FOOD AND DRUG ADMINISTRATION 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
CENTER FOR DRUG EVALUATION AND RESEARCH (CDER)
PERIPHERAL AND CENTRAL NERVOUS SYSTEM (PCNS) DRUGS ADVISORY COMMITTEE
Thursday, March 10, 2011 8:00 a.m. to 5:15 p.m.
Hilton Washington DC/Silver Spring 8727 Colesville Road
PERIPHERAL AND CENTRAL NERVOUS SYSTEM DRUGS ADVISORY COMMITTEE MEMBERS (PCNS) (Voting)
1 2 Jeffrey A. Cohen, M.D. 3 4 5 6 7 8 9 10
Professor and Vice Chairman
Director, Clinical Neurophysiology Fellowship Dartmouth Medical School
Dartmouth Hitchcock Medical Center Lebanon, NH
TEMPORARY MEMBERS (Voting)
Britt Anderson, M.D., Ph.D. (Acting Chair)
11 12 13 14 15
Assistant Professor, Department of Psychology University of Waterloo Waterloo, ON Canada Marshall S. Balish, M.D., Ph.D. 16 17 18 19 20 21 22
Assistant Chief, Department of Neurology
Washington Veterans Affairs (VA) Medical Center Associate Professor of Neurology
Georgetown University Washington, DC
Kevin E. Chapman, M.D. 1 2 3 4 5
Pediatric Epilepsy, Barrow Neurological Institute St. Joseph's Hospital and Medical Center
Phoenix, AZ Robert R. Clancy, M.D. 6 7 8 9 10
Senior Physician in Neurology and Pediatrics The Children’s Hospital of Philadelphia
Philadelphia, PA Thomas R. Fleming, Ph.D. 11 12 13 14 15
Professor, Department of Biostatistics University of Washington Seattle, WA Dean D. Kindler, M.D. 16 17 18 19 20 21 22
Executive Medical Director for Medical Neurosciences
Lily Jung Henson, M.D., M.M.M., FAAN 1 2 3 4 5 6
(Acting Consumer Representative) Medical Director, Neurology Clinic
Swedish Neuroscience Institute Medical Center Seattle, WA Andrew C. Leon, Ph.D. 7 8 9 10 11 12 13
DeWitt Wallace Senior Scholar
Professor of Biostatistics in Psychiatry Professor of Public Health
Weill Cornell Medical College New York, NY Eli Mizrahi, M.D. 14 15 16 17 18 19 20 21 22
Chairman, Department of Neurology Professor of Neurology and Pediatrics Baylor College of Medicine
Phillip L. Pearl, M.D. 1 2 3 4 5 6 7 8
Division Chief, General Child Neurology Children's National Medical Center
Professor of Pediatrics and Neurology George Washington University
School of Medicine and Health Sciences Washington, DC William H. Theodore, M.D. 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Chief, Clinical Epilepsy Section
National Institute of Neurological Disorders and Stroke (NINDS), National Institutes of Health (NIH) Bethesda, MD
TEMPORARY MEMBERS (Voting) 1 David M. Treiman, M.D. 2 3 4 5 6 7 8 9 10
Newsome Chair in Epileptology Director, Epilepsy Center
Barrow Neurological Institute Professor of Neuroscience
Research Professor of Bioengineering Arizona State University
Phoenix, AZ Hongyu Zhao, Ph.D. 11 12 13 14 15 16 17 18 19 20 21 22
Professor of Biostatistics, Statistics, and Genetics
Yale School of Public Health New Haven, CT
PERIPHERAL AND CENTRAL NERVOUS SYSTEM DRUGS ADVISORY COMMITTEE (PCNS) TEMPORARY MEMBER (Non-Voting)
1 2 3
Roy E. Twyman, M.D. (Industry Representative)
4 5 6 7 8 9 10 11
Vice President, Head of Central Nervous System Development
Johnson & Johnson Pharmaceutical Research and Development, LLC Titusville, NJ
GUEST SPEAKERS (Non-Voting) Jacqueline A. French, M.D. 12 13 14 15 16 17 18
Professor, Department of Neurology Director, Clinical Trials Consortium New York University (NYU) Comprehensive Epilepsy Center New York, NY Nancy R. Temkin, Ph.D. 19 20 21 22
Professor, Biostatistics and Neurological Surgery University of Washington
FDA PARTICIPANTS (Non-Voting)
1
Steven T. Dinsmore, D.O.
2 3 4 5 6 7 Medical Officer
Division of Neurology Products (DNP) Office of Drug Evaluation I (ODE-I) Office of New Drugs (OND), CDER, FDA
Norman Hershkowitz, M.D., Ph.D.
8 9 10
Team Leader, DNP, ODE-I, OND, CDER, FDA
Russell G. Katz, M.D.
11 12 13 14
Director, Division of Neurology Products ODE-I, OND, CDER, FDA
Xiang Ling, Ph.D. 15 16 17 18 19 20
Mathematical Statistician, Division of Biostatistics 1
Office of Biostatistics, Office of Translational Science, CDER, FDA
Diem-Kieu Ngo, Pharm.D, BCPS Designated Federal Officer
21 22
C O N T E N T S 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
AGENDA ITEM PAGE Call to Order and Opening Remarks
Britt Anderson, M.D., Ph.D. 11 Introduction of Committee 11 Conflict of Interest Statement
Diem-Kieu Ngo, Pharm.D., BCPS 15 FDA Introductory Remarks
Russell Katz, M.D. 21 Guest Speaker Presentations
Historical Control: Withdrawal to Monotherapy
Jacqueline French, M.D. 39 Historical Control for Epilepsy Conversion to
Monotherapy Methodology
Nancy Temkin, Ph.D. 74 Clarifying Questions 83 Industry Presentation
Overview of Epilepsy and Lamotrigine
C O N T E N T S (continued) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
AGENDA ITEM PAGE Brief Review of the Historical
Control Studies
LAM30055 – Design, Efficacy, and Safety Results Comparisons between LAM30055, US30/31 and the Historical Control Studies
John Messenheimer, M.D. 113
Summary and Conclusions Thomas Thompson, M.D. 148 Statistical Considerations Eugene Laska, Ph.D. 154 Clarifying Questions 166 FDA Presentation Lamictal XR (lamotrigine) Historical Control Trial Xiang Ling, Ph.D. 183
Clarifying Questions 201
Open Public Hearing 230
Panel Discussion/Questions 235
Discussion of Questions to the Committee 293
P R O C E E D I N G S 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 (8:00 a.m.)
