Slide.]

9 As a reminder, Rapamune is currently

10 indicated in use in combination with cyclosporine.

11 The currently approved indication is provided here.

12 Rapamune is indicated for the prophylaxis of organ

13 rejection in patients receiving rental transplants.

14 It is recommended that Rapamune be used in a

15 regimen with cyclosporine and corticosteroids.

16 You will see today that the results of

17 Studies 212 and 310 provide physicians with an

18 alternate dosing regimen for Rapamune which

19 provides acceptable immunosuppressive while

20 preserving renal function. As such, we seek

21 approval of an indication provided here in which

22 Rapamune is indicated for the prophylaxis of organ

23 rejection in patients receiving renal transplants.

24 It is recommended that Rapamune be used initially

25 in a regimen with cyclosporine and corticosteroids.

18

1 Cyclosporine withdrawal should be considered two to

2 four months after transplantation.

3 This concludes my introduction. I would

4 now like to introduce Dr. John Neylan, the Vice

5 President of Clinical Research and Development for

6 Wyeth-Ayerst.

7 Overview

8 DR. NEYLAN: Thank you Randy, and good

9 morning.

10 [Slide.]

11 As Mr. Brenner told you, Rapamune was

12 recommended for approval by this committee in 1999

13 in combination with cyclosporine for the prevention

14 of rejection in renal-transplant patients. The

15 registration of this product has provided new

16 opportunities to advance immunosuppressive therapy

17 and improve patient outcomes.

18 We are here today to provide additional

19 data which will allow transplant physicians new

20 opportunities to build upon this success, improve

21 graft function and potentially extend the life of

22 transplant kidneys.

23 [Slide.]

24 While the addition of new drugs has

25 decreased the incidence of acute rejection and

19

1 improved graft survival in the short term,

2 long-term outcomes remains suboptimal. Indeed,

3 most patients must continue to expect that their

4 transplants will fail within a decade.

5 In most cases, this graft failure will be

6 secondary to a deterioration, progressive over

7 time, in renal function.

8 [Slide.]

9 Calcineurin inhibition, while providing

10 effective immunosuppressive, has long been

11 associated with time and dosage-dependent

12 toxicities that may lead to chronic allograft

13 nephropathy. This nephrotoxic injury has been

14 reported in up to 65 percent of renal, liver, heart

15 and bone-marrow transplant recipients and has been

16 directly implicated in causing end-stage renal

17 disease in up to 10 percent of nonrenal solid-organ

18 recipients.

19 It is not surprising, then, that, since

20 1983 and the introduction of cyclosporine,

21 clinicians have continued in their quest to

22 eliminate nephrotoxicity. Our goal today is to

23 provide data to convince you that patients will

24 benefit from withdrawal of cyclosporine and

25 maintenance therapy with Rapamune. That is the

20

1 single objective of the current studies.

2 [Slide.]

3 Rapamune, through its distinct biologic

4 activity and non nephrotoxic profile, offers the

5 opportunity to provide a new cornerstone to

6 immunosuppressive regimens. Although many of you

7 are familiar with the mechanism of action, I will

8 briefly review it now.

9 [Slide.]

10 Rapamune is a novel drug, neither a

11 calcineurin inhibitor nor an antimetabolite. It

12 has a unique cellular target, mTOR, the mammalian

13 target of rapamycin. mTOR is a protein kinase

14 which is critical for cell-cycle progression and

15 cell proliferation. Rapamune blocks mTOR. This

16 action blocks cytokine-mediated cell proliferation

17 in T-cells, B-cells and mesenchymal cells including

18 smooth-muscle cells.

19 [Slide.]

20 All known therapeutic effects of Rapamune

21 result from inhibition of mTOR. Critical pathways

22 affected by Rapamune include the following. One,

23 activation of translation for specific messenger

24 RNAs coding for cell-cycle proteins. Two,

25 activation of cyclin-dependent kinases required for

21

1 coordinated DNA synthesis. Three, synthesis of

2 specific ribosomal proteins required for cell-cycle

3 progression.

4 The interaction of Rapamune with mTOR is

5 specific and it is reversible and, importantly,

6 Rapamune is not cytotoxic. In summary, the

7 biologic activity of Rapamune as an inhibitor of

8 cell-cycle progression is consistent with both the

9 immunosuppressive and antiproliferative effects of

10 the molecule.

11 [Slide.]

12 Next, we will review the data supporting

13 the design of the current registration trials.

14 This includes the utility and outcome seen when

15 Rapamune is administered with cyclosporine to

16 renal-transplant recipients. In addition, data

17 will be presented from clinical studies in which

18 Rapamune was utilized as a prophylactic agent in

19 renal-transplant patients.

20 Finally, data will be presented in which

21 Rapamune was utilized as primary therapy for

22 recalcitrant psoriasis.

23 [Slide.]

24 In two phase III blinded trials comprising

25 some 1300 patients, Rapamune at 2 milligrams per

22

1 day or 5 milligrams per day was coadministered with

2 cyclosporine and corticosteroids and compared with

3 either placebo or azathioprine controls.

4 The Rapamune treatment groups proved to

5 have low rates of acute rejection and twelve-month

6 patient and graft survival was excellent. However,

7 an unanticipated finding in the unblinding of these

8 studies was the somewhat higher mean serum

9 creatinines in the Rapamune-treated patients.

10 Data from other trials with Rapamune had

11 suggested that the drug was not inherently

12 nephrotoxic. Thus, the change in renal function in

13 these studies was considered to be secondary to an

14 exacerbation of cyclosporine toxicity and not

15 directly related to Rapamune.

16 [Slide.]

17 The absence of nephrotoxicity is supported

18 by data obtained from two phase II trials in which

19 Rapamune was utilized as primary therapy in the

20 absence of cyclosporine. In one trial, study 207,

21 patients were randomized to receive either Rapamune

22 or cyclosporine in combination with azathioprine

23 and corticosteroids.

24 In the second trial, study 210, patients

25 received either Rapamune or cyclosporine with

23

1 concomitant mycophenolate mofetil and

2 corticosteroids.

3 [Slide.]

4 Pooled data from these studies

5 demonstrated that Rapamune and cyclosporine had

6 similar benefits in the prevention of acute

7 rejection and two-year patient and graft survival

8 but were associated with very different effects on

9 renal function.

10 Shown here are statistically significant

11 improvements in both creatinine and calculated

12 glomerular filtration rates in the Rapamune-treated

13 patients. These improvements were seen early and

14 were sustained over 24 months of follow up.

15 [Slide.]

16 In psoriatic patients, Rapamune as

17 monotherapy similarly demonstrated no adverse

18 effects on renal function. Patients with

19 recalcitrant psoriasis were administered Rapamune

20 monotherapy at doses of 1, 3 and 5 milligrams per

21 meter squared per day and compared with

22 placebo-treated patients. There were no

23 differences seen in mean serum creatinines

24 following twelve weeks of therapy in any of the

25 treatment groups even when Rapamune was

24

1 administered at doses as high as 10 milligrams per

2 day.

3 [Slide.]

4 In summary, when Rapamune was administered

5 in two phase III trials with concomitant

6 cyclosporine treatment, low rates of acute

7 rejection but higher serum-creatinine

8 concentrations were observed compared to control

9 therapies. When Rapamune was administered to

10 renal-transplant patients as primary therapy for up

11 to 24 months in doses ranging from 6 to 9

12 milligrams per day, these patients enjoyed similar

13 patient and graft survival but had lower serum

14 creatinines and higher glomerular-filtration rates

15 compared to cyclosporine-treated patients.

16 Rapamune administered as monotherapy to

17 patients with recalcitrant psoriasis at doses of up

18 to 10 milligrams per day had no adverse impact upon

19 renal function. These collective data demonstrated

20 the clinical utility of Rapamune in a variety of

21 settings. While the combination of Rapamune plus

22 cyclosporine resulted in improved rejection

23 outcomes, the changes in renal function were in

24 clear contrast to studies in which Rapamune was

25 used without concomitant cyclosporine.

25

1 Design of Clinical Studies

2 DR. NEYLAN: These collective observations

3 led us to conduct trials of Rapamune-based therapy

4 to test the benefit of cyclosporine elimination.

5 [Slide.]

6 We worked closely with over 60

7 investigators worldwide to develop studies that

8 would test the hypothesis that Rapamune-based

9 therapy could replace long-term cyclosporine-based

10 therapy.

11 [Slide.]

12 Since the introduction of cyclosporine,

13 numerous trials have been conducted to examine

14 whether this agent could be safely withdrawn from

15 long-term maintenance regimens. Many such studies

16 were based on a classic elimination strategy in

17 which immunosuppression was maximized early on for

18 its potential benefits in the prophylaxis of acute

19 rejection with subsequent elimination of

20 cyclosporine in the maintenance phase to decrease

21 long-term toxicity.

22 [Slide.]

23 Studies 310 and 212 were modeled after

24 designs tested in previous elimination trials.

25 Specifically, all of the patients were treated for

26

1 the first two to three months with a regimen

2 consisting of Rapamune plus cyclosporine and

3 corticosteroids to maximize freedom from rejection

4 during this period of greatest immunologic risk.

5 [Slide.]

6 As we previously demonstrated in two large

7 pivotal trials, Rapamune, in combination with

8 cyclosporine, provides one of the lowest rates of

9 acute rejection in this early post-operative period

10 when compared with other immunosuppressive

11 regimens.

12 Following the period of initial risk,

13 patients in the control groups continue to receive

14 combination therapy with cyclosporine while

15 patients in the treatment arms had cyclosporine

16 withdrawn from regimen and concentration-control

17 Rapamune continued during the maintenance phase.

18 The comparison of these regimens allowed us to

19 examine the incidence of acute rejection when

20 cyclosporine was withdrawn and to identify

21 differences in the safety profiles following the

22 elimination of cyclosporine.

23 The pivotal phase III trial in this

24 application is study 310. It is supported with

25 data from Study 212, a smaller phase II trial.

27

1 Both trials were open label, controlled, randomized

2 and multicenter. Study 310 was conducted in 57

3 centers in Australia, Canada and Europe and

4 included a total of 525 patients.

5 These patients were either primary or

6 secondary recipients of renal allografts and

7 received donor organs from either cadaveric or

8 HLA-mismatched living donors. Randomization in

9 this trial occurred at Month 3.

10 In Study 212 conducted in 17 centers in

11 the U.S. and Europe, 246 patients were enrolled.

12 These patients were recipients of primary renal

13 allografts from cadaveric donors with randomization

14 occurring Days 2 through 7 following

15 transplantation. It is important to note that, in

16 both studies, all centers were required to follow

17 the patients for the full duration of the study for

18 the occurrence of acute rejection, graft survival,

19 patient survival and serious adverse events even if

20 these patients discontinued study medication.

21 [Slide.]

22 The primary endpoints of the two studies

23 differed. study 310 was powered for equivalent

24 graft survival at one year while study 212 was

25 powered to demonstrate a significant difference in

28

1 renal function in a population of patients who

2 remained rejection free and on therapy at six

3 months following transplantation. For those

4 studies, multiple secondary endpoints were

5 examined.

6 [Slide.]

7 For study 310, major secondary endpoints

8 included patient survival, the incidence of

9 biopsy-confirmed acute rejection, renal function,

10 efficacy failure and treatment failure. For study

11 212, major secondary endpoints included patient and

12 graft survival, the incidence of biopsy-confirmed

13 acute rejection, renal function beyond six months

14 and treatment failure.

15 [Slide.]

16 Exclusion criteria for randomization were

17 slightly different for the two studies. In study

18 310, all enrolled patients went on to randomization

19 at month 3 with the following exceptions. Patients

20 were excluded from randomization if they had a

21 Banff grade III acute rejection or vascular

22 rejection during the preceding four weeks.

23 Patients were excluded if they were

24 dialysis-dependent at the time of randomization or

25 had a serum creatinine in excess of 4.5 milligrams

29

1 per deciliter. Finally, patients were excluded if,

2 in the opinion of the study investigator, they had

3 the inadequate renal function to continue in the

4 trial.

5 For study 212, all enrolled patients were

6 randomized at days 2 through 7 with the following

7 exceptions. Patients were not randomized if, in

8 the opinion of the investigator, they had

9 inadequate renal function within the first 48 hours

10 following transplantation or had ongoing acute

11 tubular necrosis or delayed graft function

12 persisting at day 7 post transplant.

13 [Slide.]

14 In total, studies 310 and 212 included 771

15 patients. Of the 525 patients enrolled in study

16 310, 215 were randomized to the Rapamune plus

17 cyclosporine group and 215 were randomized to the

18 Rapamune group. 95 patients were not eligible for

19 randomization. In study 212, 246 patients were

20 enrolled and 97 were randomized to the cyclosporine

21 plus Rapamune group and 100 were randomly assigned

22 to the Rapamune group. 49 patients were not

23 eligible for randomization. However, in study 212,

24 the nonrandomized patients were permitted to

25 receive Rapamune at a dose of up to 5 milligrams

30

1 per day along with cyclosporine. These patients

2 continued to have follow up through month 12.

3 Note the color scheme used in this slide

4 and throughout the remainder of the presentation.

5 The Rapamune plus cyclosporine group is shown in

6 red and the Rapamune group is depicted in purple.

7 [Slide.]

8 In study 310, a total of 525 patients were

9 enrolled and were administered a regimen consisting

10 of a single loading dose of 6 milligrams of

11 Rapamune followed by a fixed dose of 2 milligrams

12 per day. Cyclosporine was coadministered to

13 maintain trough concentrations of 200 to 400

14 nanograms per ml for the first month followed by a

15 gradual reduction through month 3.

16 At month 3, patients were randomly

17 assigned to one of two treatment groups. 215

18 patients were randomly assigned to the Rapamune

19 plus cyclosporine group. Patients in this group

20 continued to receive fixed doses of Rapamune at 2

21 milligrams per day. Cyclosporine was gradually

22 tapered for the specified ranges for the duration

23 of the study period.

24 215 patients were also randomly assigned

25 to the Rapamune group. This group of patients

31

1 received doses of Rapamune to maintain a sirolimus

2 trough concentration range of 20 to 30 nanograms

3 per ml from the time of randomization through the

4 end of month 12. Thereafter, sirolimus trough

5 concentrations remained at 15 to 25 nanograms per

6 ml for the duration of the study.

7 After randomization, patients had the dose

8 of cyclosporine tapered by 25 percent per week and

9 cyclosporine was to be completely eliminated from

10 the regimen within four weeks time. Patients in

11 both randomized groups received standard tapering

12 doses of corticosteroids.

13 [Slide.]

14 In study 212, 246 patients were randomly

15 assigned to one of the two treatment groups. 97

16 were randomly assigned to the Rapamune plus

17 cyclosporine group. Patients in this group were

18 administered a regimen consisting of a single

19 loading dose of Rapamune followed by a fixed dose

20 of 2 milligrams per day.

21 Cyclosporine was coadministered to

22 maintain trough concentration ranges of 200 to 400

23 nanograms per milligram for the first month and was

24 gradually tapered to the specified ranges for the

25 duration of the treatment period. 100 patients

32

1 were assigned to the Rapamune group. The patients

2 in this group were administered a regimen

3 consisting of fixed doses of Rapamune at 20

4 milligrams daily for the first three days followed

5 by 10 milligrams daily through day 10.

6 Thereafter, sirolimus trough

7 concentrations were maintained at a target range of

8 10 to 20 nanograms per milligram for the duration

9 of the study period. Patients also continued to

10 receive reduced doses of cyclosporine for the first

11 month after randomization at a concentration range

12 of 100 to 175 nanograms per milligram and were then

13 tapered down to 100 to 150 nanograms per milligram

14 through month 2.

15 The dose of cyclosporine was further

16 tapered by 25 percent per week and cyclosporine was

17 to be completely eliminated from the regimen by the

18 end of month 3. The patients in this study also

19 received standard tapering doses of

20 corticosteroids.

21 [Slide.]

22 It is important to note that the efficacy

23 and safety data from studies 310 and 212 were

24 deliberately not integrated. The designs of the

25 two studies, while similar, were distinct in

33

1 several important features. Time of randomization

2 differed. Study 310 allowed us to maximize the

3 opportunity to compare like patients at the onset

4 of cyclosporine withdrawal.

5 Different target sirolimus and

6 cyclosporine trough concentrations were also

7 utilized in the two studies. Complete safety and

8 efficacy data through 12 months will be presented

9 for both studies. For study 310, cumulative safety

10 data are presented for all patients through

11 month 15 with limited data being available through

12 month 24.

13 Efficacy Review

14 [Slide.]

15 DR. NEYLAN: The efficacy comparisons in

16 each study will be now be reviewed.i

17 [Slide.]

18 This slide shows the similar distribution

19 of key demographic variables among patients

20 enrolled in study 310. Comparing the features of

21 all enrolled patients to that of the randomized

22 groups shows only a slightly higher rate of delayed

23 graft function, shown here.

24 The groups were otherwise well matched for

25 gender, ethnic origin, age, receipt of a first or

34

1 second allograft, ischemia time and degree of HLA

2 mismatch. When compared to the UNOS database, the

3 race disparity is obvious.

4 But other features are similar including

5 rates of delayed graft function in the study groups

6 that were slightly greater than that of the U.S.

7 renal transplant population. Though not shown on

8 this slide, there were also no differences observed

9 in donor characteristics including donor source,

10 ethnic origin or age.

11 [Slide.]

12 The intent-to-treat analysis of the

13 primary efficacy endpoint for study 310, graft

14 survival at twelve months, is shown here with a 95

15 percent confidence interval of the differences in

16 rates. The twelve-month graft survival was

17 equivalent and excellent in both groups. Rates

18 were high in excess of 95 percent in both cohorts.

19 There were similar rates of physical and

20 functional graft loss as well as graft loss

21 secondary to patient death. Note also that there

22 was 100 percent follow up for patients in both

23 randomized groups.

24 [Slide.]

25 Similarly, patient survival in the

35

1 intent-to-treat population was equivalent at twelve

2 months following transplantation. The survival

3 rate exceeded 97 percent in both groups.

4 [Slide.]

5 This Kaplan-Meier plot shows the incidence

6 of first-biopsy-confirmed acute-rejection episodes

7 in study 310. In the prerandomization period,

8 before month 3, there were similar rates of acute

9 rejection for all enrolled patients. For month 3

10 through 12, there was an incremental increase in

11 rejection frequency in the Rapamune arm. The

12 combined incidence of acute rejection over the

13 first twelve months was not statistically different

14 for both randomized groups, 13.5 percent for the

15 Rapamune plus cyclosporine group compared with 20

16 percent for the Rapamune group.

17 [Slide.]

18 How does the acute-rejection rate compare

19 with other registration trials? The initial

20 therapy provided low acute-rejection rates which

21 meet the standards for immunosuppressive therapy

22 for today's transplant recipient. Specifically,

23 the use of Rapamune in combination with

24 cyclosporine was associated with the rejection rate

25 of only 12 percent for the entire enrolled

36

1 population of 525 patients.

2 These rejection rates compare favorably

3 with recently published registration trials.

4 [Slide.]

5 At twelve months, acute-rejection rates in

6 all enrolled patients, not just those randomized to

7 the two treatment arms, were again equal to or

8 better than recently published registration trials

9 in which calcineurin inhibitors were included and

10 maintained in the regimen.

11 [Slide.]

12 Following month 3 and the onset of

13 cyclosporine elimination, the incremental increase

14 in first biopsy-confirmed rejection was modest at

15 9.8 percent but was significantly higher than the

16 rejection rate in the control arm at 4.2 percent.

17 Even though the rejection rates were low,

18 an important question to ask is whether outcomes

19 for those patients who had rejection episodes were

20 worse than would be expected. Importantly, for

21 patients experiencing rejection in either treatment

22 arm, there was a single death in the Rapamune plus

23 cyclosporine group and no deaths in the Rapamune

24 group.

25 Additionally, there was only one graft

37

1 loss in each group.

2 [Slide.]

3 The histologic severity of acute-rejection

4 episodes was similar in the two groups. The

5 majority of these episodes were mild and no patient

6 experienced an episode of severe acute rejection

7 following cyclosporine elimination. The use of

8 antibody therapy to treat acute rejection was also

9 similar and was utilized in only two patients.

10 [Slide.]

11 Another important variable in assessing

12 the impact of acute rejection is the potential

13 effect on subsequent graft function. This analysis

14 compares the change in glomerular-filtration rate

15 from baseline to twelve months in randomized

16 patients who subsequently did or did not experience

17 an acute-rejection episode.

18 On the left, patients without acute

19 rejection had experienced a change in renal

20 function at twelve months consistent with the study

21 as a whole. Specifically, function improved in

22 patients in the Rapamune arm while it worsened for

23 patients maintained in the Rapamune plus

24 cyclosporine group.

25 On the right are depicted patients with

38

1 acute rejections after month 3. As might be

2 expected, the GFR at twelve months was numerically

3 lower for patients in either group who had

4 experienced an episode of acute rejection.

5 However, the GFR for rejectors in the Rapamune

6 group remained stable through twelve months. This

7 stability suggests that the adverse impact of acute

8 rejection upon renal function appeared to be

9 lessened with the elimination of cyclosporine.

10 [Slide.]

11 The combination of Rapamune plus

12 cyclosporine in the first three months following

13 transplantation maintained very low rejection rates

14 which were equal to or better than those observed

15 in recent registration trials. The incremental

16 increase in acute rejection following cyclosporine

17 elimination was statistically higher in the

18 Rapamune group with an absolute difference of 6

19 percent.

20 This compares favorably with previous

21 trials in which rates of rejection following

22 elimination are equal to or greater than those

23 observed in study 310. Episodes of rejection

24 attending cyclosporine elimination were generally

25 mild and clinically manageable. Importantly, there

39

1 were no episodes of severe rejection and only one

2 graft loss was reported in the Rapamune group.

3 In addition, at twelve months, there were

4 similar rates of acute rejection in the randomized

5 groups. As expected, at twelve months, the mean

6 GFRs in the rejectors were lower than those in the

7 nonrejectors. But, importantly, there was no

8 penalty in patients in whom cyclosporine was

9 eliminated.

10 [Slide.]

11 Comparable rates of efficacy failure were

12 demonstrated. These composite rates at twelve

13 months following transplantation were primarily due

14 to the occurrence of acute rejections with very few

15 graft losses or patient deaths.

16 [Slide.]

17 Treatment failure for study 310 was

18 defined as the first occurrence of rejection, graft

19 loss, death or discontinuation of study medication.

20 The overall treatment failure at twelve months was

21 significantly higher with patients randomized to

22 the Rapamune group. This was primarily due to the

23 numerically higher rates of acute rejection and for

24 discontinuations within the group.

25 On review of the clinical dataset, the

40

1 difference in the rate of treatment failure was no

2 longer statistically significant.

3 Now let's examine what many would consider

4 to be the most important efficacy endpoint in a

5 study of cyclosporine elimination, namely the

6 impact upon long-term graft function.

7 [Slide.]

8 Shown here is the intent-to-treat analysis

9 of serum creatinine and glomerular-filtration rate

10 for patients enrolled in study 310. This

11 conservative analysis includes all enrolled

12 patients including those discontinued from therapy

13 and placed back on calcineurin inhibitors. For

14 both renal-function parameters, there was a

15 statistically significant improvement demonstrated

16 at the twelve-month time point for the Rapamune

17 group.

18 [Slide.]

19 In addition to the intent-to-treat

20 analysis demonstrating excellent patient and graft

21 survival and statistically significant improvements

22 in renal function, the on-therapy analysis also

23 showed a clear benefit for patients in whom

24 cyclosporine was eliminated and who were maintained

25 on concentration-controlled Rapamune.

41

1 This group included patients who may have

2 experienced an episode of acute rejection but

3 continued within the study and received study

4 medication. The graph on the left shows serum

5 creatinine. In the Rapamune treatment group, serum

6 creatinine was significantly lower at all time

7 points following randomization. It is also

8 noteworthy that this improvement is sustained

9 through 24 months of follow up.

10 The graph on the right shows calculated

11 glomerular-filtration rates at these same time

12 points. Again, the Rapamune-treated group had

13 significantly higher GFRs at all time points

14 persisting through month 24.

15 [Slide.]

16 The benefits of cyclosporine elimination

17 on renal function were demonstrated by all patients

18 on therapy through twelve months and longer

19 regardless of their baseline renal function.

20 A quartile analysis was performed in which

21 patients were segregated according to baseline

22 renal function at the time of randomization. In

23 all four quartiles, the change from baseline was

24 favorable in comparison to patients maintained on

25 cyclosporine including those with more advanced

42

1 degrees of renal insufficiency at baseline.

2 Notably, even those patients with normal

3 renal function at baseline benefitted by the

4 removal of cyclosporine nephrotoxicity and its

5 consequent negative impact upon long-term renal

6 function.

7 [Slide.]

8 In summary, the patients enrolled in study

9 310 were similar to that of the U.S. population

10 with the exception of fewer black patients. At

11 twelve months, following transplantation, there was

12 equivalent patient and graft survival of greater

13 than 97 percent and 95 percent, respectively. In

14 addition, a low incidence of acute rejection at

15 twelve months was similar in the two randomized

16 groups and, perhaps most importantly, there was an

17 immediate improvement in renal function following

18 cyclosporine elimination which has been sustained

19 through 24 months of follow up.

20 Next, we will review the key efficacy data

21 for study 212.

22 [Slide.]

23 Key demographic variables among patients

24 enrolled in study 212 were similar. The total

25 enrolled patient population is similar to that of

43

1 the two randomized groups. These were well matched

2 for gender, ethnic origin, age, ischemia time and

3 degree of HLA mismatch.

4 The demographics are also similar to that

5 of the UNOS population of renal-transplant

6 recipients in the U.S. except for the study's

7 exclusion of living donor recipients. Therefore,

8 while study 212 is generally representative of the

9 U.S. renal-transplant population, the 212 group was

10 also at a somewhat higher risk given the absence of

11 living-donor recipients.

12 Though not shown on this slide, the

13 patients in both groups had similar donor

14 characteristics including source, ethnic origin and

15 age.

16 [Slide.]

17 Twelve-month graft survival in study 212

18 was similar in the two treatment groups being in

19 excess of 92 percent in both. There was a slightly

20 higher rate of graft loss due to physical or

21 functional graft loss in the Rapamune plus

22 cyclosporine group compared with the Rapamune

23 group. Again, as with study 212, there was 100

24 percent patient follow up in both randomized

25 groups.

44

1 [Slide.]

2 The intent-to-treat analysis of patient

3 survival in study 212 was similar. At twelve

4 months, patient survival was excellent and was at

5 least 96 percent on both groups.

6 [Slide.]

7 This Kaplan-Meier plot shows the incidence

8 of first biopsy-confirmed acute-rejection episodes

9 in study 212. Prior to cyclosporine withdrawal,

10 there were similar rates of acute rejection in both

11 groups. Following month 2, there was an

12 incremental increase in the rate of acute rejection

13 in the Rapamune group but the difference between

14 the randomized groups never achieved statistical

15 significance.

16 The intent-to-treat analysis at month 12

17 demonstrated an incidence of acute rejection of

18 18.6 percent for the Rapamune plus cyclosporine

19 group compared with 22 percent for the

20 Rapamune-treated group.

21 As in study 310, it is important to

22 examine the outcome in those patients who

23 experienced acute rejection following the

24 elimination of cyclosporine. Following month 2,

25 there was a modest numerical increase in

45

1 first-biopsy-confirmed rejections at 14 percent

2 compared with the rejection rate in the control arm

3 of 6.2 percent.

4 Importantly, for patients experiencing

5 rejection in either treatment arm, there was a

6 single death and a single graft loss in the

7 Rapamune group and no deaths or graft losses in the

8 Rapamune plus cyclosporine group.

9 [Slide.]

10 As with study 310, the histologic severity

11 of acute-rejection episodes was similar in the two

12 randomized groups. The majority of these episodes

13 were mild to moderate with only one patient in the

14 Rapamune plus cyclosporine group experiencing an

15 episode of severe acute rejection beyond the two

16 month time point.

17 [Slide.]

18 This analysis compares the calculated GFR

19 in patients who did or did not experience an

20 acute-rejection episode following month 2 and the

21 onset of cyclosporine elimination.

22 On the left, patients without acute

23 rejection had experienced a change in renal

24 function at twelve months consistent with the study

25 as a whole. Specifically, function improved in

46

1 patients in the Rapamune arm.

2 On the right are depicted patients with

3 acute rejections after month 2. As might be

4 expected, the GFR at twelve months were numerically

5 lower than nonrejectors for both groups. These

6 findings are consistent with study 310 and suggest

7 that renal function outcomes for those patients who

8 had rejection episodes were within clinical

9 expectations.

10 [Slide.]

11 Importantly, study 212 was also consistent

12 with study 310 in demonstrating improved renal

13 function in a variety of comparative analyses.

14 Depicted here is the intent-to-treat analysis. The

15 intent-to-treat population includes all enrolled

16 patients including those who experienced an episode

17 of acute rejection or had discontinued study

18 medication.

19 In this group, calculated GFRs were

20 significantly higher at six months and at twelve

21 months in the Rapamune-treated patients.

22 [Slide.]

23 Study 212 demonstrated improved renal

24 function in the primary efficacy population, namely

25 those patients that remained on therapy and

47

1 rejection-free through month 6. The graph on the

2 left shows serum creatinine compared with the

3 Rapamune plus cyclosporine treated patients,

4 Rapamune treated patients had significantly lower

5 serum creatinines starting at month 6 and

6 persisting through month 12.

7 The graph on the right shows calculated

8 GFRs at these same time points. The Rapamune

9 group, again, had significantly higher GFRs at

10 month 6 compared to the Rapamune plus cyclosporine

11 group and this difference persisted through twelve

12 months.

13 [Slide.]

14 There was also improvement observed in

15 directly measured GFRs in a subset of the primary

16 analysis population. Patients in the Rapamune

17 group with cyclosporine elimination had higher

18 measured GFRs at both six and twelve months

19 following transplantation.

20 [Slide.]

21 Improved renal function was also

22 demonstrated in the on-therapy population. This

23 group included patients who may have experienced an

24 episode of acute rejection but continued within the

25 study and received study medication. The graph on

48

1 the left shows serum creatinine. Compared with

2 Rapamune plus cyclosporine treated patients, there

3 was a trend toward lower serum creatinine at all

4 time points in the Rapamune-treated cohort. At

5 twelve months, the improvement in creatinine

6 demonstrated statistical significance.

7 The graph on the right shows calculated

8 GFRs at these same time points. Notably, GFRs were

9 significantly higher at time point 6, nine and

10 twelve months in comparison to the control group.

11 [Slide.]

12 As in study 310, the benefits of

13 cyclosporine elimination on renal function were

14 demonstrated by the majority of patients on therapy

15 through twelve months regardless of their baseline

16 renal function. Again, a quartile analysis was

17 performed in which patients were segregated

18 according to baseline renal function just prior to

19 cyclosporine elimination.

20 The change from baseline was favorable in

21 comparison to patients maintained on cyclosporine.

22 As might be expected, patients with varying degrees

23 of renal dysfunction also showed improvement.

24 [Slide.]

25 In summary, at month 12, studies 310 and

49

1 212 are consistent in their findings.

2 Specifically, these studies demonstrated that

3 following the elimination of cyclosporine,

4 concentration-controlled Rapamune maintenance

5 therapy results in the following: equivalent graft

6 survival of 95 to 97 percent, equivalent patient

7 survival of 96 to 98 percent, an incremental

8 increase in mild to moderate acute-rejection

9 episodes following cyclosporine elimination with an

10 absolute difference of 6 to 8 percent versus

11 controlled therapy.

12 This compares favorably with previous

13 elimination trials and, perhaps most importantly,

14 both studies demonstrated an immediate and

15 sustained improvement in renal function.

16 This concludes my presentation of the

17 efficacy data for studies 310 and 212.

18 Safety Data

19 DR. NEYLAN: I will now review the safety

20 data for both studies.

21 [Slide.]

22 One-year data will be shown for graft

23 loss, patient death and discontinuation from study

24 medication. The cumulative safety experience for

25 all enrolled patients will be shown for adverse

50

1 events including infection and malignancy. The

2 cumulative on-therapy data will be presented for

3 all laboratory parameters including blood pressure.

4 [Slide.]

5 The safety assessments will be reviewed in

6 different categories including etiologies of graft

7 loss in patient death, adverse events including

8 those related to immunosuppression such as

9 infection and malignancy and, finally,

10 blood-pressure measurements and laboratory

11 parameters.

12 [Slide.]

13 I have already shown you graft survival

14 data for the randomized patients. Graft survival

15 in the randomized groups was in excess of 95

16 percent. An analysis of overall graft survival for

17 all patients enrolled in the study was also high at

18 approximately 89 percent. This group included

19 patients with severe acute or vascular rejection,

20 sustained delayed graft function and other criteria

21 that precluded randomization.

22 [Slide.]

23 The causes of graft loss in study 310 are

24 shown in this slide. An intent-to-treat comparison

25 of the randomized cohorts was conducted censoring

51

1 graft loss secondary to death. These data revealed

2 similar incidences of graft loss due to infection,

3 renal fibrosis, renal dysfunction, graft vascular

4 thrombosis or recurrent primary disease.

5 The causes of graft loss in these two

6 groups were not statistically different.

7 [Slide.]

8 This slide includes patient survival for

9 all patients enrolled in the study. Patient

10 survival in the overall population which includes

11 the nonrandomized patients was in excess of 94

12 percent.

13 [Slide.]

14 The causes of patient death are shown

15 here. An intent-to-treat analysis at twelve months

16 demonstrated no significant differences in death

17 due to cardiovascular cause or infection.

18 [Slide.]

