FOOD AND DRUG ADMINISTRATION CENTER FOR DRUG EVALUATION AND RESEARCH (CDER) JOINT MEETING OF THE PULMONARY-ALLERGY DRUGS ADVISORY COMMITTEE DRUG SAFETY AND RISK MANAGEMENT ADVISORY COMMITTEE AND THE PEDIATRIC ADVISORY COMMITTEE DAY ONE Rockville, Maryland Wednesday, December 10, 2008 (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 2 1 PARTICIPANTS: 2 PULMONARY-ALLERGY DRUGS ADVISORY COMMITTEE 3 4 Voting Members: DAREN KNOELL, PharmD The Ohio State University 5 Temporary Voting Members: 6 7 8 JOHN HOIDAL, M.D. Salt Lake City, Utah FERNANDO MARTINEZ, M.D. University of Arizona 9 10 11 MARK BRANTLY, M.D. University of Florida ANDREA HOLKA (Patient Representative) Attack on Asthma Nebraska 12 13 14 LEE NEWMAN, M.D. University of Colorado-Denver JESSE JOAD, M.D. U.C. Davis Medical Center 15 16 17 DAVID SCHOENFELD, Ph.D. Massachusetts General Hospital ERIK SWENSON, M.D. (Acting Chair) University of Washington 18 19 20 21 22 (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 3 1 PARTICIPANTS (CONT'D): 2 DRUG SAFETY AND RISK MANAGEMENT ADVISORY COMMITTEE 3 4 Voting Members: SEAN HENNESSY, Pharm.D., Ph.D. University of Pennsylvania 5 6 7 JUDITH KRAMER, M.D., M.S. Duke University SYDNEY WOLFE, M.D. (Consumer Representative) Health Research Group of Public Citizen 8 Non-Voting Member: 9 10 11 12 D. BRUCE BURLINGTON, M.D. (Industry Rep) Gaithersburg, Maryland Temporary Voting Members: SEBASTIAN SCHNEEWEISS, M.D. Harvard University 13 14 15 DEBORAH SHATIN, Ph.D. Shatin Associates, LLC JULIE ZITO, Ph.D. University of Maryland 16 17 18 DAVID MARGOLIS M.D., Ph.D. Hospital of the University of Pennsylvania EDWARD KRENZELOCK, Pharm.D. University of Pittsburgh Medical Center 19 20 JACQUELINE GARDNER, Ph.D. University of Washington 21 22 (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 4 1 PARTICIPANTS (CONT'D): 2 PEDIATRIC ADVISORY COMMITTEE 3 4 Voting Members: MARSHA RAPPLEY, M.D. (Co-Chair) Michigan State University 5 6 7 AMY CELENTO (Patient-Family Representative) Nutley, New Jersey AVITAL CNAAN, Ph.D., M.S. Children's National Medical Center 8 9 10 CARL D'ANGIO, M.D. University of Rochester MELISSA HUDSON, M.D. St. Jude Children's Research Hospital 11 12 13 KEITH KOCIS, M.D., M.S. University of North Carolina DANIEL NOTTERMAN, M.D. Princeton University 14 15 GEOFFREY ROSENTHAL, M.D., Ph.D. Children's Hospital Center Cleveland, Ohio 16 17 18 19 ELAINE VINING (Consumer Representative) Silver Spring, Maryland Non-voting Member: BRAHM GOLDSTEIN, M.D. (Industry Representative) Princeton, New Jersey 20 GUEST SPEAKER (Non-Voting) 21 22 ROBERT LEMANSKE JR., M.D. University of Wisconsin-Madison (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 5 1 2 3 4 5 6 7 8 PARTICIPANTS (CONT'D): FDA PARTICIPANTS (Non-Voting): JOHN JENKINS, M.D. Director, Office of New Drugs CDER/FDA BADRUL CHOWDHURY, M.D., Ph.D. Director, Division of Pulmonary Drug Products CDER/FDA SALLY SEYMOUR, M.D. Deputy Director for Safety Division of Pulmonary Drug Products CDER/FDA 9 10 ANN McMAHON, M.D. Acting Director Division of Pharmacovigilance II CDER/FDA 11 12 13 14 15 16 17 HENRY FRANCIS, M.D. Deputy Director Office of Surveillance and Epidemiology CDER/FDA DIANNE MURPHY, M.D. Director, Office of Pediatric Therapeutics OC/FDA ROBERT NELSON, M.D., Ph.D. Bioethicist, Office of Pediatric Therapeutics OC/FDA 18 19 * * * * * 20 21 22 (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 6 1 P R O C E E D I N G S 2 (8:30 a.m.) 3 4 DR. SWENSON: Good morning, everyone. We'll need to bring this meeting to order. 5 I'm Erik Swenson, acting chairman 6 of this rather august and large committee. 7 And co-chairing with me is Marsha Rappley 8 here. 9 We'll need to begin, and I need to 10 first read a statement here. 11 that we'll be discussing today, there are 12 often a variety of opinions, some of which 13 are quite strongly held. 14 today's meeting will be a fair and open forum 15 for discussion of these issues, and that 16 individuals can express views without 17 interruption. 18 For the topics Our goal is that So as a gentle reminder, 19 individuals are please asked to speak into 20 the record only when they're recognized by 21 the Chair, and we'll look forward to a 22 productive meeting. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 7 1 In the spirit of the Federal 2 Advisory Committee Act and the Government in 3 the Sunshine Act, we ask that the Advisory 4 Committee members take care that their 5 conversations about the topic at hand take 6 place in the open forum of the meeting. 7 We are aware that members of the 8 media are anxious to speak with FDA about 9 these proceedings. However, FDA will refrain 10 from discussing the details of this meeting 11 with the media until its conclusion. 12 should reiterate that's quite important, 13 because we're all coming to as best knowledge 14 as possible in these two days. 15 And I I would also like to identify the 16 FDA press contacts, Ms. Sandy Walsh and Karen 17 Riley. 18 stand up? And if you're here, could you please Good. 19 I would like to remind everyone 20 present to please silence cell phones and 21 other electronic devices if you haven't 22 already done so. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 8 1 And again, committees are reminded 2 to please refrain from discussing the meeting 3 topic during breaks or lunch. 4 5 Thanks very much. And I'll turn it briefly over to Kristine Khuc. 6 Before that, let's have an 7 introduction of all the Committee members, 8 and if we could start at my far right-hand 9 side. 10 DR. GOLDSTEIN: I'm Brahm Goldstein. 11 I'm a pediatric critical care physician and one 12 of the industry representatives. 13 MR. BURLINGTON: 14 consultant, and I'm a industry rep. 15 DR. KRAMER: Bruce Burlington, Judith Kramer. Excuse my 16 laryngitis. 17 at Duke University, and a general internist, 18 with a background in randomized controlled 19 trials and observational studies. 20 I'm associate professor of medicine DR. D'ANGIO: Carl D'Angio. I'm an 21 associate professor of pediatrics at the 22 University of Rochester, and I'm a neonatologist (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 9 1 and vaccine researcher. 2 DR. MARGOLIS: I'm David Margolis. 3 I'm a professor of dermatology and a professor 4 of epidemiology at the University of 5 Pennsylvania. 6 DR. HENNESSY: Good morning. My 7 name's Sean Hennessy. 8 research at the University of Pennsylvania. 9 SPEAKER: I do pharmacoepidemiology If I could just interrupt 10 for a moment. We are having a little difficulty 11 understanding your words. 12 you quite clearly. 13 coming back. 14 maybe a little echo or something. 15 people just speak more slowly and clearly into 16 the mic, that would help. Can you -- we heard But prior to that, it's not We can't understand it too well; 17 Thank you. 18 DR. BRANTLY: So could My name is Mark Brantly. 19 I'm professor of medicine at the University of 20 Florida, and a pulmonary and critical care 21 physician. 22 (202) 464-2400 DR. NOTTERMAN: My name is Daniel Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 10 1 Notterman. 2 molecular biologist. 3 Molecular Biology at Princeton University. 4 5 I'm a pediatric intensivist and a MS. CELENTO: I'm in the Department of Amy Celento, patient representative, Pediatric Advisory Committee. 6 DR. HUDSON: Melissa Hudson, pediatric 7 oncologist from St. Jude Children's Research 8 Hospital in Memphis. 9 DR. KRENZELOCK: Good morning. I'm Ed 10 Krenzelock. 11 Poison Center and Drug Information Center at the 12 University of Pittsburgh Medical Center, and a 13 professor of pharmacy and pediatrics at the 14 University of Pittsburgh. 15 I'm director of the Pittsburgh DR. KNOELL: Good morning. My name is 16 Daren Knoell. 17 Ohio State University, and appointed in pharmacy 18 and internal medicine. 19 I'm an associate professor at The MS. VINING: Good morning. 20 Vining. 21 Pediatric Advisory Committee. 22 (202) 464-2400 I'm Elaine I'm the consumer representative for the DR. SHATIN: Good morning. Beta Court Reporting www.betareporting.com Deborah (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 11 1 Shatin, consultant, Drug Safety and Risk 2 Management Committee. 3 DR. CNAAN: Avital Cnaan. I'm a 4 biostatistician. 5 at Children's National Medical Center. 6 I direct multicenter studies DR. ROSENTHAL: Good morning. My name 7 is Geoff Rosenthal. 8 cardiologist and epidemiologist at the Cleveland 9 Clinic, and I'm a member of the Pediatric 10 I'm a pediatric Advisory Committee. 11 DR. MARTINEZ: My name is Fernando 12 Martinez. 13 University of Arizona, and I'm the director of 14 the Arizona Respiratory Center. 15 pulmonologist. 16 I'm a professor of pediatrics at the DR. RAPPLEY: I'm a pediatric Marsha Rappley. I'm 17 from Michigan State University. 18 the Pediatric Advisory Committee, co-chair 19 today. 20 pediatrics. 21 22 I am chair of My area is developmental and behavioral DR. SWENSON: Erik Swenson. I'm professor of medicine in the Division of (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 12 1 Pulmonary and Critical Care Medicine at the 2 University of Washington, and have served on 3 this Committee in the past. 4 MS. KHUC: Kristine Khuc, designated 5 federal official for the Pulmonary-Allergy Drugs 6 Advisory Committee. 7 DR. SCHOENFELD: David Schoenfeld. 8 I'm a professor of medicine at Harvard Medical 9 School, and I'm a biostatistician. 10 11 DR. HOIDAL: John Hoidal, University of Utah, pulmonary critical care physician. 12 DR. GARDNER: Jacqueline Gardner, 13 professor of pharmacy, University of Washington 14 in Seattle. 15 DR. JOAD: Jesse Joad, professor of 16 pediatrics at University of California-Davis. 17 I'm a pediatric allergist and pulmonologist. 18 MS. HOLKA: Andrea Holka. I'm the 19 patient rep on the Pulmonary-Allergy Drug 20 Advisory Committee, and I am executive director 21 of Attack on Asthma in Nebraska. 22 (202) 464-2400 DR. KOCIS: Good morning. Beta Court Reporting www.betareporting.com Keith Kocis (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 13 1 from the University of North Carolina in Chapel 2 Hill. 3 pediatric cardiologist and intensivist. I'm a professor pediatrics, and I'm a 4 DR. WOLFE: Sid Wolfe. I'm a general 5 internist. 6 of Public Citizen. 7 Risk Management Advisory Committee. 8 9 10 I'm with the Health Research Group I'm on the Drug Safety and DR. ZITO: Julie Zito, University of Maryland, professor in pharmacy and psychiatry, and new member to the Drug Safety Group. 11 DR. NEWMAN: 12 University of Colorado, Denver. 13 of public health, and a professor of medicine 14 and a pulmonologist. 15 Lee Newman. DR. SCHNEEWEISS: I'm at the I'm a professor Sebastian 16 Schneeweiss. 17 medicine and epidemiology at Harvard Medical 18 School and School of Public Health, and I do 19 pharmacoepidemiology. 20 I am an associate professor of DR. FRANCIS: Good morning. I'm Henry 21 Francis. 22 of Surveillance and Epidemiology at FDA, and an (202) 464-2400 I'm the deputy director of the Office Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 14 1 infectious disease physician by training. 2 DR. McMAHON: Ann McMahon, acting 3 director, Division of Pharmacovigilance II in 4 the Office of Surveillance and Epidemiology. 5 am a pediatrician by training. 6 DR. SEYMOUR: I Thank you. Sally Seymour. I'm the 7 deputy director for safety in the Division of 8 Pulmonary and Allergy Products at the FDA. 9 DR. CHOWDHURY: I'm Badrul Chowdhury. 10 I'm the director of the Division of Pulmonary 11 and Allergy Products at the FDA. 12 13 DR. ROSEBROUGH: Curt Rosebrough, director, Office of Drug Evaluation II. 14 DR. JENKINS: 15 Jenkins. 