CDR NGO: Good morning, everyone. Before we start, I'd like to remind you to please silence
your cell phones, BlackBerrys, and other electronic devices if you have not already done so.
Ms. Sandy Walsh, if you're here, please stand up. She's the press contact. She may be coming a little bit later. Thank you.
Call to Order and Opening Remarks DR. ANDERSON: Good morning, everyone. Thank you all for coming.
At this point we usually begin by going around the table and giving our names and our
connection to the panel. So perhaps we could start on your end, Dr. Twyman, and then we'll just go sequentially around the table.
Introduction of Committee
DR. TWYMAN: Certainly. Good morning. My name is Roy Twyman. I'm the industry
representative. I'm from Johnson & Johnson.
a child neurologist at the Children's Hospital, Philadelphia and the University of Pennsylvania School of Medicine. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
DR. MIZRAHI: I'm Eli Mizrahi. I'm a child neurologist. I'm from Baylor College of Medicine in the Department of Neurology and Pediatrics.
DR. CHAPMAN: I'm Kevin Chapman. I'm a pediatric neurologist at St. Joseph's Hospital and Barrow Neurological Institute in Phoenix.
DR. LEON: I'm Andrew Leon, a professor of biostatistics in psychiatry and in public health at Cornell Medical College in New York.
DR. FLEMING: Thomas Fleming, professor of biostatistics, University of Washington.
DR. ZHAO: I'm Hongyu Zhao, professor of statistics and biostatistics at Yale University.
DR. KINDLER: Dean Kindler, neurologist, Kalamazoo, Michigan.
DR. COHEN: Jeffrey Cohen, neurologist, Dartmouth Medical School.
CDR NGO: Commander Diem-Kieu Ngo, Designated Federal Officer for the meeting.
DR. ANDERSON: I'm Britt Anderson. I'm an adult neurologist, and I'm in Waterloo, Ontario.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
DR. HENSON: Lily Jung Henson. I'm a neurologist from Swedish Neuroscience in Seattle and clinical associate professor at the UDub, and acting consumer rep.
DR. BALISH: Marshall Balish. I'm a
neurologist. I'm local at Washington, D.C. VA and Georgetown.
DR. THEODORE: I'm Bill Theodore. I'm a neurologist in the intramural program at NIH.
DR. TREIMAN: David Treiman, adult
neurology, Barrow Neurological Institute, Phoenix. DR. LING: Xiang Ling, statistical reviewer at FDA.
DR. DINSMORE: Steven Dinsmore, medical reviewer at FDA.
DR. HERSHKOWITZ: Norm Hershkowitz, medical team leader, FDA.
DR. KATZ: Russ Katz, director, Division of Neurology Products, FDA.
again. Now I have a passage to read here. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
For topics such as those being discussed at today's meeting, there are often a variety of
opinions, some of which are quite strongly held. Our goal is that today's meeting will be a fair and open forum for discussion of these issues and that individuals can express their views without
interruption. Thus, as a gentle reminder, individuals will be allowed to speak into the record only if recognized by the chair. We look forward to a productive meeting.
In the spirit of the Federal Advisory Committee Act and the Government in the Sunshine Act, we ask that the Advisory Committee members take care that their conversations about the topic at hand take place in the open forum of the
meeting.
We are aware that members of the media are anxious to speak with the FDA about these
proceedings. However, FDA will refrain from discussing the details of this meeting with the media until its conclusion. Also, the committee is
reminded to please refrain from discussing meeting topics during breaks or at lunch. Thank you.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Conflict of Interest Statement
CDR NGO: The Food and Drug Administration is convening today's meeting of the Peripheral and Central Nervous System Drugs Advisory Committee under the authority of the Federal Advisory Committee Act of 1972.
With the exception of the industry
representative, all members and temporary voting members of the committee are special government employees or regular federal employees from other agencies, and are subject to federal conflict of interest laws and regulations.
The following information on the status of the committee's compliance with federal ethics and conflict of interest laws, covered by but not
limited to those found at 18 USC Section 208 and Section 712 of the Federal Food, Drug and Cosmetic Act, is being provided to participants in today's meeting and to the public.
temporary voting members of this committee are in compliance with federal ethics and conflict of interest laws. Under 18 USC Section 208, Congress has authorized FDA to grant waivers to special government employees and regular federal employees who have potential financial conflicts when it is determined that the agency's need for a particular individual's services outweighs his or her
potential financial conflict of interest.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Under Section 712 of the FD&C Act, Congress has authorized FDA to grant waivers to special
government employees and regular federal employees with potential financial conflicts when necessary to afford the committee essential expertise.
Related to the discussions of today's meeting, members and temporary voting members of this committee have been screened for potential financial conflicts of interest of their own, as well as those imputed to them, including those of their spouses or minor children and, for purposes of 18 USC Section 208, their employers. These interests may include investments, consulting,
expert witness testimony, contracts, grants, CRADAs, teaching, speaking, writing, patents and royalties, and primary employment.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Today's agenda involves discussion of the use of historical-controlled trials for the
approval of anticonvulsant monotherapy for seizures of partial origin for antiepileptic drug products that are already approved for adjunctive therapy. The committee will also discuss how this may
specifically apply to the approval of the
supplemental new drug application for Lamictal XR, sponsored by SmithKline Beecham Corporation, doing business as GlaxoSmithKline, for monotherapy in patients 13 years of age and older with partial seizures who are receiving therapy with a single antiepileptic drug.
This is a particular matters meeting, during which specific matters related to SmithKline
Beecham's supplemental new drug application for Lamictal XR will be discussed. Based on the agenda for today's meeting and all financial interests reported by the committee members and temporary
voting members, conflict of interest waivers have been issued in accordance with 18 USC Section 208 to Drs. Jeffrey Cohen and David Treiman.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Dr. Cohen's waiver under 18 USC Section 208 is for a research grant to his employer funded by a competing firm. Dr. Treiman's waiver under 18 USC Section 208 is for a research grant to his employer funded by a competing firm. Neither Dr. Cohen nor Dr. Treiman has any personal involvement in the study. The funding for each study is between zero dollars to $50,000.