19 Next we will review the adverse-event data

20 including those events generally associated with

21 immunosuppressive therapy such as infection and

22 malignancy.

23 [Slide.]

24 The adverse events for this study were

25 similar to the safety profile observed in

52

1 previously completed pivotal trials that supported

2 the initial approval of Rapamune. What I want to

3 focus on are changes in the profile when increased

4 doses of Rapamune are utilized after cyclosporine

5 elimination.

6 As is common in all renal transplant

7 clinical trials, there were a number of reports of

8 adverse events in study 310. These data represent

9 new adverse events occurring following

10 randomization. Shown are the statistically

11 significant differences observed between the two

12 groups.

13 Statistically higher in the Rapamune plus

14 cyclosporine group were cyclosporine toxicity,

15 increased creatinine, edema, hypertension and

16 hyperuricemia. Significantly higher in the

17 Rapamune group were hypokalemia, elevated SGOT and

18 SGPT and thrombocytopenia.

19 [Slide.]

20 All patients in study 310 were followed

21 for the occurrence of serious infections including

22 those requiring hospitalization. In general, the

23 results show no difference in infections in the two

24 randomized groups and are consistent with the known

25 safety profile. The only significant difference is

53

1 an increased reporting of Herpes zoster infection

2 in the patients in the Rapamune plus cyclosporine

3 group. There was no difference in the incidence of

4 sepsis, CMV infection, pneumonia, Herpes simplex or

5 urinary-tract infection.

6 [Slide.]

7 Similarly, there was no statistical

8 difference in the reported incidence of neoplasia.

9 Specifically, the rates of skin cancer, lymphoma,

10 leukemia and other malignancies were similar and

11 not different between the randomized groups. The

12 overall rates of reporting in this study were also

13 consistent with numerous other studies in which

14 transplant recipients received similar levels of

15 immunosuppression.

16 [Slide.]

17 The next safety parameter I would like to

18 discuss is that of blood pressure. Hypertension is

19 common in renal-transplant recipients and an

20 important contributor to cardiovascular risk. In

21 the next two slides, we will review blood-pressure

22 measurements as well as the percentage of patients

23 requiring antihypertensive medications in this

24 study.

25 [Slide.]

54

1 The mean systolic and diastolic blood

2 pressures are shown here. On the left, are shown

3 mean systolic blood-pressure measurements.

4 Compared with the Rapamune plus cyclosporine group,

5 Rapamune-treated patients had significantly lower

6 systolic blood pressures at all time points

7 starting at month 6 and persisting through 24

8 months of follow up.

9 On the right are mean diastolic

10 blood-pressure measurements. Similarly,

11 statistically significantly lower diastolic

12 blood-pressure measurements were observed from

13 month 6 through 18 for Rapamune-treated patients.

14 [Slide.]

15 It is important to consider the need for

16 antihypertensive agents in these patients.

17 Although the study was not designed to capture

18 specific dosages of antihypertensive medications,

19 it was possible to analyze the need for combination

20 regimens. The cumulative requirement for multidrug

21 antihypertensive therapy was less in the Rapamune

22 group at month 12. This difference was

23 statistically significant.

24 Thus, the improvement in blood-pressure

25 management demonstrated by the lowering of systolic

55

1 and diastolic means was also attended by a

2 decreased need for multidrug therapy.

3 [Slide.]

4 We will next review several laboratory

5 parameters. The first analysis will address the

6 issue of lipid elevations, an important risk factor

7 in renal-transplant recipients.

8 [Slide.]

9 In study 310, approximate 19 percent of

10 the patients were receiving lipid-lowering

11 medications prior to transplantation including

12 statins and/or fibrates. Following initiation of

13 study medication, 73 percent of patients in both

14 randomized groups were receiving statins while up

15 to 25 percent of patients in both groups were

16 administered fibrates. The overall use of these

17 agents in both randomized groups was similar.

18 [Slide.]

19 An observation made early in the clinical

20 program was the effect of Rapamune on cholesterol

21 and triglycerides. In study 310, the median

22 fasting cholesterol concentrations in the two

23 randomized groups were similar at month 12.

24 The range of values is depicted in these

25 box-and-whisker plots. 80 percent of the patients

56

1 in each treatment group are contained within the

2 respective box-and-whisker plots. Thus, the

3 majority of patients were found to have cholesterol

4 values at or below 250 milligrams per deciliter

5 despite the fact that concentration-controlled

6 Rapamune-treated patients had increase sirolimus

7 trough levels as mandated by protocol.

8 The results observed in study 212 were

9 similar.

10 [Slide.]

11 Measurements of fasting HDL and LDL

12 cholesterol levels were also similar. For HDL

13 cholesterol, the two randomized groups were similar

14 except at month 18 when there was a statistically

15 significant increase in the Rapamune group. LDL

16 cholesterol, calculated for those patients who had

17 triglycerides below 400 milligrams per deciliter

18 was similar in the two randomized groups with the

19 exception of month 3 when there was a significant

20 increase in the Rapamune group.

21 [Slide.]

22 As with serum cholesterol, fasting

23 triglycerides were similar in study 310 in the two

24 randomized groups through twelve months of follow

25 up. Again, despite the higher sirolimus

57

1 concentrations, the Rapamune-treated patients

2 maintained fasting triglycerides in the majority of

3 patients within the 150 to 250 milligram per

4 deciliter range. The results observed in study 212

5 were similar.

6 [Slide.]

7 With regard to liver-function tests, SGPT

8 and SGOT were measured at various time intervals.

9 In the Rapamune-treated patients, SGPT was

10 significantly higher for months 12 through 24.

11 SGOT was significantly higher for months 12 through

12 18. At all other time points, these liver enzymes

13 remained similar in the two randomized groups and

14 below the upper limits of normal.

15 In study 212, the majority of patients

16 also had transaminase levels below the upper limits

17 of normal.

18 [Slide.]

19 Shown on this slide are the causes of

20 elevated liver enzymes in a small number of

21 patients with at least one SGPT value greater than

22 five times the upper limit of normal.

23 Approximately 50 percent of these patients had an

24 infectious etiology as a potential cause for the

25 SGPT elevation.

58

1 [Slide.]

2 The effects of Rapamune on

3 bone-marrow-derived cells are consistent with its

4 biologic activity in that small decreases in

5 platelets, red cells and leukocytes have been

6 observed. Most important, however, is that there

7 is no evidence of chronic or irreversible

8 bone-marrow dysfunction or depression.

9 In general, white blood-cell counts were

10 similar in study 310 with the exception of

11 statistically significant differences noted at

12 months 3 and 6. However, it is important to note

13 that the mean white-blood-cell counts remained

14 within a clinically normal range for all of the

15 patients.

16 Platelet counts for the two randomized

17 groups were also similar. While statistically

18 significant differences were observed at months 6,

19 15 and 18, mean platelet counts remained above

20 200,000 at all time points. It is also important

21 to note that platelet counts remained stable as

22 patients continued to receive Rapamune through

23 month 24.

24 Similar results were observed in study 212.

25 [Slide.]

59

1 In summary, in study 310, there was

2 equivalent patient and graft survival. In the

3 Rapamune plus cyclosporine group, there was an

4 increased incidence of cyclosporine toxicity,

5 increased creatinine, edema, hypertension and

6 hyperuricemia.

7 In the Rapamune group, there was an

8 increased incidence of hypokalemia, increased SGOT,

9 SGPT and thrombocytopenia. There were similar

10 rates of infection and malignancy. Improved blood

11 pressure followed cyclosporine elimination and

12 there were similar effects on lipid profiles and

13 hematologic parameters despite the higher

14 trough-level concentrations in the Rapamune group

15 following cyclosporine elimination.

16 [Slide.]

17 I will now review the safety data for

18 study 212. This slide includes graft survival for

19 all patients enrolled in the study. As previously

20 demonstrated, similar rates were observed in the

21 randomized group. The nonrandomized group

22 demonstrated a lower graft-survival rate not

23 inconsistent with that typically observed in

24 patients with ATN or delayed graft function.

25 [Slide.]

60

1 Causes of graft loss in this study are

2 shown here. An intent-to-treat comparison of the

3 randomized cohorts was conducted censoring graft

4 loss secondary to patient death. The data revealed

5 a similar incidence of graft loss due to rejection,

6 acute tubular necrosis and hemolytic uremic

7 syndrome.

8 [Slide.]

9 As previous presented, similar patient

10 survival was observed in the two randomized groups.

11 Patient survival in the nonrandomized group was

12 slightly lower.

13 [Slide.]

14 Causes of patient death in study 212 are

15 shown here. Analysis at twelve months following

16 transplantation demonstrated no significant

17 differences in death due to cardiovascular cause,

18 infection or pulmonary edema.

19 [Slide.]

20 Similar to study 310, there were a number

21 of reports of adverse events in study 212. Again,

22 I will primarily be emphasizing the statistically

23 significant differences. Significantly higher in

24 the Rapamune plus cyclosporine were hypertension,

25 dyspnea, edema, hypervolemia and hypomagnesemia.

61

1 Significantly higher in the Rapamune group

2 were thrombocytopenia, hypokalemia, diarrhea,

3 abnormal liver-function tests and atrial

4 fibrillation. With the exception of atrial

5 fibrillation, these types of adverse events were

6 previously observed in the pivotal clinical trials.

7 The increased incidence of atrial

8 fibrillation in the Rapamune group is discussed in

9 more detail in the next slide.

10 [Slide.]

11 In study 212, atrial fibrillation occurred

12 in a total of nine patients. This included one

13 patient in the Rapamune plus cyclosporine group and

14 an additional eight patients in the Rapamune group.

15 Six of these eight patients had episodes of atrial

16 fibrillation occurring within the first 40 days

17 following transplantation and thus prior to the

18 elimination of cyclosporine.

19 All cases resolved promptly with therapy

20 and, in the opinion of the investigators, none were

21 considered related to study medication.

22 In the larger study, 310, the incidence of

23 atrial fibrillation was 1.9 percent in the

24 cyclosporine-plus-Rapamune group compared with 3.7

25 percent in the Rapamune group. This difference was

62

1 not statistically significant. Likewise, in

2 previous registration trials, atrial fibrillation

3 was uncommon and not statistical different from

4 controlled therapies.

5 [Slide.]

6 The intent-to-treat analysis of infections

7 in study 212 is listed here. Infections were

8 typical of the general renal-transplant population

9 and the data showed no statistical difference

10 between the two randomized groups.

11 [Slide.]

12 As with study 310, the overall rates of

13 malignancy observed in 212 were also similar and

14 consistent with previously published studies in

15 transplant recipients. By twelve months, a

16 comparison of the two randomized groups showed no

17 difference in the rates of nonmelanomtous skin

18 cancer and one case of presumed post-transplant

19 lymphoproliferative disease. There was one case of

20 renal-cell carcinoma in a native kidney.

21 [Slide.]

22 In summary, in study 212, there was

23 equivalent patient and graft survival. In the

24 Rapamune plus cyclosporine group, there was an

25 increased incidence of hypertension, dyspnea,

63

1 edema, hypervolemia and hypomagnesemia. In the

2 Rapamune group, there was an increased incidence of

3 thrombocytopenia, hypokalemia, diarrhea, increased

4 SGOT, SGPT and atrial fibrillation.

5 The infrequent observation of atrial

6 fibrillation was not considered by study

7 investigators to be related to Rapamune. There

8 were similar rates of infection and malignancy and

9 there were similar effects on lipid profiles and

10 hematologic parameters despite the higher

11 trough-level concentrations in the Rapamune group

12 following cyclosporine elimination.

13 To compete the overall safety profile, the

14 next several slides will review patient outcomes in

15 those patients discontinued from treatment as well

16 as the overall success of cyclosporine elimination.

17 [Slide.]

18 The overall disposition of patients in

19 study 310 is shown in this slide. As previously

20 discussed, 525 patients were enrolled at the time

21 of transplantation. 430 patients met the

22 predetermined eligibility criteria at month 3 and

23 were randomly assigned to one of the two treatment

24 groups.

25 215 patients were assigned to each of the

64

1 groups. the overall rates of discontinuation in

2 study 310 were similar to those observed in recent

3 immunosuppressive registration trials. 18.1

4 percent of patients had discontinued by month 3 and

5 36.4 percent of patients had discontinued by month

6 12.

7 [Slide.]

8 The reasons for discontinuation in study

9 310 are listed here. A total of 95 patients, or

10 18.1 percent of the total population, were not

11 randomized and were discontinued due to a variety

12 of causes typical for patients in this early period

13 following transplantation.

14 74 percent were discontinued for adverse

15 events including infections, renal dysfunction,

16 surgical complications, laboratory abnormalities

17 and a small number of miscellaneous causes. 13

18 percent of these patients were discontinued because

19 of the acute rejection.

20 Following randomization by month 12, the

21 overall rate of discontinuation was higher in the

22 Rapamune group. Acute rejection was an infrequent

23 cause of discontinuation accounting for only 2

24 percent and 5 percent in the Rapamune plus

25 cyclosporine and the Rapamune groups, respectively.

65

1 Upon review of the cumulative dataset

2 which includes data for all patients at or beyond

3 15 months, the difference in the rate of

4 discontinuation was no longer statistically

5 significant.

6 [Slide.]

7 While the reasons for patient

8 discontinuations for the study as a whole were

9 similar to other immunosuppressive trials, it is

10 important to look at the special group of patients

11 in whom cyclosporine elimination was not or could

12 not be successfully completed.

13 Given the present availability of other

14 immunosuppressive agents, clinicians were able to

15 choose from a variety of alternative regimens for

16 these patients. Most patients remained on

17 corticosteroids plus a calcineurin inhibitor and,

18 in 26 percent of these cases, patients were

19 converted from cyclosporine to tacrolimus.

20 In many of the cases, an antimetabolite

21 was also added to the regimen. It is notable that

22 in 19 percent of these cases, Rapamune was

23 maintained while the calcineurin inhibitor was

24 reintroduced.

25 Three deaths and two graft losses occurred

66

1 in the discontinued group. By month 12, there were

2 no acute rejections reported in patients converting

3 to alternative therapies.

4 [Slide.]

5 In the majority of patients randomized to

6 the Rapamune group, cyclosporine elimination was

7 successful. 50 percent of these patients

8 accomplished this within the first 42 days and 90

9 percent were cyclosporine free by day 72 post

10 randomization. In total, 92.6 percent of the

11 patients were successfully withdrawn from

12 cyclosporine.

13 [Slide.]

14 The overall disposition of patients in

15 study 212 is shown in this slide. A total of 246

16 patients were enrolled at the time of transplant

17 and randomly assigned to one of the two treatment

18 groups. 97 patients were assigned to receive

19 Rapamune plus cyclosporine and 100 to Rapamune.

20 The overall rate of discontinuation in

21 study 212 was similar to that observed in other

22 recent immunosuppressive registration trials with

23 29.7 percent of patients discontinued by month 12.

24 In the following slides, we will review

25 the outcomes for these discontinued patients.

67

1 [Slide.]

2 The reasons for discontinuation in study

3 212 are listed here. A total of 49 patients were

4 not randomized. Of these, 28 discontinued due to

5 adverse events, acute rejection or other causes.

6 Post randomization, a total of 45 patients were

7 discontinued from the study by twelve months, 20 of

8 these in the Rapamune plus cyclosporine group and

9 25 in the Rapamune group.

10 These discontinuations were similar in

11 nature to those of study 310. Clinicians

12 participating in study 212 chose to reinitate

13 calcineurin inhibitors for most patients

14 discontinued from the Rapamune group.

15 [Slide.]

16 As in study 310, the majority of patients

17 randomized to the Rapamune group of study 212 had

18 cyclosporine successfully eliminated. On the left

19 is depicted an analysis of all patients randomized

20 to the Rapamune group. 76 percent of patients

21 randomized from the time of transplantation

22 successfully eliminated cyclosporine.

23 On the right is an analysis of these

24 patients who were eligible for cyclosporine

25 elimination at month 2. Note the similar success

68

1 rate to that of study 310 in that 93 percent of

2 these patients successfully had cyclosporine

3 eliminated from the regimen.

4 Thus, in both studies, patients maintained

5 on Rapamune plus cyclosporine for the first two to

6 three months after transplantation emerged from the

7 high-risk period and went on, in 92 to 93 percent

8 of cases, to successfully eliminate cyclosporine.

9 [Slide.]

10 In conclusion, studies 310 and 212 are

11 consistent in confirming the beneficial safety

12 profile of Rapamune-based therapy following

13 cyclosporine elimination. Both studies

14 demonstrated excellent patient and graft survival,

15 similar rates of infection and malignancy and

16 significantly lower rates of several other

17 cyclosporine-related adverse events.

18 In addition, study 310 demonstrated a

19 significant and sustained improvement in blood

20 pressure. Despite the higher concentration of

21 Rapamune required when cyclosporine is eliminated,

22 the overall Rapamune safety profile is similar to

23 that observed when it is administered as a fixed 2

24 milligram dose in combination with cyclosporine.

25 [Slide.]

69

1 In addition, rates of discontinuations in

2 these studies were similar to other

3 immunosuppressive registration trials. The reasons

4 for early discontinuation were typical of those

5 observed in renal allograft recipients including

6 surgical complications and delayed graft function.

7 Very few patients were discontinued due to

8 acute rejection. In fact, in study 310, 70 percent

9 of patients experiencing episodes of acute

10 rejection in the first three months went on to

11 randomization. As expected, various alternative

12 therapies were available for patients discontinued

13 from the studies.

14 Importantly, cyclosporine was successfully

15 eliminated in the great majority of patients in the

16 Rapamune group of both studies.

17 This concludes my presentation of the

18 safety data. At this time, I would like to

19 introduce Dr. James Zimmerman, Senior Director of

20 Clinical Pharmacokinetics at Wyeth-Ayerst who will

21 now review the pharmacokinetics of Rapamune

22 concentration-controlled trials and sirolimus

23 therapeutic drug-level monitoring in this patient

24 population.

25 Dr. Zimmerman?

70

1 Pharmacokinetics

2 DR. ZIMMERMAN: Thank you, John. Good

3 morning.

4 [Slide.]

5 In our original application, Rapamune was

6 approved for a fixed-dose administration without

7 the need for therapeutic drug monitoring or TDM.

8 TDM was recommended in certain patient populations

9 and to compensate for serious pharmacokinetic drug

10 interactions but it was not required. Today we

11 have proposed a new regimen that will require TDM.

12 This new regimen is proposed based on safety and

13 efficacy data from Rapamune

14 concentration-controlled trials that involve

15 cyclosporine elimination in which drug exposure was

16 guided by TDM.

17 [Slide.]

18 My purpose today is to show you data to

19 support the following four points. First, we have

20 a sufficient understanding of sirolimus PK to apply

21 therapeutic drug monitoring to guide treatment in

22 renal-transplant patients. Secondly, we have a

23 robust and reliable assay for sirolimus. Thirdly,

24 the concentration range for sirolimus TDM has been

25 defined and it is effective. Fourth, we have data

71

1 to show that transplant physicians can utilize TDM

2 safely and efficaciously in post-transplant

3 patients.

4 Now, before belaboring on these four

5 points, I want to remind you of the conditions

6 under which Rapamune is administered by fixed dose

7 in concentration-controlled regimens.

8 [Slide.]

9 The currently approved Rapamune regimen is

10 a fixed-dose regimen which was based on the

11 administration of Rapamune four hours after a oral

12 formulation of cyclosporine. The fixed-dose

13 regimen is recommended for most patients during

14 coadministration with cyclosporine.

15 [Slide.]

16 Concentration-controlled Rapamune

17 administration is recommended during administration

18 with cyclosporine under certain conditions; in

19 pediatric patients, in hepatic impairment, during

20 administration with strong inducers or inhibitors

21 or the CYP3A P450 subfamily and P-glycoprotein and

22 also after marked changes in cyclosporine doses.

23 Concentration control is required when

24 administered without cyclosporine and it is the

25 method of dose administration for the current

72

1 indication.

2 [Slide.]

3 Let me start with the assay methodology.

4 Whole-blood sirolimus concentrations were measured

5 during phase II and phase III clinical trials using

6 an immunoassay or a chromatographic assay as we can

7 see by the first two columns.

8 However, as shown in the third column, the

9 immunoassay is not currently available for

10 post-approval use. Instead, HPLC/UV or HPLC/MS/MS

11 are being used at local and commercial

12 laboratories. It is important to realize that the

13 two assays provide different numerical values for

14 sample analysis as shown in the column on the

15 extreme right.

16 For example, chromatographic assay values

17 are 20 percent lower than the immunoassay values.

18 Consequently, the ranges for therapeutic drug

19 monitoring are different for the two assays. In

20 this presentation, sirolimus concentrations are

21 expressed in terms of the immunoassay since the

22 vast majority of the samples for pivotal phase III

23 trials were measured by this method.

24 Turning now to the impact of sirolimus PK

25 on TDM.

73

1 [Slide.]

2 The fact that sirolimus exhibits dose

3 proportionality over a wide range and also shows

4 linear Cmin versus AUC relationship simplifies

5 concentration-controlled dosing. Dose

6 proportionality has been demonstrated for sirolimus

7 Cmax and AUC first in renal allograft patients

8 after coadministration of Rapamune oral solution

9 and cyclosporine over a dose range of 2 to 22

10 milligrams.

11 Secondly, in healthy volunteers after

12 administration of Rapamune tablets over a dose

13 range of 5 to 40 milligrams. Therefore, sirolimus

14 trough levels would be expected to increase in

15 simple proportion to the dose over a dose range of

16 2 to 40 milligrams.

17 Moreover, the correlation between

18 sirolimus Cmin and AUC in renal allograft patients

19 is excellent as shown by an r-squared value of

20 0.96. For the regression line over a concentration

21 range of approximately 1 to 30 nanogram per ml.

22 The experimental data is shown on the next slide.

23 [Slide.]

24 This figure is a plot of sirolimus 24-hour

25 troughs on the Y axis and sirolimus 24-hour AUCs on

74

1 the X axis based on the administration of Rapamune

2 oral solution in combination with cyclosporine

3 during study 301. The individual data points were

4 collected at months 1, 3 and 6 post transplant

5 after doses of 2 and 5 milligrams per day in 42

6 patients.

7 Plotted along with the individual data is

8 the regression line. These data show that troughs

9 can be used for purposes of dose adjustments during

10 sirolimus TDM and the range of concentrations is

11 wide enough to cover the sirolimus target range

12 during TDM as we will see in the final section of

13 this presentation.

14 The important outcome of this relationship

15 is that multiple samples do not have to be drawn

16 during a dose interval at steady state which

17 provides a convenience for the patient and reduces

18 the cost of TDM.

19 [Slide.]

20 Next, there are three PK parameters that

21 affect the implementation of Rapamune

22 concentration-controlled dosing. These are the

23 time to steady state, the loading dose and the

24 maximum dose per day. The mean times to read

25 steady state in renal-allograft patients during

75

1 coadministration of Rapamune oral solution and

2 cyclosporine was five to seven days. That is

3 without a loading dose although the time to state

4 was as long as thirteen days in individual

5 patients.

6 These results indicate that a blood sample

7 for the determination of a steady-state trough

8 should not be drawn for at least five to seven days

9 after the previous dose adjustment when a loading

10 dose is not administered.

11 A loading dose is necessary to quickly

12 reach steady state and the mean estimated sirolimus

13 loading dose determined in renal-allograft patients

14 during coadministration of Rapamune oral solution

15 and cyclosporine was three times the maintenance

16 dose. When a loading dose is used, it may not be

17 necessary to wait as long as five to seven days to

18 draw a sample for purposes of dose adjustment.

19 The maximum dose on any day that was

20 recommended in study 310 was 40 milligrams. It is

21 also recommended, however, that a loading dose

22 larger than 40 milligrams be administered in

23 divided doses over two days.

24 Now, in the next series of slides, I want

25 to discuss our experience with

76

1 concentration-controlled trials.

2 [Slide.]

3 Four studies provided data after one year

4 post transplant as shown in this second column.

5 Study 310, the pivotal study for the current

6 submission, study 212, the supportive study for the

7 current submission, and studies 207 and 210, which

8 were early studies directly comparing Rapamune

9 versus cyclosporine using concentration control.

10 Concentration-controlled data were

11 obtained for both the tablet and the oral solution.

12 The remainder of this presentation will focus on

13 the one-year PK data but data beyond one year has

14 also been presented to FDA.

15 [Slide.]

16 The sirolimus target ranges for

17 cyclosporine withdrawal in studies 212 and 310 were

18 set prospectively based on the results from phase

19 II studies 207 and 210. For sample analysis by an

20 immunoassay, these ranges were 10 to 20 nanogram

21 per ml for study 212 and 20 to 30 nanogram per ml

22 for study 310.

23 The adequacies of the prospective target

24 ranges were supported be efficacy results and

25 similarities in the mean sirolimus trough levels

77

1 for the two studies; that is 18 nanograms per ml

2 for study 212 and 23 nanograms per ml for study

3 310.

4 [Slide.]

5 We evaluated the implementation of

6 concentration control in four Rapamune studies by

7 estimating the percentages of patients showing

8 concentrations below, with and above the sirolimus

9 target concentration ranges. This slide shows the

10 average percentages of patients among studies and

11 ranges for the sirolimus concentration-controlled

12 treatments or Rapa groups in studies 207, 210, 212

13 and 310. These data are shown by the hatched

14 purple bars.

15 A comparison of the data in the center

16 figure with the data in the left and right figures

17 shows that large majorities of the patients in all

18 four studies fell within the target range. It is

19 important to note that the vast majority of the

20 investigators obtained these results using a

21 central lab and did not have the benefit of an

22 assay at the transplant site.

23 Based on averages among the four studies

24 as shown by the purple bars 12 percent of patients

25 were below the target range. 70 percent were

78

1 within the range and 18 percent were above the

2 target range. Overall, 88 percent were above the

3 lower limit of the target range.

4 [Slide.]

5 This figure shows the sirolimus and

6 cyclosporine trough levels over time before and

7 after randomization in the sirolimus

8 concentration-controlled treatment or Rapa group of

9 study 310. You are looking at the outcome of the

10 first Rapamune clinical trial in which

11 investigators were required to simultaneously

12 withdraw cyclosporine while increasing the dose of

13 Rapa. The vertical bar represents randomization at

14 90 days.

15 Trough concentration for cyclosporine are

16 plotted on the left Y axis and for sirolimus and

17 the right Y axis. The time is plotted on the X

18 axis. I want to reiterate that the sirolimus

19 concentrations and target range on this slide are

20 for an immunoassay as are the concentrations and

21 target ranges shown on subsequent slides.

22 Before randomization in this region,

23 cyclosporine troughs, shown as triangles, gradually

24 decreased over 90 days as doses were gradually

25 decreased and sirolimus troughs, shown as circles,

79

1 remained stable at approximately 11 nanogram per ml

2 during the fixed-dose time period.

3 After randomization, in this area,

4 cyclosporine troughs decreased rapidly to near zero

5 concentrations at 150 days as the doses were

6 reduced and sirolimus troughs rapidly increased to

7 reach the target range as doses were increased.

8 [Slide.]

9 Overall, the investigators were quite

10 successful in this first Rapamune trial that

11 required simultaneous adjustment in the dosages of

12 two drugs and cyclosporine was eliminated in 50

13 percent of patients by week 6 after randomization.

14 We can anticipate that the ability to achieve and

15 maintain the sirolimus target range using TDM will

16 improve in the future as more experience is

17 obtained with cyclosporine withdrawal.

18 [Slide.]

19 This figure provides a summary of the

20 sirolimus doses and troughs after reaching the

21 target range in study 310 between 4.5 and twelve

22 months post transplant. In the

23 concentration-controlled treatment, as shown by the

24 purple bars, a mean Rapamune dose of 8.4 milligrams

25 per day produced mean sirolimus troughs of 23.3

80

1 nanograms per milligram which was within the target

2 range for the study.

3 In the fixed-dose treatment, as shown by

4 the red bars, a mean Rapamune dose of 2.1

5 milligrams per day produced a mean sirolimus trough

6 of 10.8 nanograms per milligram. There appears to

7 be a disparity between doses and concentrations

8 since a fourfold increase in dose produces only a

9 twofold increase in concentration. The apparent

10 discrepancy between doses and troughs is due to the

11 fact that cyclosporine produces about a twofold

12 increase in the extent of absorption of sirolimus.

13 Therefore, without the coadministration of

14 cyclosporine, sirolimus troughs would be decreased

15 by one half compared to those during

16 coadministration with cyclosporine and, therefore,

17 higher doses are required.

18 [Slide.]

19 Let me tell you now what we have learned

20 about implementing sirolimus TDM. There are four

21 parameters that I want to discuss which include the

22 frequency of blood sampling for rapid

23 determinations after randomization, the number of

24 days required to reach the target range after

25 randomization, the number of dose changes required

81

1 to reach the target range after randomization and

2 the recommended target trough range for sirolimus

3 TDM.

4 I will also be commenting on the

5 availability of the sirolimus assay.

6 [Slide.]

7 In pivotal trial 310, blood samples were

8 to be drawn weekly during the first month after the

9 start of cyclosporine withdrawal, every two weeks

10 during months 2 and 3, monthly during months 4 to

11 12 and every three months after month 12.

12 The actual number of samples required for

13 the use of sirolimus TM in new patients will have

14 to be individualized since the number of samples

15 depends on the rate of CSA withdrawal and the time

16 needed for sirolimus to reach the target range in

17 the individual patient.

18 Based on an analysis of the number of days

19 to reach the target range, 50 percent of patients

20 reached the target range by approximately twenty

21 days after randomization and also 90 percent of

22 patients reached the target range by 68 days after

23 randomization.

24 Based on an analysis of the number of dose

25 changes to reach the target range, 50 percent of

82

1 patients reached the target range after two doses

2 and 90 percent reached the target range after five

3 doses--after dose changes.

4 [Slide.]

5 Turning our attention now to the sirolimus

6 TDM range, we conducted a logistic-regression

7 analysis of acute rejection using the

8 post-randomization data but the results did now

9 show significant p-values for either sirolimus or

10 various patient parameters. This result is not too

11 surprising since there were relatively few

12 rejections post randomization and a single limited

13 range of concentrations was investigated.

14 In the absence of the PK/PD model, the

15 sirolimus TDM range was established based on

16 distribution analysis of sirolimus troughs among

17 nonrejectors and rejectors and clinical outcomes

18 for studies 310 and 212.

19 The next slide shows the distribution of

20 average sirolimus trough concentrations among

21 nonrejectors in studies 310 and 212.

22 [Slide.]

23 The figure on the left shows the data for

24 study 310 and the figure on the right is for study

25 212. For study 310, the average sirolimus trough

83

1 concentrations in individual patients were

2 determined between six weeks post randomization and

3 one year, and for study 212, the averages were

4 determined between three weeks post randomization

5 and one year.

6 The lengths of the blue bars in the

7 figures represent the numbers of nonrejecting

8 patients at a given concentration as determined by

9 the SAS procunivariate statistical procedure. The

10 dashed lines represent the 5th and 95th percentiles

11 for the sirolimus distribution.

12 As you can see, the ranges for the two

13 studies showed considerable overlap although the

14 212 distribution is shifted downward due to the

15 lower protocol target range. We also observed

16 considerable overlap for rejectors in the two

17 studies, as shown in the next slide.

18 [Slide.]

19 In these figures, sirolimus trough

20 concentrations in individual patients are plotted

21 against the rejection times. The concentrations in

22 the figures are those closest to the rejection

23 time. The dashed lines are, again, the 5th and

24 95th percentiles for nonrejectors.

25 As you can see, the ranges for rejectors

84

1 were very similar for studies 310 and 212 and also

2 a large fraction of the rejectors fell within the

3 5th to 95th percentiles for nonrejectors.

4 Now, one may question whether a fixed-dose

5 regimen could be used in place of TDM. However, as

6 shown in the next slide, sirolimus TDM considerably

7 reduces the intersubject variability compared to a

8 fixed-dose regimen.

9 [Slide.]

10 This figure provides a comparison of the

11 distributions of average sirolimus troughs in

12 nonrejectors beginning a six weeks after

13 randomization in study 310. The box plot on the

14 left is for actual data and the box plot on the

15 right shows the actual concentrations normalized to

16 an 8 milligram daily dose of sirolimus.

17 If patients in 310 had received an

18 8-milligram daily regimen without TDM, the range of

19 sirolimus trough levels would have increased

20 considerably and many patients would have exceeded

21 the 95th percentile observed in study 310 and a

22 number of patients would have fallen between the

23 range of 40 to 70 nanograms per milligram. The

24 data in this slide strongly argued for the need of

25 sirolimus TDM.

85

1 The next slide provides our

2 recommendations for a TDM range.

3 [Slide.]

4 A sirolimus TDM range of 15 to 25

5 nanograms per milligram, as determined by

6 immunoassay, is recommended based on the

7 distributions of sirolimus troughs among

8 nonrejectors and rejectors in studies 310 and 212

9 and the very similar clinical outcomes in studies

10 310 and 212 with respect to graft survival, patient

11 survival and improved renal function within Rapa

12 treatments.

13 These similarities in clinical outcomes

14 were achieved in spite of the different target

15 ranges used in the two studies.

16 As the last topic under the implementation

17 of sirolimus TDM, I want to comment on the

18 availability of the sirolimus assay.

19 [Slide.]

20 Currently, there are 23 bioanalytical

21 lamps that measure sirolimus concentrations by

22 either an HPLC/UV or HPLC/MS/MS assay. Quest

23 Diagnostics in San Juan Capistrano, California, is

24 our central laboratory. Six additional

25 laboratories analyzed samples on a commercial scale

86

1 and sixteen laboratories are located in transplant

2 centers throughout the United States.