16 Drugs at FDA. 17 Good morning. I'm John I'm the director of the Office of New DR. NELSON: Robert Nelson, in the 18 Office of Pediatric Therapeutics, and I'm a 19 pediatric critical care physician and 20 neonatologist. 21 22 DR. MURPHY: I'm Diane Murphy. I'm the director of the Office of Pediatric (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 15 1 Therapeutics, and the office commissioner on 2 pediatric infectious disease specialists. 3 DR. SWENSON: 4 I'll turn it over to Kristine Khuc 5 Thank you very much. of the FDA. 6 MS. KHUC: The Food and Drug 7 Administration is convening today's joint 8 meeting of the Pulmonary-Allergy Drugs, Drug 9 Safety and Risk Management, and Pediatric 10 Advisory Committees under the authority of the 11 Federal Advisory Committee Act of 1972. 12 the exception of the industry representative, 13 all members and temporary voting members are 14 special government employees or regular federal 15 employees from other agencies, and are subject 16 to federal conflict of interest laws and 17 regulations. 18 With The following information on the 19 status of the Committees' compliance with 20 federal ethics and conflict of interest laws 21 covered by but not limited to those found at 22 18 USC Section 208 and Section 712 of the (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 16 1 Federal Food, Drug, and Cosmetic Act are 2 being provided to participants in today's 3 meeting and to the public. 4 FDA has determined that members and 5 temporary voting members of these committees 6 are in compliance with federal ethics and 7 conflict of interest laws. 8 Section 208, Congress has authorized FDA to 9 grant waivers to special government employees Under 18 USC 10 and regular federal employees who have 11 potential financial conflicts, when it is 12 determined that the agency's need for the 13 particular individual's services outweighs 14 his or her potential financial conflict of 15 interest. 16 Under Section 712 of the Federal 17 Food, Drug, and Cosmetic Act, Congress has 18 authorized FDA to grant waivers to special 19 government employees and regular federal 20 employees with potential financial conflicts, 21 when necessary, to afford the Committee 22 essential expertise. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 17 1 Related to the discussions of 2 today's meeting, members and temporary voting 3 members of these committees have been 4 screened for potential financial conflicts of 5 interest of their own, as well as those 6 imputed to them, including those of their 7 spouses or minor children, and for purposes 8 of 18 USC Section 208, their employers. 9 These interests may include 10 investments; consulting; expert witness 11 testimony; contracts/grants; cooperative 12 research and development agreements; 13 teaching/speaking/writing; patents and 14 royalties; and primary employment. 15 Today's agenda involves discussions 16 of the benefit/risk assessment of long-acting 17 beta-2 adrenergic agonists for the treatment 18 of asthma in adults and children. 19 particular matters meeting, during which 20 specific matters related to long-acting 21 beta-2 adrenergic agonists will be discussed. 22 (202) 464-2400 This is a Based on the agenda for today's Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 18 1 meeting and all the financial interests 2 reported by the Committee members and 3 temporary voting members, conflict of 4 interest waivers have been issued in 5 accordance with 18 USC Section 208(b)(3) and 6 Section 712 of the FD&C Act to Dr. Fernando 7 Martinez. 8 Advisory Board for a competing firm, for 9 which he receives between $5,001 to $10,000 10 Dr. Martinez is a member of the per year. 11 With regard to FDA's guest speaker, 12 the Agency has determined that the 13 information to be provided by this speaker is 14 essential. 15 made public to allow the audience to 16 objectively evaluate any presentation and/or 17 comments made by the speaker. 18 The following interest is being Dr. Robert Lemanske has 19 acknowledged that he is a consultant for 20 Merck, GlaxoSmithKline, Novartis, 21 AstraZeneca, and Mapp Pharmaceuticals. 22 guest speaker, Dr. Lemanske will not (202) 464-2400 Beta Court Reporting www.betareporting.com As a (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 19 1 participate in Committee deliberations, nor 2 will he vote. 3 individual to participate fully in today's 4 deliberations. 5 The waivers allow this FDA's reasons for issuing the 6 waivers are described in the waiver 7 documents, which are posted on FDA's website 8 at www.fda.gov/ohrms/dockets/default.htm. 9 Copies of the waivers may also be obtained by 10 submitting a written request to the Agency's 11 Freedom of Information Office, Room 6-30, of 12 the Parklawn Building. 13 statement will be available for review at the 14 registration table during this meeting, and 15 will be included as part of the official 16 transcript. 17 A copy of this With respect to FDA's invited 18 industry representatives, we would like to 19 disclose that Drs. Brahm Goldstein and Bruce 20 Burlington are participating at this meeting 21 as non-voting industry representatives, 22 acting on behalf of regulated industry. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 20 1 Drs. Goldstein and Burlington's 2 role at this meeting is to represent industry 3 in general and not any particular company. 4 Dr. Burlington is an independent 5 pharmaceutical consultant. 6 We would like to remind members and 7 temporary voting members that if the 8 discussions involve any other products or 9 firms not already on the agenda for which an 10 FDA participant has a personal or imputed 11 financial interest, the participants need to 12 exclude themselves from such involvement, and 13 their exclusion will be noted for the record. 14 FDA encourages all other participants to 15 advise the Committee of any financial 16 relationships that they may have with any 17 firms at issue. 18 I would also like to remind panel 19 members to please make your selections for 20 the menu. 21 collect it at the break. 22 (202) 464-2400 Please set it aside and we'll Thank you. Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 21 1 DR. SWENSON: And now I'd like to 2 introduce Dr. Chowdhury of the FDA, who will 3 proceed with some opening remarks. 4 DR. CHOWDHURY: Good morning. Good 5 morning, Honorable Co-Chairs and members of the 6 Pulmonary-Allergy Drugs Advisory Committee, Drug 7 Safety and Risk Management Advisory Committee, 8 and Pediatric Advisory Committee, 9 representatives from AstraZeneca, GSK, and 10 Novartis, and others in the audience. 11 you to this meeting on behalf of the U.S. Food 12 and Drug Administration. 13 I welcome Dear members of the Advisory 14 Committee, I particularly thank you for your 15 participation in this meeting. 16 presentation, I will introduce the objective 17 of this Joint Advisory Committee meeting and 18 the questions that you will discuss and work 19 upon. 20 In this brief There are two inhaled long-acting 21 beta agonist bronchodilators marketed in the 22 United States. (202) 464-2400 These are salmeterol and Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 22 1 formoterol. They are marketed as 2 single-ingredient products and also as 3 fixed-dose combination products with inhaled 4 corticosteroids. 5 salmeterol and formoterol are indicated for 6 use in patients with asthma, exercise-induced 7 bronchospasm, and chronic obstructive 8 pulmonary disease. 9 long-acting beta agonist bronchodilators in Products containing Important safety risk of 10 patients with asthma are asthma-related death 11 and severe asthma exacerbation. 12 Dear members of the Committees, the 13 objective of this Advisory Committee meeting 14 is to discuss risks of these drugs, their 15 benefits, and discuss the risk/benefit in 16 adult and pediatric patients with asthma. 17 As you can see on the agenda, Dr. 18 Robert Lemanske, professor of pediatrics at 19 the University of Wisconsin, will speak 20 first, giving an overview of asthma and 21 current asthma treatments. 22 fortunate that Dr. Lemanske, an expert in (202) 464-2400 We are very Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 23 1 asthma, particularly pediatric asthma, has 2 agreed to speak at this meeting. 3 4 I thank Dr. Lemanske on behalf of the FDA. 5 After Dr. Lemanske, we'll have 6 representatives presenting from the FDA for 7 the representations by sponsors of these 8 products. 9 brief presentations by the FDA, open public 10 We will reconvene tomorrow with hearing, and discussion by the Committee. 11 Dear members of the Committee, as 12 you hear the presentation, I request that you 13 keep in mind the questions that you will 14 discuss and vote on tomorrow. 15 There are a total of 10 questions. 16 I will show the questions in 10 subsequent 17 slides. 18 because there are some common themes in some 19 of these questions, and also, they're 20 available in print at this meeting. 21 22 I will not read all the questions, As you can see, questions 1 through 4 are related. (202) 464-2400 These questions are by active Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 24 1 ingredients and by age groups. 2 through 4 are not for voting. 3 Questions 1 Questions 5 through 8 are also 4 related. 5 long-acting beta agonist 6 bronchodilator-containing drug products, and 7 also by age group. 8 voting questions. 9 These questions are by individual, Questions 5 through 8 are Here is question 1, and I will read 10 this question for you. 11 of using salmeterol for the treatment of 12 asthma in patients not adequately controlled 13 on other asthma-controller medications; e.g., 14 low to medium dose inhaled corticosteroids, 15 or whose disease severity clearly warrants 16 initiation of treatment with two maintenance 17 therapies in each of the following age 18 groups: 19 or older, in adolescents 12 to 17 years of 20 age, in children 4 to 11 years of age. 21 22 Discuss the benefit In adults equal to 18 years of age Question 2 is related to Question 1. (202) 464-2400 This is also a benefit question, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 25 1 but the active moiety in this question is 2 formoterol. 3 five years, to match the lower age of 4 approval of single-ingredient 5 formoterol-containing products. 6 The lower age for children is Question 3 is related to the 7 previous two questions, with the exception 8 that the question is on risks, and the active 9 moiety in this question is salmeterol. 10 Question 4 is related to 11 Question 3. 12 but the active moiety in this question is 13 formoterol. 14 five years, to match the lower age of 15 approval of single-ingredient 16 formoterol-containing products. 17 This is also a risk question, The lower age for children is Here is Question 5, and this is a 18 different theme, and I'll read this question. 19 Do the benefits of Serevent -- salmeterol 20 xinafoate -- outweigh its risk for the 21 maintenance treatment of asthma in patients 22 not adequately controlled on other asthma- (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 26 1 controller medications; e.g., low to medium 2 dose inhaled corticosteroids? 3 disease severity clearly warrants initiation 4 of treatment with two maintenance therapies 5 in the following age groups: 6 years of age and older -- this is a voting 7 question; in adolescents to 17 years of 8 age -- again, a voting question; in children 9 4 to 11 years of age -- again, a voting 10 All host In adults 18 question. 11 Question 6 is related to 12 Question 5, where the drug product subject of 13 this question is Foradil. 14 children is five years, to match the lower 15 age of approval of single-ingredient 16 formeterol-containing product. The lower age for 17 Question 7 is related to Question 5 18 and 6, where the drug product subject of this 19 question is Advair. 