The waivers allow these individuals to participate fully in today's deliberations. FDA's reasons for issuing the waivers are described in the waiver documents, which are posted on FDA's website at www.fda.gov/advisorycommittees/
committeesmeetingmaterials/drugs/default.htm. Copies of the waivers may also be obtained by submitting a written request to the agency's Freedom of Information Office, Room 6-30 of the Parklawn building. A copy of this statement will be available for review at the registration table
during this meeting and will be included as part of the official transcript.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
To ensure transparency, we encourage all standing committee members and temporary voting members to disclose any public statements that they have made concerning the product at issue.
Dr. Jacqueline French, who is a guest speaker with us today, has acknowledged a professional and financial relationship with
GlaxoSmithKline, whose product is under discussion today. Dr. French also has a professional and financial relationship with firms that have products which closely compete with the product under discussion today. These firms include UCB, Lundbeck, Johnson & Johnson, Novartis, Pfizer, Eisai, Schwarz, SK Biopharmaceuticals, Sunovian, Supernus, and Upsher-Smith.
Dr. Nancy Temkin, who is a guest speaker with us today, has acknowledged a professional and financial relationship with GlaxoSmithKline, whose product is under discussion today. Dr. Temkin is working with GlaxoSmithKline on a manuscript for
LAM 30055, a study to determine the effectiveness of a lower monotherapy dose of lamotrigine than that currently approved.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Dr. Temkin also has professional financial relationships with firms that have products which closely compete with the product under discussion today. These firms include Novartis, Eisai, UCB, Johnson & Johnson, and Schwarz.
With respect to the FDA-invited industry representative, we would like to disclose that Dr. Roy Twyman is participating in this meeting as a nonvoting industry representative acting on
behalf of regulated industry. Dr. Twyman's role at this meeting is to represent industry in general and not any particular company. Dr. Twyman is employed by Johnson & Johnson.
We would like to remind members and
temporary voting members that if the discussions involve any other products or firms not already on the agenda for which an FDA participant has a
personal or imputed financial interest, the
involvement, and that exclusion will be noted for the record. FDA encourages all participants to advise the committee of any financial relationships they may have with the firm at issue. Thank you.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
DR. ANDERSON: Just for procedural reasons, can you give your announcement, Dr. Pearl, and your connection to the committee?
DR. PEARL: Thank you. Phillip Pearl, chief of neurology division at Children's National here in Washington. Thank you.
DR. ANDERSON: Thank you.
At this point we will move to the FDA's introductory remarks. Dr. Katz?
FDA Introductory Remarks
DR. KATZ: Thanks, Dr. Anderson. Let me just add my welcome to the committee today. And I'd like to offer a special welcome to the numerous invited guests that we have sitting on the
committee. We have a wealth of expertise on the committee today, both neurologic and statistical, and I'd like to thank everybody for coming.
Dr. Anderson, I think whose formal participation in the committee ended a while ago. I remember giving you a plaque. But he keeps coming back because we keep asking him back to serve as chair temporarily. So we appreciate very much your coming again.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So today, as the members of the committee know, we'll be asking for your advice in
consideration of Supplement 6 to NDA 22-115 for the use of Lamictal XR as monotherapy in the treatment of adults with partial seizures.
Lamictal XR is already approved for
adjunctive treatment for partial seizures, and the immediate-release Lamictal preparation is approved for both adjunctive and monotherapy use in partial seizures. Of course, the reason we're asking for the committee's advice today is because of the
unique and novel trial design that the sponsor used to study Lamictal XR for monotherapy, that is, of course, the use of an historical-controlled design.
Before I get into some of the issues we'd like you to consider with regard to the historical design questions, let me just give you a very brief
background about why the sponsor had to do a trial in the first place to establish Lamictal XR as effective as monotherapy. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
I mentioned Lamictal XR itself is already approved as adjunctive treatment, and the
immediate-release Lamictal is approved as
monotherapy. So one could argue, I suppose, that a trial to establish Lamictal XR as effective as
monotherapy is really not necessary.
But the reality is, in our division it's been the practice to require a controlled trial for a given dosage form to establish effectiveness of the treatment in a new clinical setting. So
Lamictal XR is approved as adjunctive therapy, but we would typically require a study of such a
product to show that it's effective in monotherapy. We also require controlled trials for
extended-release formulations where the immediate-release formulation is already approved for the same indication. So that again would argue for a study in monotherapy for Lamictal XR, and that's why we asked the study to do such a study.
I should point out in this particular case that the dose that the sponsor is proposing to
establish as effective in monotherapy is lower than the dose that's either approved for Lamictal XR as adjunctive therapy or lower than the dose that Lamictal immediate-release is approved for
monotherapy. So in any case, we think because of that issue, a study would need to be done.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Anyway, I provide that information just as background as to why a study was done in the first place. Those issues that I discussed really aren't the purpose that we've come to the committee today.
We are here, though, as you know, to have the committee discuss two very important topics, namely whether or not a historical-controlled trial, and a historical-controlled trial of the particular design that was employed here, can
provide substantial evidence of effectiveness as an anticonvulsant or an antiepilepsy drug, or an AED, as monotherapy, and whether the specific study that was done with Lamictal XR provides that evidence.
historical-controlled trial as the basis for granting a monotherapy claim arose out of the perception, I believe, amongst experts in the field, or the belief amongst experts in the field, that
performing a true placebo-controlled trial in the monotherapy setting was really unethical.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
There have been various other designs that have been proposed to be used in this setting. We believe that there are significant flaws in all of them. For example, again, placebo-controlled
trials are considered unethical for the most part. A popular alternative would be an active control noninferiority trial; we think those trials for the most part are uninterpretable.
But we are very interested in the
committee's view of this whole issue, that is to say whether or not it is the case that the
monotherapy studies with a placebo control are unethical or whether or not there are alternative designs that are better than the one that we're going to be talking about today. That issue more or less underlies the entire project that we're
talking about today, so it's very important for us to know what the committee thinks about that
fundamental assumption. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
But anyway, if we take as a given that other study designs are inappropriate in this setting, the unavailability of other alternative designs gave rise to a search for an adequate monotherapy trial design for a study that could be conducted and of course for a study that could be
interpreted.
So as a result of these circumstances,
Dr. Jacqueline French and her colleagues proposed a particular historical-controlled design, and of course we're going to hear a great deal about that today. In fact, we're going to hear about that from Dr. French and Dr. Nancy Temkin, who's a
statistician who's been intimately involved in the construction of the historical control. And so I want to thank them of course very much for being here.