3 The two assay methods include the ranges

4 to the 95th percentiles observed in

5 concentration-controlled studies as shown by the

6 footnotes in the table. The HPLC/UV method has a

7 range of 2.5 to 75 nanograms per milligram and the

8 HPLC/MS/MS method has a range of 1 to 50 nanograms

9 per milligram.

10 [Slide.]

11 Turning to guidance that will be provided

12 to physicians, physicians will be informed with

13 respect to algorithms for estimating both a new

14 maintenance dose and new loading dose. The maximum

15 recommended dose of Rapamune per day, time of blood

16 draws for dose adjustments, action guidelines based

17 on assay results and the limitations of TDM.

18 In conclusion, experience with sirolimus

19 TDM without cyclosporine coadministration has been

20 obtained in four clinical trials during one year

21 post transplant among 347 patients. Efficacy

22 outcomes in the TDM groups were equivalent to the

23 respective fixed-dose groups. Studies 310 and 212

24 provided data to define a range of sirolimus trough

25 concentrations for TDM in the proposed indication.

87

1 The results show that TDM can guide the

2 safe and effective use of sirolimus.

3 [Slide.]

4 For TDM without cyclosporine

5 coadministration--that is, for the proposed

6 indication--the recommended sirolimus TDM target

7 range is 15 to 25 nanograms per milligram based on

8 the immunoassay or 12 to 20 nanograms per milligram

9 based on a chromatographic assay.

10 This concludes my presentation. Dr.

11 Neylan will now close today's presentation with a

12 few final remarks.

13 Concluding Remarks

14 DR. NEYLAN: Thank you, Jim.

15 [Slide.]

16 I would like to conclude our presentation

17 today by emphasizing that within the past few

18 years, great strides have been made in advancing

19 the clinical science of renal transplantation. In

20 general, these advances have come as a result of

21 our improved understanding of the optimal use of

22 available immunosuppressive agents.

23 While calcineurin inhibitors have played

24 an important role in the past twenty years,

25 long-term patient and graft survival remain

88

1 suboptimal and the persistent nephrotoxicity

2 associated with maintenance cyclosporine continues

3 to take its toll.

4 The emergence of Rapamune as a new

5 therapeutic option has provided clinicians new

6 opportunities to individualize therapies. Based on

7 the data presented this morning, it is clear that

8 we have made further progress still.

9 [Slide.]

10 The combined safety and efficacy data from

11 studies 310 and 212 are consistent and provide

12 compelling evidence that Rapamune may be utilized

13 to spare the inherent nephrotoxicity long

14 associated with chronic cyclosporine

15 administration.

16 The benefits of concentration-controlled

17 use of Rapamune with cyclosporine elimination

18 include excellent patient and graft survival, a low

19 rate of acute rejection following cyclosporine

20 elimination and an acceptable safety profile.

21 [Slide.]

22 A regimen of maintenance Rapamune is

23 associated with several distinct advantages when

24 compared to long-term use of cyclosporine. These

25 include significantly better renal function that is

89

1 sustained over time, significantly lower blood

2 pressure that is also sustained and significantly

3 lower incidence of several other

4 cyclosporine-related adverse events.

5 [Slide.]

6 Based upon the population of

7 renal-transplant recipients included in these two

8 trials, it is reasonable to expect that these

9 benefits can be realized by most patients now

10 awaiting transplantation in the United States.

11 Specifically, by initiating Rapamune plus

12 cyclosporine and corticosteroids, clinicians can

13 anticipate that most patients can be successfully

14 withdrawn from cyclosporine.

15 In the current studies, greater than 90

16 percent of patients eligible two to four months

17 after transplantation successfully completed

18 cyclosporine elimination. Therefore, only a small

19 number of patients will not be able to accomplish

20 this goal because of complications in their

21 clinical course or intolerance of the

22 immunosuppressive regimen.

23 For these patients, alternative strategies

24 are at hand and may be utilized according to

25 clinical judgment.

90

1 [Slide.]

2 We are excited about these data and their

3 implications for the transplant community. We

4 believe that utilization of Rapamune in the

5 proposed indication may significantly improve the

6 practice of clinical transplantation and enhance

7 the lives of transplant recipients.

8 In conclusion, I would like to acknowledge

9 the patients and investigators who participated in

10 these trials. Their diligence and their commitment

11 has made all of this possible.

12 Thank you for your attention. We will now

13 be pleased to address any questions you may have.

14 DR. ENGLUND: At this point, I would like

15 to ask if there are any clarification questions,

16 just clarification only. We will having the

17 discussion questions later.

18 DR. HUNSICKER: I had a couple, just one

19 clarification question.

20 DR. ENGLUND: Go ahead.

21 DR. HUNSICKER: One of the things that you

22 said earlier is that a certain fraction of patients

23 were removed or permitted not to be randomized

24 because of basically physician judgment that their

25 creatinine was too high. Could you tell us how

91

1 many and what the creatinines were? The issue has

2 to do with what we actually about the group of

3 patients who were randomized and on whom we have

4 effective data.

5 DR. NEYLAN: Yes. Let's see if we can

6 call up a slide looking at the nonrandomized

7 patients.

8 DR. HUNSICKER: That is in study 310,

9 primarily.

10 DR. NEYLAN: You want to look at study

11 310?

12 DR. HUNSICKER: Yes.

13 DR. NEYLAN: Let's show this first.

14 [Slide.]

15 To begin, in study 310, there were 95

16 patients who did not meet the randomization

17 criteria at or before month 3. The reasons for

18 discontinuation in study 310 are listed in the next

19 slide.

20 [Slide.]

21 74 percent of those patients were

22 discontinued because of adverse events prior to the

23 randomization. These adverse events included

24 issues of renal function like ATN, potentially

25 renal-vein or renal-artery thrombosis, cyclosporine

92

1 toxicity. Another category listed as renal

2 dysfunction, and then a host of the other

3 complications that are not out of the usual sort in

4 the more immediate post-operative period.

5 [Slide.]

6 The next slide shows that, in addition to

7 this 74 percent, there were twelve of the 95 that

8 were discontinued because of rejection. These were

9 early rejections prior to the month-3

10 randomization. Nine of these patients had mild to

11 moderate, one severe and one graft loss. Notably,

12 70 percent of the patients within the enrolled

13 population that experienced rejection within the

14 three-month period actually went on to

15 randomization.

16 [Slide.]

17 Then finally, the remaining thirteen

18 patients of this 95 nonrandomized group were

19 discontinued for these listed reasons.

20 DR. HUNSICKER: If I can just clarify my

21 question a bit. I think this is something that is

22 going to have to actually eventually be dealt with

23 by the FDA, the patients in whom we have a

24 comparison are those who were randomized. That is

25 the only group in whom we can make any judgment

93

1 about the relative efficacy.

2 We have to know very precisely what those

3 randomized patients were so that we will be able to

4 tell the public in the future what group of

5 patients there is now data that you could possibly

6 remove the cyclosporine. I think that I would not

7 want to come across that we could remove

8 cyclosporine in all patients because there are a

9 substantial number of patients who never really had

10 this tested.

11 DR. NEYLAN: We would certainly agree with

12 that. So, in addition to the patients who declared

13 themselves, if you will, in this early time point

14 with either a severe rejection or a prolonged or

15 more severe delayed graft function, we have those

16 patients who emerged from this period at month 3,

17 and it is those patients, indeed, in which the

18 decision should be made.

19 We had a slide previously which I wanted

20 to show.

21 [Slide.]

22 It shows the patients who came to month 3

23 and, at that point, were discontinued. I think

24 this, perhaps, more aptly addresses the question

25 you had asked originally which was what number of

94

1 the 95 actually, through physician decision at this

2 three-month time point, elected not to, then, be

3 put through the randomization. We see that there

4 were five patients that fit the bill of a

5 creatinine greater than 4.5, five patients that had

6 either severe renal dysfunction or were on

7 dialysis.

8 The remainder of the patients at this

9 three-month visit mark, which was the time in which

10 physicians decided whether to go on to

11 randomization or discontinue, had these other

12 issues for which the physicians decided not to

13 continue them in the study.

14 DR. ENGLUND: Dr. Auchincloss?

15 DR. AUCHINCLOSS: A couple of reasonably

16 quick questions. The steroid dose you mentioned as

17 being the standard taper. Did that sort of

18 typically end at 15 milligrams a day or were people

19 going even lower?

20 DR. NEYLAN: The tapering went down to

21 lower than 15 milligrams and we have the steroid

22 dosing for the studies. In general, it came down

23 to the range of about 10 milligrams per day.

24 Would you like to see that data?

25 DR. AUCHINCLOSS: No; I don't need to see

95

1 it. I just need to get a sense of it. Secondly,

2 your S15 slide showing the remarkable similarity of

3 use of lipitore in the two groups despite the fact

4 that one is using a four-times-higher dose of

5 rapamycin in the right-hand panel there. Were they

6 using much more lipitore or dose doesn't matter

7 when you get onto rapamycin?

8 DR. NEYLAN: Unfortunately, these studies

9 were not designed a priori to collect actual

10 dosing, so I am afraid I can't answer that

11 question. The choice of lipid-lowering agents

12 certainly included lipitore but it also include

13 other HMG co-A-reductase inhibitors.

14 As you see, 73 percent of both groups were

15 receiving some form. We are certainly interested

16 in this and we are collecting these data now in

17 other trials and trying to get an assessment of the

18 dose response, if you will, to these agents. But

19 we don't have that information for you, these

20 studies, today.

21 DR. AUCHINCLOSS: Can I do one more?

22 DR. ENGLUND: One more.

23 DR. AUCHINCLOSS: The third one is that

24 212 is the one trial that actually had a number of

25 black patients. I believe it was fifteen. And

96

1 then we had a slide later that showed rejectors

2 just near the very end, and there were five spots

3 for black rejectors. So five out of the fifteen

4 rejected at some point in the rapamycin group; is

5 that true?

6 DR. NEYLAN: Yes.

7 DR. AUCHINCLOSS: I know the numbers are

8 small, but is there reason to think that blacks

9 would handle this less well?

10 DR. NEYLAN: Well, actually, I think what

11 I would like to do is, if I might, run through a

12 couple of slides on this issue because, to give you

13 the conclusion first, we think that, although the

14 number of black patients was somewhat small within

15 the collected database of these two studies, the

16 results, in general, mirrored the expectations that

17 might be seen in general clinical practice for

18 these patients and, most importantly, the benefits

19 seen with the cyclosporine elimination are also

20 demonstrated in this group.

21 If I could have the first slide.

22 [Slide.]

23 We see that, indeed, in study 310

24 conducted in non-U.S. countries, the number of

25 black patients was very small but was

97

1 representative of their representation within those

2 general populations. We really won't touch on any

3 of these data since the numbers are, indeed, too

4 small to make much of them.

5 [Slide.]

6 Within study 212, 19 percent of the

7 enrolled population was of black ethnicity. The

8 distribution of their enrollment in the two

9 randomized arms is shown here, 18.6 percent

10 randomized to the control group of 212 and

11 15 percent to the treatment arm. 28.6 percent were

12 not randomized.

13 [Slide.]

14 In the 212 Rapamune group, the

15 cyclosporine elimination arm, as I said, there were

16 fifteen that were enrolled. There were three that

17 were eligible for cyclosporine taper by month 2.

18 Two had experienced acute rejection episodes prior

19 to that.

20 Of those thirteen eligible for

21 cyclosporine taper, all completed the cyclosporine

22 taper. Three had rejection episodes following the

23 cyclosporine withdrawal at days 35, 64 and 122

24 following that elimination.

25 [Slide.]

98

1 The rates of rejection over time are shown

2 here, are shown for black and non-black patients

3 within 212. You will recall that month 2 was the

4 point in this study at which patients went on the

5 cyclosporine discontinuation or were maintained in

6 the control treatment strategy.

7 Four black patients, at month 2 and,

8 again, prior to cyclosporine elimination, not

9 unexpectedly, we saw higher rates of acute

10 rejection in black patients than nonblack patients

11 in both treatment arms. By month 12, now following

12 these patients on through the period of

13 cyclosporine elimination for the Rapamune treatment

14 arms, you see that black patients in the Rapamune

15 treatment, as contrasted with the control, had

16 similar rates of acute rejection, both 33 percent

17 by month 12, this in contrast to the nonblack

18 patients where we see results essentially mirroring

19 that of the study as a whole with a slightly higher

20 rate of acute rejection for the nonblack patients

21 in the Rapamune treatment arm.

22 [Slide.]

23 Most importantly, though, the effect on

24 blood pressure was also confirmed in black patients

25 in the 212 study. We see that, in black patients,

99

1 these are now calculated GFRs at months 2 through

2 12, that there was a trend towards improvement in

3 the Rapamune arm for black patients enrolled that,

4 by month 12, was now statistically significantly

5 different.

6 In fact, this represents a roughly 48

7 percent improvement.

8 [Slide.]

9 There was also a trend in mirroring the

10 results in blood-pressure management as well for

11 black patients although, again, with the small

12 numbers, we don't achieve statistical significance.

13 But, again, we see that four black patients, the

14 systolic and diastolic pressures tended to be lower

15 for black patients in the Rapamune arm than the

16 Rapamune plus cyclosporine arm.

17 [Slide.]

18 Finally, in the last slide, we see that

19 overall patient and graft survival at one year is

20 essentially the same for black and nonblack

21 patients in these two treatment arms, the black

22 patient survival being 100 percent for the Rapamune

23 arm, 94 percent for the Rapamune plus cyclosporine

24 arm, and comparable to that of nonblack patient

25 survival.

100

1 Graft survival is also comparable, 93

2 percent for the Rapamune arm compared with 94

3 percent for the control arm, again similar to the

4 nonblack groups and none of these showed any

5 statistical difference.

6 So, in sum, although the numbers are

7 small, the outcomes in black patients in study 212

8 do mirror the study as a whole and, importantly,

9 also show the same benefits in terms of renal

10 function.

11 DR. ENGLUND: Dr. Abernethy?

12 DR. ABERNETHY: I have a couple. Looking

13 at the severity of rejection in both studies across

14 groups, do we have a chi square or some sort of

15 analysis looking at the mild rejectors and the

16 moderate rejectors? Just looking at the numbers,

17 it would appear that the Rapamune-only group had

18 more severe rejection.

19 DR. NEYLAN: If we could show again the

20 310 rejection histology slide.

21 [Slide.]

22 In the presentation I showed you, the

23 rejections that we saw following randomization

24 actually had no episodes of severe rejection in

25 either of the two treatment groups. What we have

101

1 in the group randomized to the Rapamune was a

2 predominance of mild rejections, 66.7 percent, and

3 moderate rejections, either 2a or 2b, but, again,

4 no severe rejections.

5 These were fairly similar to the severity

6 seen of the rejectors in the control arm of 77.8

7 mild and then there are two types of moderate.

8 DR. ABERNETHY: I suppose one could do a

9 chi-square analysis and see if that is different?

10 DR. NEYLAN: I would have to ask one of my

11 statisticians. Robert, could you speak to that?

12 DR. GOLDBERG-ALBERTS: I am Robert

13 Goldberg-Alberts, Rapamune project statistician.

14 With the sparse numbers there, I wouldn't have done

15 a chi square but I would be happy to get you an

16 exact p-value for the difference in the

17 distribution. I could have that for you after

18 lunch, if you wish.

19 DR. NEYLAN: Thank you, Robert.

20 DR. ENGLUND: One more.

21 DR. ABERNETHY: What was your definition

22 of hypokalemia and thrombocytopenia, just the

23 numbers?

24 DR. NEYLAN: Yes. The definitions are

25 slightly different depending on whether we are

102

1 looking at it from the listing of laboratory values

2 or we are listing it as an investigator-initiated

3 spontaneous adverse-event report.

4 In the case of the laboratory parameters,

5 they simply are those of the laboratory standards.

6 However, in the case of the spontaneous reporting

7 of adverse events, we are simply relying on the

8 investigator's personal view.

9 If I could have the potassium through time

10 for study 310, what I would like to show is that,

11 indeed, we saw in patients in whom cyclosporine was

12 eliminated, that the cyclosporine effect in

13 retarding potassium secretion was demonstrated on

14 those patients and, in addition, the mild kaluretic

15 effect that we have seen with Rapamune was also

16 seen.

17 [Slide.]

18 This summary experience, while it created

19 statistical difference between the treatment arms,

20 did not bring patients down below the lower limits

21 of normal for potassium. So, again, to reiterate,

22 at month 3, as you would expect, these two groups

23 are similar and then, as they proceed through the

24 period in which cyclosporine is eliminated in the

25 Rapamune arm, you begin to see statistical

103

1 difference which is maintained here at month 12 and

2 here at month 24. Statistical difference, yes; but

3 the Rapamune-treated patients are still maintaining

4 potassiums above the lower limit of normal.

5 MR. LAWRENCE: To be absolutely precise

6 about that, you are showing SEMs there. You are

7 not showing standard deviations. What you really

8 need to show is the fraction of patients that are

9 below the level to say that, John.

10 I am not calling for another slide. I

11 think that it is probably fine. But don't say that

12 the potassiums are all fine because the mean is

13 fine.

14 DR. NEYLAN: We brought 1500 slides, just

15 to warn you.

16 DR. ENGLUND: Let's go on. Dr.

17 Suthanthiran?

18 DR. SUTHANTHIRAN: John, I wanted to ask

19 you about acute rejection. It is true at the end

20 of the twelve months, both groups seemed to have a

21 nonsignificant difference in the incidence of acute

22 rejection. But if you look at post randomization,

23 excluding the first three months when the patients

24 are on cyclosporine, there is, in fact, an increase

25 in the incidence of acute rejection.

104

1 I wonder, in your cyclosporine, you

2 actually have three phases, an induction phase, a

3 taper and a discontinuation. Is there a place in

4 the taper time that there is a particular level of

5 cyclosporine at which, when it goes below a certain

6 threshold, you start seeing acute rejection?

7 DR. NEYLAN: First, as we are looking for

8 the slide that I would like to show you showing the

9 changing cyclosporine levels, we can first look at

10 this.

11 [Slide.]

12 310, as you say, shows that, up to the

13 point of randomization, there were identical and

14 very low rates of acute rejection that were seen

15 for all the patients enrolled in the study.

16 But, subsequent to the point of

17 randomization and, with that, the onset of

18 cyclosporine elimination in the Rapamune-treatment

19 arm, you see an increment difference in the rates

20 of rejection statistically significantly different

21 here comparing new rates but in cumulative

22 accounting, not statistically different there.

23 What I want to find is the histogram that

24 shows the cyclosporine levels as they go through--I

25 believe it is in your slide packet, Jim. What we

105

1 saw was that, not unexpectedly, with the attendant

2 decrease in cyclosporine exposure, there was an--at

3 the beginnings of the increase in these incremental

4 rejection episodes following the randomization.

5 There was a window of time, in showing

6 this histogram, between the elimination of

7 cyclosporine completely.

8 Yes; this is the slide. Thank you.

9 [Slide.]

10 What we see here in study 310 are, in the

11 red bars, the mean cyclosporine trough levels.

12 Here is day 90, the point of randomization, the

13 point at which cyclosporine is beginning to be

14 tapered by the investigators for patients in the

15 Rapamune arm.

16 In these line drawings, you see the rates

17 of acute rejection for the patients randomized to

18 the Rapamune arm and the patients randomized to the

19 control arm. So, following the cyclosporine

20 troughs, you can see that, at this point, things

21 are fairly similar and there begins an incremental

22 increase at or about the time that cyclosporine is

23 being completely eliminated.

24 This incremental increase appears to

25 continue a bit longer beyond the point at which, at

106

1 least for the mean, the cyclosporine has been

2 completely eliminated. This may relate to, also,

3 the rapidity at which the investigators were

4 achieving the target ranges for Rapamune.

5 So, again, we have two moving targets

6 here. We have cyclosporine coming down and

7 Rapamune, of course, being adjusted upward to

8 achieve the new target ranges.

9 DR. ENGLUND: Dr. DeGruttola had a

10 question.

11 DR. DeGRUTTOLA: I just had a question on

12 a similar point. You made a statement in the

13 summary that there are similar incidents, similar

14 rates of acute rejection, between the two groups,

15 the 13.4 and the 20 percent with a p-value of 0.08.

16 I am just wondering what the definition of similar

17 rates is there.

18 Usually, statistically, when you describe

19 something as similar, we are saying we can reject a

20 difference of a certain amount or define a window

21 of equivalence. I was wondering if that is how

22 similar is defined or is it just reflecting the

23 fact that the p-value doesn't happen to be below

24 0.05?

25 DR. NEYLAN: I see Jim Burke shaking his

107

1 head. I think I will ask him to address this

2 question. Jim, if you could first identify

3 yourself at the microphone.

4 DR. BURKE: Jim Burke, Wyeth-Ayerst

5 Research. It is the latter that is true, that we

6 call them similar because the p was not less than

7 0.05.

8 DR. DeGRUTTOLA: Another question that I

9 had was regarding the analyses of cholesterol

10 values and triglycerides and so on. Are those done

11 on an intent-to-treat or on an on-therapy

12 population?

13 DR. BURKE: These are on-therapy.

14 DR. ENGLUND: Dr. Shapiro?

15 DR. SHAPIRO: John, that was a really nice

16 presentation. I have a couple of questions. As

17 you know, most patients entered into trials tend to

18 be somewhat selected. And then you selected again,

19 throwing out 18 percent of the patients in the 310

20 trial and 20 percent of the patients in the 212

21 trial. These were the nonrandomized patients.

22 Then you end up with patients who have

23 extremely good outcomes. What were the patient and

24 graft survival rates, rejection rates and resistant

25 rates in the nonrandomized patients in both 310 and

108

1 212?

2 DR. NEYLAN: Let's show this slide while

3 we are getting that data for you.

4 [Slide.]

5 This is first to look at the study 310 and

6 compare the demographic features of the patients

7 who were not randomized against those patients who

8 went on to randomization. They are actually the

9 same, or at least similar, with two exceptions.

10 As you might expect, the nonrandomized

11 patients had a higher percentage of delayed graft

12 function and a higher percentage of acute rejection

13 than the patients who went on to randomization.

14 And that addresses your point that, from a

15 clinical-utility standpoint, these are both studies

16 in which patients are enrolled but then followed

17 through a critical window of time, a high-risk

18 window of time.

19 Those patients who get to that subsequent

20 time point are the ones that are logically

21 candidates for this kind of strategy.

22 [Slide.]

23 This next slide shows the breakdown of the

24 histologic grade of rejections by twelve months

25 comparing the two randomized groups to that of the

109

1 nonrandomized group. To walk you through it is to

2 say we have this period of time prior to the point

3 of actual randomization. These patients went on

4 to, of course, be randomized but their rejection

5 episodes occurred in that early period of time.

6 As I say, 70 percent of patients that had

7 acute rejections within the first three months

8 actually went on to randomization. So that is the

9 first point. We have mandated by protocol that

10 only the severe rejection episodes would be

11 disallowed from being considered for randomization

12 subsequently at three months.

13 In contrast, we have, during this same

14 window of time, this early three-month, the types

15 of rejection, the histologic grades of rejections

16 seen for the nonrandomized group. Being

17 nonrandomized, then, we have only follow up for

18 those. You see a small number of patients that, in

19 the follow-up period, had rejection episode within

20 that time frame.

21 Does this address your question?

22 DR. SHAPIRO: It doesn't discuss the

23 patient and graft survival.

24 DR. NEYLAN: All right. Show this slide,

25 please.

110

1 [Slide.]

2 What we saw for the treatment arms in

3 study 310 was the overall one-year graft survival

4 that was comparable, actually numerically superior,

5 for the Rapamune treatment arm. These are the

6 causes of graft loss within these groups. In

7 comparison, we see the 95 patients who, again, were

8 not randomized at the three-month mark and the

9 causes of graft loss in this group.

10 DR. ENGLUND: Dr. Mannon?

11 DR. MANNON: My question relates more to

12 the TDM aspect. I guess these results are based on

13 the immunoassay and, in your conclusion, you

14 related both either targets towards the immunoassay

15 or the HPLC. Is the expectation that the

16 immunoassay may be eventually available and, if

17 not, do you think we could obtain comparable

18 results if we stuck with HPLC?

19 DR. NEYLAN: I think I can just tackle

20 this, Jim, if you don't mind. I think what we have

21 seen is that there is a clear correlation between

22 the immunoassay and the HPLC methodology so we can

23 readily adapt values and put them in the context of

24 what we have seen with these studies and the

25 immunoassay.

111

1 Those centers are available now and they

2 include both the central laboratories as well as,

3 in some cases, on site within the transplant

4 centers. As to the future, yes; an immunoassay is,

5 indeed, in our future. At long last, I am happy to

6 report that we are now working hand-in-hand with a

7 company who will in, I hope, the very near future

8 have a immunoassay out and available in a manner

9 similar to the assays available for other

10 immunosuppressants.

11 DR. MANNON: My last question again

12 relates to levels. Were patients in either of

13 these studies required or encouraged to be on a

14 particular diet for the morning meal or was there

15 any follow up or guidance regarding their diet?

16 DR. NEYLAN: No; there was no specific

17 dietary restriction.

18 DR. ENGLUND: Dr. Ebert?

19 DR. EBERT: A couple of questions related

20 also to TDM. First of all, it appears from your

21 serum-concentration ranges that you have

22 established, certainly there appears to be some

23 evidence for the lower level, not going below a

24 certain level, based on the fact that you had a

25 higher number of rejectors.

112

1 But I am curious if you have any evidence

2 in your upper level that you are looking for from a

3 target range. Were there any adverse events that

4 were correlated with exceeding that value.

5 DR. NEYLAN: Jim, do you want to say just

6 very briefly? We did, indeed, look at that.

7 DR. ZIMMERMAN: We did look at several lab

8 parameters. We looked at potassium. We looked at

9 liver-function tests and I believe triglycerides

10 and cholesterol and we did not find any trends for

11 patients above 25 nanograms per milligram that

12 would lead us to believe that there is a

13 relationship there.

14 DR. EBERT: The second question is I

15 realize you had to do a number of serum

16 concentrations to titrate your regimens. Were

17 there any population parameters, age, preexisting

18 liver disease, et cetera, that might have helped

19 you to more closely predict the ultimate

20 maintenance dose?

21 DR. NEYLAN: We don't think so because we

22 conducted the logistic regression analysis for the

23 time period after randomization up to one year. We

24 looked at factors such as HLA mismatch,

25 donor-related--can we bring up that slide? I don't

113

1 have all the parameters. We looked at about five

2 or six different parameters in that regression,

3 also sirolimus concentrations But we could not

4 find the relationship.

5 [Slide.]

6 This is for 310. As you can see, we have

7 both drug concentrations there, gender, increasing

8 recipient age, cadaveric HLA mismatch, increased

9 ischemia time, increased donor age and number of

10 rejections. Except for increasing donor age, there

11 were no significant p-values.

12 DR. EBERT: These are things that predict

13 rejection; is that correct?

14 DR. NEYLAN: That's correct.

15 DR. EBERT: I am looking at were there

16 patient-related variables that predicted the drug

17 clearance, the final dose that was required to be

18 achieved in those patients.

19 DR. NEYLAN: We didn't do it in this

20 population but, from all of our previous data with

21 the tablet submission and the oral-solution

22 submission, we did not find any patient-related

23 factors that would help.

24 DR. ENGLUND: I think with that, we are

25 going to actually take a break now. There is going

114

1 to be time for questions after lunch, after the FDA

2 proposal. So let's take a break now. We are going

3 to start at ten minutes after 11:00, fifteen

4 minutes.

5 [Break.]

6 DR. ENGLUND: We will now hear from the

7 FDA Presentation.

8 FDA Presentation

9 DR. TIERNAN: Good morning.

10 [Slide.]

11 My name is Rosemary Tiernan and I work in

12 the Division of Special Pathogens and Immunologic

13 Drug Products. I would now like to begin the FDA

14 presentation of our review of Rapamune for the

15 indication of cyclosporine withdrawal in renal

16 transplantation.

17 [Slide.]

18 Before I begin, I would just like to

19 acknowledge the efforts of the members of the

20 Rapamune review team who are listed on this slide.

21 I would especially like to thank our statisticians

22 Dr. Cheryl Dixon and Dr. Karen Higgins.

23 [Slide.]

24 The presentation will cover the following

25 areas; background information regarding the initial

115

1 approval of Rapamune in 1999 and the phase IV

2 commitments that were negotiated. They will be

3 briefly reviewed. I will highlight certain issues

4 regarding the design of the clinical studies

5 submitted in the current NDA to support a labeling

6 change.

7 Efficacy and safety considerations will be

8 discussed. Finally, our Division Director, Dr.

9 Renata Albrecht, will present the questions to the

10 advisory committee

11 [Slide.]

12 The basis of the initial approval for the

13 prevention of acute rejection in renal

14 transplantation included two randomized,

15 double-blind, phase III studies, study 301 and 302,

16 comparing Rapamune, 2 milligrams and 5 milligrams

17 to azathioprine or placebo. Both studies

18 demonstrated noninferiority with respect to

19 12-month patient and graft survival and a

20 significant reduction in the incidence of rejection

21 at six months.

22 Despite a lower rate of acute rejection at

23 six months post transplant, renal function, as

24 measured by serum creatinine, and calculated GFR

25 was decreased at twelve months in the

116

1 Rapamune-treatment groups compared to controls.

2 [Slide.]

3 As a phase IV commitment, the applicant

4 agreed to report long-term follow-up safety and

5 efficacy data from studies 301 and 302. It was

6 requested the data pertaining to GFR and serum

7 creatinine be included as follow-up information and

8 be collected throughout the entire duration of the

9 study whether or not patients remained on study

10 drug.

11 Based on 24-month data of only those

12 patients who remained on assigned therapy, renal

13 function continued to be decreased in the Rapamune

14 treatment groups compared to controls.

15 [Slide.]

16 It had been noted in the double-blind

17 studies 301 and 302 that mean and median

18 whole-blood cyclosporine concentrations had

19 remained at or above the upper limit of the

20 specified target concentration ranges. An

21 additional commitment was to evaluate the optimum

22 therapeutic range for sirolimus and the value of

23 reduced cyclosporine concentrations in combination

24 with sirolimus.

25 Proposed sirolimus concentration ranges

117

1 were based on preliminary PK/PD analyses on a

2 subset of patients in the phase III studies. The

3 concentration ranges were evaluated prospectively

4 in subsequent controlled trials including those

5 that we will be discussing today.

6 [Slide.]

7 The applicant is proposing to amend the

8 label to include a consideration of cyclosporine

9 withdrawal at two to four months after

10 transplantation and the use of

11 concentration-controlled sirolimus adjusted to 15

12 to 25 nanograms per milligram when used without

13 cyclosporine.

14 [Slide.]

15 The application for the labeling change is

16 supported by two studies that utilize cyclosporine

17 withdrawal with Rapamune in

18 concentration-controlled regimen. Study 310 was an

19 open-label non-IND study conducted in Europe,

20 Canada and Australia with randomization at month 3

21 post transplant. Study 212 was an open-label study

22 conducted in the U.S. and Europe with randomization

23 at days 2 to 7 post transplant and we are in

24 general agreement with the applicant's description

25 of these studies and the reported results.

118

1 [Slide.]

2 In the cyclosporine-withdrawal arm, the

3 dosage of sirolimus was increased after withdrawal

4 and was adjusted to maintain whole-blood

5 concentrations by immunoassay. Study 310 targeted

6 trough levels of 20 to 30 nanograms per milligram

7 while study 212 targeted trough levels of 10 to

8 20 nanograms per milligram.

9 [Slide.]

10 The strengths of these studies include the

11 randomized controlled design, the quality of the

12 concentration control of cyclosporine and sirolimus

13 and the quality of follow up for patient and graft

14 survival. Weaknesses of the study include the

15 open-label study design which creates a potential

16 for bias in the assessment of acute rejection

17 episodes were comparative safety, the lack of

18 adequate representation of subpopulations of

19 interest such as African-Americans and Hispanics

20 and the early randomized in study 212 allowed for

21 dropout before reaching the time of cyclosporine

22 withdrawal.

23 [Slide.]

24 We would now like to briefly cover the

25 following efficacy considerations; the patient

119

1 population, discontinuations during treatment,

2 patient and graft survival at twelve months, acute

3 rejection after cyclosporine withdrawal and renal

4 function at twelve months.

5 [Slide.]

6 Study 310 excluded high-risk transplant

7 recipients from randomization to cyclosporine

8 maintenance or withdrawal at two to four months

9 after transplantation. Based on protocol-specified

10 criteria which included Banff grade III

11 acute-rejection episodes or vascular rejections

12 occurring four weeks before random assignment,

13 dialysis dependency, serum creatinine greater than

14 400 micromoles per liter or inadequate renal

15 function in the opinion of the investigator to

16 support cyclosporine elimination.

17 [Slide.]

18 In study 212, patients were randomized at

19 an earlier time than in study 310. Patients with

20 adequate renal function, as determined by the

21 investigator, were randomly assigned within 48

22 hours after transplantation to cyclosporine

23 maintenance or withdrawal. The remaining patients

24 were eligible for randomization if their acute

25 tubular necrosis or delayed graft function had

120

1 resolved sufficiently by the seventh day to allow

2 them to receive cyclosporine A. Patients whose

3 acute tubular necrosis or delayed graft function

4 had not resolved by day 7 after transplantation

5 were not randomized.

6 [Slide.]

7 Discontinuation after randomized

8 assignment to treatment is problematic in

9 open-label studies and it is difficult to determine

10 if the actual regimen led to the discontinuation or

11 if it was due to patient or physician concern over

12 randomized treatment. More patients discontinued

13 during assigned treatment in the Rapamune arm

14 compared to the Rapamune plus cyclosporine arm.