20 Question 8 is related to Question 5 21 through 7, where the drug product subject of 22 this question is Symbicort. (202) 464-2400 The lower age Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 27 1 for children is 12 years, to match the lower 2 age of approval of single-ingredient 3 formeterol-containing product. 4 Advair and Symbicort are combination products 5 containing a long-acting beta agonist and an 6 inhaled corticosteroid. 7 Of note, Question 9 and Question 10 are 8 different, and I'll read these for you. 9 Question 9: Based on your discussion, are 10 there further labeling changes or risk 11 mitigation strategies for individual LABA 12 products or the class as a whole that would 13 be advisable? 14 Question 10: What further studies, 15 if any, would clarify important unanswered 16 questions of safety and efficacy for 17 individual LABA products or the class as a 18 whole? 19 We look forward to an interesting 20 meeting. 21 members, for your time, effort, and 22 commitment to this important public health (202) 464-2400 I thank you again, the Committee Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 28 1 service. 2 I now turn the podium over to Dr. 3 Henry Francis, deputy director, Office of 4 Surveillance and Epidemiology, for him to 5 make some brief opening remarks. 6 Dr. Francis, please. 7 MR. FRANCIS: Thank you, Badrul. 8 Chairpersons, distinguished members of the 9 Committee, and guests, the Office of 10 Surveillance and Epidemiology wants to welcome 11 all of you and give a couple of simple messages. 12 First, from the office level, we 13 have not made a final conclusion on this very 14 complex subject, which is dealing with a very 15 important public health issue. 16 forward to the insights, advice, and 17 recommendations from the Committee, and look 18 forward to working with our colleagues and 19 FDA in making an informed decision on the 20 things that we have to do. 21 22 We look Have a good morning, and we look forward to a very successful meeting. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 29 1 DR. SWENSON: All right. At this 2 juncture then, Dr. Lemanske will give us an 3 introductory lecture here on asthma and its 4 control. 5 DR. LEMANSKE: Well, thank you very 6 much. I greatly appreciate the opportunity to 7 address this incredible audience. 8 seen so many bright people in one room before. 9 And it's not just because it's Christmas. I've never It's 10 because obviously, there's a lot of people with 11 a lot of talents. 12 With the time that I have, I would 13 like to address some clinical studies that I 14 have been part of for the last 15 years that 15 have been generated by two NHLBI-funded 16 networks: 17 Network and the Childhood Asthma Research and 18 Education Network. 19 The Asthma Clinical Research I'm going to focus on studies that 20 we have dealt with beta agonists, both 21 short-acting and long-acting beta agonists, 22 to sort of give you a historical perspective (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 30 1 of what the field was thinking in the last 10 2 to 20 years; what studies we put together to 3 try and address what we consider to be 4 important questions; and then the answers 5 that we got to the questions, and how these 6 answers generated some more questions, and 7 where we need to go in the future. 8 9 Before I go over some of these trials, I thought it would be important for 10 me to just give you a little bit of a 11 background on some of the things that we have 12 done also with the EPR-3 group to -- which 13 relies on some of the data that has been 14 generated by these networks to give treatment 15 recommendations to the community. 16 17 Can I advance this from here? it's this one. 18 Oh, Sorry. So first of all, to set the stage, 19 when I was training, we were taught that 20 asthma was primarily a bronchospastic 21 condition of lungs. 22 the advent of bronchial biopsies in asthmatic (202) 464-2400 Then in the 1990s, with Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 31 1 patients, we began to recognize that it was 2 truly an inflammatory condition. 3 questions that we still need to answer are, 4 first, when does this inflammatory process 5 start, particularly in the pediatric 6 population? 7 that, can we prevent it from occurring? 8 we can't, can it be reversed? 9 is it possible that certain of the therapies Some of the If we can figure the answer to If And finally, 10 that we use to treat asthma could actually 11 make the inflammatory process worse? 12 As far as the guidelines are 13 concerned, this inflammatory process was the 14 major focus for the first National Asthma 15 Education and Prevention Program in 1991. 16 1997, based on two studies -- one in the 17 adult population and one in children -- we 18 began to realize the importance of early 19 recognition and diagnosis of asthma and 20 appropriate treatment intervention in order 21 to conserve lung function over time. 22 (202) 464-2400 In In 2002, we didn't revamp the whole Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 32 1 guidelines, but we focused on five important 2 questions, and hopefully answered them. 3 this then led to the current document which 4 was released in 2007. 5 And Now, the new concepts in this 6 current document are on this slide. First, 7 the recommendations that we made are now made 8 using three age ranges, not two. 9 purposes of this Committee, I think this And for the 10 grouping will be very convenient: 0 to 4, 5 11 to 11, and greater than 12 years of age. 12 We also advanced the concept of not 13 only asthma severity, but much more detail in 14 terms of asthma control, and evaluated both 15 severity and control using two domains: 16 Impairment and risk. 17 Now, severity of asthma is really 18 the intrinsic intensity of the disease 19 process, and is most easily and directly 20 measured in patients not receiving long-term 21 therapy. 22 decisions during the initial evaluation and (202) 464-2400 It basically guides clinical Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 33 1 prior to the start of controller therapy. 2 The other concept, control, is the 3 degree to which asthma-related symptoms, 4 functional impairment, and the risk of 5 untoward events are minimized and the goals 6 of therapy are met. 7 clinical decisions to either maintain or 8 adjust therapy once the therapy is initiated. 9 Control really guides And finally, a concept which many 10 different clinical research groups have been 11 interested in is the concept of 12 responsiveness. 13 better to certain types of medications while 14 others do not? 15 phenotypic characteristics of the 16 patient -- gender, for example, race -- or is 17 it related to certain genetic characteristics 18 or the patient's genotype? 19 Why do some patients respond Is it related to certain Both severity and control include 20 the domains of current impairment and future 21 risk. 22 presentation this afternoon will go into this (202) 464-2400 And I think Dr. Stoloff in his Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 34 1 in much detail. 2 is it's basically a cross-sectional 3 evaluation of the patient at any point in 4 time, and it looks at the frequency and 5 intensity of symptoms and functional 6 limitations reflected by evaluations of 7 pulmonary function and quality of life. 8 9 The importance of impairment The risk domain is a more longitudinal look at the patient in terms of 10 the frequency and severity of asthma 11 exacerbations; progressive decline in lung 12 function, which is somewhat difficult for us 13 to assess, but we need to think about it; or 14 the risk of adverse effects from various 15 medications; and in children, of course, 16 inhaled corticosteroids and their effect on 17 growth. 18 Now, it's quite possible that 19 therapy with a certain medication may do a 20 wonderful job in controlling impairment, but 21 not do a very good job in controlling risk. 22 So when we look at clinical trial data, I (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 35 1 think it's important for us to understand 2 what outcomes are being evaluated, and what 3 drugs effect impairment risk or potentially 4 both. 5 The primary goal of asthma therapy 6 is to enable a patient to achieve and 7 maintain control over their asthma to 8 eliminate impairments, including symptoms, 9 functional limitations, poor quality of life, 10 and other manifestations of asthma. 11 This is impairment, and then reduce 12 future risks of exacerbations, emergency 13 department use, and hospitalizations. 14 treatment goals, it's important to recognize, 15 are identical for all levels of asthma 16 severity. 17 The With that as a background, I'd now 18 like to review with you some of the trials 19 that have been generated by the Asthma 20 Clinical Research Network and also the Child 21 Asthma Research and Education Network. 22 it's important for you to understand that (202) 464-2400 Beta Court Reporting www.betareporting.com And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 36 1 these trials that I'm going to show you have 2 been funded by the National Heart, Lung, and 3 Blood Institute and not by the pharmaceutical 4 industry. 5 Therefore, we feel that we have 6 been very dispassionate, if you will, in 7 terms of designing these trials, to hopefully 8 answer questions that we felt were important 9 in terms of filling in on some of the 10 evidence that we thought was lacking, so we 11 could better make recommendations for asthma 12 treatment in national and international 13 guidelines. 14 Let's look first at short-acting 15 beta agonists. 16 in the early 1990s, there was debate as to 17 whether or not taking a short-acting beta 18 agonist on a regular basis was bad or good. 19 And there was a -- when I would go to 20 scientific meetings, there was heated 21 discussions about this. 22 (202) 464-2400 I'm sure many of you remember So the Asthma Clinical Research Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 37 1 Network felt that this is an area that we 2 should try to address. 3 have acronyms, so you'll have to bear with me 4 on this. 5 And all of our trials And the first trial was called the 6 Beta Agonist Study, or the BAGS trial. And 7 the question we wanted to answer here: Is 8 treatment with regularly scheduled albuterol 9 safe? We took a group of mild asthmatics, 10 treated them with two puffs four times a day 11 of albuterol or matching placebo, and then 12 took them off for a five-week period to see 13 if there were differences during treatment, 14 and when we took the treatment away, if the 15 albuterol-treated group got worse because 16 albuterol was doing something bad to their 17 lungs. 18 The answer we got from our trial 19 was it was neutral. 20 harmful nor was it beneficial. 21 published these results in the New England 22 Journal of Medicine in 1996. (202) 464-2400 This therapy wasn't And we Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 38 1 At the time we were working on this 2 trial, Steve Liggett and his colleagues began 3 to uncover the importance of beta adrenergic 4 polymorphism, receptor polymorphisms. 5 asked the question, is it possible that there 6 may be a subgroup of patients based on their 7 genotype that might actually do better with 8 beta agonist therapy or potentially do worse? 9 And so we did a retrospective ancillary 10 And we study. 11 And the question we wanted to ask 12 here is do beta receptor polymorphisms 13 influence the response to chronic treatment 14 with beta agonists? 15 we uncovered was that patients with the 16 Arg/Arg genotype at codon 16 may be at 17 increased risk of loss of asthma control. 