They unfortunately cannot sit at the table and participate fully as committee members during
the course of the day's discussion, but they are available. They will present their work and their thinking on the issue and why they did what they did and what they did, and they'll be available for questions. So we're very lucky to have them here.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Let me just very briefly discuss the critical elements of the design that the sponsor employed and that we'll be talking about today. Dr. French and her colleague searched for all previously performed monotherapy studies of a particular design.
Over the course of the past 20 years or so, sponsors did do monotherapy studies that were
interpretable. In these studies, patients on one or two other AEDs were randomized to receive either what was believed to be an effective dose of the study drug or a subtherapeutic dose, either of that drug or of some other standard AED.
Then over time, patients were withdrawn from their background AEDs, and patients were then
treated with monotherapy, either with the active treatment or the so-called pseudo-placebo, the low
dose, of the other drug. And they were followed until they either completed the study -- and the study period varied across drugs -- or until they met one of four exit criteria. So these studies had in broad strokes a very similar design. They all had very, very similar exit criteria.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Anyway, there were ten studies that were done with this design. And over time, people began to feel that these studies were probably unethical as well because patients were randomized -- half the patients were randomized to this
pseudo-placebo, a low dose of an anticonvulsant, and the whole purpose of the study was to show a difference between the active dose and the pseudo-placebo. So people felt that that wasn't really ethically much superior to randomizing patients to placebo. So people stopped doing those studies.
But in any event, there were ten studies of that similar design that were done, and Drs. French and Temkin and colleagues included eight of them, for reasons you will hear, eight of the ten in the construction of their control group. And what they
did is to look at the escape rates in the eight pseudo-placebo groups from those studies and
calculate, in a sense, an overall escape rate, and then around that escape rate put a prediction
interval, which we'll hear more about from the statisticians. But it's analogous to the
confidence interval. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
The idea was that in a new study of a new anticonvulsant, where patients were treated without a concurrent control, the escape rate for the new drug is compared to the escape rate or the
prediction interval from this combined, calculated, constructed historical control. And the idea would be to show statistically significant superiority of the new drug in the new study to the historical control rate. That's an important point. This is a historical control, but it's a superiority
design.
In any event, we've told sponsors that such a study in form would be acceptable for providing evidence of effectiveness for their drug as
has already been shown to be effective in adjunctive therapy; that is, we had to have
assurance that the drug, that the moiety, was an effective anticonvulsant, albeit in a slightly different clinical setting.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Our requirement that a drug must already have been shown to be effective in another epilepsy setting reflects our view that the data generated in a historical-controlled trial is potentially very problematic. As I say, we've agreed that
these studies can be done, but we are well aware of the numerous problems attendant to the
interpretation of historical-controlled studies, and amongst other issues of course, we are very interested in the committee's views about these generic problems related to historical-controlled trials.
In brief, I think the great weakness of historical-controlled trials in large part derives from, of course, the lack of randomization to
treatment, because randomization ensures that
equitably or equally distributed across the groups. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Sometimes we think we know what some of those factors are, and we can try to match the patients in a new study on those factors to the factors in the patients who made up the historical control. Of course we can't ever do that
perfectly, but we can at least try to match on factors that we know about or believe are related or could be related to treatment.
But of course there are many factors that randomization equally distributes to the treatment groups that may be related to treatment
responsiveness that we are completely unaware of. And of course we can try again to match on factors we know about, or think might be related to
treatment response, but of course it's impossible to match the treatment groups on factors that we are unaware of but that might be important.
In addition, another important point about construction of a historical control would be that patients in the new study will be aware that they are on treatment; patients, investigators,
everybody will be on treatment because there is no concurrent adequate control in such a study. And of course that can have very important effects on the reliability of the data collected, the data reported. So that's an issue that we'll need to consider as well. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So in addition to sort of the generic issues related to historical controls that we want the
committee's input on, we are very interested in the committee's view about the specific
historical-controlled design that was employed here.
As I said, there were eight studies that were combined to produce the historical control. These studies were of very similar design, but they were not identical, and we need to know whether or not the committee thinks that they were
sufficiently similar to be able to reliably construct a historical control group.
There are many differences. For example, the period of time during which the patients were withdrawn from their background AED to get to the monotherapy stage was different between studies.
The studies, as I said before, all used very
similar exit criteria, but the exit criteria were not identical. And even in those cases where the description of a particular exit criterion might have been the same across the eight or ten studies, they had been analyzed differently across certain studies. So that's something that we need to think about. Patients didn't always have exactly the same seizure types in these studies, so there were many differences. Of course those are on a few that we've been able to identify.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
I should point out that Dr. Ling of our biometrics group, who you will hear from later this morning, has performed numerous analyses similar to those performed by the sponsor, and largely
obtained very similar results. But we'll hear more about that later.
Even if we were convinced that the trial design, the historical control that was
constructed, was adequate, it would be also important to know in this sort of meta-analysis approach that every trial that was relevant to the
question was a identified and included in the
analysis or excluded for good reasons. As I said, two of the ten that were identified were excluded.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
I should point out it's our view that Dr. French and her colleagues made all reasonable efforts to identify which studies to include, and she'll probably tell you about that. We are not aware from our records that there are other trials of similar design that would have been relevant to include.
In any event, as you know, the sponsor did perform a trial utilizing this design. We'll be talking today about the results of that trial. Ostensibly, on face, the trial demonstrated the effectiveness of both drugs; patients in that study were, as you know, randomized to one of two doses which were very close, 300 and 250. No true
contemporaneous control group.
But results on face notwithstanding, in addition to the generic issues about historical control and in addition to the specific issues related to the specific historical-controlled
design that was used, there are issues related to the specific study that was performed by the
sponsor, and we'd like for the committee to discuss those as well. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Many of the specific issues that we think are critical to discuss again relate to the
comparability, or the lack of comparability, between the sample that was included in the
sponsor's study and the samples that were used to construct the historical control.
For example, an important one seems to be that only about 25 percent of the patients in the Lamictal XR study were from the U.S., whereas essentially all patients that constituted the
historical control were from the U.S. So of course this raises significant questions about the
comparability of the two groups. They of course may be very different in ways that we couldn't even begin to characterize.