15 This difference is statistically significant in

16 study 310.

17 However, all patients were followed

18 through twelve months for rejection, graft loss

19 and death whether they continued assigned treatment

20 or not and the majority also had retrievable

21 renal-function information.

22 [Slide.]

23 This table depicts the reasons for

24 discontinuation in study 310. Although the overall

25 rate of discontinuation in study 310 is

121

1 significantly higher for the Rapa treatment arm,

2 comparison of the individual reasons for

3 discontinuation fail to show any noteworthy

4 differences.

5 [Slide.]

6 We are in general agreement with the

7 applicant's description and report of patient and

8 graft survival at twelve months after

9 transplantation. As the applicant discussed

10 earlier, patients and graft-survival rates were

11 high, well over 90 percent, despite the difference

12 in discontinuation from study drug between

13 treatment groups in study 310, patient and graft

14 survival among those in the Rapa arm was not

15 inferior to those in the Rapamune plus cyclosporine

16 arm.

17 [Slide.]

18 This slide presents the rates of acute

19 rejection following cyclosporine withdrawal for the

20 two studies. There was an excess of

21 acute-rejection episodes observed in the Rapa arm

22 compared to the Rapamune plus cyclosporine arm.

23 This was consistent across both studies.

24 The excess in acute rejection, however,

25 was not associated with a detectable decrease in

122

1 patient or graft survival at twelve months after

2 transplantation as show in the previous slide by

3 the high patient and graft survival rates.

4 [Slide.]

5 Renal function at twelve months post

6 transplantation was measured by serum creatinine

7 and GFR as calculated by the Nankivell method.

8 Rather than performing an on-therapy analysis, the

9 analysis of renal function that we will present

10 attempted to include all patients with a

11 functioning graft at twelve months including those

12 who discontinued study drug.

13 There was a small amount of missing data

14 reflected by the numbers of subjects included in

15 the following tables. Overall renal function is

16 better for patients in the Rapa arm. However,

17 patients who experienced an episode of rejection

18 had worse renal function regardless of which

19 treatment group they were assigned.

20 [Slide.]

21 This slide presents the mean GFR at twelve

22 months post renal transplant. In both studies,

23 significant increases in GFR are noted for the Rapa

24 treatment arms when compared to the Rapamune plus

25 cyclosporine arm.

123

1 [Slide.]

2 This slide presents similar results for

3 serum creatinine and creatinine results are

4 significantly better in the Rapa arm.

5 [Slide.]

6 The next two slides present that and serum

7 creatinine results by post-transplantation

8 rejection status. In patients who have not had a

9 rejection within the first twelve months post

10 transplant, the improvement in GFR in the Rapa arm

11 compared to Rapa plus cyclosporine remains.

12 However, patients who experience a rejection have

13 decreased GFR regardless of treatment.

14 [Slide.]

15 This slide presents similar results for

16 serum creatinine. In patients who have not had a

17 rejection within the first twelve months post

18 transplant, the improvement in serum creatinine in

19 the Rapa arm compared to Rapamune plus cyclosporine

20 remains and, once again, patients who experience

21 rejection have decreased renal function regardless

22 of treatment.

23 [Slide.]

24 Safety considerations that we will present

25 will include defining the exposure to sirolimus, a

124

1 review of the original Rapamune NDA adverse-event

2 profile for the 5 milligram dose compared to the 2

3 milligram dose and then we will highlight specific

4 adverse events that occurred in the current two

5 pivotal trials.

6 [Slide.]

7 The mean trough concentration for

8 sirolimus following 2-milligram and 5-milligram

9 doses in the original NDA, study 310, are depicted

10 on this slide. Note that the observed sirolimus

11 trough concentrations in the current study 310, in

12 the sirolimus concentration arm, are comparable to

13 those observed in the 5-milligram arm of study 310.

14 [Slide.]

15 Trough concentrations were determined

16 using an immunoassay method in the clinical trials

17 and the applicant is proposing a validated HPLC

18 methodology for therapeutic dose monitoring. This

19 involves sending samples to analytical centers,

20 laboratories, for determining the trough

21 concentrations.

22 [Slide.]

23 The original Rapamune NDA was approved in

24 September of 1999 and, at that time, when

25 considering treatment-emergent adverse events that

125

1 occurred at a frequency of greater than 20 percent,

2 a significantly higher incidence of fever,

3 diarrhea, anemia, leukopenia, thrombocytopenia and

4 hyperlipidemia occurred with the use of the higher

5 5-milligram dose of Rapamune when compared to the

6 2-milligram dose.

7 Consequently, our safety review focused on

8 ascertaining whether these side effects would be

9 more problematic in the current studies which

10 utilize concentration-controlled Rapamune with

11 higher drug exposure and, indeed, diarrhea in study

12 212 and thrombocytopenia in both studies 212 and

13 310 occurred at a significantly higher incidence in

14 the Rapa treatment arm.

15 The incidence of hypercholesterolemia and

16 hypertriglyceridemia and the use of lipid-lowering

17 agents was not significantly different across the

18 two treatment arms in study 212 and 310.

19 [Slide.]

20 Now, considering treatment-emergent

21 adverse events that occurred in the original NDA at

22 a frequency of greater than 5 percent and less than

23 20 percent, one notes a significantly higher

24 incidence of chills, face edema, hypotension,

25 hypokalemia, increased LDH, skin ulcer,

126

1 lymphocoele, tachycardia, insomnia and epistaxis

2 with the use of the higher 5-milligram dose of

3 Rapamune when compared to the 2-milligram dose.

4 In the present studies, 310 and 212,

5 hypokalemia occurred in a significantly greater

6 frequency in the Rapa arm.

7 [Slide.]

8 There were discontinuations for elevated

9 liver-function test in the Rapa arm in study 310.

10 Hepatitis B virus and hepatitis C virus data was

11 not available on all patients. There was an

12 increased incidence of elevated LFTs again in the

13 Rapa arm versus the Rapamune plus cyclosporine

14 treatment arms of both studies. There were no

15 deaths in study 212 or 310 which were due to

16 hepatic failure or attributable to study drug.

17 [Slide.]

18 The majority of the patients in the two

19 studies were at lower risk to develop CMV

20 infection. Approximately 12 percent of patients in

21 study 310 were high risk with CMV-donor positivity,

22 recipient-negative for CMV. There were no

23 significant differences in the incidence of

24 infection across treatment arms except for the

25 higher incidence of Herpes zoster in the Rapamune

127

1 plus cyclosporine arm in study 310 and a higher

2 incidence of fungal dermatitis in the Rapa arm in

3 study 212 which Wyeth has already discussed.

4 There were no detectable differences in

5 the treatment arms related to malignancy or

6 post-transplant liver proliferative disease.

7 [Slide.]

8 To summarize, finally, please consider the

9 risks and benefits of utilizing

10 concentration-controlled Rapamune in a cyclosporine

11 withdrawal regimen for renal-transplant patients.

12 The risk of cyclosporine withdrawal include the

13 surge of early mild rejection seen in these studies

14 coupled with higher exposure to sirolimus and the

15 associated adverse events such as thrombocytopenia,

16 hypokalemia and elevated liver-function tests.

17 The benefit of cyclosporine withdrawal

18 include the less cyclosporine-associated toxicities

19 and mean renal function was improved in those

20 patients who did not experience rejection.

21 That's the conclusion for the FDA review.

22 Fairly brief.

23 DR. ENGLUND: Questions?

24 DR. ABERNETHY: With your review of the

25 data, what do you believe the definition of

128

1 hypokalemia and thrombocytopenia was? I am just

2 trying to understand. Is it less than the other

3 group?

4 DR. TIERNAN: It is less than the other

5 treatment arm; right.

6 DR. ABERNETHY: But we are really not

7 talking about below 3.5 or below 50,000?

8 DR. TIERNAN: No. It is more of a

9 relative--

10 DR. HUNSICKER: One thing I didn't get

11 from the rapid thing. I, of course, have the

12 advantage of the briefing document from

13 Wyeth-Ayerst and only a brief thing from you. When

14 you did the analysis for creatinine on an

15 intent-to-treat basis rather than on a, whatever

16 they called it, the basis that excluded patients

17 who were not still on drugs. If you include all

18 the patients, including the patients who rejected

19 and whatever, what was the difference at the last

20 analysis at one year? What was the difference in

21 creatinine between those that were on the Rapamune

22 and those that were on the Rapamune plus

23 cyclosporine?

24 DR. TIERNAN: Dr. Cavaille-Coll, do you

25 want to--

129

1 DR. CAVAILLE-COLL: I think we want to

2 look again at slide 20, please.

3 [Slide.]

4 I have to first apologize that these

5 analyses are not in the briefing package we gave

6 you. We had to have our briefing package prepared

7 a month ago and we only received the data that

8 allows us to do these within the last few days.

9 The numbers, the n's, we see here show the

10 numbers of patients for whom we were able to

11 retrieve data. We believe that we have data on

12 practically all the patients that still had a

13 functioning graft. This represents, basically, the

14 serum creatinine in micromoles per milliliter at

15 twelve months for the different groups. This did

16 not separate them out for whether they rejected or

17 did not reject.

18 DR. HUNSICKER: This includes rejectors

19 and nonrejectors.

20 DR. CAVAILLE-COLL: Yes.

21 DR. HUNSICKER: So long as they still have

22 a functioning graft.

23 DR. CAVAILLE-COLL: Yes.

24 DR. HUNSICKER: And we have the problem of

25 the loss because of a nonfunctioning graft and we

130

1 would have to deal with that if they were uneven.

2 But they are relatively even so we are going to be

3 able to ignore that.

4 DR. CAVAILLE-COLL: Actually, since these

5 were very low-risk patients already, there were

6 very few graft losses and deaths.

7 DR. HUNSICKER: I want to say this now as

8 sort of a preparation to what I would like to say

9 later on about the relationship between rejection

10 and creatinine that, at the end of the day, taking

11 all the patients, the patients assigned to Rapamune

12 on an intent-to-treat basis wound up with about a

13 13, which is about--what does that translate, about

14 1 milligram per deciliter difference?

15 DR. ENGLUND: Who could translate the

16 micromoles into milligrams per deciliter?

17 DR. HUNSICKER: It is about 0.1. It is

18 about a 0.1 milligram per deciliter difference.

19 DR. CAVAILLE-COLL: Yes.

20 DR. HUNSICKER: In the favor of Rapamune

21 even taking into account the increased numbers of

22 rejections.

23 DR. CAVAILLE-COLL: Do you want to also

24 see the next slide, 22, which will show you how it

25 breaks down by rejector and nonrejector?

131

1 DR. HUNSICKER: Yes.

2 [Slide.]

3 I actually did see that one and what I

4 noticed was that amongst the rejectors, there is no

5 difference meaning that--well, I will just simply

6 say there is no difference whereas there is a

7 substantial difference in the nonrejectors. But at

8 least it is not worse in the rejectors.

9 DR. CAVAILLE-COLL: I think that is what

10 the slide says; yes.

11 DR. ENGLUND: Other questions? Dr.

12 Suthanthiran?

13 DR. SUTHANTHIRAN: In both these studies,

14 this is a concentration-controlled trial keeping

15 sirolimus levels at 15 to 25. Do we have any data

16 in terms of whether these levels are actually

17 therapeutic? Is there any relationship between

18 these levels and the absence or presence of acute

19 rejection because when I looked at earlier data

20 when it was presented, it appeared that the

21 majority of patients, rejectors or nonrejectors,

22 fell within this 15 to 25 nanograms per milligram,

23 because we are going to place a lot of emphasis on

24 keeping patients at these levels.

25 I wonder whether keeping them at this

132

1 level really has a clinical benefit in terms of

2 either absence or presence of rejection or in terms

3 of creatinine levels or in terms of clearance. I

4 don't know whether the FDA looked at it.

5 DR. ENGLUND: Could the FDA respond to

6 that?

7 DR. CAVAILLE-COLL: We didn't look at that

8 specifically. Again, I must say that the

9 information that we had on the retrievable

10 information on twelve-month data for creatinine

11 clearance, for creatinine and GFR really we have

12 only had for less than two weeks. The company made

13 a very good effort to try to retrieve that since

14 that was not something that they had planned to

15 collect originally under their protocols.

16 DR. ENGLUND: So we don't have, really,

17 that much intent-to-treat pharmacokinetics at

18 twelve months?

19 DR. ABERNETHY: I think that the issue at

20 least some of us are feeling is that there has been

21 no rationale presented yet for therapeutic drug

22 monitoring with this drug. I think we are seeking

23 that rationale.

24 DR. ENGLUND: We certainly want to discuss

25 that after the FDA presentation. So, be

133

1 forewarned.

2 Do we have any other questions concerning

3 the FDA presentation specifically that was given to

4 us here?

5 DR. HUNSICKER: I guess I would like to

6 ask the FDA, as they discussed with the sponsor the

7 planning of this trial, there are two things that I

8 find surprising. The first is that a lot of the

9 analyses, the toxicity analyses, which are really

10 the basis on which a superiority is being proposed,

11 were not done on an intent-to-treat basis making it

12 very difficult to understand.

13 Was this an understanding that you all had

14 beforehand?

15 DR. CAVAILLE-COLL: The FDA had very

16 little input in the planning of these studies.

17 Study 310 was conducted outside the U.S. and not

18 under the U.S. IND. Most of the planning of study

19 212, FDA had very little input on that

20 DR. HUNSICKER: Okay.

21 DR. CAVAILLE-COLL: As far as analysis for

22 safety, it is customary to do an analysis in the

23 population of all patients who received at least

24 one dose of study drug. Another variation, though,

25 is to do an analysis only based on patients who are

134

1 still on the study drug up to a certain number of

2 days after discontinuation of study drug.

3 DR. HUNSICKER: Yes. I guess the reason I

4 am coming down on this though is that the role of,

5 in quotations now, toxicity here is very different

6 in this application from the typical one in which

7 you have a major comparison in which you are

8 showing superiority and you just want to make sure

9 you are not killing people or doing something nasty

10 on the side.

11 There the toxicity is really supportive of

12 the major conclusion. In this particular

13 situation, the whole world has been turned upside

14 down. You are showing equivalence for what we

15 consider to be--or looking at the question of

16 equivalence--for what are the major outcomes and

17 you are justifying this new agent on the basis of

18 less toxicity.

19 Under those circumstances, it seems to me

20 that there is a real requirement that the toxicity

21 analysis be done the same way that we would have

22 done any other analysis for a major outcome; that

23 is to say, on an intent-to-treat basis. We have to

24 see all of the data.

25 DR. ENGLUND: Are there any more

135

1 questions?

2 DR. DeGRUTTOLA: A brief follow up on that

3 question. I thought that was an excellent point

4 and I think one of the issues here is whether

5 toxicities are likely to persist after therapy has

6 been discontinued.

7 On the one hand, there is the issue of

8 whether comparisons are interpretable because they

9 are based on the randomized populations which I

10 think the previous speaker mentioned and the other

11 issue I think pertains to the persistence of

12 toxicity. So I think reconsidering this issue in

13 the discussion about how to interpret the toxicity

14 results with those issues in mind--

15 DR. HUNSICKER: I do have another question

16 for the FDA when it is my turn again.

17 DR. ENGLUND: What I would like to propose

18 is to finish up FDA questions and then, since we

19 have a little bit of time, to go back to our

20 pharmacokinetics questions yet before lunch. So,

21 if we have any other questions, if this is an FDA

22 question having to do with this presentation.

23 DR. HUNSICKER: This is a--I am almost

24 embarrassed to say it is probably a legal question

25 but there is an issue here about the requirement

136

1 for a sponsor to show sufficient numbers of major

2 subpopulations of the United States for us to be

3 able to say anything.

4 My question is--here, I will tell you in

5 advance my opinion that we don't have enough

6 information about blacks or hispanics to be able to

7 say anything very substantial about them. We just

8 simply don't have the data. I don't think that the

9 small numbers of patients that were randomized to

10 the 212, I guess it was, trial are sufficient

11 really to give us any confidence about where things

12 are going to be, particularly if you take it from

13 the point of view that this is a group in which we

14 know the risks, both acutely and longer term, are

15 much higher.

16 The question is what do we have to say at

17 the end of the day about the entire application

18 when it does not have enough information about

19 subpopulations? Can we say that this is a

20 reasonable proposal for people who are in the

21 population, that they were studied but that we

22 don't have information, or do we have to say, "You

23 really have to show information about your

24 subpopulations before you come to us." I don't

25 know the answer to that.

137

1 DR. ALBRECHT: I would like to say that

2 what we are looking for you to say to us, from a

3 patient-management scientific approach, is is the

4 absence of that data so critical that, in fact, it

5 is not possible to recommend whether there is a set

6 of patients that can responsibly be managed with

7 this regimen or whether the absence of that

8 information is such that, in fact, it precludes

9 putting the drug on the market because of possible

10 risks for patients by not having that information.

11 In the end, when we approve a regimen,

12 what we need to do is be able to provide labeling

13 that can be followed by clinicians and others to

14 manage patients. If, after deliberation, you

15 believe that labeling cannot be written which can

16 overcome some of these limitations that you are

17 identifying, then it would be good if you were to

18 let us know that so that we can then proceed

19 accordingly.

20 DR. HUNSICKER: My shy partner over here

21 who is the representative of the public interest

22 has shoved over to me just the single datum that

23 currently on the UNOS renal waiting list,

24 African-Americans constitute 35 percent of the

25 population.

138

1 DR. ENGLUND: Dr. Shapiro?

2 DR. SHAPIRO: Can I ask a corollary

3 question. The pivotal trial here is entirely

4 non-USA patients, the 310. What are the

5 implications of that in terms of approving a change

6 in the labeling for USA patients?

7 DR. ALBRECHT: The regulations do allow

8 the FDA to take into consideration data from

9 foreign trials when making a decision about

10 marketing and approving a drug product. However,

11 the caveats to that are that the foreign data are

12 of the quality and caliber that would be requested

13 to be provided from US patients in addition to

14 which the results of such studies must be

15 applicable to populations within the United States.

16 If those parameters are met, then we are

17 to consider foreign data in making a decision.

18 DR. ENGLUND: Dr. Johnson.

19 DR. JOHNSON: I have another question

20 about the labeling. What are the federal

21 limitations on what the label can say in respect to

22 ethnic populations? Is there such a thing?

23 DR. ALBRECHT: Are you asking whether, if

24 there is an absence of data, we can put such

25 information into the package insert?

139

1 DR. JOHNSON: I guess that is my question.

2 DR. ALBRECHT: Just wanted to make sure.

3 Again, we can put into the labeling information

4 that factually reflects studies that were conducted

5 and the results from such studies with the caveat

6 that such labeling should then be able to direct

7 physicians to properly use the drug in managing the

8 patients that they would encounter in their

9 practice.

10 Again, to follow up Dr. Hunsicker's

11 question, we will look to you to give us guidance

12 on whether the absence of certain subsets of the

13 population are such that they would actually

14 preclude clinicians being able to effectively use a

15 particular drug regimen.

16 DR. JOHNSON: I guess my question is a

17 little bit more to the point and that is I am not

18 really asking whether or not somewhere within the

19 insert that we can place that, "This drug was not

20 studied in the subpopulation." I guess what I am

21 asking specifically in the labeling statement, can

22 we have limitations upon which groups this drug

23 should be approved for for the current labeling

24 indications.

25 DR. ALBRECHT: I think the short answer is

140

1 yes.

2 DR. ENGLUND: With that, we have a little

3 bit of time. I really think now would be a good

4 time to go back. We have such good pharmacologic

5 expertise on the panel and with Wyeth-Ayerst.

6 Perhaps, if you would like to, Dr. Abernethy, just

7 rephrase briefly your one sentence and we could

8 have a response from the company.

9 DR. ABERNETHY: I think the issue is that,

10 with the data from these two studies presented, we

11 really didn't see good data suggesting that a

12 better outcome could be obtained by bracketing

13 concentration ranges. If that data is absent, then

14 the clinician part of my says it is easy. If there

15 is a question, you just give a higher dose because

16 there is no toxicity to pay for that.

17 In the FDA presentation, there was some

18 data from historical studies that did suggest some

19 dose relationship to some of the side effects. I

20 am just trying to get a feel because the data we

21 are seeing here is at a higher concentration range

22 than any of the stuff that that came from.

23 DR. BURKE: I am Jim Burke with

24 Wyeth-Ayerst Research. I have a slide coming up.

25 [Slide.]

141

1 This is a slide of the PK/PD analysis

2 during the first 75 days following transplantation.

3 It is up to 75 days. We looked at all the

4 different possible explanatory factors that could

5 lead to rejection.

6 Here is a simplified diagram showing only

7 the effect of cyclosporine and sirolimus. So one

8 can see that, indeed, there is a concentration

9 effect between the concentrations of cyclosporine

10 and the concentrations of sirolimus in outcome.

11 This was done in all patients during the

12 first 75 days. So we have 525 patients in a fairly

13 large range of concentrations. If one looks at the

14 data after randomization and one wants to look at

15 those that went on to Rapamune therapy, the number

16 of acute rejections have gone down considerably and

17 also the sample size has gone down to 215 patients.

18 So the power of doing an analysis of the

19 relationship between effect and concentration after

20 randomization is limited by those factors. Indeed,

21 one should remember that we only studied a single

22 concentration range after randomization. Although

23 you have a few outliers, you should consider all of

24 the outcome as part of the population.

25 So we defined the concentration as the

142

1 distribution of the concentrations in that

2 population that was studied. Could we have used

3 higher concentrations? Should you worry about

4 higher concentrations? For that, I think you

5 should go back to two earlier studies that were

6 done, studies 207 and 210.

7 We started off on concentrations targeted

8 at a mean of 30 milligrams per milliliter in the

9 first two months. In those studies, although the

10 overall safety and efficacy was acceptable, if one

11 looks at toxicities at those higher concentrations,

12 cholesterol, triglycerides, hypokalemia, they were

13 considered unacceptable for chronic maintenance.

14 So when we designed study 310, we had

15 those data available so we chose a lower range of

16 concentration rather than retesting a higher

17 concentration where we had observed toxicities.

18 DR. HUNSICKER: My recollection is that

19 there was a slide shown, I think at the end of the

20 pharmacokinetic section, which dealt with the

21 values of sirolimus levels that were observed and

22 what would have been observed if there had not been

23 dose correction. That showed predominantly that

24 there was an excess--the imputed, the presumed,

25 levels would have been higher. There were very few

143

1 lower levels.

2 That is my recollection of that study;

3 that is to say, using the non-dose-adjusted thing,

4 you had very few people who were below the lower

5 limits.

6 DR. BURKE: What we have heard now is that

7 there is a weak relationship between the sirolimus

8 levels above that and toxicities. There is

9 probably some but we haven't seen strong

10 relationships. So the argument from your data that

11 you present, as I see it, is that the advantage of

12 the dose monitoring is primarily to avoid

13 excessively high doses for which we don't have very

14 much toxicity demonstrated to us as opposed

15 to--this is the slide over here--the possibility of

16 having excessive low levels which would be

17 associated with rejection.

18 I am aware of some things that I can't

19 cite to you because they are in the literature.

20 One was a regression in the earlier pivotal trials

21 of the actual achieved levels with rejection that

22 showed that people who were higher than, I guess it

23 was 8 or something like that, very rarely had

24 rejection episodes.

25 I believe that there are other data in the

144

1 literature that show, with low-dose cyclosporine,

2 that also there is a critical relationship between

3 the lower end, that you need to get above a certain

4 level to avoid rejection.

5 But the question, I think, that is being

6 implicitly put is whether we really are achieving

7 anything on the low end here with the TDM as

8 opposed to just simply avoiding the high end for

9 which we have not yet defined toxicities.

10 DR. BURKE: Certainly, this slide does

11 demonstrate the preference of doing therapeutic

12 drug monitoring over giving a fixed dose. If one

13 goes back to the toxicity and the data from the

14 previous studies, actually the concentration-effect

15 relationships on study 310 that I just showed you

16 were very similar to the pooled data analysis of

17 301 and 302.

18 So we have reproduced that. What is the

19 cutoff on the lower end? Well, in this early

20 period where we do have sufficient rejections and a

21 sufficient distribution of data, we were able to do

22 an analysis where we dichotomized the data based on

23 cutoffs of the lower end of recommended levels.

24 That was at 5 for sirolimus and 150 for

25 cyclosporine. Indeed, we do find that, if they are

145

1 below those levels, they have a significant

2 increase of the incidence of acute rejection. We

3 can do that during that early period. I will admit

4 that, in the later period, in the maintenance

5 period, we don't have sufficient evidence to do

6 that.

7 But I think the ranges that we are

8 recommending will avoid clinicians treating

9 patients with too low levels. We have seen that

10 there are a few additional rejections and we

11 certainly don't want to increase that number.

12 DR. HUNSICKER: Getting back to what is up

13 there, and I am going to throw in a little

14 bit--believe it or not, I take care of patients and

15 I also have noticed that sometimes the levels are

16 much lower than you expect. I have used sirolimus

17 levels to adjust that.

18 But what you have here is a predicted--the

19 range that you would get if you did TDM as opposed

20 to what you would have had had you used an

21 8-milligram fixed-dose regimen and you would make

22 the adjustments based on the proportionality of

23 dosing levels.

24 What you see is that, at the bottom level,

25 which is the risk for rejection where I think that

146

1 the data are fairly solid, there isn't a hell of a

2 lot of difference. What you are really seeing is

3 that you are avoiding higher levels with your drug

4 monitoring. That is where--at least, I have taken

5 the argument from that side of the table. There

6 isn't a hell of a lot of evidence that there is

7 much toxicity there.

8 It does bring in complexity. So the

9 question is does the avoidance of those higher

10 levels really justify the complexity of the issue.

11 DR. BURKE: I will go back and did see a

12 slide showing the relationship between

13 concentration and lipids and I think there is

14 another parameter during those earlier phase II

15 studies. You can put that up.

16 [Slide.]

17 To repeat the design of this study, we

18 compared cyclosporine direction to sirolimus from

19 the time of transplantation. There were about 40

20 patients in each group. As I say, the sirolimus

21 concentrations were targeted at 30 during the first

22 two months. After two months, the concentrations

23 were to be reduced to a target concentration of

24 about 15. You can see they are slightly higher

25 than that.

147

1 Let's take a look at this early period

2 when the concentrations are high, the average got

3 as high as 35. You can see, in the yellow, the

4 triglycerides that got up to over 4 millimole. I

5 think that is over 400 milligrams per deciliter.

6 Cholesterol; the average was up to 8, which is--I

7 am trying to convert that. That is about 300

8 milligrams per deciliter. So it would not be

9 reasonable to treat a population at those high

10 concentrations for a maintenance therapy.

11 When you see that the sirolimus

12 concentrations have been increased to levels very

13 similar to those they were recommending, a mean

14 slightly less than 20, you can see that there was

15 an improvement in these laboratory parameters.

16 Here I have shown two parameters. I could also

17 show others that are affected by sirolimus. This

18 is platelets.

19 So I think there was reasonably

20 justification in the study design not to study much

21 higher levels of concentration. Indeed, there is

22 reasonable evidence that we should put that in our

23 labeling today to avoid toxicities.

24 I have one more I will show you here, the

25 SGPT values.

148

1 [Slide.]

2 You can see, once again, higher levels in

3 the beginning and lower levels later when the

4 concentrations are decreased. It is not quite as

5 evident. I know they were very nice on these

6 platelets. So there is evidence for us to instruct

7 clinicians not to target very high levels.

8 On the lower end, to go back to the one

9 slide we showed, you saw, whether you had given it

10 on dose or whether you had given it on therapeutic

11 drug concentration, there are a number of values

12 that are low.

13 You have to realize that that presentation

14 is an intent-to-treat presentation, that it

15 includes data on patients, even those that

16 discontinued a few days after randomization and did

17 not have time to have their target concentrations

18 increased.

19 So it is an extremely vast population. If

20 one went out further, one would find very few

21 patients that are below what we are recommending.

22 So you shouldn't confuse that intent-to-treat

23 population with what patients are actually

24 receiving beyond six months, twelve months, and so

25 on.

149

1 DR. HUNSICKER: Let me just do one last

2 stab as sort of a provocateur here, the issue

3 having been raised. Then I am going to cede to the

4 pharmacologists who raised this question in the

5 first place.

6 I can imagine three policies. One is you

7 just give a fixed dose and you ignore what is

8 happening. The second is you give what you have

9 got, you would get therapeutic dose monitoring.

10 The third is that you give a fixed dose and, as

11 long as you stay out of trouble, you do what you

12 are doing and, if you find that you have got some

13 more toxicities, you go back and check your dose.

14 Or, if you find that you are having a rejection,

15 you recheck that dose.

16 What I am trying to get across is that I

17 am not sure that we need to absolutely, in the

18 indication, nail people to the requirement for this

19 kind of therapeutic monitoring. I think that it

20 might be sufficient to advise them that you can

21 have levels that are lower than you expect and

22 there is a lower level that you should be achieving

23 and that you can find out about this. Of you can

24 have toxicity and you can find out about the level

25 with a TDX or with whatever measurement you are

150

1 using, rather than require that it be done in every

2 case.

3 I think that--I am imputing to you what

4 your question was, but I think that is really the

5 issue that we are raising. We have to tie this to

6 therapeutic dose monitoring.

7 DR. SUTHANTHIRAN: May I make a point. My

8 question has been rephrased and I have been called

9 a pharmacologist. I don't find anything bad about

10 it, but the issue I was trying to make, I think

11 your first slide made the point that, when you use

12 different levels of sirolimus and different

13 concentrations of cyclosporine, if the sirolimus

14 concentration is high, you can reduce the incidence

15 of rejection even with the lower levels of

16 cyclosporine. There is a synergy between the lower

17 levels of cyclosporine and high trough levels of

18 sirolimus.

19 That point is very clear and you had

20 enough cases in the first three months. My concern

21 was, after the patient is randomized, when we

22 suggest certain levels, 15 to 25, there is really

23 not much data to support that 50 to 25 levels, in

24 fact, prevents acute rejection because the number

25 of patients who had acute rejection were in the 15

151

1 to 25 nanogram level. In fact, 16 out of 23

2 patients who had acute rejection were within this

3 suggested target.

4 It appears to me a higher target level may

5 be problematical from the toxicity perspective and

6 the current data doesn't tell us what is the actual

7 level we need to keep the patient at in order to

8 prevent an acute rejection episode.

9 I wonder whether we could, in fact, go a

10 little bit under the level. Maybe we will avoid

11 some of the toxicity and have the same therapeutic

12 benefit. This was the point I was trying to make,

13 whether there is any data you analyzed or the FDA

14 analyzed that tells us that a particular level of

15 sirolimus is therapeutic in terms of preventing an

16 episode of acute rejection.

17 DR. BURKE: The data that we do have is

18 simply the quartiles that we presented. We know

19 that, beyond a certain point, those 207 and 210

20 patients are now out to five or six years, about a

21 quarter of those patients. They haven't lost their

22 grafts. They haven't had an increase in their

23 creatinine. They haven't had a rejection.

24 That doesn't mean that additional work

25 does not need to be done, and this is always very

152

1 difficult when you are talking about long-term

2 outcome, how do you target levels. Indeed,

3 additional work probably needs to be done in that

4 early post randomization period, or after three

5 months, to learn how to better adjust those

6 concentrations.

7 So additional work does need to be done

8 but the evidence we have today does support the

9 concentrations that we are recommending.

10 DR. NEYLAN: I don't know if this would

11 help so I need to ask permission first. But we

12 have additional data for 310. As you know, this is

13 a five-year study. Most of these patients are now

14 approaching the three-year mark. So, on this issue

15 of the relationship between the suggested target

16 range and the incidence of acute rejection, we do

17 have data that is subsequent to the twelve-month

18 mark on rejection frequency in these randomized

19 arms.

20 I will again remind you that the

21 randomized arm in 310 to the Rapamune maintenance

22 therapy was downregulated in the Rapamune exposure

23 to approximately the range that we are suggesting

24 today.

25 So the question is, first, would that data

153

1 be of any use in addressing your question and, if

2 so, would we be allowed to show it.

3 DR. SUTHANTHIRAN: I think so. If you can

4 show that patients who are kept at the levels you

5 suggest had a lesser incidence of acute rejection

6 subsequently compared to patients who had lower

7 than that level, I think it will support the idea

8 that keeping the sirolimus at a particular level

9 would be of benefit.

10 DR. ENGLUND: Yes; if you are going to be

11 showing levels and rejection after the twelve-month

12 period.

13 DR. NEYLAN: Let me show you, then, the

14 trough levels first.

15 DR. ENGLUND: Wait. I think we need to

16 hear from the division.

17 DR. NEYLAN: Oh; I'm sorry.

18 DR. ALBRECHT: I just wanted to comment.

19 I don't believe that information has been submitted

20 to the FDA for our review.

21 DR. NEYLAN: No; it hasn't.

22 DR. ALBRECHT: So we would be hearing your

23 viewpoint, but we could not comment on it from the

24 division.

25 DR. ENGLUND: Are we allowed to see it?

154

1 DR. ALBRECHT: Yes.

2 DR. HUNSICKER: Can they show it is the

3 question.

4 DR. ALBRECHT: Having said what we said,

5 certainly you can show it.

6 DR. NEYLAN: Do I have permission to show

7 it? First, let's see the rejection slide. Then we

8 will go back to that slide.

9 [Slide.]

10 This is the follow up then beyond the

11 twelve-month mark onto 24 months for study 310.

12 What we have seen in that, after the twelve-month

13 mark, there have been no rejections in the Rapamune

14 maintenance group and only two rejections in the

15 Rapamune plus cyclosporine group.

16 The Rapamune maintenance group, again, is

17 a group of patients that are receiving Rapamune

18 doses at the suggested target range. I should also

19 comment here that there were a handful of

20 rejections seen in both of these groups at the

21 twelve-month mark because of protocol biopsies.