18 And let me show you the data on which this 19 was based. 20 And lo and behold, what The major outcome in this trial was 21 a.m. peak flow. 22 looking at the weeks of the study. (202) 464-2400 The yellow bar here, this is Beta Court Reporting www.betareporting.com This is (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 39 1 16 weeks of treatment followed by 5 weeks in 2 the run-out period. 3 And you can see the homozygote 4 Arg/Arg, when they were treated with placebo, 5 did fine. 6 Gly/Gly, when they were being treated with 7 regularly scheduled albuterol, they also did 8 fine. 9 represents about 16 percent of the Caucasian The opposite genotype, the But the Arg/Arg genotype, which 10 population and about 25 percent of the 11 African-American population, when they were 12 given regularly scheduled albuterol, their 13 peak flows went down. 14 off, they went down considerably further. 15 This strongly suggested to us that there may 16 be a genotype-attributable effect with 17 chronic beta agonist therapy that was 18 important for us to try to learn more about. 19 And when we took them Because this was a retrospective 20 study, and retrospective studies can be 21 hypothesis-generating but really don't do 22 much for the evidence base, we decided to (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 40 1 design a prospective study which we called 2 BARGE. 3 here is does this B16 loci influence asthma 4 outcome measures other than peak flow? 5 And the question we wanted to answer And what -- as far as we know, this 6 was the first asthma clinical trial in which 7 patients were randomized by genotype. 8 what we found and published in Lancet in 2004 9 was, yes, the subjects with the Arg/Arg And 10 genotype had better control when regularly 11 scheduled albuterol was stopped and subjects 12 with the Gly/Gly genotype had improved 13 control with regular use. 14 Now, there's much more data to this 15 trial than I have time to discuss, but this 16 confirmed in our minds that at least for 17 short-acting beta agonists, there may indeed 18 be a genotype-attributable effect that we 19 needed obviously to learn more about. 20 With that as a background, the 21 long-acting beta agonists came on the 22 marketplace as we were designing BAGS and (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 41 1 thinking about other trials that would be 2 important for us to be able to do. 3 focused on long-acting beta agonists by 4 designing two companion trials: 5 SOCS, and the other called SLIC. 6 patients who were enrolled in these trials 7 were treated with a medium dose of inhaled 8 corticosteroids for six weeks. 9 that six-week time period, if they were And we One called The At the end of 10 well-controlled, they went into SOCS. 11 they were not well-controlled, they went into 12 SLIC. 13 And if Let me cover SOCS first. The question we wanted to answer in 14 SOCS: 15 asthma, who are well-controlled on inhaled 16 corticosteroids, can salmeterol replace the 17 inhaled corticosteroid and be used as 18 monotherapy? 19 were coming into our clinics and saying, Doc, 20 you started me on this new medicine, 21 salmeterol. 22 and do I have to keep taking these inhaled (202) 464-2400 In patients with mild persistent There were clearly patients who It makes me feel really good, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 42 1 corticosteroids on a daily basis? And what 2 we found is once they were entered into SOCS, 3 they continued their inhaled corticosteroid, 4 they went off their steroid and went on 5 monotherapy with salmeterol, or they went on 6 to placebo. 7 in JAMA was that, no, salmeterol monotherapy 8 would increase the risk of loss of asthma 9 control and actually increase the risk of And what we found and published 10 asthma exacerbations. 11 confirmed what many of us thought, that 12 salmeterol should not be used as monotherapy. 13 So this really What about combination therapy? In 14 the mid-1990s, two groups published some 15 very, very surprising data. 16 from United Kingdom, and the late Ann 17 Woolcock from Australia asked the question, 18 in adult patients who are not doing well on 19 low doses of inhaled corticosteroid, which is 20 the next best treatment step to pursue: 21 Giving them more steroid, thinking that this 22 was an inflammatory condition of the lung, (202) 464-2400 Andrew Greening Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 43 1 and that would be helpful or giving 2 adjunctive therapy in the form of a 3 long-acting beta agonist? 4 what both groups were able to demonstrate is 5 that adding the long-acting beta agonist was 6 actually significantly better than giving 7 more steroid for a number of different 8 outcomes. 9 And lo and behold, Then came the Facet Study, 10 published by our European colleagues, in 11 which they took a group of patients on 12 budesonide, 200 micrograms per day. 13 took another group and they quadrupled the 14 dose of the steroid. 15 other intervention where they kept the 16 steroid dose the same, the low dose the same, 17 added the long-acting beta agonist 18 formoterol, or took the high dose and added 19 the long-acting beta agonist as well. 20 They And then they did one And I think you can see here when 21 they looked at severe exacerbation per 22 patient year, that as you added a long-acting (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 44 1 beta agonist, you got improvement, a 2 reduction in exacerbation. 3 steroid, you also got improvement in terms of 4 exacerbation rates. 5 got the greatest effect. 6 If you gave more And if you did both, you So the data from Greening and 7 Woolcock, along with the Facet Study, 8 suggested that the addition of a long-acting 9 beta agonist could not only influence 10 impairment significantly, but also 11 potentially the risk domain as well. 12 This then led us to SLIC. And 13 remember, the patients going into SLIC were 14 those who were not well-controlled on a low 15 dose of -- or, I'm sorry, a medium dose of 16 inhaled corticosteroid. 17 what the patients were going to do is if we 18 put them on a long-acting beta agonist based 19 on the data from Greening and Woolcock, if 20 they were uncontrolled and then they became 21 controlled, or their control was improved by 22 the addition of the long-acting beta agonist, (202) 464-2400 And we felt that Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 45 1 the patients would then come back to our 2 clinic and they would say, Doc, I'm so much 3 better now that you added this long-acting 4 beta agonist. 5 steroid? 6 Can I reduce the dose of my And if the answer to that question 7 was yes, we thought they would come back 8 again and ask us, can I eliminate the steroid 9 altogether? 10 And what we found and published in 11 JAMA in 2001 is we could reduce the dose by 12 about 50 percent with impunity, demonstrating 13 that LABAs have the potential of being 14 inhaled corticosteroid-sparing. 15 eliminated the drug altogether or the steroid 16 altogether, the treatment failure rate went 17 up significantly. 18 monotherapy with the long-acting beta 19 agonists, at least salmeterol, was not a good 20 idea, but it did offer the potential of 21 steroid reduction in patients on higher doses 22 of inhaled steroid. (202) 464-2400 But when we So like we found in SOCS, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 46 1 Well, you're probably asking, he 2 told me about this Arg/Arg genotype stuff 3 with short-acting beta agonists. 4 the long-acting beta agonists? 5 looked at a retrospective analysis of our 6 SOCS and SLIC data. 7 this, this is hypothesis-generating. 8 9 What about Well, we And remember, when we do And the questions we wanted to ask -- or answer: Are the adverse effects of 10 chronic albuterol treatment in patients with 11 the Arg/Arg genotype at this loci also 12 demonstrable with the long-acting beta 13 agonist salmeterol? 14 concomitant treatment with an inhaled 15 corticosteroid alter these effects? And if so, does 16 And what we found and we published 17 in the Blue Journal in 2006 was that similar 18 effects were seen with salmeterol in the 19 Arg/Arg patients, and also the effect was not 20 prevented by concomitant inhaled 21 corticosteroid treatment. 22 then to thinking about designing another (202) 464-2400 So this led us Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 47 1 trial, prospective trial, to try and look 2 very specifically at genotype-attributable 3 effects, with the addition of a long-acting 4 beta agonist with an inhaled steroid 5 backbone, if you will. 6 So this is the stud, LARGE. And I 7 have to tell you that this data was presented 8 at the AAAAI (?) meetings and the ATS 9 meetings, so it has been in the public forum. 10 But the manuscript is currently in 11 preparation. 12 So I've been asked by my NIH colleagues that 13 you please look at this data, but don't quote 14 it until we have a chance to get it in press. 15 It has not been peer reviewed. The question we want to answer in 16 LARGE: 17 adverse effects on control due beta receptor 18 polymorphisms in patients receiving LABAs in 19 combination with inhaled corticosteroid? 20 what we found, to our surprise, was the 21 answer to this question was no. 22 salmeterol to inhaled corticosteroids for 18 (202) 464-2400 Are there genotype-attributable Beta Court Reporting www.betareporting.com And Addition of (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 48 1 weeks produced similar improvements in airway 2 caliber in both Arg/Arg and Gly/Gly genotype 3 groups, and exacerbation rates were also 4 similar. 5 There have been a number of other 6 groups within industry who have also looked 7 at this particular question regarding beta 8 receptor polymorphisms and long-acting beta 9 agonists. And I don't have time to quote all 10 of them, but this is some work published by 11 Gene (?) Bleecker, who looked at a salmeterol 12 study, and his results were that the response 13 to therapy was not dependent on B16 genotype 14 or haplotype. 15 does not vary between adrenergic genotypes 16 after chronic dosing with an inhaled 17 corticosteroid. 18 And the response to salmeterol So the majority of the evidence at 19 this point would suggest that the addition of 20 a long-acting beta agonist, if it has adverse 21 consequences, we are having a very, very 22 difficult time attributing this to the (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 49 1 patient's genotype in terms of the beta 2 receptor polymorphisms that we have looked at 3 thus far. 4 So why did we see these differences 5 with SABAs, or short-acting beta agonists, 6 and we can't see them with long-acting beta 7 agonists? 8 possibilities. 9 There may be a number of First, genotype-specific 10 differences only occur with short-acting beta 11 agonists, but not with long-acting beta 12 agonists when used with inhaled 13 corticosteroids. 14 Second, higher doses of inhaled 15 corticosteroids could possibly blunt a 16 genotype-specific effect of the therapy. 17 Third, higher doses of inhaled 18 corticosteroids could delay a 19 genotype-specific effect of salmeterol. 20 And finally, genotype-specific 21 effects may be more prominent in 22 subpopulations that were under-represented in (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 50 1 our LARGE study. 2 What about combination therapy in 3 children? Some of the work that led to some 4 of the current approval of this class of 5 drugs in this age group was some work first 6 by Dr. Russell and colleagues, who looked at 7 4 to 16-year aged children who were currently 8 taking inhaled corticosteroids, whose peak 9 flow was less than or equal to 90 percent 10 predicted, and had diurnal variation at peak 11 flow of greater than 15 percent. 