In the Lamictal XR study, patients were only permitted to have been on one background AED before they were enrolled in the study, whereas in many of
the historical-controlled studies or the studies that constitute the historical control, patients were allowed to have been on two AEDs. Does that mean that those patients were different from these patients? It certainly raised the question.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
About 80 percent of the patients that constitute a historical control were on
carbamazepine before being switched to monotherapy, which is an enzyme-inducing AED. In the Lamictal XR study, enzyme-inducing AEDs as background AEDs were excluded.
So those are just some of the differences. There were different baseline seizure rates in some of the studies that constitute the historical
control compared to the Lamictal XR studies. And these are only some of the differences that we have been able to identify because, as I said before, of course there may be many, many other issues that are relevant that we are not able to identify.
I should also say again, and you'll hear about this later, Dr. Ling from our biometrics
try to get at this question of comparability
between the historical control and the Lamictal XR sample. You'll hear about those. Most of those have been fairly reassuring. They've given us the same results as the overall result. But of course, as reassuring as those analyses are, they can't definitively address the potential bias that might have crept in from differences between the groups that we have not identified.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
There are some other issues of interest also with regard specifically to the Lamictal XR study and its conduct that raise some questions. The number of exits, or the number of patients who met exit criteria, in the study as it was being
performed was actually quite low. Only after folks went back and looked at records did a more
realistic or a true picture of how many exits really had occurred emerge. So that of course raises questions about the conduct of the trial.
Another interesting point I think is that if we look at the exit rate in this study, the
the active control arms in the eight historical studies, the exit rate in this study was amongst the lowest. It's not the lowest, but it was amongst the lowest escape rate, or exit rate, of all the eight studies that were included in the construction of the historical control. So that of course again raises the question that maybe the fact that people were aware that they were on
active treatment might have affected the recording of the data. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So all of these issues -- and certainly, I'm sure, others -- raise questions that we think need to be discussed about the appropriateness of
constructing the historical control as Dr. French and her colleagues did it; the propriety of relying on such a study to provide definitive evidence of effectiveness as monotherapy in this setting; and of course the acceptability of the results of the specific study that we'll be talking about today. And so, of course, these are the issues that we would like the committee to discuss.
relevant issues that we haven't identified that you think need to be discussed, of course we are very eager to know what those are and of course eager to hear what you think about them.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So with that, again let me thank the
committee. Thanks for coming. Thanks for the work that you've done in preparation for the discussion today and for the work that you will do during the day. So again, thanks very much. I'll turn it back to Dr. Anderson.
DR. ANDERSON: Thank you.
At this point in the schedule we're ready to begin hearing from the guest speakers. And the
first one on the list is Dr. French.
Guest Speaker Presentation – Jacqueline French DR. FRENCH: Good morning. Thank you for allowing me to present in front of the committee. This has been a long time coming, so it's very nice to be able to discuss this in front of the FDA.
So I first want to very much thank my
collaborators on this work -- Nancy Temkin, who is here and will present after me; several other
statisticians -- because I am not a statistician -- Steve Wang, who works for Johnson & Johnson, and Bob Warnock, who used to work for GlaxoSmithKline, and amazingly now works for Genzyme, nothing to do with epilepsy whatsoever, and yet gave willingly and enormously of his time and efforts in order to finish off this work.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So this is just to remind you that there have been numerous antiepileptic drugs that have been approved over the last 20 years for use in epilepsy. All of the green drugs that you see on this slide are drugs that were approved for use after 1980, and in fact we have 17 drugs that we can use for the treatment of epilepsy, which is a wonderful advance from the time that I started to treat epilepsy when we have five major drugs for use in the treatment of epilepsy.
But unfortunately, approval of these drugs as monotherapy has lagged behind the initial
approval that all these drugs get, which is adjunctive therapy for treatment of partial -- usually partial-onset seizures. And just to
illustrate this, of the 12 new drugs that were approved for use over the last two decades, only four are approved for any use in monotherapy at all. These are oxcarbazepine, felbamate,
lamotrigine, and topiramate. And only two are approved for initial monotherapy, and those are oxcarbazepine and topiramate.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
The older antiepileptic drugs, phenytoin, carbamazepine, phenobarbital, and valproate, are approved for monotherapy, but they were not
approved based on any standardized or randomized or controlled design demonstrating effectiveness in monotherapy. They were essentially grandfathered into that indication. And the other illustrative point of how difficult it's been to get a
monotherapy indication is that there have been no new monotherapy approvals of any antiepileptic drug since 2005, despite the fact that four new
antiepileptic drugs have been released since 2005. You all -- many of you, at least -- are familiar with our standard design by which we get approval of these drugs as adjunctive therapy.
They are interpretable. They have an internal
placebo control. Patients with treatment-resistant partial epilepsy receiving usually one to two or maybe three background drugs, who are having four or five seizures a month, go through a baseline evaluation and observation for seizure frequency. They're then titrated onto one or two doses of the drug under evaluation, and this is compared to an internal placebo control. Very interpretable.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
We've been very successful, as you can see, with getting indications for the drugs based on that design. So why aren't we just happy with this add-on treatment? Why do we want to treat patients with monotherapy?
Well, I think that most of you will
understand that many patients are receiving drugs that have potential disadvantages for those
individual patients. And particularly, we have developed 12 new drugs because we feel that for some patients, those 12 new drugs will represent some advantage. And there are many advantages, on an individual basis, that they represent.
Obviously, you can't get those advantages in most cases and leave people on the drugs that
they're already taking, particularly when it comes to improvement of side effects, long- and short-term; elimination of drug interactions, and many of the drugs that were approved before the new crop were hepatic enzyme inducers, and that's already been referred to by Dr. Katz.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Hepatic enzyme induction does many things long-term to the human being, including increasing cholesterol levels, increasing cardiovascular risk factors, interfering with oral contraceptive pills, interfering with many other drugs that people take on a day-to-day basis. We can't eliminate hepatic enzyme induction without eliminating the hepatic enzyme inducer.
Reducing people's medication loan improves compliance, obviously improves cost. And one
major, major issue, particularly in women with epilepsy, is that some of the older drugs -- and I would point to valproate as the biggest offender -- has been demonstrated to be a serious teratogen.