22 If we could go to the next slide.

23 DR. HUNSICKER: Were those protocol biopsy

24 rejections clinically manifest?

25 DR. NEYLAN: No; they were not.

155

1 DR. HUNSICKER: So we don't even know they

2 are rejections other than by histological criteria.

3 DR. NEYLAN: Right. Exactly so.

4 DR. HUNSICKER: Just so that some of the

5 nonnephrology and nontransplant people are aware of

6 that, there has been a lot of debate about what

7 "rejection" on histology means. There has been a

8 lot of debate about the meaning of rejection found

9 on histology without clinical correlates.

10 I don't take a side on that but I think

11 that does put a very different picture on that

12 little cluster of rejections that happens, if they

13 are not clinically manifest but simply the

14 consequence of protocol biopsies. It is not ever

15 clear that they are rejection.

16 DR. NEYLAN: Right. But, again, let me

17 emphasize the point that, at the twelve- to

18 24-month mark, there were no subsequent rejections

19 in the Rapamune maintenance group. This group was

20 receiving, now, on average, 6 milligrams of

21 Rapamune today and maintaining mean sirolimus

22 trough concentrations as measured either by the MS

23 or by the immunoassay within this suggested target

24 range today.

25 So, again, I just wanted to add that in

156

1 case it sheds any additional light on the

2 discussion.

3 DR. CAVAILLE-COLL: May I ask a question,

4 since we have not seen this data. The previous

5 slide, please, that graph.

6 [Slide.]

7 Does this represent all patients

8 randomized or does this just represent those

9 patients who are still on study therapy at up to

10 month 24 and, if so, what proportion are still on

11 study therapy at month 24?

12 DR. NEYLAN: Jim, since you have access to

13 the 310.

14 DR. BURKE: This is all randomized

15 patients so that we are counting 215 patients in

16 both groups. The number of patients on therapy is

17 nearly identical, 145 and 146.

18 DR. CAVAILLE-COLL: Thank you.

19 DR. ENGLUND: Dr. Ebert?

20 DR. EBERT: Another question that relates

21 to these two graphs that I have, the second graph

22 that you showed I believe showed the mean

23 concentrations over time. But I am assuming there

24 was probably a pretty wide variation in the

25 concentrations over a given period of time.

157

1 I think this really relates to my

2 questions about what was your strategy for dosing

3 and adjusting doses after randomization and did

4 you, in fact, perhaps, have--and I don't know if

5 you did or not, but did you have a group where

6 maybe the adjustment took longer, you had a longer

7 period of time where concentrations were low and

8 whether that early adjustment period might have

9 contributed to the fact that you saw rejections

10 early on in the trial.

11 If you went back to that three-line graph

12 with the cyclosporine and the sirolimus

13 concentrations, as you start to drop off on your

14 cyclosporine concentrations, you do somewhat

15 compensate by increasing the sirolimus

16 concentrations, but I am not sure if you do that

17 completely.

18 So, the bottom line is I am wondering if

19 maybe just not being aggressive enough early on may

20 have contributed to some of the rejections that you

21 saw.

22 DR. NEYLAN: If we could show the core

23 slide from the pharmacokinetics showing the

24 divergence of cyclosporine taper and sirolimus

25 concentration ranges. Yes; this slide.

158

1 [Slide.]

2 This is the slide I believe you were

3 referring to that shows the overlap period in which

4 the cyclosporine is coming down. These are the

5 mean trough levels of cyclosporine for the group

6 and the sirolimus concentrations are coming up and

7 are, at this point, just entering into the target

8 range.

9 Yes; there is a window of time here in

10 which that overlap is occurring and it is at least

11 possible, from a clinician's standpoint, that some

12 of these patients may have been experiencing

13 rejection because there was, at the time, a

14 relative decrease in net immunosuppression.

15 We have those two studies which both

16 sought, at a time point post-transplant, to have

17 clinicians change these two important variables in

18 the immunosuppressive regimen. Both of these

19 studies were somewhat groundbreaking. So I think

20 it is not surprising that clinicians were

21 exhibiting some degree of caution in making these

22 changes.

23 I believe that, as this is better

24 understood, that the rapidity of this change can be

25 improved upon.

159

1 DR. ENGLUND: One more question?

2 DR. AUCHINCLOSS: It is actually a subject

3 that I want to come back to this afternoon at some

4 length, but if you could just put up D10. There is

5 all this talk about how we are changing multiple

6 drugs at the same time, but that wasn't true in

7 study 212, was it? They were already, from day 10,

8 on high-dose sirolimus.

9 When they withdraw their cyclosporine in

10 the withdrawal group, that is a month or two later;

11 right?

12 DR. NEYLAN: That's correct. The only

13 difference is the target range of the sirolimus.

14 DR. AUCHINCLOSS: Oh; I understand. It is

15 a slightly lower target range.

16 DR. NEYLAN: Which was slightly lower.

17 When you adjust that for HPLC--

18 DR. AUCHINCLOSS: But there is only one

19 adjustment at the time of cyclosporine withdrawal

20 in this group of patients.

21 DR. NEYLAN: That's correct.

22 DR. AUCHINCLOSS: The other thing that I

23 didn't understand, and this is what I want to talk

24 about this afternoon, is that these two groups are

25 completely different from early on. The top group,

160

1 that never had cyclosporine withdrawn, was the

2 low-dose sirolimus and moderately high-dose

3 cyclosporine whereas the group that eventually gets

4 withdrawn is the low-dose cyclosporine from the

5 beginning with high-dose sirolimus from the

6 beginning; right?

7 So there is no comparison that you can

8 make between these two groups when it comes time

9 for the cyclosporine withdrawal in group No. 2.

10 Events have already happened in the group above,

11 and we will look at that this afternoon, that are

12 completely separate from what--that don't have

13 anything to do with cyclosporine withdrawal.

14 So I am interesting in looking at what

15 happens in the second group, the

16 cyclosporine-withdrawal group. I can only compare

17 what has happened up until that time in that group

18 with what happens to it afterwards. It is a very

19 strange trial design.

20 DR. NEYLAN: You are right in pointing out

21 that the phase II trial, 212, was asking a slightly

22 different question than the pivotal trial upon

23 which, obviously, the bulk of this indication is

24 resting.

25 This question specifically about whether,

161

1 right from the beginning, lower exposures to

2 cyclosporine coupled with the combination of a

3 concentration-controlled use of Rapamune might be

4 beneficial was one of the questions that was being

5 asked by this study.

6 DR. AUCHINCLOSS: If you put up the E21

7 results, it looked to me like you got a great

8 protocol there.

9 DR. NEYLAN: If you are about to show the

10 rejection rates--is that what this is? Yes.

11 DR. AUCHINCLOSS: At the time that you

12 came to cyclosporine withdrawal, you have got a 6

13 percent rate of accumulated rejections.

14 [Slide.]

15 What you did, when you showed these

16 results, is you compared the cyclosporine arm to

17 the red arm and you said, "Gee; you know it all

18 comes out the same." The red arm was bad to begin

19 with, or certainly less good. What I see when I

20 look at that slide, is you have a 6 percent rate of

21 rejection up until the moment of cyclosporine

22 withdrawal and now, suddenly, you are 20 percent

23 within six months afterwards.

24 I think you get 10 to 15 percent

25 acute-rejection rates when you withdraw

162

1 cyclosporine. Don't look at the red bar. Just

2 look at blue bar. That is what happens when you

3 withdraw cyclosporine.

4 What I find most amazing is that the

5 levels of cyclosporine at the time of withdrawal

6 were only 100 to 150.

7 DR. NEYLAN: Right.

8 DR. AUCHINCLOSS: You have got a fantastic

9 synergy. Why do you want to tell people to

10 withdraw cyclosporine? Tell them to go to low-dose

11 cyclosporine.

12 DR. NEYLAN: What we are trying to do with

13 these two studies is basically define the margins,

14 if you will, of how to use cyclosporine and

15 sirolimus. On the one hand, we have the pivotal

16 trials--

17 DR. AUCHINCLOSS: And you have defined it.

18 DR. NEYLAN: On the one hand we have the

19 pivotal trials that were approved in '99.

20 DR. AUCHINCLOSS: Well, I think the

21 pivotal trial shows pretty clearly that you get a

22 10 to 15 percent acute-rejection hit if you

23 withdraw cyclosporine.

24 DR. HUNSICKER: I actually calculated the

25 difference and it is--well, we will do it later

163

1 this afternoon.

2 DR. AUCHINCLOSS: This one goes from 5 to

3 20. That one went from 10 to 20, something like

4 that.

5 DR. NEYLAN: What we have with these two

6 sets of trials is, on the one hand, with the

7 original trials, rejection rates that were in the

8 range of 15 to 20 percent and the potential

9 detrimental impact upon renal function when the

10 combination was used in relatively full dosage for

11 both in the long term.

12 On the other hand, we have now these sets

13 of studies which define, if you will, a different

14 limit where we can see similar rates of rejection,

15 in this case in the range of about 20 percent, and,

16 with that, the elimination of cyclosporine, a

17 vastly different outcome in terms of renal

18 function.

19 I think what you are suggesting is that

20 there may also be opportunities to explore

21 variations in between these two margins; that is,

22 the combination in some lower dose or

23 concentration-controlled mediated fashion, of both

24 of these drugs in a maintenance regimen. I

25 certainly would not discount that.

164

1 The goal, though, today is to convince you

2 that these two studies also represent a safe and

3 effective way to use Rapamune and that safe and

4 effective way is that, in fact, in many patients,

5 we can eliminate the calcineurin inhibitors.

6 DR. AUCHINCLOSS: There is no doubt about

7 that. Probably about 80 percent of them, maybe

8 even 90 percent, of them. But you have portrayed

9 to us, and you intend to portray in the intended

10 labeling, the notion that there is not going to be

11 any increase in acute rejection. To me, your data

12 strongly indicate otherwise, that you will, in

13 10 percent of your patients, pay a price with an

14 acute-rejection episode that wouldn't have occurred

15 otherwise.

16 DR. NEYLAN: I would not want to argue

17 with you that there is not an incremental increase

18 in rejection.

19 DR. AUCHINCLOSS: Shouldn't that go into a

20 labeling change, that when you consider

21 cyclosporine withdrawal, it is quite likely that

22 there is a 10 percent or some finite risk, some

23 measurable risk, to your patient population?

24 DR. NEYLAN: I am reasonably confident

25 that, when all of this gets to the stage of

165

1 labeling discussion, that the data will be a part

2 of that label. The data clearly demonstrates that,

3 in fact, that incremental increase is there, yes.

4 One other--

5 DR. ENGLUND: Final sentence, or

6 sentences.

7 DR. NEYLAN: I was just going to--very

8 quickly, then, if we could show this next slide.

9 [Slide.]

10 I was just going to raise the point that,

11 even with lower doses of cyclosporine, in

12 combination, there is potentially a penalty to pay

13 in terms of renal function. This is a study that

14 was done in psoriatic patients, so non-transplant

15 patients. It looks at mean creatinine over a

16 period of treatment in which these patients either

17 received cyclosporine at relatively conventional

18 doses for transplantation or received sirolimus as

19 monotherapy.

20 The middle group is a group receiving

21 low-dose cyclosporine and this same dose of

22 sirolimus. You can see the spectrum of renal

23 function.

24 DR. AUCHINCLOSS: I agree with you. I

25 know you want to go to lunch, so save E29 for me.

166

1 We will come back to that this afternoon.

2 DR. ENGLUND: Good. We are going to break

3 now for lunch.

4 [Whereupon, at 12:15 p.m., the proceedings

5 were recessed to be resumed at 1:10 p.m.]

167

1 A F T E R N O O N P R O C E E D I N G S

2 [1:15 p.m.]

3 DR. ENGLUND: We are now back from lunch.

4 I would now like to open the meeting for the open

5 public hearing. We have one registered speaker who

6 is going be talking to us, Dr. Alan Wilkinson. He

7 has some slides, too.

8 Open Public Hearing

9 DR. WILKINSON: I didn't realize I was, in

10 fact, the entire joint public but I am pleased to

11 be there and I would like to commend the

12 presentations teams on the thoroughness of the

13 presentation.

14 I am here really to provide both, I

15 suppose, an experienced and a naive viewpoint on

16 the studies. I am a nephrologist, transplant

17 nephrologist, at UCLA. I am here in part as a

18 consultant to Novartis and they have paid for my

19 trip here.

20 I have also done studies for all of the

21 companies that make immunosuppressant drugs and

22 have, in fact, lectured and received honoraria for

23 speaking for both Novartis, Wyeth-Ayerst, Fugisawa,

24 Abbott. I don't think I am too selective in my--

25 [Slide.]

168

1 What I wanted to talk about was my

2 perception of where study 301 stands in terms of

3 what we do, and also where we stand as transplant

4 physicians with regard to cyclosporine and

5 withdrawing cyclosporine. I know that when John

6 presented the data, he used still the half-life of

7 transplants for about ten years.

8 I think it is true, but I think we just

9 need to remind ourselves of this paper from Harry

10 Hiriharan that appeared in the New England Journal

11 of Medicine where, if you took out people who had

12 died--and, of course, we include death as an

13 endpoint in many of these things and that is not

14 necessarily fair to the transplanted organ.

15 If you took out people who died and looked

16 at living donors, the recipients of living donors,

17 then the half-life is approaching forty years.

18 This is in a calcineurin-inhibitor-rich

19 environment. For cadaveric transplants, where the

20 donor characteristics, of course, are less certain

21 and there is pre-death injury presumably that we

22 think affects the kidney, even in kidneys that are

23 set up to be very subject to the effects of

24 calcineurin inhibitors, even there, the half-life

25 is approaching twenty years.

169

1 This is the UNOS data that was used. The

2 USRDS data is, perhaps, a little less optimistic

3 than that. But I think we have to accept that,

4 during the calcineurin-inhibitor period, we have

5 improved transplant survival dramatically. That

6 isn't to say that the TOR inhibitors, Rapamune and

7 potentially Certicam are not advances in what we

8 do, but I think we have to place them in context of

9 where we are coming from.

10 I wanted just to start off with saying, in

11 addition, that I am not somebody who is particular

12 in favor of using calcineurin inhibitors in high

13 dose. I have written quite extensively on the

14 effects of calcineurin inhibitors, or rather, on

15 renal dysfunction in recipients of heart and liver

16 transplants and, in fact, have just done a big

17 review on liver-transplant recipients and renal

18 function and dysfunction in those patients, large

19 parts of which are, of course, due to calcineurin

20 inhibitors. Some of it is due to injuries to the

21 kidneys separate from that in liver recipients.

22 But, certainly, I am not in favor of

23 keeping calcineurin inhibitors there if we can

24 avoid having them. I also wanted to talk a little

25 bit before I went further on sort of the power and

170

1 authority of this committee before us here today.

2 I think it is true that the committee here has

3 enormous power in terms of deciding what drugs are

4 approved and how they are used to some extent.

5 But I think the labeling confers authority

6 on the usage of drugs which goes beyond, in a

7 sense, the power of committee. So, if you, as a

8 committee, say that a drug should be used in a

9 different way, that confers authority on that usage

10 and, to some extent, we have to look at your

11 fairness to the producer of the drug, in this case,

12 Wyeth. Is it fair? Is the data they are bringing

13 to you such that it is fair to them to change the

14 labeling.

15 But, at the same time, I think you have to

16 be fair to both physicians and patients in this and

17 make sure that labeling doesn't put physicians,

18 particularly, in a difficult circumstance when they

19 choose to use different protocols in patients

20 because, if we have labeling that says that, for

21 example, the use of cyclosporine with Rapamune

22 beyond three months in low-risk patients is

23 something that is not recommend, if we continue to

24 do that, that, to some extent, I think, puts us at

25 some risk.

171

1 So I think we have to be very careful as

2 you make determinations about labeling what impact

3 that has on clinical practice or what impact that

4 has on standard of care and what impact that has on

5 the legal liability of physicians who are

6 prescribing these drugs.

7 Remember that you have approved sirolimus

8 for use with cyclosporine and prednisone.

9 Sirolimus is used in large numbers of patients with

10 tacrolimus. Although you are debating today

11 whether, in fact, it is feasible to withdraw

12 cyclosporine from patients on sirolimus, there are

13 many patients out there on whom that has already

14 been done in circumstances where physicians thought

15 that was a sensible thing to do.

16 So, really, what you are looking at here

17 is a trial which has addressed that. But we have

18 to remember what clinical practice is achieving in

19 the community and remember that the labeling of

20 drugs and their usage are, in a sense, two separate

21 things, whether the FDA likes that or not. But I

22 would like to believe that the labeling of drugs

23 should make it as simple as possible for the

24 prescribers within the safety of those agents.

25 I also think the question before you here

172

1 today is different from the question before the

2 European committee that addressed this issue

3 because, in that case, they had actually refused,

4 and I thought it was the wrong decision--they

5 refused to approve sirolimus when it was first

6 presented to them and then only approved it when it

7 was presented to them with the improvement in renal

8 function.

9 I think that the analysis actually

10 misconstrued what was shown by the study, by the

11 withdrawal study, because one comment they made in

12 their scientific analysis of that data was that

13 they recommended that sirolimus not be used with

14 cyclosporine because there was evidence of additive

15 nephrotoxicity when the two were used together.

16 As it happens in that study, there is no

17 arm which shows whether there is additive toxicity

18 when you have cyclosporine and sirolimus used

19 together. If you had had an arm in that study

20 where you had actually withdrawn sirolimus, you

21 might have shown that. But you don't actually have

22 that to show in that study.

23 If we go back to the 301 and 302 studies

24 and look at the comparator arms in both of those

25 studies, the GFRs in the comparator arms--in the

173

1 American study, azathioprine was used. In the

2 European study a placebo was used.

3 But if you look at the GFRs at twelve

4 months in the control arms of both those studies,

5 they are as robust as the GFRs in the

6 Rapamune-withdrawal study before you today. So, I

7 think when we look at GFR and look at outcome, we

8 have to be very careful not to jump from GFR to a

9 recommendation about the usage of drugs.

10 I don't think any of us would go back to

11 say that the correct protocol to use today is

12 cyclosporine, prednisone and azathioprine. I think

13 there would be few people who would argue for that

14 although many centers may still be doing that. So

15 I think that is an important thing to recognize.

16 I also think if you look at the change in

17 GFR--let me get to that in a moment. Can you move

18 on one?

19 [Slide.]

20 The other thing which I think is important

21 in all the data, and Dr. Hunsicker, I am sure, will

22 talk to this at length later this afternoon, is

23 that rejection is one of the best predictors of a

24 less-good long-term outcome. If you look at the

25 patients in whom the half-life has improved, it is

174

1 those patients who have not had a rejection. So

2 rejection is a very profound effector of long-term

3 graft function.

4 We shouldn't trivialize that, I think. I

5 know, in this study, it didn't reach statistical

6 significance. But we should not trivialize the

7 effect of rejection on long-term graft outcome.

8 Remember, for each patient, their graft is the only

9 one. In these venues, we discuss large trials and

10 lots of numbers but, for each patient, their graft

11 is the only one.

12 The other thing which may be addressed

13 later is the predictability using the serum

14 creatinine at one year or at some time period in

15 terms of long-term graft function. I would like to

16 remind you that that data holds best for patients

17 that were on calcineurin inhibitors because that is

18 the population in which that study was done.

19 [Slide.]

20 We have no really good data long-term in

21 these studies. So I think my concerns are that we

22 don't really know the effects of late acute

23 rejection in this group yet. The data is still

24 very early. Even the two-year data is still early

25 compared to the long-term data.

175

1 The improved renal function certainly is

2 there but, in any study in which you withdraw

3 cyclosporine, you are going to get improved renal

4 function. In fact, the delta GFR in this study is,

5 perhaps, surprisingly small. If you look back at

6 some of the old studies done by Curtis and Luke and

7 some other studies, they had bigger improvements in

8 renal function when they switched from

9 cyclosporine-prednisone to prednisone-azathioprine

10 which suggests that the effect of cyclosporine at

11 this point is less than it maybe was in those

12 studies.

13 The other thing which I think we should

14 realize is that the patients who had rejection, if

15 we look at their renal function subsequent to

16 rejection, it was brought down to a greater extent

17 than the patients who were on the

18 cyclosporine-sirolimus arm, that the end result for

19 the two groups was equivalent but the starting

20 point was actually better for the Rapamune group.

21 So the effect of rejection in patients--I

22 know it is small numbers but we are, in fact,

23 arguing from small numbers, the effect of rejection

24 was greater in those patients on sirolimus and

25 prednisone only.

176

1 The other thing which I think is important

2 in that data is that the GFR in the

3 cyclosporine-prednisone-treated patients was, in

4 fact, stable, that there was no decline. So when

5 we talk about additive toxicity and progressive

6 toxicity in those patients who were kept on

7 cyclosporine, there was no proof of that in that

8 study.

9 The GFRs were certainly lower. We would

10 expect that in patients treated with cyclosporine.

11 We don't know if those patients were taken off

12 cyclosporine now at two years whether, in fact,

13 their GFRs would improve to the same extent and

14 that they would have GFRs equivalent to those

15 patients maintained on cyclosporine because the

16 effect on GFR of cyclosporine is, of course,

17 twofold.

18 There is the hemodynamic effect of

19 cyclosporine which affects the flow of blood into

20 the glomerulus, the afferent arteriolic

21 constriction, so the pressure in the glomerulus is

22 reduced. I am going to show a slide at end of a

23 blood-pressure study which is interesting at this

24 context.

25 So cyclosporine has an effect on the

177

1 glomerulus by affecting flow in, and cyclosporine

2 also has an effect because of its tissue toxicities

3 which the experts on this committee are on as well.

4 So what we don't know in the study is whether the

5 continued reduction in GFR compared to the

6 sirolimus group is, in fact, occasioned by injury

7 to the kidney or whether it is occasioned just by

8 perpetuation of the hemodynamic effect of

9 cyclosporine.

10 You might even argue, and I have actually

11 wondered about this for the TOR inhibitors,

12 whether, because they affect intimal hyperplasia,

13 perhaps reduce that, and whether they, in fact,

14 might be protective against some of the fibrosis we

15 see so that a combination of a TOR inhibitor and a

16 calcineurin inhibitor might actually mitigate some

17 of the long-term toxicities even though, when you

18 just look at the GFR and the creatinines, that may

19 not, at first blush, be apparent.

20 So I think we just don't know that data

21 and, for that reason, I am anxious about us moving

22 along too fast. So I think the relationship you

23 have between renal function at a given time and

24 long-term outcomes, we don't know. I have covered

25 my concern that labeling shouldn't be too directly

178

1 prescriptive, that it should allow us a great deal

2 of freedom in using these drugs.

3 [Slide.]

4 The other issue I think before us is that,

5 because of the way studies are done, the comparator

6 drug here is cyclosporine. That is not the only

7 calcineurin inhibitor. The FDA would rule here on

8 one agent within a class of drugs. I think that,

9 to me, again, is not something I would like to see

10 done because we don't have comparable data using

11 tacrolimus. There are many people, I think, right

12 across this room who, I think, have favored

13 tacrolimus over cyclosporine and who believe that

14 you can, very effectively, use low-dose

15 cyclosporine and TOR inhibitor regimens to achieve

16 excellent outcomes.

17 Of course, these studies, too, don't

18 always include an anti-R2 inhibitor and the

19 rejection rates on those studies are very low and

20 the increase in rejection in this study may be

21 unacceptable in that context.

22 A lot of the discussion here I think is

23 reverberating now about where you can or couldn't

24 discontinue cyclosporine. I would be concerned if

25 every transplant nephrologist and surgeon in this

179

1 country did not know the data that we have

2 presented here today. I would be dismayed if

3 people were making adjustments to immunosuppression

4 and yet didn't know this data.

5 It has been published. It ought to be

6 known. So I don't think there is any question that

7 this ought to be known by people changing the doses

8 and the way in which we use drugs.

9 But I, for example, am an African. I

10 don't look like an African at first sight but, in

11 one definition, I am an African. I was born in

12 South Africa. When I get my citizenship, I will be

13 an African-American. To some extent, the decision

14 as to whether or not you are African-American or

15 not is your own decision.

16 There is also, in a sense, the prejudicial

17 decision in this country of who is and who isn't an

18 African-American. I am a South African and so I am

19 very sensitive to these issues. At the height of

20 apartheid in South Africa, if you did HLA typing

21 and looked at genetic mix within the white

22 Africaner race, about 40 percent of them showed

23 evidence of African parentage.

24 So when we talk about subgroups and

25 cleanly dividing subgroups of patients up so it is

180

1 safe in this group, it is not safe in that group, I

2 think we have to be very careful in what we are

3 doing.

4 I wanted, also, just to remind you of the

5 steroid-withdrawal studies where we have had

6 studies that have looked quite good in the short

7 term where the five-year data, perhaps, doesn't

8 look quite as good. So, again, I think we have to

9 be careful.

10 I would also like to just mention again

11 the potential cost. You have to use considerably

12 more Rapamune to get an adequate level when you

13 take cyclosporine away. Of course, you don't have

14 to pay for the cyclosporine anymore.

15 [Slide.]

16 Then, finally, if I could just show you

17 one last slide, just to go back to the GFR, I

18 wanted to show you this slide because I like to

19 think of kidney transplants as, in every patient

20 with a kidney transplant, to some extent, there is

21 some renal, chronic kidney, disease. I think we

22 can presume that most kidney transplants have had

23 some injury.

24 If you look at how we treat patients these

25 days with chronic kidney disease, particularly

181

1 patients with proteinuria, the recommendation is

2 that we use ACE inhibitors aggressively. We use

3 ACE inhibitors aggressively even though we know

4 that the GFR falls. The GFR falls, not because you

5 are doing anything to the afferent arteriole

6 leading into the glomerulus, but because you are

7 opening up the efferent arteriole.

8 But the net effect is a reduction in

9 glomerular pressure. Now, the other effects of ACE

10 inhibitors, I am not going to get into that in too

11 great detail here, but in all the metaanalyses of

12 the protection of kidneys in patients with chronic

13 kidney disease, the dihydropyridine, the nifedipine

14 family, has been shown to be less good in

15 protecting kidneys than ACE inhibitors. The

16 reduction in proteinuria and the maintenance of GFR

17 has been less good.

18 The title of this paper was Sustained

19 Increase in Glomerular Filtration Rate in Kidney

20 Transplant Patients with Hypertension Treated with

21 Nifedipine. You can see here--unfortunately, the

22 baseline was post treatment so they don't actually

23 have a baseline before they were put on nifedipine.

24 But nifedipine is a calcium channel

25 blocker and the argument for why this was good was

182

1 that it counteracted some of the afferent

2 construction of cyclosporine. These patients were

3 treated with cyclosporine and azathioprine and

4 prednisone.

5 When placed on nifedipine, the GFR rose

6 over twelve months to 56 compared to 46 where as

7 those on lisinopril, an ACE inhibitor, remained the

8 same. The take-home message that the authors put

9 into this paper that, therefore, we should be

10 treating patients with hypertension who have renal

11 transplants with nifedipine and not with ACE

12 inhibitors because the GFR is better.

13 In fact, in this presentation today, there

14 has been discussion about the lower blood pressures

15 in patients on sirolimus. But patients on

16 sirolimus don't have the afferent construction that

17 cyclosporine confers on the patients we give it to.

18 When you treat somebody with the

19 dihydropyridine for blood pressure, you lower the

20 blood pressure, but you also open up the afferent

21 arteriole. So, if you don't drop the mean arterial

22 pressures sufficiently, the actual pressure

23 reflected on the glomerulus may actually be higher

24 than it was when the afferent arteriole was

25 constructed and the mean arterial pressure was

183

1 higher.

2 So we don't know if you have got a

3 slightly lower blood pressure, not at the target

4 level we would recommend now for patients with

5 kidney disease, a slightly lower systemic blood

6 pressure, mean arterial blood pressure, but a

7 wide-open afferent arteriole, whether, long-term,

8 that will be good or bad for the kidneys.

9 That is true for this study, to some

10 extent, and it is true for the sirolimus studies as

11 well. Over the short term, it certainly looks

12 good. The GFRs are higher.

13 There is also a paper recently published

14 in the Journal of Urology I wanted to bring to

15 committee's attention, and that was a paper that

16 looked at the long-term GFRs in transplant donors.

17 It was a patient that had actually twenty years,

18 so, of course, much longer than this. But the GFRs

19 in those patients were actually, for the men, I

20 think roughly 73. Corrected for age, they ran at

21 about 68 to 67.

22 So the GFRs we are achieving with

23 sirolimus and with azathioprine and with placebo

24 were actually almost as good as you can get with a

25 single kidney. You have a mild reduction in the

184

1 GFR with cyclosporine, that's true. But, provided

2 the calcineurin inhibitors are not actually

3 injuring the kidney over long-term, and we don't

4 know that yet. I am not pretending we know that.

5 But, with low doses, it may be that we could

6 successfully use both combination of calcineurin

7 inhibitors and the TOR inhibitors and actually

8 achieve long-term GFRs which are very good,

9 long-term creatinines that are very good.

10 So, if you could go back one.

11 [Slide.]

12 I just wanted to say we have, in fact,

13 many studies now that are being published and are

14 underway looking at combinations of either

15 sirolimus or certicam with low-dose cyclosporine or

16 tacrolimus in which the outcomes, in terms of

17 rejection, are very good and which the outcomes in

18 renal function appear to be better than when the

19 higher doses of calcineurin inhibitor were used.

20 The doses of calcineurin inhibitor in

21 these studies, which are called low-dose, are

22 actually still quite high-dose in the context of

23 those studies. I think there was a question

24 earlier about that in terms of what we do.

25 I think I would be hesitant at this point

185

1 with what we know from what is front of us today

2 to, in a sense, change the prescription boundaries

3 of this drug to an extent beyond which I think the

4 current evidence actually allows us to do.

5 Thank you very much.

6 DR. ENGLUND: Thank you.

7 For the committee, are there any questions

8 regarding this presentation?

9 For the sponsor, any comments or

10 questions?

11 DR. NEYLAN: No.

12 DR. ENGLUND: Are there any other speakers

13 that wanted to say anything at this point in

14 time--not from the table. At this point in time,

15 then, I would like to close the Open Public Hearing

16 and I would like to ask Dr. Albrecht to give us the

17 charge.

18 Charge to the Committee

19 DR. ALBRECHT: We would like to ask you to

20 discuss three questions, and specifically to vote

21 on the first one. So, while we are waiting for the

22 slide to go up, let me go ahead and start the first

23 question.

24 [Slide.]

25 Do the data presented support the

186

1 effectiveness or efficacy and safety of

2 cyclosporine withdrawal and

3 concentration-controlled sirolimus two to four

4 months after kidney transplantation in patients

5 treated initially with a regimen of sirolimus,

6 cyclosporine and corticosteroids?

7 If I could elaborate a little bit on that

8 question. We heard from Dr. Neylan the results

9 from these studies where the patient survival

10 graft-loss rates were reported as comparable. Then

11 we did see presentations of slides, for example

12 slide E8 in which acute rejection was reported to

13 be statistically significantly different in favor

14 of the Rapamune and cyclosporine, for example slide

15 E13 where treatment failure showed a difference of

16 25.6 versus 37 percent.

17 So we would appreciate it if you could

18 discuss the significance of those kinds of results

19 within these studies. In addition, for example, if

20 we think about slides E15 and E27, as was noted

21 before, some of these analyses represent

22 on-treatment patient subsets, not the

23 intent-to-treat population, so, therefore, do not

24 take into consideration all the patients that were

25 randomized. We would appreciate you addressing

187

1 that as well.

2 Briefly, as far as during your

3 deliberation of safety, again, which sets are

4 presented and, for example, for slide S33 where we

5 learned that discontinuation was 18 percent versus

6 27 percent and, again, the lower number in favor of

7 the Rapamune plus cyclosporine arm.

8 If we can go to the next slide.

9 [Slide.]

10 If, after you consider these factors, the

11 answer to the first question you believe is yes,

12 should this consideration for this regimen be

13 restricted to a particular subpopulation or,

14 conversely, is there a particular subpopulation for

15 which cyclosporine withdrawal should not be

16 considered.

17 I think this has already been touched on

18 during the earlier discussions so, specifically,

19 the factor that between 18 to 20 percent of the

20 patients in these studies, in fact, did not go on

21 to randomization and how they reflect the patients

22 that could not participate.

23 We have already heard that 94 percent of

24 the patients, for example, in study 310 were white

25 and a relative underrepresentation of other

188

1 patients.

2 Then, to continue, if the answer is no,

3 what additional studies would be needed to support

4 approval of such a maintenance regimen.

5 [Slide.]

6 On the question that I just finished

7 speaking about, we would actually like a formal

8 vote. On the following two, we are looking

9 basically for your suggestions, namely, what

10 additional phase IV studies would you recommend. I

11 say phase IV because the drug Rapamune, of course,

12 is already approved and, therefore, we have asked

13 for some phase IV studies but others may be

14 appropriate based on today's meeting.

15 [Slide.]

16 Finally, the last slide, and this is an

17 area that is of great interest to us and we would

18 like to ask if you have any comments or

19 recommendations regarding study design and/or

20 endpoints for controlled clinical trials that are

21 intended to support the safety and efficacy of

22 maintenance immunosuppressive regimes in renal

23 transplantation.

24 DR. ENGLUND: Thank you.

25 Subcommittee Discussion and Vote

189

1 DR. ENGLUND: This is the discussion

2 phase. I would like to give everyone around the

3 table a chance to--why don't we go around the

4 table. It will be easier. Dr. Mannon?