12 So these kids were taking inhaled 13 corticosteroids, still had some abnormalities 14 in pulmonary function that the group felt was 15 noteworthy and deserving of something else. 16 And so what they did is they gave 17 half the group salmeterol powder and the 18 other half placebo, and they treated them for 19 12 weeks. 20 Now, this is a summary of what they 21 found. 22 significantly improved morning peak flow, (202) 464-2400 The addition of salmeterol powder Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 51 1 reduced asthma symptoms, and reduced daytime 2 rescue albuterol use. 3 theme that we see with long-acting beta 4 agonists. 5 the impairment domain. 6 Now, this is a common They're very good at addressing Evening peak flow of nighttime 7 asthma symptoms and nighttime rescue 8 albuterol use followed a similar pattern, but 9 were not significant after the first four 10 weeks. 11 studies in which the difference between LABAs 12 and placebo seems to be most noteworthy 13 within the first four to eight weeks. 14 then whether it's due to placebo effect or 15 the children are taking their medicine more 16 on a regularly scheduled basis, the inhaled 17 corticosteroid monotherapy group tends to 18 catch up. 19 This has also been seen in other And The overall incidence of adverse 20 events was similar in both groups. 21 headaches were more common in the salmeterol 22 group. (202) 464-2400 Beta Court Reporting www.betareporting.com However, (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 52 1 What about the addition of 2 salmeterol versus doubling the dose of 3 beclomethasone in children with asthma? 4 is some work published by our Dutch 5 colleagues, Dr. Verberne et al., in the Blue 6 Journal in 1998, in which they looked at 177 7 children, ages 6 to 16 years, with mild to 8 moderate asthma: 9 90 percent predicted, and using low to medium This Pulmonary function 55 to 10 doses of inhaled steroids for at least three 11 months. 12 These kids had demonstrable 13 reversibility with bronchodilator and 14 methacholine hyper-responsiveness. 15 were then put into a six-week run-in period 16 at a constant dose of inhaled steroid, and 17 then treated for an entire year with the 18 addition of a placebo to this inhaled steroid 19 background. 20 plus additional steroids. 21 800 micrograms per day of inhaled 22 corticosteroid. (202) 464-2400 And they This inhaled steroid background So this gave him And finally, the low dose Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 53 1 was continued and salmeterol was added twice 2 daily. 3 And there was a lot of data 4 generated from this trial, and I don't have 5 time to summarize it all, but when they 6 looked at pulmonary function for the entire 7 54 weeks, there was really no difference 8 between the three treatments. 9 when they looked at side effects, this bottom Importantly, 10 line here, this is the change in height from 11 baseline and this is going down about .3 12 sonometers (?). 13 It appears as if the higher dose of 14 the inhaled corticosteroid significantly 15 reduced growth velocity compared to the other 16 treatment groups. 17 And I think this is an important 18 thing to keep in mind, that when we look at 19 patients who are not well-controlled on low 20 doses of inhaled corticosteroid, what is 21 better to give them more steroid, 22 particularly in the pediatric population, or (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 54 1 add a long-acting beta agonist? 2 would suggest that adding a long-acting beta 3 agonist may actually be the safer way to go 4 in terms of producing adverse effects. 5 This data The problem with that thinking, 6 though, is that there really was no 7 difference between these two treatment 8 groups, suggesting -- or begging the 9 question: What greater amount of efficacy 10 does adding a long-acting beta agonist do in 11 children to a backbone of inhaled 12 corticosteroids, and in children who are 13 enrolled in a clinical trial and probably 14 taking their medicine on a regular basis? 15 The CARE Network has also done some 16 work with beta agonists that I'd like to 17 briefly review with you. 18 that I'm going to present to you is not from 19 our group alone, but from a group of 20 wonderful colleagues at National Jewish 21 Medical and Research Center, University of 22 California, San Diego, Washington University (202) 464-2400 And again, the data Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 55 1 at St. Louis, and Fernando Martinez in the 2 audience here from the University of Arizona. 3 The Data Coordinating Center is 4 located at Penn State University. 5 the information that I'm going to share with 6 you has been funded by the National 7 Institutes of Health. 8 And again, A number of years ago, when we were 9 trying to make our recommendations for EPR-3 10 in terms of step 2 care for children, we did 11 not have a lot of comparative -- in fact, we 12 had no comparative studies in which an 13 inhaled corticosteroid was compared to 14 montelukast or compared to combination 15 therapy, all of which were in the 16 marketplace. 17 And we needed to learn more about 18 what is the best choice of therapy for the 19 treatment of mild persistent asthma in 20 children. 21 corticosteroid monotherapy, combination 22 therapy with LABA plus an inhaled steroid, or (202) 464-2400 Three choices: Inhaled Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 56 1 monotherapy with montelukast. 2 The way the trial was designed, 3 there were three parallel groups. 4 got fluticasone, 100 micrograms in the 5 morning, 100 micrograms in the evening. 6 second group got -- we wanted to see if the 7 long-acting beta agonist salmeterol could 8 potentially be steroid-sparing. 9 One group A We treated the kids with 100 10 micrograms of fluticasone in the morning, 11 also salmeterol, and salmeterol at night. 12 The final group got a leukotriene receptor 13 antagonist at night. 14 three groups: 15 combination therapy because the steroid was 16 not given twice a day -- we couldn't because 17 of the formulations available to us at the 18 time -- and finally, a leukotriene receptor 19 antagonist group. 20 And so these were the Monotherapy ICS; packed The primary outcome of packed was 21 asthma control days during the 12-month 22 treatment period. (202) 464-2400 And using self-reported Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 57 1 diary data, an asthma control day was defined 2 as a day without albuterol use, it was 3 permitted pre-exercise; the use of non-study 4 asthma medications; daytime or nighttime 5 asthma symptoms; unscheduled health care 6 provider visits for asthma; and school 7 absenteeism for asthma. 8 9 Now, we generated a lot of data impact, and I don't have time to go over it 10 all with you, but this is a summary slide 11 looking at a number of different outcomes: 12 Asthma control days, of course, was the 13 primary outcome; asthma control 14 questionnaire; asthma treatment outcomes; 15 time to prednisone burst (?); first 16 prednisone burst; time to treatment failure; 17 number of treatment failures. 18 And then looking more at impairment 19 or pulmonary function outcomes: a.m. and p.m. 20 peak flow; FEV1 and FEV1/FEC ratios; exhaled 21 nitric oxide, which we think is a biomarker 22 for airway inflammation; PC-20, which is a (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 58 1 reflection of airway responsiveness, a major 2 physiologic characteristic of asthma; and 3 maximum bronchodilator response. 4 Now, this first column, on a 5 statistically significant basis, fluticasone 6 was favored over montelukast for every single 7 outcome we evaluated. 8 fluticasone over combination primarily were 9 in the pulmonary function domain, including The outcomes favoring 10 exhaled nitric oxide and PC-20. Also, in 11 terms of the different outcomes that favored 12 combination over montelukast, asthma control 13 days were better compared to montelukast. 14 And a number of other pulmonary 15 function outcomes were better on combination 16 than montelukast. 17 we felt very confident when we were putting 18 the stepwise approach for managing asthma in 19 children 5 to 11 years of age, that clearly 20 in the majority of patients, the preferred 21 step two therapy should be low dose of 22 inhaled corticosteroids as opposed to (202) 464-2400 Based on the packed data, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 59 1 combination, or as opposed to leukotriene 2 receptor antagonists. 3 We have also now begun to design 4 trials that are looking at the questions of 5 adjusting therapy based on asthma control. 6 In the children who are not well-controlled, 7 what is the best way to step them up? 8 9 Secondly, in the children who are doing well, the parents are coming to us and 10 they say, Doc, my kids have been on this 11 therapy for years, and is it possible for me 12 to be able to take them off? 13 take -- and if you can take them off, can you 14 do it safely? 15 And if I can And what are my options? So the two trials we're doing right 16 now are, one, looking at the importance or 17 the question of stepping up, and the other 18 the question of stepping down. 19 protocol is called BATGER. 20 you a little background, there is one 21 principal investigator that's responsible for 22 putting together a trial. (202) 464-2400 The step up And just to give And I think -- if Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 60 1 you know that the University of Wisconsin 2 mascot is the Bucky BATGER and I'm from the 3 University of Wisconsin, you can imagine who 4 put this trial together. 5 hard with Penn State University because they 6 were not really too happy about this. 7 I fought long and So I told them, look, you guys, if 8 you can think of a trial with an acronym of 9 Nittany Lions, go for it. We still don't 10 have that trial. 11 BATGER is. 12 for "best add-on therapy giving effective 13 responses." 14 attempting to answer here -- and all the kids 15 are enrolled in BATGER; we should finish this 16 trial next spring -- is as follows: 17 patients receiving daily low-dose inhaled 18 corticosteroid treatment who are not 19 well-controlled, what are the next best 20 treatment options? 21 22 So let me tell you what BATGER is an acronym which stands And the question we're In And the trial is a very unique study design. (202) 464-2400 It's a three-way crossover in Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 61 1 which the children have to demonstrate a lack 2 of control on a low dose or 1X ICS therapy. 3 They're then in a randomized sequence given 4 one of three treatments. 5 depict for you in yellow what one particular 6 subject might undergo through the trial. And I'm going to 7 First, they're doubling the dose of 8 inhaled corticosteroids to see if giving more 9 steroid is better. 10 Second, they then cross into 11 receiving the same dose, but now LABA is 12 added to see if that would be a better 13 therapy for them. 14 And finally, the dose of inhaled 15 steroids is kept the same at this low dose 16 and a leukotriene receptor antagonist is 17 added. 18 Now, obviously, the kids aren't 19 going to be random -- in a very random 20 sequence here. 21 this order. 22 periods is for 16 weeks. (202) 464-2400 They're not all going to get And each of the treatment Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 62 1 Now, the trial has a very 2 interesting design in terms of statistical 3 analysis, in that we are looking at the 4 tendency for a child to have a differential 5 response. 6 therapy versus another. 