Obviously, if there are drugs that are less teratogenic, then we would very much want to
eliminate the teratogen, and why should we not want to? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So scenarios where monotherapy is clearly an imperative is when a patient is having a side
effect from their current therapy; a new drug is added and is effective. Obviously, the drug that's producing the side effect should be withdrawn. That's good clinical care. I mean, I can't tell you how many patients come in and they feel bad on the drug that they're currently receiving. They feel like -- just to give an example, I've had patients who had to sleep 10 to 11 hours a night because they were on one of the older antiepileptic drugs. They want to get on a drug that's less
sedating. There's no benefit to adding a new drug if you can't take the sedating drug away; and
again, when a woman is contemplating pregnancy. So I think that you will all agree that we need to use drugs in monotherapy. But the second question one might ask is, do we need to use them
on label? Do we really need a monotherapy
approval? We can use the drugs. You know, it's not illegal to use a drug off-label, but there are problems with using drugs off-label.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
First of all, when you talk about patients with newly-diagnosed epilepsy, this limits our choice of drugs for patients with newly-diagnosed epilepsy. Antiepileptic drugs, unlike drugs in other categories, are very individual, and we pick them on a very individualistic basis based on
patient characteristics. Some make you gain
weight. Some make you lose weight. Some give you kidney stones. Some interfere with oral
contraceptives, as I've already said. So you look at the characteristics of the patient and you look at the best fit of the drug. And the insurance companies may not let us try what we think is best drug fit if that drug doesn't have an approval for initial monotherapy.
I have been contacted personally by I can't tell you how many attorneys asking me if I would -- I don't do medical legal work, but they wanted to
know if I would be involved in a case where a terrible doctor withdrew a patient to
monotherapy -- for example, to lamotrigine
monotherapy, very, very common -- to do that. But it's very easy to convince a jury that doing it is -- since there is no FDA indication for use of Lamictal as monotherapy, this doctor must be a rogue doctor who's practicing, you know, something very in left field. So there are liability issues.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
There are many doctors who are comfortable in a field who feel comfortable using a drug off-label and don't have any problem with it. But there are also many doctors who don't feel comfortable in a field and will follow FDA
direction word by word. And this is particularly true for general neurologists and primary care physicians. So there is a difference between having an indication and using a drug off-label.
So why do we have difficulty in obtaining monotherapy approval for antiepileptic drugs? This arises from the fact that the FDA, in accordance with the International Conference on Harmonization,
prefers -- and we've already heard this from Dr. Katz -- trials that contain an internal
interpretable control group, that is one where the test drug shows superiority to some internal
control. And there's a difficulty with
demonstrating superiority in epilepsy patients.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So you have two populations, basically, to choose from. You can choose newly-diagnosed
patients, or you can choose treatment-resistant patients. Treatment-resistant patients are already on a drug because they're treatment-resistant. So in order to do a monotherapy, you have to withdraw them to monotherapy.
Now, you can do a comparative study and try to compare your drug to some other drug and see whether your drug is superior. But the problem is -- there are many problems. One is that our new drugs don't necessarily show superiority to other drugs. They commonly show that they are equally as effective and have other advantages.
Even if there was a possibility of a drug showing superiority, and it would have to show
superiority to a number of different other drugs, most of the patients who are treatment-resistant have tried most of the other drugs before, so you'd already know that they had failed probably the
comparator drug. And so it would be very, very difficult to do that kind of a study.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
With newly-diagnosed patients, we know that in fact two-thirds of patients respond to
essentially any drug that you put them on and usually at a relatively low dose. So when two
drugs have been compared, as Dr. Katz mentioned, in an active control comparison, it almost always
comes out -- well, no, not almost always -- it does always come out to a no-difference outcome in terms of efficacy.
So, as I said, our new drugs represent improvements in safety, tolerability,
pharmacokinetics, much more than they do in efficacy. And definitely that is true in the population of newly-diagnosed patients who, as I said, are sensitive.
active control study -- because these studies are done in Europe for monotherapy approval -- was between carbamazepine and levetiracetam, a drug that's actually commonly used in monotherapy here and abroad. And the results were that I think 73.6 percent of the patients on carbamazepine were
seizure-free versus 73.4 percent of the patients on levetiracetam. So you just can't show a
difference. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So the response to that, if you want to show superiority, then you have to drive the dose of
your comparator down to almost imperceptible levels so that you can show this artificial difference between drugs. And this is where this concept of pseudo-placebo, which is just a wonderful word, came from.
The question is whether placebo-controlled or very-low-dose-controlled trials are an ethical or practical alternative for individuals with established diagnosed epilepsy, which is a
potentially life-threatening condition, when those patients have 17 other drugs to choose from, or
16 other drugs. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
I would like to quote Bob Temple when he was talking about add-on trials in his 2000 article in the Annals of Internal Medicine, where he said, "This add-on design is common in trials of therapy for cancer, heart failure, and epilepsy, in which omitting standard therapy would generally be
unacceptable."
So there is a big question of whether placebo-controlled or very-low-dose-controlled
studies are an ethical or practical alternative for individuals who have established, diagnosed
epilepsy, which is a potentially life-threatening condition. Again, this was addressed initially by a strategy which was basically the only alternative under the circumstances, using a pseudo-placebo.
So what is a pseudo-placebo trial? In a pseudo-placebo trial, you assign patients to receive study drug and compare that to what you hope to be -- this is what is rather interesting -- what you hope to be a suboptimal dose of either the same drug or an effective approved drug at a
suboptimal dose. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So often valproic acid, 1000 milligrams per day or the initial starting 10 milligrams per
kilogram, I think it was, of valproic acid were used as the comparator. And the idea initially was a good one, which was, we're going to stop the
serious, life-threatening seizures but allow some seizures to come through, and that will allow us to establish the superiority of giving an effective dose.
In order to keep people safe, we will use escape criteria so that when we see that they are worsening, we're not going to let them get worse for a long period of time. We'll immediately rescue them and take them out of the trial.
So these were endpoints which were described as therapeutic failure or escape criteria. And the problematic thing for the epilepsy community is that these trials were intended to demonstrate clinical worsening of a subject, not improvement, but worsening, and your therapeutic drug was
demonstrate that the people who were randomized to the subtherapeutic dose had less seizure control.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So this is what these trials looked like. Patients came in on either monotherapy, or
monotherapy and usually a small or subtherapeutic dose of a second drug. Again, they went through the same type of baseline that they go through with the add-on drug to establish seizure frequency.