5 DR. MANNON: Do you want me to address

6 each of these questions in turn?

7 DR. ENGLUND: No, no. I think we should

8 just address question 1. I think we should address

9 question 1, just the first part here because then

10 we are going to have to go further on.

11 Yes?

12 DR. JOHNSON: May I ask a question. I

13 thought, after lunch, we were going to have an

14 opportunity to ask the sponsor some additional

15 questions before discussion. Is that not true?

16 DR. ENGLUND: There is, but my intention,

17 although we can talk about that, was as it relates

18 to each of these three questions. The sponsor is

19 here and they are available to answer our

20 questions. So this is not voting. This is

21 discussion.

22 DR. MANNON: Let me pass to Dr. Hunsicker

23 first and then come back to me.

24 DR. ENGLUND: We don't have to do it

25 around the table. If we have people that want to

190

1 respond to somebody else on the committee, then we

2 can do that, too.

3 DR. HUNSICKER: It will surprise nobody

4 that I have some thoughts on these issues and I

5 have something that I have sort of organized to

6 day.

7 You would like us to address these

8 questions one at a time, but they are interleaved

9 and if you don't mind, madame chairman, I would

10 like to have permission to interleave them to some

11 extent.

12 DR. ENGLUND: To some extent is fine.

13 DR. HUNSICKER: Okay. I want to start out

14 with that we are in a new category here. I have

15 already said this. The usual thing that we have

16 looked at is to show that an agent, a drug, is more

17 effective than either a placebo or a comparator and

18 that it is relatively safe. The emphasis has been

19 on the type I kind of analysis, can we be sure that

20 this is better than what the alternative is. And

21 the safety stuff has, to some extent, been

22 supportive.

23 What we have today is the first of what I

24 suspect is going to be a series of studies that

25 really turn this paradigm upside-down entirely.

191

1 The efficacy issue is one of equivalence. The

2 sponsor is not trying to convince us that the new

3 regimen is superior to the old regimen in terms of

4 the traditional hard outcomes but, rather, they are

5 arguing that it is as good as that and that the

6 side effects which will come down under the area of

7 toxicities, if you will, are better.

8 I think it is important for to move down

9 this line, but I think we have to do some things

10 that are different from what was done today in

11 order to go down this line.

12 Let me turn first to the issue of

13 equivalence. The nature of an equivalence trial is

14 basically that it is looking for type II error

15 rather than type I error. You are trying to show

16 that there is no real likelihood that there is a

17 difference greater than a certain amount that would

18 have happened with your new drug compared to the

19 others or, perhaps, that it is superior.

20 To do that, what you really need to look

21 at is confidence intervals. P-values are utterly

22 meaningless in dealing with a type I error. No

23 significant difference doesn't mean that there

24 isn't a difference. It just means that you can't

25 determine that there is a difference. You all know

192

1 that.

2 So what I would like to urge the sponsor

3 today, if he does more along this line, or other

4 sponsors in the future, is to phrase their analysis

5 of equivalence in terms of confidence intervals and

6 we ought to have, in advance, a statement of how

7 much of a difference makes a difference.

8 So, for instance, if we say that

9 equivalence is that the treatment is no more than

10 10 percent worse than whatever, we can come to

11 agreement that if, in fact, the confidence interval

12 doesn't include 10 percent that they have shown

13 equivalence. But we need to have agreement before

14 we start that that 10 percent is an appropriate

15 number.

16 My own personal opinion is that 10 percent

17 would be a reasonable number for acute rejection

18 but it would not be a reasonable number any longer

19 for graft survival. A 10 percent difference in

20 graft survival between two regimens is clearly

21 clinically meaningful.

22 So I found myself--what I, in fact, had to

23 do, I went back when I got the briefing document

24 and went through and calculated confidence

25 intervals for all of these things. In fact, the

193

1 sponsor does relatively well for some of them but

2 clearly not well for others.

3 The fact of the matter is that numerically

4 the new regimen did better than the comparator

5 regimen, the Rapamune plus cyclosporine regimen,

6 with respect to graft survival and, because of

7 that, the confidence intervals, in fact, are

8 reasonable and don't suggest that there is a high

9 likelihood that the new regimen is going to be

10 worse within the period of time that we are looking

11 at with respect to graft survival.

12 But it would have been a whole lot easier

13 had these things been all explained in advance and

14 clearly so we knew what we were accepting as

15 equivalence. Now, with respect to rejection, it is

16 clear that the new regimen is not as good as the

17 old regimen. I have to say here that rejection, in

18 my community--I don't know what FDA thinks about

19 it--rejection has had sort of a dual life because

20 it is a clinically meaningful outcome on its own.

21 And I don't want ever to forget that.

22 So the fact that the new regimen is

23 clearly less good than the old regimen with respect

24 to rejection episodes can't be washed away, but it

25 also has been used in our community as a predictor

194

1 of what is coming downstream and I have to talk

2 about that separately.

3 When I say that the rejection episode was

4 clearly higher, if you look at pivotal study 310, I

5 think is the number--if you look at the number of

6 rejection episodes following randomization, it is

7 clearly higher in the patients that were assigned

8 to the withdrawal of cyclosporine.

9 Is that disastrous? No; I don't know that

10 that is disastrous, but it can't be ignored and we

11 have to have that clearly stated up front.

12 Then, when we turn to the issue of the

13 toxicity things, traditionally, it has been done

14 that toxicity is based on treated patients or

15 something like that. But today, now, we are really

16 basing our long-term judgment on the acceptability

17 of this regimen, on what it promises to us in terms

18 of toxicity. For that, it seems to me, we have to

19 insist on intent-to-treat analyses, across the

20 board.

21 We have to understand what is--if we are

22 going to say that this is a better way to go

23 because of less toxicity, we have to understand

24 that that is true for the entire randomized

25 population.

195

1 I also think we have to distinguish

2 between what I would call clinically apparent and

3 numerically apparent toxicities. What I mean by

4 clinically apparent toxicity is that an infectious

5 episode, a pneumonia, or whatever, is clinically

6 apparent but changes in blood pressures and changes

7 in creatinines are not important today. They are

8 important for what they may mean for the future and

9 there is a smaller degree of certainty as to what

10 their significance is for the future and we have to

11 look at these in terms of what they mean for the

12 future.

13 So I would like to see all of these

14 analyses within intent-to-treat analyses and I

15 would like to see a distinction between the

16 clinically evident things today and the long-term

17 outcome. This is because what I see as the issue

18 before us today, the tradeoff of an increased

19 frequency of rejection when you withdraw

20 cyclosporine, which is as clinically meaningful

21 outcome, increased today, for which you receive as

22 compensation better serum creatinine and the hope

23 of long-term better outcome with respect to graft

24 survival.

25 Turning to that, I have already spoken

196

1 informally to the sponsor and said that I think

2 that there is a more appropriate analysis than the

3 analysis that we have of the renal function and

4 progression over time.

5 First of all, to look at the patients at

6 risk at each time point and take the average over

7 time is statistically not an appropriate way to

8 look at what is happening over time. There is a

9 different group of patients at risk in each pool

10 and you really can't compare the values from time

11 to time.

12 The issue here is critical. Is there, in

13 fact, a difference of creatinine over time, an

14 analysis which I would like to suggest is a

15 reasonable one. There may be other ways of doing

16 this, to do a GEE analysis on the delta from

17 baseline, the baseline being the time just

18 immediate before randomization.

19 So what you are looking for is whether

20 there is a stepped decrease in the first period of

21 time and what is the trend of the creatinine after

22 that time, or clearance or whatever other measure

23 that you are having.

24 Most of my colleagues here, both in the

25 audience and around this table, know that I have

197

1 done an analysis of what I call intercepts and

2 slopes on creatinine clearance following renal

3 transplantation. The results of this analysis

4 which involved some 48,000 patients from the UNOS

5 database are, in essence, that you can, on average,

6 treat the progression of renal disease over time as

7 linear loss of renal function, of GFR or creatinine

8 clearance over time, just as you can with native

9 kidneys.

10 If this is the case, if my analysis is

11 correct, which I believe it is and it represents

12 the reality--and I would like to just call Alan

13 Wilkinson's caveat into consideration here; this

14 analysis was done virtually entirely on patients

15 who were receiving calcineurin inhibitors. So

16 there is some question of whether it would be

17 extrapolatable across.

18 If there is a difference in serum

19 creatinine today and if there is no difference in

20 slope--that is to say, if there is a step

21 decrease--that step decrease will translate into

22 longer graft life. The term that I have there is

23 that about 2.5 milliliters of GFR is equivalent, on

24 average, all other things being equal, to one year

25 of graft life.

198

1 So if you have an improvement, a step

2 improvement, of somewhere between 5 and 10

3 millimeters per minute better GFR estimate at one

4 year or six months or whatever the time is, the

5 anticipation is that that would lead to a two to

6 four year improvement in graft life for the

7 patients that were on the Rapamune-only regimen.

8 But this is conditional that the trends of

9 serum creatinine or creatinine clearance following

10 that time don't converge. That we don't really

11 know. We have no idea what is happening to the

12 difference over time. So we have a promissory note

13 in exchange for a payment of an increased rejection

14 rate which is a clinically important event and we

15 need to know how solid that promissory note is

16 before we can know whether this is a reasonable

17 bargain or not.

18 I am going to go off that to the second

19 series of questions that I have about this

20 application. Section 2 in my little list of notes

21 here has to do with approval and indication. I

22 constantly annoy my friends at the FDA by pointing

23 out that most of the transplant community pays no

24 attention to what goes into an indication anyway.

25 We never read the damned things and we do whatever

199

1 we please.

2 So the question comes up, then, what is

3 the impact of approval and what is the impact of

4 the indication. I, for one, believe that it is

5 essential that our community continue to explore

6 the issue of calcineurin-free regimens. I think

7 that there is, from these data and other data that

8 I am aware of with respect to sirolimus, the

9 suggestion that, in fact, there may be major

10 long-term improvements--may be. But it is a long

11 way from saying that we have to continue these

12 things, to say that we should say that they have

13 now met the standard of use everywhere.

14 So the question comes up how right are we

15 for calcineurin withdrawal and who should be doing

16 it. I am lucky because I don't get to vote today,

17 you know. I just get to express my opinion and

18 raise the questions and let the rest of the

19 committee decide to vote.

20 I think we have to explore this but I am

21 not sure that I want this explored primarily in the

22 least-expert groups of patients. If I ask myself

23 where the approval of the FDA and where the

24 indication would have the greatest effect, it is

25 likely to have the greatest effect amongst the

200

1 people who are not as thoroughly involved in all of

2 these issues themselves; i.e., in the less expert

3 people.

4 That troubles me because I would like to

5 see these issues addressed first in the most expert

6 group of people. I have a feeling that what I am

7 telling you I think it is still investigational. I

8 am not sure we know the long-term impact.

9 This is complicated by the fact that we

10 have a major limitation in the population about

11 which we could say anything. We have already

12 discussed the fact that there are no

13 African-Americans. There are no Hispanics. Some

14 of the groups in whom our problems are greatest are

15 not represented with sufficient numbers, in my

16 opinion, for us to be able to say anything.

17 I want to make sure that that does not say

18 that there is not a benefit. I just don't think

19 that it is at all established that there is a

20 benefit or a harm. I think we do not know what

21 would happen in African-Americans. I don't think

22 we would know what would happen in Hispanics.

23 I also don't think we really know what

24 would happen in people with initial ATN because,

25 largely, those people are delayed graft function.

201

1 Those people were not randomized and so we have

2 really no idea what would happen in this group.

3 In fact, the group that wound up getting

4 randomized is still very fuzzy in my mind. One of

5 the charges I would put to the FDA is it has got to

6 be very clear what were the patients in whom this

7 experiment was really done because, clearly, we

8 don't know anything beyond the patients in whom the

9 experiment was done.

10 Now, if an indication can be drafted that

11 says that this should be done only in patients who

12 don't have initial graft dysfunction, have not had

13 a type III rejection within the first six months,

14 whose creatinine is less than thus and such, and so

15 forth, and who, by the way, are neither

16 African-American nor Hispanic because we can't say

17 anything about that.

18 If you can come up with an indication,

19 that would be fine but I think it is going to be so

20 complicated that I am not quite sure where you are

21 going to wind up.

22 So my issues here are first methodologic.

23 I want to have the way we present these kinds of

24 studies changed so that we know exactly what at

25 cost is, the potential cost, when we are talking

202

1 about equivalence and then exactly what the benefit

2 is that we would see on the other end from the

3 reduction in toxicity.

4 In this case, this means, what can we

5 extrapolate to in terms of long-term graft

6 survival. I have problems with whether this has

7 reached a state of ripeness that I really want to

8 have the least expert people in our community begin

9 doing it which is what I think is implied by

10 approval and by the indication and I really have

11 some reservations about what the population is in

12 whom we could say that this has now been

13 established as safe and effective.

14 DR. ENGLUND: Did you have any specific

15 questions for the sponsor?

16 DR. HUNSICKER: No. I was giving a

17 philosophic tirade and I am sorry for that, but I

18 am asked what my opinions are about these things

19 and you now know my opinions. I feel good about

20 this because I have to leave at 3:30 because I have

21 got to make a plane to get home.

22 I know that the sponsor--I have spoken

23 with them about some of these things in

24 between--has some slides that they would like

25 eventually to show that relates to the question of

203

1 whether there is a trend in creatinine or clearance

2 or something over time that can be established. If

3 you want them to show that, that would be fine with

4 me.

5 I am happy to tell my folks at the FDA

6 that they have got to establish that there is as

7 reasonable likelihood that a short-term delta

8 creatinine is going to translate into a long-term

9 graft survival before I am going to feel that that

10 is a benefit that will balance the increased rate

11 of rejection early.

12 DR. ENGLUND: Let's go on and see. I

13 heard you say you didn't have any questions, so

14 let's go on. If someone has a question, or I might

15 have a question--

16 MR. LAWRENCE: First I would like to thank

17 the FDA for inviting me to participate in this. It

18 is always reassuring to the patient community to

19 know that at least somebody was there with their

20 best interests up front. Even though the

21 physicians and the pharmaceuticals are laboring on

22 our behalf all the time, we still like to be there,

23 so thank you for that.

24 I agree with everybody here. Everyone has

25 said things that are intelligent and compelling

204

1 but, coming at this from a lawyer's viewpoint, the

2 question that hasn't been precisely answered for me

3 is what, exactly, are we supposed to be doing here.

4 What words are we supposed to be changing?

5 In the stuff that you sent out several

6 weeks ago that we all go to review before this, it

7 says that the application is proposing to modify

8 the indication that says that Rapamune shall be

9 used in concert with cyclosporine. This says that

10 the applicant is proposing to modify that to allow

11 consideration of cyclosporine withdrawal.

12 Then I see the slides that were presented

13 by Wyeth and it says cyclosporine withdrawal should

14 be considered. This is much more directive. I

15 think that there are probably a large number of

16 patients who would benefit by having cyclosporine

17 withdrawn. I take cyclosporine, myself. I am not

18 unaware of the renal implications of taking this

19 drug.

20 I also gather from comments that have been

21 made by all of the knowledgeable people here that

22 there are probably some patients in whom it should

23 not be withdrawn or the jury is certainly still

24 out. I am not here representing UNOS, who is my

25 employer, but on the UNOS website, anybody can pick

205

1 up these data that I am about to give you.

2 The current waiting list which is

3 tragically approaching 90,000 or something--it is a

4 lot of people waiting for organs in this country.

5 Caucasians represent 42.3 percent of the current

6 renal waiting list. This is renal waiting list.

7 Hispanics, 14.5, Asians, 5.6 and blacks, 35.1 So

8 the data that we have seen today actually applies

9 most directly to 42.3 percent of the waiting list.

10 That is simply an insufficient

11 representation to support language which is direct,

12 saying that cyclosporine withdrawal should be

13 considered. I think that the use of the word

14 "should" would be of much more interest to my

15 fellow lawyers than it would be to physicians, most

16 of whom--I spoke to a number of them before coming

17 here and they said, "We don't care what they say

18 because we are going to do what we feel is right

19 for our patient anyway."

20 That may be, in reality, how medicine is

21 practiced, but I don't think that a case has been

22 made to be as directive as it should be. I would

23 like to see something along the lines of

24 cyclosporine withdrawal "may" be considered

25 because, obviously, it would be in the interest of

206

1 many patients that cyclosporine, in fact, be

2 withdrawn. I think that is conclusively true for

3 many patients, but it is also conclusively true to

4 me that that does not apply to all patients.

5 Therefore, saying that cyclosporine

6 withdrawal should be considered is too strong a

7 statement. I would just suggest that I would agree

8 with Wyeth that withdrawing cyclosporine, where

9 that can be done without any deleterious effect,

10 should be done, in fact, and probably that is a

11 majority of patients although what that means, I

12 don't know.

13 So I would suggest simply reconsidering

14 the terminology we are using here. Thank you.

15 DR. ENGLUND: Thank you.

16 DR. MANNON: I let Dr. Hunsicker go first

17 because I knew he would--not that I knew that he

18 had his plane but because I knew that he would have

19 a lot of things to say.

20 Just a couple of things that may be in

21 agreement with him and may not be totally in

22 agreement with him, and the comments that I heard

23 earlier today is that the question always comes as

24 to who is doing this. Yes; I think that transplant

25 nephrologists and surgeons do have ways of using

207

1 drugs in different fashions that may not be on the

2 label, necessarily.

3 How it is being done is also important.

4 The issue is that, if the label goes in a certain

5 way, it means that anybody who has that kind of

6 certification can and it may not be in a large

7 academic center. It may be in a smaller transplant

8 center. I think that is one of the concerns is

9 that if this labeling goes as black and white, will

10 everybody be doing it that way or is that on the

11 entree for people to go ahead and do.

12 Clearly, there are caveats to doing that

13 therapy. I do have questions about the

14 applicability. Again, I think the race issue is

15 one that was again reiterated by a number of people

16 around this table. The issues of children are

17 obviously not addressed in this and that is a small

18 population. But, again, that should be addressed.

19 I also wanted to point out that in these

20 studies, this was a very large population of

21 cadaverics. In fact, living transplants were a

22 minority of about 60 patients that were in the 212

23 study. Again, should the indications--I know in my

24 practice, when we see living transplants, we tend

25 to ease off on immunosuppression based on their

208

1 long-term outcomes.

2 I think the issues, again, that were

3 brought up regarding delayed graft function and

4 ATN, we don't know enough, I guess, based on the

5 randomization about how to manage them. Along

6 those lines is should there be indications

7 regarding ischemic time. Can we tease apart the

8 patients that had those rejection episodes based on

9 maybe they had more prolonged hold time.

10 PRA or highly sensitized patients, how are

11 they in this population and how are they thrown in

12 and is there a way of going back and looking at the

13 data collected by the sponsor to say that maybe

14 that would be an indicator of someone that you

15 would not really choose.

16 I think if I went around this room and

17 said, "You have a PRA of 90 percent," the majority

18 of us would probably not choose to put that person

19 as a withdrawal patient, per se, but maybe there is

20 data available.

21 My last, I guess, sort of point is about

22 the monitoring. I have a lot of practical clinical

23 experience about monitoring in this drug. Although

24 there are eighteen centers available, I want to

25 point out that, for most of us, we Fed-Ex our

209

1 samples or UPS our samples, so there is a 24-hour

2 delay to get the sample to be monitored and another

3 24 hours, about a one-day turnaround time. So you

4 are talking about a total of 48 hours which,

5 although the drug has a fairly long half-life, it

6 is sometimes difficult to monitor.

7 I think the availability of the more

8 rapid, less labor-intensive, test would be--there

9 are two issues. One is should we be monitoring

10 these patients. I know that was brought up. The

11 other issue is if we are going to monitor them,

12 what is the best way to do that.

13 I think if you are going to have a mass, a

14 large number of centers doing numbers of these

15 tests, it will become a very important issue as far

16 as the turnaround time and documenting--I think it

17 would be helpful--I know that they talked about

18 doing an algorithm on the labeling. It would be

19 important for us to look at that algorithm,

20 perhaps, and sort of decide if that would be of any

21 help in the long-term monitoring of the patients.

22 DR. ENGLUND: We are going to go around

23 the room, but I think in terms of the monitoring

24 issue, perhaps could we spend a minute or two? Are

25 there any other comments about the monitoring as we

210

1 go around? It really is implied in the part of the

2 question that it would be part of the approval to

3 do it, as has been done in the study.

4 Dr. Shapiro?

5 DR. SHAPIRO: I would just have a comment

6 about monitoring. I guess two-and-a-half years

7 ago, the position was that this drug did not

8 require monitoring. We learned, at least in my

9 case, the hard way that that was not correct. Even

10 now in the context of this particular protocol, we

11 found that we have not been able to use sirolimus

12 safely without close monitoring.

13 I think that is probably a consensus among

14 most people who are involved in transplantation.

15 DR. AUCHINCLOSS: I was going to say the

16 same thing. I can't imagine trying to use this

17 drug in any protocol at this point without

18 monitoring.

19 DR. MANNON: It is difficult. We can't

20 even agree about what the--I know one question was

21 can you predict the level based on the patient or

22 the race or the age or the weight. I can tell you,

23 in my limited experience, it has been difficult to

24 tell when you use a loading dose of 15 and then go

25 on 5. We have been trying to look at peak--post

211

1 load doses to see if we can predict.

2 So I agree. I think you need--monitoring,

3 for me, has been essential.

4 DR. HUNSICKER: Very briefly, a comment

5 about monitoring. I agree with the people who have

6 spoken who say that, in fact, we do monitor the use

7 of this drug. The question I would have is whether

8 the specific recommendations as to monitoring are

9 based on anything other than grabbing some numbers

10 out of the air.

11 I would not, at all, mind if this is an

12 indicated drug, having an indication saying that

13 there is a high variability of bioavailabilty and

14 that it might be wise to check the levels. But to

15 tie yourself to a specific monitoring program, as

16 was described to us, on the amount of information

17 we have to say that that makes sense would be

18 difficult for me.

19 DR. ENGLUND: Dr. Abernethy?

20 DR. ABERNETHY: I would support that

21 assertion, to simply say that, in my clinical

22 experience, monitoring is essential. One can say

23 that about many drugs. However, then, at a later

24 point in time, one looks at the data, it sometimes

25 turns out that the data support that that was a

212

1 correct statement and other times it turns out that

2 that just was a clinical impression that doesn't

3 stand up to scrutiny.

4 I think, at this point, I don't know. I

5 haven't seen data either in the material we were

6 provided or this morning that told me that we know

7 that there is a therapeutic index such that,

8 particularly at the high end of the concentration

9 range, that we know where we should put a cutoff on

10 that.

11 If that is correct, clinically as well as

12 with the data, then the correct response, I

13 believe, is that one simply increases the dose when

14 there is a question about whether things are

15 happening the way they should. If that is

16 incorrect, then I think we need more data in order

17 to assert that it is incorrect.

18 DR. SHAPIRO: There was the figure that

19 the sponsor had shown showing that there were lower

20 rejection rates with higher sirolimus levels and

21 this interacted with the amount of cyclosporine

22 patients were receiving also. So it is not

23 completely pulled out of the air.

24 DR. ABERNETHY: That is very complicated

25 because when you have those two drugs together,

213

1 they are interacting both pharmacokinetically with

2 each other as well as pharmacodynamically. So that

3 was an interesting chart without confidence

4 intervals and without data points. I will have to

5 say, I would have to really look at that data a

6 long time before I could come to any conclusion

7 about what it was trying to tell me.

8 I am not saying it is incorrect. I am

9 just saying I can't look at a slide like that and

10 say, "Oh; right."

11 DR. ENGLUND: Dr. Ebert?

12 DR. EBERT: Just maybe a short addendum to

13 that. Again, most of the association that we are

14 seeing here is largely, at least in my opinion,

15 kind of a post-hoc analysis where patients were at

16 least initially dosed on the drug, subsequently or

17 retrospectively, were found to have certain

18 outcomes associated with certain serum

19 concentrations.

20 I think that differs from what might be

21 considered to be a concentration-controlled

22 prospective study where patients are randomized or

23 targeted different target concentrations and then

24 looking at outcomes. I am not sure that the two

25 are equivalent as far as the conclusions that we

214

1 can draw.

2 DR. ENGLUND: Back to general comments

3 about question No. 1?

4 DR. AUCHINCLOSS: The question is do the

5 data support the safety and efficacy of

6 cyclosporine withdrawal. I think, in a general

7 sense, the answer to that question is yes. But the

8 problem is, well, yes, it is apparent that that

9 would be true for some patients, that there would

10 be some associated risk and that there would be

11 some associated benefit.

12 The problem is that both from limitations

13 of numbers and from study design, it is very hard

14 for us to answer precisely any of those aspects of

15 where this efficacy applies. I think it is clear

16 that we are talking about a group of patients that,

17 in general, are doing well and I would second the

18 comments of others that there are distinct

19 populations including African-Americans about whom

20 I would have tremendous concern.

21 What is the risk? I have no doubt that

22 there is, indeed, a risk of acute-rejection

23 episodes precipitated by cyclosporine withdrawal.

24 It looks to me like it is about 10 percent. I am

25 sure there are other side effects of high-dose

215

1 sirolimus. We saw dose-response curves for

2 cholesterol, et cetera. So there is some

3 additional risk by going to this protocol.

4 What are the benefits? Clearly, you are

5 going to get rid of some side effects of

6 cyclosporine. I have no doubt that there will be

7 an improvement in renal function and I believe

8 those data. What I don't know is what the

9 long-term consequences of that are.

10 So what does all that mean to me as a

11 clinician? From the data that I have seen today, I

12 think I would consider cyclosporine withdrawal in a

13 group of patients who are on sirolimus who are

14 generally doing well but who are tolerating

15 cyclosporine in some fashion very poorly and who

16 demonstrated the capacity to tolerate Rapamune

17 without side effects, or without major side

18 effects.

19 I am not sure exactly how you turn that

20 into a label. I am sure that the labeling words

21 "should--" the word "should" should not be the one

22 that is used. Frankly, I really think overall, at

23 this point, that the data that we have are

24 insufficient and premature to define the answers to

25 these kinds of questions that make a labeling

216

1 change appropriate at this point.

2 DR. ABERNETHY: I really don't have much

3 to add. I think that we, saying it slightly

4 differently, are handicapped by trial design and

5 that we are looking at a very selected group. I am

6 struggling with how to generalize that effectively

7 or if it, perhaps, should be generalized.

8 DR. ENGLUND: Dr. DeGruttola?

9 DR. DeGRUTTOLA: I have a number of

10 comments. I think what Dr. Hunsicker and Dr.

11 Auchincloss were referring to is what statisticians

12 refer to as a surrogate-endpoint problem. We have

13 evidence that there is adverse effect on acute

14 rejection which is not, apparently, a clinical

15 event but indicative of potentially future higher

16 risk of clinical event. And we have apparent

17 benefit on some measures of kidney function

18 although it is not clear whether those would

19 translate into longer-term benefits.

20 In addition, there is a concern about

21 whether creatinine levels measured at a particular

22 time have the same meaning regardless of the

23 treatment that a patient is on. In other words,

24 does a benefit in creatinine levels that results

25 from a treatment have the same impact as naturally

217

1 having better creatinine level.

2 I think, to answer those questions

3 generally requires longer-term follow up to

4 understand the relationship between treatment, the

5 surrogates of creatinine or measures of acute

6 rejection and the longer-term clinical benefit in

7 the absence of compelling evidence that we can

8 infer longer-term benefit from the shorter-term

9 outcome. That is difficult.

10 I think that the problem of interpretation

11 of results would exist anyway, but it is compounded

12 by the fact that we have been presented with a lot

13 of as-treated analysis and, as Dr. Albrecht

14 pointed out, the analyses that we saw of creatinine

15 and GFR looked at the as-treated population or

16 on-therapy population and such results are harder

17 to interpret.

18 The FDA analysis provided us with

19 intent-to-treat comparisons showing a benefit of

20 the Rapamune alone which was useful. But the FDA

21 analysis just gives us the two-by-two tables. The

22 sponsor's analysis gives us the time trends which

23 are really valuable to know for the reasons that

24 Drs. Hunsicker and Auchincloss pointed out. We

25 would really like to have some sense of whether

218

1 these are persisting or increasing.

2 It is precisely when you are trying to

3 evaluate time trends in these effects that the

4 difference between an intent-to-treat and an

5 as-treated population would be so important to know

6 because, in an as-treated or on-therapy population,

7 where the populations are changing, it is hard to

8 interpret the time trends.

9 So I think that it would certainly be

10 useful to able to see the intent-to-treat analysis

11 at least to give us a sense of whether the effects

12 are increasing as they appear to be from the

13 on-therapy analysis, the effects of benefit of the

14 Rapamune alone on creatinine.

15 I believe that there may be additional

16 evidence in support of a relationship between

17 markers like creatinine and longer-term outcomes.

18 I would be very interested in seeing such results

19 from the sponsor if we can request that.

20 DR. ENGLUND: Would you like to show them

21 now?

22 DR. NEYLAN: Yes; I would. Thank you.

23 What I wanted to do was to show you some of the

24 longer-term data and also look at some of the

25 different analyses that address some of the

219

1 concerns.

2 [Slide.]

3 First, just a reminder, this first slide,

4 we have the intent-to-treat analysis of renal

5 function which concurs with the FDA analysis that

6 the patients in 310 had enjoyed an improvement in

7 both the mean serum creatinine and the calculated

8 GFRs which was statistically significant.

9 What I would like to do is call up the

10 slides that look at the slope intercept analyses.

11 This is, I think, an analysis that is probably

12 somewhat near and dear to Dr. Hunsicker.

13 [Slide.]

14 Calling up this first slide, looking at

15 UNOS data, this recent publication from Johnson and

16 colleagues looked at over 100,000 renal-transplant

17 patients within the UNOS database between 1988 and

18 1998. As Dr. Hunsicker has pointed out to us both

19 today and in his prior publications, it is

20 important to consider not only where you are

21 starting from but how quickly you are getting to

22 the next place.

23 So the baseline creatinine as well as the

24 rate of change in that creatinine are important

25 measures when determining the likely success or

220

1 lack thereof of a kidney transplant. Indeed, in

2 the best-case scenario, looking at this large

3 database, when you start off with a great

4 creatinine and you have a very small change in that

5 creatinine from the six-month to the twelve-month

6 mark, you can expect a half-life of 11.6 years.

7 If, on the other hand, you see a more

8 rapid change, and, by change, I mean increase in

9 serum creatinine, over this time point from six to

10 twelve months, then that half-life is decreased and

11 so on down the way. If you start off at baseline

12 with a poorer functioning graft, you will have a

13 reduced half-life even if your rate of change is

14 relatively minor.

15 The worst-case scenario, of course, is

16 when you start off with a poorly functioning graft

17 and see a rate of change that is greater. There,

18 the half-lifes are, of course, the worst. Taking

19 this kind of approach, we looked at our own data

20 and if we can show the next slide.

21 [Slide.]

22 What we looked at was a kind of similar

23 slope-intercept analysis and looked at, in the case

24 of the 310 patients, the patients who had the serum

25 creatinines that were either excellent or greater

221

1 than 1.5. We looked at the rate of change between

2 six and twelve months.

3 Is this correctly labeled, this

4 creatinine? Is that at twelve months?

5 DR. BURKE: That is creatinine at twelve

6 months.

7 DR. NEYLAN: Looking, then, at this

8 baseline and the rate of change of getting there,

9 you can see the following. You see in the Rapamune

10 group that there is a preponderance of the patients

11 who fit this bill--namely, excellent creatinines

12 and a small rate of change. Again, the

13 six-to-twelve-month mark is relevant because, as

14 was alluded to, with the relief of cyclosporine and

15 the relief of that vasoconstriction, one would

16 expect that the short-term change up to six months

17 might one thing but, subsequent to that, rate of

18 change may well be related to other factors.

19 So we see this rate of change being the

20 least in the Rapamune group compared to roughly

21 half as many patients in the control group and so

22 on down the way. Conversely, at the bottom, we see

23 more patients, or twice as many, in the control

24 group that start off with a worse baseline and have

25 a more rapid rate of change.

222

1 [Slide.]

2 We also did another analysis that is a

3 slope analysis of patients in the next slide

4 who--this is 24 months. Is this data part of the

5 package? Could you turn that slide off for a

6 second?

7 DR. BURKE: That includes data that is not

8 part of the package. It is creatinines after

9 twelve months.

10 DR. CAVAILLE-COLL: May I ask you a

11 question about the previous slide where you were

12 showing--you were applying Johnson's analysis to

13 your data. Was that submitted to the application

14 and does that analysis include all patients treated

15 or just the information on patients on therapy?

16 DR. NEYLAN: That is an on-therapy

17 analysis, I believe.

18 DR. CAVAILLE-COLL: An on-therapy

19 analysis? Was that analysis submitted to the

20 application?

21 DR. NEYLAN: No; excuse me. Was that--

22 DR. BURKE: The analysis was not

23 submitted. The data that was used for that

24 analysis is in the application.

25 DR. CAVAILLE-COLL: So it is the

223

1 on-therapy analysis. It is the on-therapy data

2 that you submitted to the application. It is not

3 an intent-to-treat analysis.

4 DR. BURKE: No.

5 DR. CAVAILLE-COLL: And it is not an

6 analysis that you have submitted to the FDA for

7 review.

8 DR. BURKE: That's correct.

9 DR. CAVAILLE-COLL: Okay. Thank you.

10 DR. NEYLAN: So, rather than show you the

11 other slope intercept which includes 24 months,

12 what I would like to show it completor's analysis.