7 defined three outcomes in which we're going 8 to gauge or evaluate this differential 9 response. 10 That is, they do better with one And we have a priori First, in the risk domain, 11 exacerbations. 12 differential response occurs when the total 13 amount of prednisone prescribed to control 14 asthma symptoms is at least 180 milligrams 15 less on one treatment than on either of the 16 other two treatments. 17 And we're going to say that a The reason we did this is because 18 in our previous trials, exacerbations were 19 really a dichotomous variable in which it 20 either occurred or didn't occur. 21 our kids in our trials get prolonged courses 22 of prednisone, which means that their (202) 464-2400 Beta Court Reporting www.betareporting.com And many of (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 63 1 exacerbations are different because they're 2 lasting longer and perhaps more severe. 3 by doing it in this way, we're going to be 4 able to evaluate a continuous as opposed to a 5 dichotomous variable. 6 So FEV1 we're defining as like a 7 differential response is occurring when the 8 FEV1 changes at least five percent higher on 9 one treatment than on the other two. 10 And finally, asthma control days, 11 another impairment domain. Characteristics 12 we're defining as occurring when the number 13 of annualized asthma control days achieved is 14 at least 31 days more on one treatment than 15 on either of the other two treatments. 16 Now, at the end of the trial, we're 17 going to, hopefully, have a group of kids who 18 are going to be put in different silos (?), 19 if you will. 20 better with more steroid, a group that's 21 going to do better with the same dose of 22 steroid adding a long-acting beta agonist, (202) 464-2400 A group that's going to do Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 64 1 and a group that's going to do better on 1X 2 ICS and the addition of montelukast. 3 If children demonstrate this 4 preferential response to one treatment, they 5 will then be evaluated using secondary 6 outcomes to determine if there are phenotypic 7 and/or genotypic characteristics that are 8 associated with this positive response. 9 goal of BATGER is obviously to help us decide The 10 up front, which is better therapy, to 11 individualize the therapy for our asthmatic 12 children. 13 So in summary, what I've gone 14 through so far is in terms of the ACRN and 15 CARE input into the guidelines, we started 16 with BAGS trying to see if chronic beta 17 agonist therapy was good or bad, at least 18 with short-acting beta agonists. 19 found that the answer to that question was at 20 least in the general population, it wasn't 21 harmful, nor was it beneficial. 22 (202) 464-2400 And we But when we did the retrospective Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 65 1 analysis, it was hypothesis-generating, 2 suggesting that there may be a specific beta 3 receptor polymorphism genotype that may 4 confer increased risks. 5 designed a prospective study, BARGE, and 6 found out that that may well be the case, at 7 least for short-acting beta agonists in the 8 context of no concomitant inhaled 9 corticosteroid administration. We, therefore, 10 We then designed SOCS. 11 trying to establish whether or not at step 12 two CARE people could come off their inhaled 13 corticosteroids and just treat themselves 14 with the long-acting beta agonist salmeterol. 15 And we clearly found that that was not a good 16 idea. 17 We're SLIC was very helpful in moving 18 from step two to step three. 19 were not well-controlled on inhaled steroids, 20 we demonstrated that we could clearly 21 replicate the findings of Woolcock and 22 Greening, that they got better when we added (202) 464-2400 The people who Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 66 1 the long-acting beta agonist. 2 further than their trials in that we were 3 able to demonstrate that we could reduce the 4 dose of steroids by at least 50 percent in 5 the majority of patients. 6 to take the steroids away completely, the 7 treatment failure rate was significantly 8 increased. 9 But we went But when we tried LARGE, to our surprise, when we did 10 a prospective study to look at long-acting 11 beta agonists with the inhaled corticosteroid 12 being administered at the same time, we were 13 not able to demonstrate that there was a 14 genotype-attributable effect. 15 industry colleagues in their retrospective 16 look at their data sets have not been 17 demonstrate this as well. 18 And our Pat, we looked at step two. I 19 think I -- I hope I convinced you the data 20 strongly suggests that inhaled corticosteroid 21 monotherapy is clearly the treatment of 22 choice for the majority of children at this (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 67 1 juncture. 2 BATGER, we're in the process of 3 doing. 4 some very, very needed information as to what 5 we do between step 2 and step 3 in this 5 to 6 11 population. 7 with -- I'm sure this is a topic that you 8 might not want me to say anything about and 9 this is off-label. 10 And this will, hopefully, give us So finally, I'd like to end I guess I have to say that. 11 But the potential for beta agonists 12 and inhaled corticosteroids are both 13 maintenance and reliever therapy, our 14 colleagues from Europe and Canada have done a 15 lot of work with this, and I believe that the 16 drug is approved, the combination therapy is 17 approved for this approach in a variety of 18 different countries. 19 Now, I don't have time to go over 20 all this data with you, but this is some work 21 that was summarized by Paul O'Byrne (?) in 22 the Blue Journal in 2005 from the state study (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 68 1 in which patients, both children and adults, 2 were treated with four times the dose of 3 budesonide plus a short-acting beta agonist 4 as a rescue. 5 and the long-acting beta agonist formeterol 6 plus the short-acting beta agonist does the 7 rescue. 8 9 The combination of budesonide And the maintenance and the lever group who were treated with budesonide, 10 formeterol as maintenance therapy, but when 11 they had symptoms they also used this 12 combination as a reliever instead of a 13 short-acting beta agonist. 14 And the importance of this is that 15 when they're using this beta agonist for 16 symptoms, they're not only getting a 17 bronchodilator. 18 steroid. 19 was the time to first exacerbation. 20 total number of exacerbations were 21 significantly reduced in the group that was 22 receiving combination therapy both for (202) 464-2400 They're getting some more And lo and behold, what they found Beta Court Reporting www.betareporting.com The (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 69 1 maintenance and reliever. 2 This begs the question in my mind: 3 If combination therapy or LABA had something 4 to do with increasing exacerbation 5 rates -- and I know you're going to be 6 discussing this -- this data would suggest, 7 well, that may not be the case in all patient 8 populations. 9 a certain way, this combination -- That in certain groups used in 10 maintenance and reliever -- may have some 11 significant benefit. 12 Now, let me be a little more 13 controversial for you. 14 reliever medication in this context to be a 15 long-acting beta agonist? 16 that we could design an inhaler that would 17 have albuterol in it plus an inhaled steroid, 18 so when the patient was -- when they needed 19 some symptom relief, they would be getting a 20 short-acting beta agonist, but they would 21 also be getting an increased dose of inhaled 22 steroid. (202) 464-2400 Is it necessary for Is it possible Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 70 1 So this is some work that was 2 published by Dr. Poppy (?) et al., in the New 3 England Journal of Medicine in which they 4 looked at four treatment groups. 5 these patients had mild asthma; important to 6 keep that in mind. 7 months. 8 a.m. peak flow. 9 four treatment groups. 10 All of They were treated for six The primary outcome variable was And you can see here the There was a scheduled and there was 11 an as-needed. Group A received placebo. 12 when they needed relief, they were given not 13 only albuterol, but also a dose of 14 beclomethasone, the inhaled steroid. 15 A second group got placebo 16 scheduled, but they only got albuterol for 17 relief. 18 And A third group got steroid 19 maintenance every day and they got 20 albuterol-only rescue. 21 got both the inhaled steroid and the 22 short-acting beta agonist daily for (202) 464-2400 And the third group Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 71 1 maintenance, but they got albuterol only for 2 rescue. 3 And if you look at the data, 4 Group A, which is not getting daily 5 therapy -- remember these patients are 6 mild -- but they are getting albuterol plus 7 the steroid when they need it. 8 a.m. peak flow, the major outcome, and 9 exacerbations, Group A was equal to Group C, 10 was equal to Group D, all of the groups that 11 are getting some form of steroid, either 12 daily or intermittently. 13 better than the patients who are not 14 receiving inhaled steroids at all. 15 In terms of And they did much The cumulative dose of inhaled 16 corticosteroids was actually lower in Group A 17 compared to that that was received in Group C 18 or Group D. 19 begs the question, again, is this benefit 20 that we see with maintenance and relief 21 therapy, is it related to the additional of 22 LABA, or is it related to the fact that (202) 464-2400 This data is very intriguing and Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 72 1 patients are getting more inhaled 2 corticosteroids when they're having symptoms? 3 So with that in mind, the final 4 study I'm going to show you is also a CARE 5 study. 6 children to prevent exacerbations of asthma." 7 And as I just showed in this adult population 8 who had mild asthma, they could potentially 9 get by with just intermittent therapy with The title of this is TREXA, "treating 10 the inhaled steroid, plus a short-acting beta 11 agonist for relief. 12 So the question we're trying to 13 answer in TREXA: 14 low dose inhaled corticosteroid treatment, 15 who are well-controlled, can inhaled 16 corticosteroid doses be reduced? 17 possible, what is the best strategy for doing 18 so? 19 In patients receiving daily And if The design of TREXA is a run-in 20 period in which the children have to 21 demonstrate control on low dose of inhaled 22 corticosteroid. (202) 464-2400 And they are then randomized Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 73 1 into 1 of 4 treatment groups for a total of 2 44 weeks. 3 inhaled corticosteroid twice a day, but gets 4 inhaled corticosteroid and albuterol now, the 5 short-acting beta agonist, for relief. 6 Group B continues to receive the inhaled 7 corticosteroid on a daily basis, but now gets 8 placebo inhaled steroid and albuterol for 9 rescue. 10 Group A continues to receive the Group C receives now no maintenance 11 therapy except placebo. 12 stepping down their maintenance therapy, but 13 they're receiving combination therapy, if you 14 will, with ICS and albuterol as a reliever. 15 And Group C is getting, again, a step down 16 from the inhaled steroid to getting placebo 17 for maintenance. 18 therapy is placebo-inhaled corticosteroid 19 plus albuterol. 20 We're obviously now And now their reliever This trial is currently about 21 two-thirds enrolled. 22 measure that we are looking at in TREXA is (202) 464-2400 And the major outcome Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 74 1 asthma exacerbations. 