They are then randomized to either the active treatment or what is called the
pseudo-placebo, and then their active drugs are removed so that by the end of the withdrawal period, they
remain only on the treatment or the pseudo-placebo. And then for three months, they are maintained in that condition, and they then escape if they meet an escape criterion.
At this point, I just wanted to comment on the issue of blinding and randomization and what patients knew at the time that they were
randomized. Many of the drugs that were used in these trials were actually approved for use at the time that these monotherapy trials were done
because they were a secondary indication for the drug. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
I find it hard to imagine, in my own mind, that a patient would willingly agree to be
randomized to a suboptimal dose of a drug for the purposes of a trial when they could get the drug at the optimal dose just by asking for a prescription or by their physician writing them a prescription.
So I actually went back and looked at the consent forms for a number of these trials to see what in fact the patients were told. And
basically, in the consent form, there is nothing to say that this is not basically an active control equivalence trial. In the consent form, it says, you will be randomized to one of two drugs. One will be a low dose, one will be a high dose, and we don't know which one is safer and more effective, essentially.
So in terms of whether the patients were prepared for -- I mean, any time you switch from drug A to drug B, we know that there's risk. But whether they were actually prepared for we're
waiting for you to fail is another story altogether. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So what were the escape criteria? Doubling of average monthly seizure rate; doubling of
highest consecutive two-day seizure rate; emergence of a new, more severe seizure type; and clinically significant prolongation of generalized
tonic-clonic convulsions.
Now, when you look at these four exit
criteria, I think that it's reasonable to say that if somebody has a temporary doubling of seizure frequency, it's not a good thing for them, but it's not an urgently life-threateningly bad thing. You know, you then put back the meds that they were on before, and they'll probably go merrily on their way; the same thing for doubling of the highest consecutive two-day seizure rate.
On the other hand, the two on the bottom, emergence of a new, more severe seizure type or clinically significant prolongation of generalized tonic-clonic convulsions could in fact be life-threatening. And the more we learn about, for
example, SUDEP being associated with the occurrence of generalized tonic-clonic convulsions, the more we don't want generalized tonic-clonic convulsions to happen if at all possible.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
In fact, probably the reason that I became so passionate about eliminating the pseudo-placebo trials is a patient of mine, who was in a trial very early, gabapentin trial -- and the drug was only available through being enrolled in this
trial. And that was at a time when there were very few drugs available for epilepsy, so I felt that the benefit was worth the risk. But in the
trial -- and I don't know which arm she was randomized to.
It turned out in that particular trial it didn't really matter because all of the arms failed to demonstrate significant efficacy; whereas before she had had only complex partial seizures, which were sort of graceful posturing for a few seconds, she began to have seizures where she would fall to the ground and strike her head. And she ended up in the emergency room, which was very upsetting to
me. But what was even more upsetting is that this young lady, after she got reestablished on her background medication, never stopped falling from that point on, and ended up with multiple
surgeries, which she probably wouldn't have had had she not been in the trial.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So this is demonstrating a result of one of those studies, oxcarbazepine 300 versus 2400. And you can see -- so you can see that this is the withdrawal to monotherapy pseudo-placebo arm, and this is the active arm. And in this particular trial, 100 percent of the patients met one of those four exit criteria when they were randomized to the pseudo-placebo, whereas only 60 percent of the
patients met an exit criterion when they were on the active arm of 2400 milligrams.
Now, you can say, well, look. People worsen even when they're converted to an active arm, and that certainly is true. Conversion from one drug to another doesn't come without risks. There's always some chance of worsening. And in addition, we know that epilepsy generally is a variable
condition. And just by chance, even if you had had an arm where people had not done anything but just continued with their typical medications, there would have been some percent of patients that would have hit one of those exit criteria, just because there's variability in seizure frequency over time.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So just to give you another example of why there is some question about why these
pseudo-placebo trials should no longer be performed, just looking at the serious adverse events that were observed in previously performed pseudo-placebo withdrawal to monotherapy trials will give you some sense. I'm going to -- I'm sorry to the sponsor -- pick on the lamotrigine trial. It's just one of the trials that was done withdrawal to monotherapy.
There were five serious adverse events that were recorded in the active treatment arm of that trial -- and this is all in the manuscript -- and five in the pseudo-placebo group. But my concern is that they were a very different type of serious adverse events. The patients in the lamotrigine group experienced chest pain, pneumonia, rash,
Stevens-Johnson syndrome, and suicidal ideation. And certainly at least the last three could have some relationship to being started on an active drug, and all active drugs have the potential for having adverse events, and that's why you do a trial to figure that out.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
The patients in the pseudo-placebo group, on the other hand, these were their serious adverse events: exacerbation of convulsions, increased seizures, paranoid ideation, status epilepticus, and one sudden unexplained death. I have a problem when the adverse events are related to the trial design rather than to the therapeutic intervention.
So because of all of these things, we held a meeting in March 2001. The epilepsy community was very concerned about the lack of monotherapy
indications, so they brought together the physician community, the NINDS, the FDA, to discuss possible trial designs. And after looking at all of the possible trial designs and all of the data, we came to a consensus conclusion, and this is what that consensus conclusion was.
"We understand that the ideal methodology to prove the effect of an antiepileptic drug in
monotherapy would be a comparison trial designed to show a difference. This could be in a
newly-diagnosed or refractory population. But there are convincing reasons why such a design is problematic in the epilepsy population, and therefore, we will investigate historical control in the outpatient refractory monotherapy situation." So that is where the concept came from.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So many of you know what historical control is. But just to reiterate, instead of using
placebo as the control group, historical control is using the expected behavior of placebo. And you only know the expected behavior of placebo if you have prior trials where there was a placebo, or in this case a pseudo-placebo.
It is not the same as a natural history, and nobody would want to use natural history as a
comparator in a clinical trial because historical control must be linked, as Dr. Katz already said, to a very specific trial design and a very specific
population in whom you think you understand the expected behavior of placebo.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So the question we asked at the time in 2001 was, can we look at existing data from trials to generate a historical control and eliminate the need for placebo or pseudo-placebo?
So here you have the standard situation. If you have an active arm and let's say a comparator arm, let's say in an active control study, and this is the standard situation where you might use
historical control, you want to know that there actually is assay sensitivity, that these patients, if they had been randomized to placebo, would do worse than they would do on the active arm, and you want to know where that placebo dot is. So the historical control is a nice, comforting dot that says, this is where they would be. This is where placebo would be, and now we know, if the active arm behaves in such a fashion, then it is superior to placebo.