13 One of the problems that we have in fulfilling the

14 more rigorous statistical requirements of

15 intent-to-treat is, in this case, the problematic

16 return to calcineurin inhibitors which can occur in

17 patients discontinued from the treatment group

18 which then creates a kind of convergence. That

19 makes it sometimes challenging to discern important

20 clinical differences.

21 [Slide.]

22 One way to get around that is to do a

23 completor's analysis. Here we have, again, an

24 analysis that is taken from the dataset that FDA

25 has received, although this particular analysis

224

1 that was--rather the dataset is within your hands.

2 The analysis that we did was separate.

3 DR. CAVAILLE-COLL: Do we have the dataset

4 up to 24 months?

5 DR. NEYLAN: Yes; you do.

6 DR. CAVAILLE-COLL: That does not include

7 all the subjects on the study, then?

8 DR. NEYLAN: What this shows, working

9 backward, is a completors' analysis so it includes

10 only those patients who, from the starting point on

11 through, are successfully treated in either group.

12 So it takes away that bias of patients who are

13 dropping out along the way in an on-therapy

14 analysis.

15 What we see here with the mean creatinines

16 is, again, data which is representative of the

17 other datasets that we have shown you, namely that

18 serum creatinines in the control group stabilize or

19 slightly increase over this time period whereas the

20 slope of the treatment arm is stable or, in fact,

21 slightly downward.

22 I would certainly be open to any inquiries

23 about that.

24 DR. ENGLUND: Dr. DeGruttola?

25 DR. DeGRUTTOLA: I just wanted to comment

225

1 on a couple of points. I think that the

2 intent-to-treat analysis is valuable even if

3 patients do end up crossing over to another

4 treatment because that is the information that you

5 really want, what is the outcome when you intend to

6 treat a patient in a particular way but the reality

7 is you may not necessarily be able to treat them in

8 the way that you want to, and finding out whether

9 there is, in fact, a benefit, in terms of

10 creatinine, over time for patients who are intended

11 to be treated with Rapa is exactly what you want to

12 know.

13 If you do something like a completors'

14 analysis, you are getting sort of a filtration

15 effect in the populations. You are taking out the

16 people that are having difficulty, so you may see

17 an effect that is increasing but that may be purely

18 artifact of who is left in that population.

19 While I think that there are questions of

20 interpretation when you do the intent-to-treat

21 because patients are switching therapy, you can do

22 analyses that will tend to indicate whether the

23 fact that the curves are coming together results

24 from the changes in therapy for the population who

25 must change therapy or loss of an effect in the

226

1 patients who remain on therapy.

2 You can do additional analyses to help

3 with the interpretation, but the most directly

4 interpretable analysis will be the intent-to-treat.

5 The fact that patients have to change therapy is a

6 result. It is an important outcome of the study

7 and I don't think that you can solve the problem by

8 doing the completors' analysis.

9 DR. NEYLAN: I apologize if I meant to

10 suggest that we were solving the problem. But, in

11 addition to the intent-to-treat analysis which

12 shows the benefit, I was just hoping to provide

13 some additional analyses which, while not perfect,

14 can help to address some of the issues of patient

15 dropout and, again, the challenges of comparing

16 these groups of patients when the alternative to

17 not staying within the study is most typically a

18 return to the calcineurin inhibitor.

19 DR. DeGRUTTOLA: I think the problem is

20 that we have the intent-to-treat analysis for the

21 two-by-two table but we don't have the

22 intent-to-treat analysis over time, which means

23 that we can't get a valid estimate of the time

24 trend. I think that is concern.

25 If you want to do completor analysis as an

227

1 additional analysis in order to help with the

2 interpretation, that is okay. But I think that it

3 would be valuable to be able to see the time trend

4 for the intent-to-treat analysis to see what is

5 going on.

6 DR. NEYLAN: I will ask the group. Do we

7 have any time-trend analysis?

8 DR. BURKE: We are unable to provide an

9 intention-to-treat time analysis at this time. We

10 recently gathered the intent-to-treat at twelve

11 months. We will be gathering additional time

12 points but, at this present time, we cannot provide

13 time analysis on intent-to-treat.

14 DR. NEYLAN: We may be able to very

15 shortly.

16 DR. ENGLUND: I am going to interrupt as a

17 prerogative here. In the specific slides that I

18 would be interested in as intent-to-treat are E15

19 and E28.

20 DR. NEYLAN: Could we call those up.

21 DR. ENGLUND: You are showing me here

22 improved renal function and these are really nice

23 slides, but it is not intent-to-treat.

24 DR. HUNSICKER: There are two things.

25 First of all, it is not intent-to-treat and the

228

1 second thing is that the people at risk are

2 different at different times. They have got to do

3 a proper analysis. I think that it would not serve

4 Wyeth-Ayerst. It would not serve the FDA and it

5 wouldn't serve reality for us to try to squeeze out

6 an analysis between now and two hours from now.

7 I thoroughly second your comment about

8 intention to treat and I am not going to say

9 anything further. I think that this is a given.

10 We have solved these problems long since. We don't

11 have to resolve them. This is the standard.

12 I do want, for the purposes of the record,

13 to put in a comment about the timing from which you

14 are measuring slope and why I am so insistent upon

15 that. There is, as has already been said by I

16 guess it was Alan, a strong understanding that the

17 acute effect of administering a calcineurin

18 inhibitor is that you get a vasospasm in the kidney

19 and that results in an acute decrease in renal

20 function.

21 When you take off the calcineurin

22 inhibitor, if you do it within a short period of

23 time, that returns. So you have an acute effect

24 that is vasomotor. You then have, we think, as a

25 result of calcineurin inhibitors, progressive

229

1 fibrosis and other long-term changes of the kidneys

2 that are not likely to be reversed when you reverse

3 the cyclosporine.

4 The reason I make this comment is that if

5 you are going to do a slope analysis, you have to

6 make sure that your slope finishes after you have

7 had the completion of your acute effect or the

8 acute effect will be bundled in with your chronic

9 effect.

10 That you showed, for instance, in the

11 two-by-two analysis that the creatinines were still

12 superior at 23 months or whatever the last time

13 period was, doesn't really answer the slope

14 question because that buries into that delta the

15 effect of taking off the cyclosporine acutely. So,

16 what we need to do is to get an estimate of what

17 has happened acutely with the removal of the

18 cyclosporine and then what the trends are

19 long-term, independent of changes in cyclosporine

20 dosing.

21 DR. NEYLAN: We have an analysis that is,

22 again, based on the dataset that has been submitted

23 to FDA but the analysis, itself, was not part of

24 the packet and that is a slope analysis at a later

25 time point.

230

1 Would it be all right to show that? I

2 think it helps to address some of the questions you

3 are relaying about the acute versus chronic

4 effects.

5 DR. CAVAILLE-COLL: Is this the on-therapy

6 analysis or intent-to-treat analysis?

7 DR. NEYLAN: This is an on-therapy, is it

8 not?

9 DR. CAVAILLE-COLL: Because you submitted

10 two datasets to us. You submitted to us very

11 recently which I think is a closer intent-to-treat.

12 Then there is the original data set with the

13 application which was just on-therapy.

14 DR. NEYLAN: So this is on-therapy. Jim,

15 I might ask you, again, since this is your data, to

16 speak to it.

17 DR. BURKE: I will be showing two slides.

18 [Slide.]

19 They are slopes of creatinine over time.

20 This shows data between six and 24 months, but,

21 indeed, any patient for which we could determine a

22 slope after six months was included in this

23 analysis. So, even if they didn't complete twelve

24 months, if they had a slope between six and twelve

25 months, they are included. So this is sort of

231

1 between an on-therapy and a total intent-to-treat.

2 Let's look at two things, first of all,

3 the slopes. If one looks at the slopes, one can

4 see that, in both cases, they are significantly

5 different from zero. We see that here. One takes

6 a look at the slopes. One is negative for the

7 group. That still includes cyclosporine which

8 means their renal function is decreasing. In the

9 patients for which had they had cyclosporine

10 limited, the slope is positive showing that their

11 renal function is improving.

12 If one takes at a look at the difference

13 between those two, it is significant. So the two

14 slopes are not converging. The time at which this

15 was done; at six months, the initial effect of

16 eliminating cyclosporine is no longer there, so we

17 are looking at a true evolution after that. But if

18 one wants to be even more conservative, I would to

19 show the next slide.

20 DR. ABERNETHY: If I could interrupt and

21 show my statistical ignorance here, but, doesn't

22 this get us back to this issue of looking at

23 equivalence of these slopes versus looking at

24 differences between these slopes and are these

25 5 percent confidence intervals--I mean, am I headed

232

1 in the right direction with that thought?

2 DR. DeGRUTTOLA: Is the question the fact

3 that the confidence intervals don't overlap imply a

4 difference in the slopes?

5 DR. ABERNETHY: Apply nonequivalence,

6 which is what is being suggested, I think.

7 DR. DeGRUTTOLA: I think that, on the face

8 of it, it does appear that those slopes are

9 different and the fact that you are rejecting zero

10 implies that there are differences between those

11 slopes.

12 DR. HUNSICKER: Not only are the

13 individual slopes different from zero but the

14 difference in the slopes, the slopes, themselves,

15 differ by class.

16 DR. BURKE: That's right. Once again, in

17 one group, those that are on cyclosporine, their

18 renal function is decreasing. Those that

19 cyclosporine has been removed, their renal function

20 is improving.

21 So, one more slide.

22 [Slide.]

23 This is more conservative. We are looking

24 at a slope after twelve months. One can see that

25 the slope analysis for those who remain on

233

1 cyclosporine is still negative and significantly

2 different from zero. On the other hand, the slope

3 for Rapamune is, one would say, not significantly

4 different from zero. So it is neither--it is flat.

5 Once again, the difference between the two

6 treatments is significant.

7 DR. CAVAILLE-COLL: On those slides, do

8 you have the actual n's of the numbers that are

9 included in each one of those analyses? Which

10 subset of the study is being looked at here?

11 DR. BURKE: I don't have them on the

12 slide. Obviously, in the first set, to be

13 included, they would have to be on-therapy at six

14 months and have at least two points to be able to

15 determine the slope. So, in the first analysis, if

16 one looks at the rate of discontinuation before six

17 months, let's say there were about 190 or 200

18 patients, approximately, in each group. So it is

19 not an intent-to-treat analysis. We do exclude

20 those that discontinued before we could establish

21 the slope. But they didn't have to complete twelve

22 months to be included.

23 DR. NEYLAN: Thank you for the opportunity

24 to present that data.

25 DR. ENGLUND: Are there specifically more

234

1 questions relating to what has just been discussed

2 now?

3 DR. HUNSICKER: Could I just say, from my

4 point of view as an amateur statistician, that

5 these are very encouraging data but this is still

6 not the definitive analysis. It needs to be done

7 right and we shouldn't try to rush this.

8 I am willing to trust that, between the

9 company and FDA, that they can look at this and

10 make sure that they have got the best possible

11 analysis. But this is not a trivial issue. This

12 really is at the nub of where I said--we are asking

13 whether we are paying for the increased number of

14 rejection episodes with something substantial. It

15 has got to be convincing.

16 DR. ENGLUND: Dr. Ebert, would you like

17 to--

18 DR. EBERT: I don't really have a lot to

19 add to what has been discussed. I think, maybe to

20 summarize my own thoughts, it appears that the

21 efficacy really depends on the definition that one

22 uses. If you are talking about acute rejection,

23 obviously, there is a difference. If you are

24 talking about renal function at a later time, at

25 least from the twelve-month data, it appears that

235

1 there may not be a difference.

2 One is kind of, I think, challenged to

3 decide whether the early rejections are more of a

4 bump in the road or are they considered to be

5 failures. I agree that the analysis needs some

6 improvement. I would like to see the

7 intent-to-treat analysis of renal function over

8 time to be able to try to get an overall

9 determination of the efficacy with this particular

10 intervention.

11 With regards to the monitoring, again, I

12 think there is some evidence for concentration

13 versus effect. I don't know that it is strong

14 enough that, as noted earlier, if we should get

15 into the "should" say side of "should" monitor

16 versus perhaps making a statement and saying that

17 the majority of patients who experienced rejection

18 had a concentration below X and that elevated lipid

19 concentrations were associated with a concentration

20 above X and then maybe leave it at that and to try

21 to enable the clinician to use those serum

22 creatinines in patients where it is indicated.

23 DR. ENGLUND: Thank you.

24 DR. SUTHANTHIRAN: I think, over the last

25 decade, we have been basically adding on

236

1 immunosuppressive drugs. We went from one to two

2 to three to four drugs. So I think

3 philosophically to try to keep transplant patients

4 on a lesser number of immunosuppressives is a very

5 attractive option.

6 I actually had a lot of difficulty with

7 this question about what should be the answer, like

8 many other members here. I think the data clearly

9 shows that the patient survival and the graft

10 survival are very similar. You don't pay a price

11 by holding back cyclosporine. The creatinine and

12 creatinine clearances are better. We don't know

13 what the significance is.

14 Clearly, some of the complications suggest

15 the number of hypertensive drugs you may need.

16 Blood pressure seems to be better with cyclosporine

17 withdrawal. These all I would put on the paucity

18 side of supporting cyclosporine reduction and

19 withdrawal.

20 On the other hand, I do think that the

21 incidence of acute rejection is increased. If you

22 see the data from the point of randomization, then

23 the increase is real, both in the 310 study as well

24 as the 212 study. Almost all of the drugs we have

25 approved in the last four to five years,

237

1 mycophenolate, Rapamycin, IL2-receptor antibodies,

2 were all approved on the basis of the ability to

3 reduce acute rejection in the six months.

4 In fact, that was the endpoint we all

5 used. Now, we are coming up with a strategy that

6 actually increases the acute rejection. On the

7 other hand, this acute rejection doesn't seem to

8 extol a very heavy price in terms of a one-year

9 graft-survival rate. So I am very concerned about

10 this acute-rejection increase.

11 The other issue is that, in both the

12 studies, about 20 percent of the patients were

13 nonrandomized. In other words, this kind of an

14 approach is probably not applicable to broad-based

15 patients but, perhaps, to more of a lower-risk

16 patient population that are much more a selected

17 patient population.

18 So this is all the data we have. I think

19 it is very difficult for us to make a very strong

20 case, either to vote no or to vote yes. But, as an

21 advisory committee, we have to come up with some

22 calculation and we can't take the Larry Hunsicker

23 route saying, "I have got a flight at 3:30." So we

24 need to make some recommendation.

25 I kind of lean towards a qualified yes. I

238

1 am not at all comfortable with the way the

2 sponsor's proposed indication of how this should be

3 changed. I share the view that the word "should"

4 be done. I think it is a very important point.

5 The proposed indication, I probably would

6 be more comfortable about would be to remove the

7 word "initially." Here it says, "It is recommended

8 Rapamune be used initially." I don't think we need

9 that word and just leave it as it was originally.

10 Then, the second part of the statement

11 where it says, "Cyclosporine withdrawal should be

12 considered two to four months after

13 transplantation," maybe--I don't know whether this

14 is feasible. One way of defining it may be

15 cyclosporine reduction or withdrawal may be

16 considered in a selected patient population.

17 I think, of all the protocols that were

18 used today and the data we saw, the best protocol

19 was the patients in 212 in group B who were on

20 low-dose cyclosporine and a good dose of rapamycin.

21 They had a very nice, about an 8 to 10 percent,

22 acute-rejection rate before they were randomized

23 and then they went into what was intended in the

24 study.

25 Mechanistically, there is some good data

239

1 to support synergy between cyclosporine and

2 rapamycin. I am not sure we need to have an

3 abstinence protocol, complete elimination. We may

4 get the best bang for the buck by having a smaller

5 dose of cyclosporine and have the option rather

6 than have the recommendation that it should be

7 eliminated.

8 So my suggestion would be to consider this

9 indication statement that would say something like

10 not just withdrawal but, "Cyclosporine reduction or

11 withdrawal may be considered." I think it is very

12 important to point out that this is in a subset of

13 patients, that we simply don't have the data to

14 recommend it universally, given the patient

15 population we have studied.

16 Also, it is very clear in the 18 percent

17 discontinuation in 310 and the 20 percent

18 nonrandomized in 212 that we need to focus it on a

19 very select population of patients. That would be

20 my thoughts at this time.

21 DR. ENGLUND: Dr. Shapiro?

22 DR. SHAPIRO: I don't have a lot to add to

23 what Dr. Suthanthiran said. I guess if I had to

24 answer the question it would be, "Yes, but." I am

25 not convinced that having, as the pivotal trial of

240

1 a large non-U.S. population makes it remotely

2 applicable to U.S. populations which I think are

3 more heterogeneous.

4 The analysis has done a lot of selecting

5 out, 18 percent in 310, 20 percent in 212, and the

6 selected-out patients did very poorly, as I guess

7 one would have expected. And then we had 37

8 percent failure in the 310, in the group randomized

9 to Rapa. So you are dealing with sufficient

10 winnowing that you get close to a cherry-picking

11 situation with relatively small numbers of patients

12 who, in fact, did quite well.

13 I share all the concerns about

14 intent-to-treat but the reality is that there are

15 some very convincing data about how well the

16 patients who made it through the gauntlet of

17 getting randomized and not having an efficacy

18 failure, they did quite well but they certainly do

19 not represent the mainstream of the patients that I

20 transplant and I don't think they represent a great

21 deal of the mainstream of the patients waiting for

22 kidneys right now.

23 If you are going to say that this is okay

24 to do, you are going to have to word it in a very

25 careful way because, otherwise, you will open the

241

1 door to having a lot of kidneys ruined by people

2 that are doing this in the wrong way and applying

3 this to the wrong patients.

4 I think there are some very narrow

5 indications for patients who have sailed through

6 their transplant and are doing quite well who may

7 be able to tolerate the increased risk of rejection

8 who will do well without a calcineurin inhibitor.

9 What is less well defined is who those

10 patients are. I am not sure we have enough data to

11 say with confidence who those patients are and who

12 those patients are not.

13 DR. ENGLUND: Dr. Johnson?

14 DR. JOHNSON: I would like to make a

15 comment and, perhaps, Dr. Neylan can respond if he

16 it appropriate, but I am somewhat troubled by the

17 data with respect to the charge that the committee

18 has given us. My difficulty is that, as a

19 practitioner, I would utilize the drug very much

20 similar as the sponsor has suggested in many

21 instances.

22 But, as a committee member, in respect to

23 evaluating the data, particularly in regards to

24 safety, I am having some difficulty. The

25 difficulty, really, revolves around the question of

242

1 consistency with respect to the fact that the data

2 that is presented is not consistent with the target

3 population.

4 I don't want to beat the dead horse but

5 the main emphasis of the data is the preservation

6 of renal function with the removal of the

7 calcineurin inhibitor. However, in previous

8 studies that were shown here a few years ago when

9 the drug was first approved, the sponsor showed

10 that the African-American population was not

11 comparative to the Caucasian population with

12 respect to rejection, particularly at the lower

13 doses.

14 We are now presented with data that really

15 shows, or a study that really shows, no data with

16 respect to this subgroup. It is pretty easy to

17 say, "Okay; well, let's just exclude that subgroup

18 in the labeling," but I don't think it is that

19 easy. As was mentioned, demographic data that was

20 given in the presentation stated that 23 percent of

21 kidney recipients in the United States are black.

22 But, in reality, as was noted, 35 percent

23 of the waiting list is black and UNOS is dealing

24 with that disparity by lessening, and maybe even

25 eliminating, HLA matching with regards to kidneys

243

1 in the future and, therefore, that population will

2 likely expand and, in some areas, may represent 50

3 percent or half of the patients who are going to be

4 transplanted.

5 We also showed that the benefit in

6 eliminating the calcineurin inhibitor, to some

7 degree, is eliminated in those patients who have a

8 rejection episode. Therefore, if you have half of

9 the group, just hypothesizing, that may be eligible

10 for a protocol such as this, who you know are a

11 higher responder group, who may have a higher

12 incidence of rejection, those folks may, indeed,

13 have very little benefit from this regimen and, in

14 reality, may be harmed by this because we don't

15 know.

16 Maybe the rejection rates in this group

17 are going to be zero. Maybe 10. Maybe a third.

18 Maybe a half. We just don't know and so it is very

19 troubling for me to sit back and think about how

20 you would label this given the data that we have to

21 evaluate and given the demographics of the United

22 States renal-transplant population and what that

23 population is likely to look like a few years from

24 now.

25 DR. NEYLAN: I would be happy to respond

244

1 if you like.

2 DR. ENGLUND: Actually, I doubt that you

3 could.

4 DR. NEYLAN: I would like to. Thank you.

5 First, we certainly don't want to give the

6 impression that the data from 310 and 212 should be

7 universally applied or rather one-size-fits-all

8 kinds of thinking. In fact, in the proposed

9 labeling document that we have sent to FDA, we have

10 said that the data in black patients in

11 insufficient to make a specific recommendation.

12 The current labeling for Rapamune has

13 looked, as you mention, quite thoroughly at the

14 potential difference that black patients might well

15 require a different dosing strategy and, indeed,

16 the 301 study is supportive of that idea in that,

17 from the efficacy standpoint, the acute rejections

18 were statistical lower for black patients in the

19 5-milligram dosing arm as opposed to the

20 2-milligram.

21 However, we realize that that, in itself,

22 is not enough and we have continued additional

23 studies and we have postmarketing agreements with

24 FDA to continue in these efforts to expand our

25 understanding of how Rapamune is best utilized in

245

1 black patients.

2 I think one of the overriding concerns for

3 us in presenting this data on top of the previous

4 data is that we want to afford clinicians the

5 opportunity to optimize and individualize treatment

6 strategies. I think Dr. Hunsicker has intimated

7 earlier that we are long past the early days of

8 transplantation where we can look at a kind of

9 one-size-fits-all approach. I dare say, also as

10 Dr. Hunsicker mentioned earlier, that in

11 near-future applications to this committee, many

12 other groups may be proposing strategies which look

13 at the long-term maintenance to start from a sort

14 of nonequivalence standpoint and say, "Okay; that

15 is the bench where we have to stay level but, from

16 there, what can we do to reduce long-term

17 toxicities?"

18 So what we have shown you today is a

19 balance of some tradeoff. We will agree with

20 everything you have said that this isn't meant to

21 fit all patients. But we want to get this out

22 there because we think it represents a potential

23 viable option for some patients.

24 We studied the patients we did because

25 that is who we had at hand. But we know our job

246

1 isn't finished. We have additional studies to do

2 and we want to take those on.

3 DR. ENGLUND: While I have you up there,

4 my question is what is being planned or actually

5 done in terms of pediatric studies?

6 DR. NEYLAN: I'm glad you asked. We have

7 three pediatric studies ongoing now. Two of them

8 are being done in concert with Napratix and NIH.

9 The first is a large-scale study of some 400

10 patients looking at the use of Rapamune in

11 combination with cyclosporine to determine whether

12 steroid withdrawal is feasible in this group of

13 patients in which corticosteroid complications are

14 especially problematic.

15 We have a second large-scale study also

16 being done in concert with napratix looking at the

17 potential efficacy of Rapamune in combination with

18 either of the calcineurin inhibitors for high-risk

19 pediatric recipients, those being defined as

20 patients who have had at least one prior episode of

21 acute rejection.

22 There, we are looking at not only

23 longer-term graft survival but also examining

24 histology at later dates. Finally, we have a study

25 being done through an NIH grant looking at

247

1 calcineurin-inhibitor-free regimens in the

2 pediatric population. So we fully recognize our

3 responsibilities in that area as well and we are

4 moving forward.

5 DR. ENGLUND: Do you have any studies

6 ongoing that haven't been mentioned here in terms

7 of African-American and Hispanic populations?

8 DR. NEYLAN: Yes; we do, and we have

9 ongoing discussions with FDA about future trials as

10 well. One of those is, indeed, a postmarketing

11 commitment that stems from the original submission

12 of the 301 and 302 data.

13 DR. ENGLUND: Dr. Hunsicker?

14 DR. HUNSICKER: On a slightly different

15 tack, I noticed, of course, that you had a lot of

16 biopsies at twelve months. Do you have anything to

17 say about what you found in the biopsies in terms

18 of fibrosis?

19 DR. NEYLAN: We, unfortunately, have run

20 into much the same problem that other protocols

21 have in their attempt to incorporation protocol

22 biopsies into the regimens. If we could show this

23 next slide.

24 [Slide.]

25 What we found, in asking all principal

248

1 investigators to obtain protocol biopsies in all

2 patients enrolled in 310 was that many of them were

3 fairly good at getting the baseline biopsies, those

4 biopsies at the time of transplantation. But,

5 unfortunately, we had a much lesser number, roughly

6 a third of the study population obtaining

7 twelve-month biopsies as dictated by the protocol.

8 So our ability to analyze the paired

9 biopsies in these two treatment groups is severely

10 limited by the small numbers. What we found with

11 those small numbers is that the composite score,

12 the chronic allograft damage index, which is a

13 summation of individual elements 0 through 3 for

14 the six categories and can be ranked, therefore,

15 from a summation score of 0 to 18, was, for both

16 groups, on the order of about 3.5 and not

17 statistically different.

18 We found that there was probably a little

19 bit of sampling bias as well in obtaining these

20 biopsies in the net slide.

21 [Slide.]

22 In that, again, of these very small

23 numbers of patients, the renal function at the time

24 in which these biopsies were obtained was somewhat

25 different for the yes/no of obtaining biopsies

249

1 between these two treatment arms so that, for the

2 Rapamune group, the biopsies were more likely to be

3 obtained. These are numeric trends, not

4 statistically significant. But the GFRs tended to

5 be slightly lower for those that got biopsies as

6 opposed to those that did not whereas, for the

7 control group, the GFRs were just the opposite.

8 They tended to be slightly more than those that did

9 not.

10 Given that this is an open-label study and

11 clinicians knowing full well that patients are

12 going to have an important element of the regimen

13 removed, it is not surprising that there was as

14 sort of differential predisposition to this kind of

15 behavior.

16 I should add that, as I said, this study

17 is five years. We held an investigator's meeting

18 in the fall just at the time now where the patients

19 are starting to enter the three-year mark. We have

20 exhorted, extolled and badgered in any way we can

21 the investigators to obtain three-year biopsies on

22 as many patients as possible because we, again, do

23 have a number of baseline biopsies.

24 So, even if these investigators haven't

25 gotten the one-year biopsies, we are hoping they

250

1 will get the three. It may be that the difference

2 in function that we are seeing may be more easily

3 expressed in the histology with a longer period of

4 time on these two separate regimens.

5 So we are certainly anxious to see those

6 three-year biopsies and certainly, as the data

7 becomes available, they will be brought before the

8 FDA.

9 DR. HUNSICKER: I guess I find myself with

10 another question for my FDA hosts over here, and

11 specifically Dr. Cavaille-Coll, I have spoken with

12 you about this before. Let us assume that they

13 submit, and you agree to, an analysis of the

14 creatinines over time that is very rigorous and

15 that shows something similar to what we have seen

16 here which is a diverging trend, a trend for the

17 creatinine to be rising or to be more negative, if

18 you will, in the continued cyclosporine and Rapa as

19 opposed to the rapamycin, itself.

20 Let's assume that the qualified yesses I

21 heard some of turn out to be the majority opinion.

22 I don't know quite what I am asking here but what I

23 am trying to get across is that it would seem to me

24 this is one area where it is absolutely crucial

25 that these patients be followed long-term in an

251

1 intent-to-treat fashion so we find out whether

2 these early changes do, in fact, mature into a

3 difference in graft survival, which is what we are

4 looking for.

5 I think that this is one of the places

6 where whether you speak about this in terms of

7 accelerated approval with ultimate validation later

8 on or however you want to term it, we have got to

9 assure that if there is an indication given, we

10 have to confirm what this means in the long haul.

11 DR. CAVAILLE-COLL: Are you suggesting

12 that, before we take any kind of decision, that we

13 should be looking at the analysis you are proposing

14 at these folks up to 24 months as they are entering

15 their third year and that that should be the

16 intent-to-treat analysis including all the patients

17 that were in the study that still have a kidney to

18 generate.

19 DR. HUNSICKER: It is not going to be a

20 trivial issue because there are some patients who

21 are going to be lost because they have failed, and

22 that is obviously an informative failure and you

23 have got to figure out how you are going to analyze

24 that, whether you do it by medians, or whatever.

25 But I believe that. I am going to assume

252

1 you are going to get some recommendation. All of

2 what we do, all of what my colleagues do because I

3 don't vote today, is a recommendation to you

4 anyway. What I am suggesting is that, however this

5 comes out, my feeling is that I would not want to

6 act, were I voting, until I saw the results of a

7 really well-done slope analysis because I think we

8 are trying to bet a known, maybe not too great,

9 negative today against a promise of something that

10 may be substantial and I want to have that promise

11 of what is substantial in terms of creatinine be

12 tied down as best I can.

13 But, no matter how you do it or we do or

14 anybody does it today or tomorrow or the day after,

15 the proof of the pudding is going to be in what

16 happens five years from now. I think that one of

17 the things that is essential is that there be an

18 understanding that, if there an approval given of

19 some form, that this approval has to be validated,

20 if you will, downstream by seeing whether these

21 differences in function, in fact, translate

22 ultimately into differences in graft survival.

23 DR. ALBRECHT: If I may go ahead and sort

24 of paraphrase what I think you said and, really, in

25 a sense, review some of the options that actually

253

1 are available to us. I think the issue you raise

2 about, let's say, five-year follow-up data in the

3 setting of a regulatory decision earlier than that,

4 under the regulatory options available to us, we

5 could take such a course if we were to approve and

6 indication and then request a phase-IV commitment

7 for data long-term.

8 That is certainly one approach and that

9 would be the kind of approach where we felt that a

10 decision at this time was based on adequate

11 information and one that we could comfortably

12 reach. Clearly, this is why we are asking you to

13 assist us with making this decision and that, in

14 fact, the long-term data is just to confirm and

15 make us comfortable that the hypotheses and

16 decisions we made early are, in fact, confirmed.

17 However, if, to take it to the next stage,

18 if we are dealing with--we are construing surrogate

19 endpoints where we believe they are likely to

20 predict the long-term outcome but we really don't

21 have the data on which to make that conclusion,

22 then there is, under the regulation, a section

23 called Subpart H in 314.500 where what we say is

24 this is an approval based on a surrogate which we

25 believe will have predictive value long-term

254

1 clinically but we are not certain.

2 As part of that action, the company is

3 required to continue clinical trials--in this case,

4 the long-term follow up for example--and provide

5 such information to, in fact, confirm or show that

6 these results are not consistent over time and then

7 the regulatory action would follow based on those

8 results.

9 Having gone over those two, I think what

10 we would like to ask you, as the committee, as you

11 are discussing and voting on this, is to provide us

12 your best advice on whether you believe the

13 findings now, the likelihood is that what we would

14 be doing long-term is confirming--or whether the

15 information is such that, at this point, it would

16 premature for you to expect that these results are

17 predictive.

18 In fact, the final option really would be

19 to say the information we have is so preliminary

20 that we do need further data before we can even

21 reach a decision. So I think those are the three

22 options before us and we look to you for guidance

23 on which of those really you believe scientifically

24 and clinically are supported by the data presented.

25 DR. ENGLUND: Are there any comments

255

1 before we proceed with voting on No. 1?

2 DR. SHAPIRO: I have a question for John.

3 Do you think that, if you had more time, more

4 follow up, maybe an additional trial that would

5 strengthen and sort of amplify in the data that you

6 have presented, that that would make your position

7 a little bit stronger but, perhaps, also, more

8 generalizable and would it be worth it from Wyeth's

9 point of view to try to do that to increase

10 everybody's confidence in your claim?

11 Right now, everybody is sort of saying,

12 "Yeah, well, for a very small subset of patients

13 who are doing really well, this is probably a good

14 thing but they represent not a huge number of

15 patients whom we are transplanting today in real

16 life."

17 The question is, if you had more

18 information, would it be stronger from the

19 company's point of view to have a stronger

20 indication.

21 DR. NEYLAN: Let me address that in two

22 parts. One, yes. Wyeth is, in fact, even now,

23 undertaking a variety of studies which further

24 explore this issue, the issue of the use of a

25 reduced calcineurin inhibitor, the issue of

256

1 continued exploration of a withdrawal strategy.

2 In fact, in that latter point, we are now

3 initiating one of the largest clinical trials in

4 the maintenance population that has ever been

5 undertaken and that is a randomized comparative

6 analysis for the maintenance population of a

7 continuance of calcineurin inhibitors versus a

8 conversion and taking patients with all ranges of

9 renal function.

10 We are building into that protocol

11 biopsies and a variety of what I believe are going

12 to be very important elements to help the community

13 better understand these issues as they relate to

14 the long-term care of recipients.

15 So the commitment is there. It is

16 ongoing. What we have with these two studies,

17 however, is now two-year data for 310, emerging

18 three-year data, and a commitment to go to five.

19 At each of these time points, these twelve-month

20 time points, we are seeing a consistency or a

21 confirmation, if you will, of the elements that

22 have come before.

23 So, while the commitment to continue this

24 study and continue the reporting of its results is

25 there, I think it would be difficult for us to

257

1 start from scratch at this point already having put

2 in so much time and effort. I think it would be a

3 disappointment if we were not able to move forward

4 with the indication today.

5 DR. ENGLUND: Dr. Hunsicker?

6 DR. HUNSICKER: Could I respond to your

7 comments, which were very clarifying for me. You

8 know that I am not going to vote but I can still

9 give you my opinion and that is, if I can start

10 with the last one and move forward, I think it

11 would be unjust to the sponsor to say that we just

12 don't know anything.

13 There are issues of how to apply what we

14 have here. We don't know quite who the population

15 is at this point and that has got to be addressed

16 as a separate issue. But if you take the

17 population that we have seen, the data that we have

18 are fairly convincing that the cost is small but

19 real and it appears that the long-term benefit is

20 going to be real.