2 BATGER trying to give us more data in terms 3 of how we step children up and the -- along 4 with TREXA in terms of how we can best step 5 kids down, if we can do this at all in 6 children who are controlled, will be 7 extremely helpful information for us in terms 8 of putting together future treatment 9 recommendations in this age group. 10 The combination of So in conclusion, I'd like to leave 11 you with a final -- my final thoughts. 12 the data that I have been privileged to be 13 part of, I'll have to say that I could not 14 recommend long-acting beta agonists to be 15 used as monotherapy. 16 From Second, combination therapy 17 significantly improves asthma control in both 18 the current impairment and future risk 19 domains. 20 Third, our responses to therapy 21 based on beta adrenergic receptor genotype 22 different with short-acting beta agonists (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 75 1 than with long-acting beta agonists. 2 really not a conclusion, but a question that 3 I think we need to think about answering. 4 Fourth, another question. This is Did 5 children respond differently to long-acting 6 beta agonists? 7 question should be not in terms of adverse 8 outcomes, but from a therapeutic standpoint. 9 Do they perform the same as adults? And here, I think the And I 10 showed you some data to suggest that that 11 might not be the case. 12 harming the kids, but they're not exactly 13 doing a lot more than monotherapy with 14 inhaled steroids. 15 much help us address this question. 16 Not that they're And BATGER should very And finally, I left you with 17 something to think about in terms of the 18 concept of using combination therapy, both 19 for maintenance and reliever, and whether or 20 not this combination in terms of reliever 21 needs to be a long-acting beta agonist or 22 whether or not it can be replaced with a (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 76 1 short-acting beta agonist as long as the 2 patient is getting additional inhaled 3 steroids. 4 And finally, I would like to thank 5 all of my colleagues in the Asthma Clinical 6 Research Network and the CARE Network. 7 has been a tremendous opportunity for me. 8 It's been one of the highlights of my career 9 to work with such an incredible group of 10 This people, such a dedicated group of scholars. 11 And finally, the data that I showed 12 you is obviously generated by some very 13 hardworking coordinators at all of our 14 centers. 15 And I don't want to forget the 16 patients who have participated in our trials 17 and many other trials that have helped us 18 learn a lot about asthma. 19 And with that, I'd like to thank 20 you all for your attention and it's been a 21 great privilege to be here. 22 (202) 464-2400 Thank you. Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 77 1 DR. SWENSON: Dr. Lemanske, thank you 2 for that fantastic review. We're just a little 3 bit ahead of time, so I think we should just 4 proceed with Dr. Seymour's presentation from the 5 FDA. 6 have time for questions. And then after her talk, I think we'll 7 DR. SEYMOUR: Good morning. My name 8 is Sally Seymour, and I'm the deputy director 9 for safety in the Division of Pulmonary and 10 Allergy Products. 11 discuss the risk/benefit assessment of 12 long-acting beta agonists for the treatment of 13 asthma. 14 And we're here today to In order to have a vigorous 15 discussion today, I think it's important to 16 understand the background and regulatory 17 history of these products. 18 focus of my presentation. 19 And that is the Here's an outline of my 20 presentation. 21 majority of my time on the regulatory history 22 of these products, specifically focusing on (202) 464-2400 I'm going to spend the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 78 1 two areas: the basis of approval of these 2 products; and then a chronological walk 3 through the important milestones in the 4 regulatory history, including some safety 5 studies, advisory committees, communications 6 issued by the Agency, and the labeling 7 history for these products. 8 Towards the end, I'm going to spend 9 a few moments on the current labeling so that 10 everyone is clear on the language regarding 11 the risks of these products that's in the 12 current labels. 13 summary. 14 And I'll close with a brief As you've heard, there are two 15 inhaled long-acting beta 16 agonists -- salmeterol xinafoate and 17 formoterol fumarate. 18 as Serevent Diskus, which is a dry powder 19 inhaler, or DPI. 20 marketed as Serevent Inhalation Aerosol, 21 which is a meter dose inhaler, or MDI, but 22 this product is no longer currently marketed (202) 464-2400 Salmeterol is approved Salmeterol also used to be Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 79 1 due to the phase-out of the propellant 2 chlorofluorocarbons. 3 There are two combination products 4 containing salmeterol and the corticosteroid 5 fluticasone propionate. 6 Diskus, which is a dry powder inhaler, and 7 Advair HFA inhalation aerosol, which is a 8 meter dose inhaler. 9 These are Advair Formoterol fumarate is marketed as 10 a dry powder inhaler, Foradil Aerolizer. 11 There is also another dry powder inhaler 12 marketed -- I'm sorry -- that's approved but 13 not currently marketed, and it's Foradil 14 Certihaler. 15 combination product formoterol, and the 16 corticosteroid budesonide Symbicort 17 inhalation aerosol, and this is a meter dose 18 inhaler. 19 In addition, there is a I do want to mention that there are 20 two other long-acting beta agonist products 21 marketed: Perforomist Inhalation Solution and 22 Brovana Inhalation Solution. (202) 464-2400 Perforomist is Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 80 1 formoterol and Brovana is aformoterol, the 2 R-enantiomer of formoterol. 3 products are only approved for COPD, and they 4 are not the focus of discussion today. 5 But these So now that you have an idea of the 6 products we're discussing, I'm going to walk 7 through the regulatory history. 8 shows the approval dates of the long-acting 9 beta agonist products for asthma in patients This slide 10 12 years of age and older, for asthma in 11 patients less than 12 years of age, for 12 exercise-induced bronchospasm, or EIB, and 13 for chronic obstructive pulmonary disease, or 14 COPD. 15 date back to the mid-1990s. 16 And as you can see, these products Several products are also approved 17 for COPD and exercise-induced bronchospasm. 18 And this is important to note, because one of 19 the issues for discussion today is withdrawal 20 of the asthma indication from these products. 21 However, if the asthma indication is removed, 22 there would be still available several (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 81 1 products on the market containing these 2 long-acting beta agonists. 3 I'd also like to take this 4 opportunity on this slide to make the 5 following point. 6 drug device combinations that act locally in 7 the lungs, and each product is considered 8 unique and is required to have a complete 9 development program to support approval. 10 Inhalation products are It's a nice segue to my next slide. 11 What is the basis of approval for these 12 products? 13 adolescents and adult patients 12 years of 14 age and older. 15 I'm going to start in the LABAs are bronchodilators, which 16 means that they dilate the airways, and these 17 products have a well-established clinical 18 development program. 19 program includes a dose ranging study to 20 inform dose selection and to evaluate the 21 pharmacodynamic effects on safety variables 22 such as glucose, potassium, and heart rate. (202) 464-2400 Typically, a Phase II Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 82 1 The Phase III program consists of at least 2 two clinical trials in patients 12 years of 3 age and older. 4 down to 12 years of age is acceptable since 5 the pathophysiology of asthma in response to 6 bronchodilators is expected to be similar in 7 adults and adolescents. 8 9 An inclusion of adolescents The duration of these trials is typically 12 weeks or longer, and the primary 10 efficacy variable is the forced expiratory 11 volume in one second, or FEV1. 12 to discuss that more in a moment. 13 And I'm going There are supporting efficacy 14 variables, including peak expiratory flow, 15 rescue medication use, symptom scores, and 16 nocturnal awakenings, which are included. 17 But these trials are not powered for these 18 secondary variables. 19 In addition to the safety and 20 efficacy of trials, a long-term safety trial 21 is required. 22 year in duration for a new product. (202) 464-2400 This trial is typically one Beta Court Reporting www.betareporting.com And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 83 1 although the purpose of this trial is safety, 2 efficacy is also usually assessed. 3 And this is a very abbreviated 4 outline of the clinical development program. 5 I can assure you there are many other studies 6 conducted throughout these products' 7 development. 8 9 So what about the basis of approval in pediatric patients -- patients less than 10 12 years of age? 11 pediatric program, the efficacy and safety in 12 patients 12 years of age and older must be 13 evaluated to ensure it's appropriate to study 14 the pediatric population. 15 Prior to considering a And because the pathophysiology of 16 asthma in response to bronchodilators is 17 expected to be similar in patients in this 18 age group, the pediatric program is not as 19 extensive as the adult program. 20 Nevertheless, we expect clinical trials to 21 establish appropriate dosing with safety and 22 efficacy measurements in pediatric patients. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 84 1 A dose ranging study to inform dose 2 selection is expected, as well as safety and 3 efficacy trials in asthma patients less than 4 12 years of age. 5 these trials depends on the novelty of the 6 drug substance and product, but efficacy and 7 safety are assessed in these trials. 8 addition, the efficacy assessment in 9 pediatric patients is supported by the The number and duration of In 10 efficacy data in adults and adolescents. 11 In addition to the development 12 program, the Agency can request additional 13 studies for pediatric data under the Best 14 Pharmaceuticals for Children Act. 15 this, this is called a written request, and 16 the Agency's pediatric group provides input 17 on these requests in the design of these 18 studies. 19 When we do Based upon this process, a written 20 request was issued for salmeterol for 21 children less than four years of age, and for 22 formoterol for children less than five years (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 85 1 of age. Additional studies in children older 2 than four or five years of age were not 3 requested. 4 There are several additional 5 important points to make regarding the basis 6 of approval of these products. 7 to mention the combination products. 8 have a development program that must 9 establish the contribution of each component, First, I need They 10 and because of this, these programs are more 11 complex. 12 and efficacy of each component has been 13 established, and the Division views the 14 combination products as products of 15 convenience. 16 But keep in mind that the safety Next, long-acting beta agonists are 17 bronchodilators, and the primary basis of 18 approval is FEV1. 19 are supportive, but these programs were not 20 powered for these secondary efficacy 21 variables. 22 describe is really not unique, as the basis (202) 464-2400 All the other endpoints The development program I Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 86 1 of approval of long-acting beta agonist 2 products is really similar to all other 3 asthma medications, including inhaled 4 corticosteroids, with the primary endpoint of 5 FEV1 and similar supportive secondary 6 endpoints. 