So as you heard, ten studies had been performed in which patients were randomized to a
pseudo-placebo or active study drug and then withdrawn towards monotherapy. And as I just explained, these studies are no longer considered to be ethical, and we felt that if anything was appropriate for historical control, this data set was appropriate for historical control because
similar methodology was used. We were fortunate in that basically the same group of people designed every single study and they used very similar methodology. So we performed a meta-analysis of all these pseudo-placebo arms to determine whether they could serve as a historical control for future monotherapy studies.
So in order to create a valid historical control, it was essential to obtain all similarly designed trials -- you already heard this -- that have been performed. So what did we do? We
reviewed the literature, searching for controlled clinical trials, epilepsy, and monotherapy, in which antiepileptic drugs were withdrawn, leading to monotherapy. We also queried our colleagues to determine if there were other trials that had been
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
performed with this methodology. We did not find any others. We directly inquired of all companies with antiepileptic drugs in development to
determine if such a trial had been performed and we had heard about it.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So in the literature, there were seven
randomized controlled trials in which patients with refractory partial seizures were randomized to
active compound or pseudo-placebo, before in six trials or after in one trial being withdrawn towards monotherapy. We also found three
additional unpublished trials that were discovered by direct inquiry of the companies.
Once all the trials were identified, each company was asked to provide primary individual patient data from pseudo-placebo arm only of the study in question, and that data was submitted directly to the statisticians, Bob Warnock and
Nancy Temkin, and Steve Wang at one point in time. Further information was obtained from study protocols where they were available, information from the FDA summary basis of approval, and study
publications. So other than the individual patient data, we tried to get as much information about the trial and the trial conduct as possible.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
All the trials were randomized,
double-blind, parallel-group design with a baseline phase, as I showed you, followed by a double-blind phase, divided into the conversion, and then a strict monotherapy phase. In all but one of the trials, patients were randomized prior to withdrawal of the baseline drug. One trial, the patients were
randomized -- they were put on the active drug and then withdrawn from their antiepileptic drug,
randomized to removing the active drug. The
conversion phase ranged from 4 to 10 weeks, and the monotherapy phase ranged from 11 to 16 weeks across the trial.
So the famous exit criteria, and how did they differ, and how were they similar. So the first exit criteria was a twofold increase in partial seizure frequency in an 28-day period compared to baseline. All studies used this criteria, although study 1 used the highest 4
consecutive weeks in baseline and the others used the average frequency in baseline. So there was a slight change there.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Criteria number 2 was a twofold increase in the highest consecutive 2-day seizure frequency that occurred during the baseline phase. All the studies used this criterion, although study 2
required three seizures if the highest daily count in the baseline was one seizure, and study 8 did not use this criterion if the highest daily count was one seizure.
The third was occurrence of a single generalized tonic-clonic convulsion if none had occurred in the previous 6 months in study 6, in the previous 2 years in study 1, and during the baseline in studies 3, 5, 7, 8, and emergence of a more severe seizure type, which would include
generalized tonic-clonic convulsions, in study 2. Exit criterion 4, which was sort of the grab-bag exit criterion, was a prolongation or worsening of seizure duration or frequency considered by the investigator to require
intervention. And this was used for all trials, although studies 2, 5, and 7 required that
worsening seizure to be generalized, and episode of serial seizure status epilepticus for studies 3, 7, and 8, and an episode of status epilepticus for study 1. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So all trials allowed either one or up to two baseline drugs, but always the second drug had to be at a subtherapeutic level. Only in one
trial, trial 5, the baseline drug had to be
carbamazepine. For the five trials that allowed two drugs, four required that one of the baseline drugs be taken at less than 50 percent of the minimum recommended dose, or that the serum concentration be less than 50 percent of the minimum effective serum dose.
The outcome measure for all studies was percent of patients exiting the trial. And the patients exited, as I mentioned, if they
experienced seizure worsening, as identified by predetermined exit criteria.
used in nine out of the ten trials. The tenth trial used only three exit criteria, and only one matched a criterion used in the other studies. And therefore, we had to exclude that trial because it did not match the others in terms of its design.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
These were the studies. The test drugs were gabapentin, lamotrigine, topiramate, felbamate,
oxcarbazepine. And from the published studies, we were able to get some patient characteristics that had been gathered by the statisticians. And you can see that the patients were pretty similar
across all the studies. So the mean age was in the 30s. The epilepsy duration, when it was commented on, was long, which is typical for these trials. The baseline seizure frequency ranged from
minimum -- that's a median -- five a month up to about ten a month. But in all cases it was high, which is consistent with these studies.
The percent taking carbamazepine at baseline varied, but you can see that in the days of these studies, about half of the patients, up to about 60 percent of them, were taking carbamazepine at
baseline. And that I think will come in as an important factor later. And the majority of
patients had complex partial seizures. Since these are all patients with partial epilepsy, the
remainder would either have had simple partial seizures or generalized tonic-clonic convulsions.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
So what's the advantage of this data set? Many, many trials with similar designs and exit criteria. All the trials randomized patients with similar demographic characteristics. And a lot of people worry with historical control that you
can't -- I can't remember the phrase, but you can't put your toe in the river twice and have it be the same water because the water flows. So over time, we know that background treatments change and
diseases change, and therefore perhaps the historical control would no longer be valid.
I have two things to say about that. Number one, it doesn't matter in this case whether
background treatments change because what we're doing -- we presume that, if anything, background treatments are getting better. And what we're
doing here is we're stripping the background
treatments away and see how much worse people get. It's unlikely that background treatments are
getting less effective, and therefore stripping away their active medication, they would get less worse than they did in the past.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
But the second thing, which I think is very reassuring, is that these trials were performed over the course of a decade. And during that decade, it was a very active decade, and many new drugs were introduced over the course of that decade. And the most recent trials actually had the highest exit rates, which is kind of what you would expect if treatments are getting better.
The only thing that differed, and we'll get to that in a minute, as Dr. Katz said, is the rate at which they withdrew, but not the fact that
ultimately they did withdraw. So here is, in fact, the famous Kaplan-Meier curves of withdrawal in all of these studies that we used. And you can see that in the end, the number of patients exiting, getting worse when you take their medicine away,