21 That remains to be qualified by what I

22 have said. I want you people to do a proper, and

23 to agree on a proper, analysis of this issue that

24 is--intention-to-treat and all that. I have said

25 that, so I don't have to go over it again.

258

1 But if, in fact, the outcome were to show

2 that there is this initial improvement in function

3 and that, in fact, over time, that difference in

4 function between the cyclosporine and the other arm

5 widens rather than constricts so that there is a

6 presumption that the creatinine is getting better

7 in time relatively speaking, the sirolimus-only

8 arm, it seems to me you have as good data as you

9 are going to have that there is likely to be a

10 long-term benefit short of actually doing the

11 experiment.

12 So I would not feel, given the

13 restrictions that I have said about what the

14 population is, that it would be just to say that

15 these folks haven't shown you anything.

16 On the other hand, if we go to the other

17 extreme, is this already cold-cocked? No; it can't

18 be because no one yet has done an interventional

19 study in which they have said, "I am going to do

20 something to lower the creatinine," some

21 intervention to lower to creatinine, and show that

22 this transforms ultimately into prolonged graft

23 survival.

24 I think the rationale behind it is strong.

25 I have written about that. I have talked about

259

1 that. I believe it. I think that it is reasonable

2 to think that but it has never been shown. I would

3 go on further to say that our obligation to make

4 sure we know what the outcome is of an intervention

5 that lowers the serum creatinine or preserves GFR

6 acutely and to see whether this translates is very

7 important because, as Dr. Neylan said, you are

8 likely to see a whole mess of this coming down the

9 pike an we have got to settle this once and for

10 all.

11 Is the presumption that a lowering of

12 creatinine and widening of things leads to better

13 graft survival in fact supported--in fact

14 supported--by our data at the end of time. So I

15 don't think that you can say that it has been shown

16 because it hasn't. Nobody has shown that.

17 So I find myself very much in the middle

18 here, as Marc knows that I have for years. I think

19 that this is something for which there is very

20 strong presumption, a basis, perhaps, for early

21 approval but with the requirement that this

22 assumption that an early improvement in function

23 will translate into longer graft survival must be

24 documented.

25 MR. LAWRENCE: A point of clarification,

260

1 if I could. We are about to vote on Question 1 but

2 I am not sure what Question 1 says. The company is

3 asking for language that says that cyclosporine

4 should be withdrawn, or should be considered to be

5 withdrawn, after two to four months. Is that what

6 we are voting on, that they have shown us

7 sufficient data to say that that is--

8 DR. ENGLUND: We are not voting on the

9 wording. We are not voting on the "should" or

10 "may." We are voting on does the--and we can ask

11 for clarification, but we are voting, does the data

12 support the contention that withdrawing

13 cyclosporine is safe and effective.

14 DR. ALBRECHT: That's correct. The

15 question is not how we should label the product but

16 whether the committee does believe that the data

17 that Wyeth has presented show that this regimen is

18 safe and is effective.

19 Actually, if I may comment a little

20 further, having heard the discussion, I find that

21 it would probably be reasonable to paraphrase the

22 second part of that question to, if the answer is

23 yes, not just the population or subpopulation that,

24 perhaps, could be discussed but I also got the

25 sense that, perhaps, as part of that question, if

261

1 the committee does believe yes is the direction in

2 which the members would like to vote, what

3 additional information would be needed before that

4 yes could take place.

5 DR. ENGLUND: So the FDA is letting us ask

6 for more information.

7 DR. ALBRECHT: The more information could,

8 of course, be more analyses.

9 DR. ENGLUND: From what we already have.

10 At this point, what I would like to do is briefly

11 summarize. I am putting, perhaps, my perspective

12 but I will try to be global. Then, at this point,

13 I would like us to vote on question 1 because I

14 think we have to vote on question 1 before we can

15 decide if we are going to answer a. or b. We are

16 not going to answer both of them because it depends

17 on how question 1 goes.

18 I think, at this point, I have several

19 comments. Number one, I think we need to

20 congratulate the pharmaceutical company for

21 proposing and carrying out a relatively complicated

22 study in the withdrawal of immunosuppressives. To

23 my knowledge, this is the first study that I have

24 seen that has been carried out with 100 percent

25 compliance. In the era--in my field of more

262

1 antivirals, I never get that. I think this is

2 amazing and they are to be congratulated and that

3 we appreciate the work that has gone to give us

4 this kind of numbers.

5 I also think that the theory and the

6 theoretical concerns as to what they are using as

7 our endpoints are good. The fact that they can't

8 tell us for sure what elevated creatinine means at

9 one year is not--they should not be penalized for

10 that because that is the state of the art.

11 So I think we, on the committee, recognize

12 some of the good work that has gone into this but,

13 in reviewing the comments from the different

14 speakers, I think we have, as a group and as a

15 committee, certain sincere difficulties and I am

16 going to just briefly go over them.

17 We have, as a group, a very big issue with

18 the population. To my knowledge, we did not show

19 living related donors in Americans. That is,

20 perhaps, going to be a very big population that we

21 would be concerned about. We have different

22 ethnicities that have not been addressed, at least

23 in our country, and these are big issues from my

24 point of view that the pediatric data, of course,

25 is still barely getting started. I feel that is an

263

1 issue.

2 So we have patient-population concerns.

3 And then I think we have some big analysis concerns

4 that, with the help of our statisticians and

5 pharmacology colleagues, we really have some

6 concerns about what is intention-to-treat, what is

7 a really appropriate comparison.

8 I have concerns about the toxicity and

9 safety. I mean, what is a low potassium? I don't

10 care of people's platelet count is 10,000 less. I

11 care if they are thrombocytopenic and I wasn't

12 able to get good values as to what some of our

13 toxicities really were. So I think that there is

14 some more analysis, that I think the data is here.

15 I think the committee as a whole has raised some of

16 these issues.

17 Last, but not least, is the use of

18 surrogate endpoints which we, as a committee, and

19 with our specialties, have to realize that that is,

20 in fact, the state of the art today. I think that

21 is important for us to realize. As much as we do

22 want more, that is what we have today.

23 So, with that, I have tried to summarize a

24 lot of people's concerns and comments, at the risk

25 of adding a little bit of my personal

264

1 interpretation.

2 With that, I think I would like to take a

3 vote and I would like to start at this end of the

4 table because we have started at that end of the

5 table first. For this vote, we really have to say

6 yes or no to question 1, or abstain, I guess.

7 But, do the data presented support the

8 effectiveness and safety of cyclosporine withdrawal

9 two to four months after kidney transplantation in

10 patients treated originally with a combination

11 regimen of sirolimus, cyclosporine and

12 corticosteroids?

13 Could you please say your name before you

14 vote so it can be taken down in the minutes.

15 Dr. Johnson?

16 DR. JOHNSON: Lynt Johnson. I guess I

17 would say no with a qualifier. But I guess it gets

18 registered as a no and it relates to the lack of

19 data representing the African-American population

20 which may, in turn, be a group that has benefit

21 from this regimen. As I got the sense of it, it

22 seems like it was more leaning towards yes with the

23 restriction of a population. I would hate to

24 restrict that population which may have benefit. I

25 just don't know.

265

1 So, with the absence of that data, it is

2 very hard for me to support question No. 1.

3 DR. ENGLUND: Dr. Shapiro?

4 DR. SHAPIRO: Ron Shapiro. Yes, but with

5 many of the same qualifications.

6 DR. SUTHANTHIRAN: Suthanthiran. I

7 actually think you can't split question 1 from 1a

8 because we are really saying yes or no to the first

9 question. I would say that I would say a qualified

10 yes in the sense--if the proposed indication is as

11 stated by the sponsor, we can't vote yes. I can't

12 vote yes on it.

13 But, if that is modified to say that "may"

14 be considered for withdrawal in certain low-risk

15 patients, I would vote yes. So I think, in my

16 mind, Question 1 and 1a and 1b are so inextricably

17 linked, I would find it difficult to--

18 DR. ENGLUND: Okay; so you are a yes

19 qualified as opposed to a no qualified.

20 DR. SUTHANTHIRAN: With the modification

21 in the proposed indication.

22 DR. ENGLUND: Steve Ebert?

23 DR. EBERT: Steve Ebert. To the question

24 that is posed, my answer is yes. I will hold off

25 on comments with the follow up.

266

1 DR. ENGLUND: Dr. DeGruttola.

2 DR. DeGRUTTOLA: Victor DeGruttola. I

3 would say that this is a gray zone. It appears

4 that there are patients who would benefit from this

5 regimen. It also appears the answer to the

6 question depends on the clinical importance of

7 acute rejection which, I understand, has been used

8 as an endpoint in some trials.

9 Given that concern, I would give this a

10 qualified no but, again, emphasize that there does

11 appear to be benefit in some populations and if

12 that can be further specified, then I think that

13 that fact should be taken into consideration when

14 FDA makes its decision.

15 I know that is a long vote, but--

16 DR. ABERNETHY: Darrell Abernethy. No. I

17 need more data and more analysis. So, at this

18 point in time, I cannot say anything other than no.

19 DR. ENGLUND: Dr. Auchincloss.

20 DR. AUCHINCLOSS: Auchincloss. No. I

21 think study 212 might as well be thrown out. I

22 don't ever think it is going to be useful. I think

23 they need longer and more analysis of the 310

24 study. I think they are going to need some

25 additional data from an additional study. So, no;

267

1 I don't think that this is ready for a label change

2 even though, as I have indicated, I will probably

3 go home and do it on a patient.

4 DR. ENGLUND: Mr. Lawrence?

5 MR. LAWRENCE: William Lawrence. My vote

6 would be yes but with the same reservations

7 expressed by Dr. Suthanthiran and Dr. DeGruttola.

8 I have serious reservations about applying this too

9 broadly but I think the answer is more yes than no.

10 DR. ENGLUND: I am sitting hedging because

11 what I am hearing is yes, not all, but we are

12 hearing a lot of yes, buts and no, buts, which is

13 difficult. But I think I would say no, but. The

14 reason for that is that if I were having to say

15 what would be the patient population to select, I

16 can't do it.

17 If they are going to expect my help, our

18 help, but my help, in designing who would benefit

19 from it, I know it is good. I know it is going to

20 work. But I don't know who to give it to and I

21 feel that is, at this point--and, perhaps, further

22 analysis could help us with that. So I am a no,

23 but.

24 But, having said that, there are four

25 yesses, five nos.

268

1 DR. ABERNETHY: So it was a tie-breaker.

2 DR. ENGLUND: The problem is I would say

3 that questions 1a and 1b are actually closely tied

4 in with question No. 2 in the sense that we need

5 more studies. I don't care what they are called,

6 but we need more studies.

7 For the yes people, how would you define

8 the patient population, if we could just briefly go

9 through the people who said yes. How would you

10 define it based on the data available?

11 DR. SUTHANTHIRAN: I am strictly going by

12 the data that I have. I know what patients to

13 exclude from entering in the study which would

14 include patients who had advanced to vascular

15 rejection. It would include patients who have

16 dialysis dependency. And it would include patients

17 who have more than 400 micromoles of creatinine.

18 These four patients, the four groups of

19 patients, cannot be entered into the study at this

20 time because we have no data to support that these

21 patients can be weaned off from cyclosporine. So

22 those patients can be excluded from the study.

23 We have no data on African-Americans so

24 those patients should not be included in the study.

25 DR. ENGLUND: I'm sorry; you mean--

269

1 DR. SUTHANTHIRAN: In cyclosporine

2 withdrawal. I am listing the group of patients for

3 whom we do not have data to make a recommendation.

4 So I have five groups of people who should not be

5 entered in a cyclosporine-withdrawal protocol at

6 this time.

7 I also know that related and living-donor

8 transplants are not--well, I am not that worried

9 about that patient population because usually, if

10 you can treat cadaveric patients, you can usually

11 get away in a living donor. So that is not an

12 exclusion criteria for me.

13 So, for my qualified yes, I would call all

14 these patients as high-risk patients, these

15 patients for whom I have no data, and I would allow

16 other patients to be entered in this. Potentially,

17 we can consider it for this protocol.

18 But I want to go back to what was said by

19 Mr. Lawrence about--I wouldn't put the word

20 "should." This is why I thought 1 and 1a are

21 inextricably linked. I think "should" gives a very

22 different connotation to the clinician. I think it

23 should be "may" or "might" be considered and I

24 would also add the line "in certain low-risk

25 patients."

270

1 DR. ENGLUND: Mr. Lawrence?

2 MR. LAWRENCE: Dr. Hunsicker and I were

3 discussing this point. This is who we thought

4 should be included. You say who should not. "May

5 be considered in low-risk patients," with an

6 asterisk to define that. "No delayed graft

7 function. No type III rejection. Adequate renal

8 function. There is too little data to address

9 blacks, Hispanics, Asians."

10 So, when I voted yes, I was voting yes

11 based on these people. If I had a chance to vote

12 no on the rest, I would vote no on the rest. But I

13 want to encourage withdrawal of immunosuppressive

14 drug. The flavor of that is a very attractive

15 flavor to people like me.

16 DR. ENGLUND: Dr. Shapiro?

17 DR. SHAPIRO: I wouldn't have much to add.

18 First, maybe second, transplant patients who have

19 kept their first kidney for a long time, low PRA,

20 low panel-reactive antibody level, if they have had

21 a rejection and easily treated, mild or

22 mild-to-moderate rejection with complete reversal,

23 preferably either no delayed graft function or

24 minimal delayed graft function, I think those

25 patients would fit into the category of patients

271

1 who might be candidates for a successful

2 calcineurin inhibitor withdrawal.

3 I guess, like everybody else, I would be

4 more concerned without more data.

5 DR. ENGLUND: I guess just as a response

6 to you is I would be very concerned about putting

7 something like--putting some of the things that

8 people have said here on a label when there is

9 actually nothing known about it. I agree, I think

10 that is what people will do and will do it at my

11 institution, but to put it on the label or to say

12 that that is who we are giving it to without any

13 data--we don't even have much living related data

14 even though I believe it is good. So this is my

15 comment.

16 DR. ENGLUND: I just want to echo what you

17 just said in that--and part of my comment was just

18 that. I tried to answer this as directly as

19 possible whether the data support the effectiveness

20 and that was the basis for my vote. But, as you

21 said, whether you are going to try to put

22 something--translate that into modifying the

23 labeling, I think is a much slipperier slope.

24 Whether this is something that should be

25 noted by practitioners and should be incorporated

272

1 into their daily practice as a "off-label" use in

2 selected individuals or whether this should be, as

3 you said, something that would be incorporated into

4 the labeling. I think those are two very different

5 actions.

6 DR. ENGLUND: We have two other questions

7 actually that are not to be voted upon but really I

8 think we should bring up for discussion.

9 DR. AUCHINCLOSS: Did the FDA feel like

10 they got their answer to the question? Do they

11 feel like they know what question we were

12 answering? I was just struck by the comment that I

13 heard over here because I think you were a yes

14 vote.

15 DR. ENGLUND: Yes was "yes, but."

16 DR. AUCHINCLOSS: But you wouldn't rewrite

17 the label?

18 DR. ENGLUND: Again, I agree with the

19 chair in that I don't think that there is enough

20 information available in a wide enough patient

21 population that I would feel strongly enough to

22 modify the labeling.

23 DR. AUCHINCLOSS: Because it was really

24 sort of that question that I used as the way to

25 hinge my vote one way or another.

273

1 DR. ALBRECHT: I think as I heard those

2 last comments about perhaps uncertainty whether the

3 information available was such that some of the

4 members felt comfortable about putting them in

5 labeling. The question that came to my mind is

6 what would be the information that you would

7 recommend or would like to see that would actually,

8 then, make you comfortable about this kind of

9 information being part of the label.

10 Again, not to go into the specific wording

11 but some of the ideas that you were voicing about

12 certain patient subsets or populations, what would

13 be the data that you would like to see before you

14 would be comfortable having this in the label?

15 Again, I ask that really just for completeness of

16 discussion, not to try to actually pin down the

17 criteria because, again, this is a very hard issue.

18 DR. ENGLUND: I would just like to

19 summarize. I think intention-to-treat data would

20 be the echo of our committee, without having to

21 call on everyone.

22 MR. LAWRENCE: Dr. Albrecht, I just have

23 to ask you again. When you say what would you have

24 to see before this would go in the label, I don't

25 know what "this" is. I have been trying to clarify

274

1 what "this" is. If "this" is that this should be

2 done generally, my answer is no, we haven't

3 seen--if the answer is that this is that this can

4 be contemplated by physicians based on the clinical

5 picture of the patient that they see and in certain

6 circumstances, then the answer is yes.

7 But you are not going to put that on the

8 label. So question 1 is not actually crafted in

9 terms of getting a yes or a no answer because the

10 qualifications are so manifest that everybody at

11 the table voted yes, but or no, but. I am not sure

12 the vote that you got today is worth much.

13 Obviously, there are serious reservations.

14 And, obviously, in some cases, it is appropriate

15 and should be encouraged. If you are going to put

16 that on the label, I will look for that.

17 DR. ALBRECHT: I think, as I said earlier,

18 there are the two aspects of making a regulatory

19 decision. The first is the burden of is the drug

20 safe and effective and that is what is specified in

21 the Food, Drug and Cosmetic Act.

22 Then the second aspect is how the

23 information about the safety and efficacy of the

24 drug is placed into the package insert so that it

25 can be understood by the individuals that would be

275

1 using the product.

2 I think we really are just asking you to

3 vote on the first issue of is the drug safe, is it

4 effective, and then the details of the words that

5 we will use to communicate that information, I

6 think, is the next level and some of the comments

7 that we are hearing, I think, indicate to me that

8 that is going to be a very challenging area.

9 DR. ABERNETHY: In terms of the further

10 information, I would suggest that--I think I need

11 to see a U.S. study that reflects the U.S.

12 transplant population. I would like to see

13 patients randomized at the time of transplant so

14 that we get a much better feel for where these risk

15 stratifications should be with regard to benefit

16 and then an intention-to-treat analysis.

17 DR. AUCHINCLOSS: Why is that? If the

18 issue is cyclosporine withdrawal versus no

19 withdrawal, why don't you randomized at the moment

20 of withdrawal? Then you can do all the

21 stratification you want at that point. It seems to

22 me they terribly muddied their picture by

23 randomizing up front and then withdrawing two

24 months later. By then, a whole series of events

25 had happened to the alternate population that

276

1 weren't comparable.

2 DR. DeGRUTTOLA: I would echo that. If

3 you could randomize at the time you would withdraw,

4 if that were a consideration, then I think that

5 would get most directly at the question.

6 I think one of the issues we are

7 struggling with is that, as Dr. Albrecht mentioned,

8 the regulations talk about the effectiveness and

9 safety of a drug, but we are really talking about

10 the effectiveness and safety of a strategy to

11 withdraw a drug which is a little bit more

12 complicated. I think that Dr. Auchincloss'

13 suggestion of randomizing at the time that you

14 would reduce the cyclosporine, that choice is an

15 interesting one to try and get most directly at

16 the--

17 DR. ENGLUND: That was the 310.

18 DR. AUCHINCLOSS: It is the 310. But we

19 are now asking for a United States study.

20 DR. ENGLUND: Right.

21 DR. ALBRECHT: May I actually add a few

22 more comments about foreign studies because I think

23 that this is an issue that is important to us. As

24 you know, drug companies do conduct studies in the

25 United States and abroad. As I indicated earlier,

277

1 the Code of Federal Regulations does allow foreign

2 studies to be used.

3 But if I understand, your concern is that

4 in the United States, there are a substantial

5 number of patients who have living, related-donor

6 transplants. In the studies that have been

7 submitted--in fact, it was 90 to 100 percent

8 cadaveric. So the concern is that we cannot

9 extrapolate the data from those studies to U.S.

10 patients--because I think it is just as important

11 to recognize that, in the area of international

12 drug development, we try not to stymie development

13 across the world but, rather, the concern is when

14 the patient population studied abroad cannot be

15 extrapolated to the patient in the United States

16 and, therefore, we cannot, then, effectively label

17 products.

18 So was that the concern, that the patients

19 being studied in these studies would not reflect

20 the U.S. population?

21 DR. ABERNETHY: I think you are saying it

22 in a, perhaps, too gracious way. The concern I

23 believe is that it is clear that when this study

24 was conducted, there was no commitment or no

25 possibility of including a population that is of

278

1 great interest here. Then, secondly, the concern

2 is is the practice pattern going to be the same in

3 one setting versus another.

4 I understand what you are saying about

5 harmonization on the one hand. On the other hand,

6 we are talking about getting an appropriate

7 practice for patients in the United States.

8 MR. LAWRENCE: A point of information.

9 Last year was the first time in this country that

10 living donors outnumbered cadaveric donors for

11 kidneys. So that is a material question.

12 DR. AUCHINCLOSS: I agree that there are

13 lots of potential concerns about an abroad

14 population. I don't think that the living-donor

15 issue is my primary one. I really would fairly

16 strongly believe that if this kind of thing works

17 for your cadaver-donor population, it is going to

18 be okay for your living-donor.

19 DR. ENGLUND: I believe that. Wouldn't

20 you like to see one or two patients?

21 DR. AUCHINCLOSS: Yes; I would like to see

22 one or two patients but that is not, by any means,

23 my primary concern about the patient population

24 abroad.

25 DR. SHAPIRO: Actually, the living donor

279

1 would be sort of a positive in that those are

2 patients who tend to do, as a group, better. I

3 think there is a sense that the American transplant

4 recipient population is more heterogeneous and that

5 the need for doing a study in the United States to

6 reflect that would be important.

7 DR. SUTHANTHIRAN: I think there is

8 another issue we need to address in terms of

9 randomization because we are going to be asked a

10 question, whatever the regimen or whatever the

11 protocol, is it safe and effective. The question

12 is being asked is it safe and effective for

13 transplant patients.

14 Now, at the time of randomization, certain

15 groups of patients are excluded from randomization

16 and they go into Arm C or nonrandomized. We are

17 always going to have this problem. We always have

18 this problem, it is safe but we cannot comment

19 about population A, B or C who were not excluded in

20 the randomization plan.

21 The only way to avoid the problem is to

22 really enter all patients into randomization.

23 Otherwise, we are going to revisit the issue all

24 the time. It is kind of a Catch 22. If you start

25 patients at zero time, all your transplant

280

1 patients, and then, let's say, at one month or two

2 months, you randomize, but you are really not

3 randomizing A B. You also have a group C because

4 of whatever notion that group C is a high-risk

5 patient population.

6 Now, when we are asked to answer the

7 question, is this protocol fine for transplant

8 patients, we are always going to say it is fine for

9 the transplant-patient population minus group C.

10 DR. AUCHINCLOSS: But that is true of many

11 studies. You have exclusion criteria and then the

12 results apply only to those that were not excluded.

13 DR. SUTHANTHIRAN: Right.

14 DR. ENGLUND: But the problem here is we

15 have not just those exclusion criteria but we also

16 have the exclusion criteria for all the people that

17 they didn't even--that weren't even enrolled in the

18 first place.

19 DR. DeGRUTTOLA: I think the point is

20 exclusion criteria, per se, shouldn't necessarily

21 be a concern. If it is not appropriate to withdraw

22 cyclosporine from some patients, then it is

23 appropriate to exclude them both from the study and

24 mention that in the label.

25 I think the issue is do you always want to

281

1 do the reduction of cyclosporine at three months so

2 that is specified then, or could there be a more

3 variable time at which you say, now is the time we

4 might consider reducing the cyclosporine. It might

5 be later, for example, in some patients. I think

6 maybe you could get at the issue that way in

7 allowing the randomization only to happen at the

8 time that you want to consider withdrawing it, not

9 necessarily fixed at three months by protocol but

10 allowing some latitude with that.

11 DR. SHAPIRO: If I could just defend the

12 selectivity on the part of the sponsor, you

13 want--this is pretty radical to stop a calcineurin

14 inhibitor and you want to load the dice to come up

15 with a trial that is going to give you--one, that

16 is going to give you a trial that is going to be

17 relatively safe to do and one that is not going to

18 fall on its face.

19 I think that they have succeeded extremely

20 well in doing a study like that, at least at a

21 first pass, in a relatively low-risk population. I

22 think to have done an allcomer study at the

23 beginning, one would have risked a result that

24 would have been harder to understand and, two,

25 would have been very difficult to do.

282

1 So I think that the rationale for looking

2 at selected populations initially was the right

3 one.

4 DR. ENGLUND: I would like us to move on

5 to what additional studies would we want, would we

6 ask for. I have heard from Dr. Abernethy.

7 DR. SHAPIRO: I would echo that. You

8 would need to do a large-scale American trial with

9 both living donor and cadaveric recipients and not

10 restrict entry on the basis of ethnic group. You

11 might want to restrict entry in terms of transplant

12 number and PRA to at least give it a shot of being

13 reasonable, just from a tactical point of view.

14 DR. ENGLUND: Would you consider

15 randomization at time of transplant or at a period

16 following transplant?

17 DR. SHAPIRO: The ideal thing would be to

18 randomize pretransplantation. But you are going to

19 increase your dropout rate enormously if you do

20 that. At some level, that is the cleanest. The

21 way to stack your trials so that they come out the

22 way you want them to is to randomize after you know

23 that you have got kidneys that are functioning in

24 patients who are doing well.

25 The ideal thing would be to randomize

283

1 pretransplantation.

2 DR. ENGLUND: Dr. Johnson?

3 DR. JOHNSON: I am not sure that I would

4 agree that they would have to redo the entire study

5 in the U.S. population. I think that--you know, my

6 main concern is that the African-American

7 population represents such a large proportion here

8 in the United States and I think that we need to

9 have some data. That data may come back and say,

10 yes, it is okay in certain conditions and they may

11 say it is worse.

12 But that is the information that I would

13 like to have because I think that, from the

14 question that was asked, can we extrapolate this

15 data, there is some extrapolation that I can do

16 with this data but I can't, based upon prior data

17 and based upon my knowledge of this group,

18 extrapolate it to that subpopulation.

19 So, in my opinion, I am not sure they

20 would need to redo the whole study. But I think

21 that they need to provide supplemental data in

22 African Americans in the United States in some

23 fashion so that we can make a judgment one way or

24 the other whether or not we need to provide a basis

25 to exclude or include them in this labeling in some

284

1 degree.

2 DR. ENGLUND: Any other comments about the

3 phase IV studies? I have heard that the pediatric

4 studies--I have heard about those and I think those

5 sound good and sufficient and I am pleased to see,

6 actually, the numbers that are being discussed for

7 the pediatric studies.

8 DR. SUTHANTHIRAN: I would add a biomarker

9 to the study. I think it would be terrific if

10 there is nice improvement in creatinine, there

11 appears to be a nice improvement in renal function,

12 it seems to hold out over time--I think it would be

13 very nice of a biopsy is really part of the

14 protocol and the patients get biopsied at defined

15 times.

16 I know logistically there are some

17 problems associated with it, but I think the study

18 would be improved so much if a biopsy is done in

19 all the patients and we can see a structural

20 correlation and a structural counterpart to this

21 improved renal function.

22 DR. ENGLUND: Dr. Shapiro?

23 DR. SHAPIRO: I would also point out that,

24 while protocol biopsies are very nice and I have

25 written about them and we have performed them and

285

1 we have published on them, and I have also been

2 slammed for them, or our paper has been, as being

3 of uncertain significance.

4 That is the problem with protocol biopsies

5 in the world today. The transplant community has

6 some people who are very interested in them and

7 think that they are wonderful and a large number of

8 people who think that they are nonsense.

9 DR. ENGLUND: I would just like to add,

10 from my experience in clinical trials, that it

11 makes recruiting in minority populations greatly

12 difficult. It makes some of the recruiting more

13 difficult in some of the populations, and I think

14 that is something to consider.

15 DR. SHAPIRO: It depends how you sell it.

16 DR. ENGLUND: You are better at it than we

17 are.

18 DR. ENGLUND: I don't really want to open

19 a can of worms on this, but I think, if additional

20 phase IV studies were going to be done, one might

21 also want to consider having a subset of patients

22 which, rather than doing prospective

23 concentration-controlled dose modification may want

24 to just start out immediately at a dose of, whether

25 it is 8 milligram a day, 10 milligrams a day,

286

1 whatever have you, and try to see whether, in fact,

2 doing dose titration really, in fact, does improve

3 on outcomes compared with just arbitrarily giving a

4 certain dose.

5 DR. AUCHINCLOSS: I know that the company

6 is thinking about different ways of using their

7 drug in combination with other drugs and they have

8 thought, not just about cyclosporine withdrawal but

9 they have thought about steroid withdrawal, et

10 cetera, et cetera, et cetera.

11 But I do think it is worth their while to

12 rethink again whether this is really their top

13 priority, cyclosporine withdrawal. To me, as I

14 looked at the 212 data which I found interesting

15 even though I don't think it is a good study, it

16 seems to me the message there is that high-dose

17 sirolimus with very low-dose cyclosporine is a

18 fantastic combination that is destroyed when you

19 withdraw the cyclosporine.

20 So I just wonder whether they want to

21 think again about whether their endpoint actually

22 should be cyclosporine withdrawal or cyclosporine

23 minimization.

24 DR. ENGLUND: Let me go, then, to question

25 No. 3 which I think we have kind of addressed, but

287

1 let's make sure we have gone through all of our

2 questions. Question No. 3 states, do we have any

3 additional comments or recommendations regarding

4 the study design and/or endpoints for controlled

5 clinical trials intended to support the safety and

6 efficacy.

7 In particular, one of the things which we

8 have, I think, discussed, but for a maintenance and

9 a maintenance withdrawal regimen which is going to

10 be coming up before this committee again, one

11 hopes, what comments do we, as a committee have?

12 What would we like to be seeing in these trials?

13 Any comments in addition to what has

14 already been stated? Perhaps our statistician?

15 DR. DeGRUTTOLA: I think points have

16 already been made about longer-term follow up and

17 the ability to relate some of the markers to longer

18 follow up. I actually think that randomizing at

19 the point when people would reduce the dose rather

20 than up front is probably better for the reason

21 that was mentioned, if you are going to have

22 dropouts and people that can't be entered into the

23 study. So I think that the design is actually more

24 appropriate.

25 DR. ENGLUND: Any comments or questions

288

1 from the FDA?

2 DR. ABERNETHY: It may have already been

3 said abundantly, but I think viewing this kind of a

4 study as essentially an equivalence study, your new

5 regimen of having one less medicine is really--you

6 are testing equivalence to the currently accepted

7 regimen and taking that point of view from day 1

8 and really understanding what that means would

9 certainly make the interpretation at this end much

10 better.

11 DR. DeGRUTTOLA: Yes; prespecifying what

12 equivalence means. I guess it is as little

13 confusing in that this study was intended to show

14 equivalence for some outcomes but superiority for

15 other outcomes, I guess prespecifying what the

16 criteria are for equivalence or noninferiority, as

17 was mentioned.

18 I also thought that Dr. Hunsicker made an

19 interesting comment about doing intent-to-treat

20 analyses of some of the toxicity results which is

21 not standard. Typically, that is done on-therapy

22 or as-treated. But, for the reasons that were

23 discussed, I think that that is something that

24 should be considered here as well.

25 DR. AUCHINCLOSS: You make a comment about

289

1 how you balance two what are, in effect, surrogate

2 endpoints when we are not sure that either is okay.

3 One is going to go up and one is going to go down.

4 I think that the outcome here was pretty much as

5 you might have predicted, a slight increase in

6 acute rejections and an improvement in renal

7 function. We are not quite sure how important

8 either one of those things are.

9 DR. DeGRUTTOLA: I think that that is

10 always a challenge in any study and there are a

11 couple of ways to approach it. One is if you

12 believe that you can predict or develop a

13 predictive model, as I believe Dr. Neylan gave one

14 example, so that you can essentially weight the

15 improvement or lack of improvement by the expected

16 clinical consequences.

17 What that presupposes is that you have

18 information to allow to relate the markers to the

19 clinical consequences and you know that that

20 relationship is not affected by the drug that

21 people are on because, in fact, that relationship

22 could differ by drug. So, it is a challenging

23 thing to do but I think that that is probably the

24 only way, really, to evaluate what the consequence

25 is going to be for the patient.

290

1 Other kinds of analyses that people might

2 do in this setting are quality of life. But,

3 because the surrogates that are being discussed

4 don't seem to have a direct clinical impact. At

5 least the acute rejection, from what I understood

6 did not. The creatinine, I am not so sure. It

7 wasn't discussed. Presumably not.

8 But, having some way to relate these

9 endpoints to their clinical effect on patients I

10 think is the only way to really address that

11 question.

12 DR. ENGLUND: With that, I would--

13 DR. CAVAILLE-COLL: One moment please. I

14 would like to get as much as we can out of this

15 question 2. I know this is about the last time we

16 are going to hear recommendations as well as

17 clinical-study designs and endpoints.

18 This has been going on the past year at

19 different meetings organized by AST and ASTS. But

20 I would still like to, before we leave here, get

21 the panel's opinion about whether they believe or

22 not that bettering renal function is important in

23 clinical studies in renal transplantation and

24 should every effort be done to attain

25 intent-to-treat information on renal function in

291

1 patients who discontinue study drug, for example,

2 as well as patients who stay on study? This is for

3 future studies.

4 DR. SHAPIRO: Yes.

5 DR. ENGLUND: Yes.

6 [Chorus of yesses].

7 DR. ENGLUND: With that, I would like to,

8 once again, thank the committee, nonvoting guests.

9 I thank everyone for their participation. Thank

10 you for your presentation. And I adjourn this

11 meeting.

12 [Whereupon, at 3:50 p.m., the meeting was

13 adjourned.]

14 - - -