7 And one may argue how these 8 endpoints translate into benefits for the 9 patients, but these efficacy variables are 10 important, and are the same variables used to 11 measure asthma control severity as outlined 12 in the Asthma Guidelines such as the NHLBI, 13 an AEPP that Dr. Lemanske described. 14 So the long-acting beta agonist 15 development programs do include endpoints 16 that are important to patients, including 17 FEV1. 18 discussing this particular efficacy variable. 19 Some of you may recognize this type of 20 figure, which is a flow/volume curve during 21 maximal exhalation as part of a pulmonary 22 function test. (202) 464-2400 And I'd like to spend a little time Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 87 1 FEV1 is the volume of air exhaled 2 in the first second of forced expiration. 3 the X axis, you have the volume, and on the 4 Y axis you have the flow. 5 beginning of forced exhalation -- this is the 6 flow curve -- until there is no longer any 7 flow of maximal exhalation. 8 expiratory volume in one second is the volume 9 exhaled in one second. 10 On And at the And the forced This is a measurement of air flow 11 that's useful for obstructive airway diseases 12 such as asthma and COPD. 13 measurement of air flow is a direct 14 clinically meaningful endpoint. 15 bronchodilator drugs, this is really not a 16 surrogate endpoint. 17 such as asthma with symptoms of wheezing, 18 beta agonists relax the airway's smooth 19 muscle and dilate the airways to relieve the 20 cardinal symptoms of asthma that are from 21 airflow obstruction. 22 (202) 464-2400 And this And for For pulmonary disease And you can see, this particular Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 88 1 figure has two curves: one pre-bronchodilator 2 and one post-bronchodilator. 3 of a bronchodilator is a method used to 4 evaluate airway reversibility. 5 can see in the figure, there is an increase 6 in FEV1 following bronchodilator 7 administration. 8 9 Administration And as you An FEV1 increase of greater than 12 percent and 200 milliliters suggests a 10 significant bronchodilation, but lack of this 11 increase does not preclude a clinical 12 response to bronchodilator therapy. 13 response to bronchodilator provides evidence 14 of reversibility, and is one of the criteria 15 used to establish a diagnosis of asthma. 16 clinical trails in asthma patients usually 17 include a response to bronchodilators as an 18 inclusion criterion for the diagnosis of 19 asthma. 20 A So Finally, for spirometry, there are 21 standards from professional organizations for 22 measurement of FEV1, and we consider the FEV1 (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 89 1 a reproducible, reliable, well-established 2 clinically meaningful endpoint in clinical 3 trials for bronchodilators. 4 FEV1 is used to grade severity of impairment 5 for obstructive pulmonary diseases such as 6 asthma and COPD. 7 In addition, In the NAEPP Guidelines, FEV1 is a 8 component of asthma severity, along with 9 symptoms, nighttime awakenings, rescue 10 medication use, interference with normal 11 activity and exacerbations. 12 As Dr. Lemanske mentioned, there 13 are two concepts within the NAEPP 14 Guidelines -- the concept of asthma severity 15 and asthma control. 16 classify patients in both of these concepts. 17 So now that I've given you kind of And FEV1 is used to help 18 an overview of the development programs for 19 these products, I'd like to provide some data 20 regarding the efficacy of these products. 21 Over the next few slides, I'm going to show 22 you some data from the development programs (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 90 1 for these products. 2 product labels or the FDA review of the 3 clinical studies conducted to sort approval. 4 The data comes from the These are by no means complete 5 efficacy data. For LABAs, there are many 6 additional studies conducted not only by the 7 sponsors but by other sources such as the 8 NIH, and Dr. Lemanske showed you many of 9 these studies. And these studies have 10 further established the efficacy of these 11 drugs, and the NAEPP has reviewed these 12 published studies along with the NDA studies 13 to form the basis of their recommendations. 14 The first data I will show you is 15 for Serevent Diskus, and the results of two 16 12-week pivotal trials in patients 12 years 17 of age and older with asthma are shown. 18 primary efficacy variable was FEV1, and the 19 change of a placebo was a 20 percent increase 20 in percent predicted FEV1 at 12 hours. 21 in mind that this is at the end of the dosing 22 interval. (202) 464-2400 The Keep And the results for secondary Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 91 1 endpoints were supportive of efficacy, even 2 though these trials were not powered for 3 these endpoints. 4 And I'd just like to highlight a 5 couple of the secondary endpoints. Note that 6 there is a change -- a decrease in rescue 7 inhalation use of 1.8 puffs per day. 8 Patients take rescue medication when they 9 have symptoms, so this suggests that symptoms 10 are decreased. 11 the other endpoints of percent days without 12 symptoms and nights without awakening, which 13 also indicates that symptoms are improved. 14 And this is consistent with And these results show that 15 Serevent Diskus is an effective 16 bronchodilator in the secondary efficacy 17 variables, showing more days and nights 18 without symptoms, and a decrease in rescue 19 medication use support the benefit of this 20 drug. 21 22 For Foradil Aerolizer, the results are shown from one of the 12-week pivotal (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 92 1 trials. 2 in liters, and at the end of the dosing 3 interval, which was 12 hours. 4 of the dosing interval, there's a 300 5 milliliter difference compared to placebo, 6 with a peak difference of 400 milliliters. 7 The FDA review also noted a statistical 8 difference in terms of peak expiratory flow 9 and rescue inhalation use compared to 10 In this slide, the FEV1 is expressed And at the end placebo. 11 Symptom scores improved and the 12 percent of nights patients required rescue 13 medication use also was less in the Foradil 14 Aerolizer group compared to the placebo. 15 These data show that Foradil 16 Aerolizer is an effective bronchodilator. 17 The secondary endpoints such as rescue 18 medication use and fewer nights awakened 19 support the benefit of this drug. 20 And these are figures from the 21 Serevent Diskus and Foradil Aerolizer product 22 labels which show the post-dose serial FEV1 (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 93 1 measurement on the last treatment day, week 2 12, in adult and adolescent patients with 3 asthma. 4 Y axis on the left is percent predicted and 5 on the right is FEV1 in liters. 6 message is the same. On the X axis is hours, and on the But the 7 You can see that the placebo group, 8 shown here and here, really have no change in 9 the FEV1 over the 12-hour dosing interval. 10 The long-acting beta agonist Serevent Diskus 11 and the two different doses of Foradil 12 Aerolizer show an increase in FEV1 that is 13 sustained throughout the 12-hour dosing 14 interval. 15 In both of these trials, albuterol 16 was also included and administered four times 17 daily. 18 curves for albuterol inhalation aerosol. 19 you can see that compared to albuterol, the 20 long-acting beta agonists have a sustained 21 bronchodilator effect compared to those 22 products. (202) 464-2400 And these figures also show the Beta Court Reporting www.betareporting.com And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 94 1 Moving on to the pediatric 2 population. 3 the Serevent Diskus. 4 And here are some results from The results for one pivotal 12-week 5 trial are shown. The 12-hour data shows an 6 increase in percent predicted of 3.3 percent 7 compared to placebo. 8 noted a decrease in rescue medication use of 9 0.5 inhalations per day, as well as an The FDA review also 10 improvement in symptoms and more nights 11 without awakenings in the salmeterol group 12 compared to placebo. 13 This, taken with the robust adult 14 data in the other clinical trials in the 15 development programs, these data show that 16 Serevent Diskus is an effective 17 bronchodilator in children. 18 For Foradil Aerolizer, the results 19 from a one-year study are shown in which 20 efficacy was assessed at three months. 21 was a 150 milliliter improvement of FEV1 area 22 under the curve compared to placebo, and an (202) 464-2400 Beta Court Reporting www.betareporting.com There (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 95 1 improvement in peak expiratory flow as well. 2 Many of the other secondary variables, 3 including rescue medication use and symptom 4 scores, showed improvement. 5 the robust adult data and the other trials in 6 the development program, these data show that 7 Foradil Aerolizer is an effective 8 bronchodilator in children. 9 And taken with Now on to the combination products. 10 The first combination product approved was 11 Advair Diskus. 12 pivotal 12-week and efficacy and safety 13 trials is shown. 14 several points from this slide. 15 factorial design is typical of a combination 16 program so that we can assess the 17 contribution of each component and we can 18 look at several different comparisons. 19 the focus today are the long-acting beta 20 agonists. 21 22 Data from one of several And I'd like to make This type of But So first, I want to point out that compared to placebo, Advair Diskus shows an (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 96 1 improvement in all of the endpoints: FEV1, 2 peak flow rate, percent rescue-free days, 3 days without symptoms, and awakening-free 4 nights. 5 Life Questionnaire, which I will mention in a 6 moment. 7 And then also the Asthma Quality of Looking at the salmeterol treatment 8 group showed an improvement in FEV1, percent 9 rescue-free days, and percent days without 10 symptoms compared to placebo. 11 way to look at the contribution of salmeterol 12 is to look at the difference between Advair 13 Diskus and fluticasone. 14 product shows an improvement in all of the 15 endpoints compared to fluticasone. 16 shows the contribution of salmeterol to the 17 combination product. 18 But another The combination And this This slide also introduces a new 19 endpoint, the Asthma Quality of Life 20 Questionnaire, or AQLQ, which measures the 21 impact of asthma on patients' perception of 22 health. (202) 464-2400 This is a Juniper questionnaire with Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 97 1 32 items representing four domains: activity 2 limitation, emotion, symptoms, and exposure 3 to environmental stimuli. 4 scored on a seven-point scale, where one is 5 maximum impairment and seven is no 6 impairment. 7 good thing. 8 9 Questions are So an increase in the score is a The minimum important difference is thought to be 0.5. And you can see that 10 Advair Diskus improved more than the minimum 11 important difference even when compared to 12 placebo. 13 of fluticasone and the known efficacy of 14 salmeterol, the data submitted in the 15 clinical program established the efficacy of 16 Advair Diskus in patients 12 years of age and 17 older. 18 And based upon the known efficacy In the pediatric population for 19 Advair Diskus, patients less than 12 years of 20 age, this slide slows data from a 12-week 21 clinical trial with Advair Diskus. 22 discussed in the overview of the clinical (202) 464-2400 Beta Court Reporting www.betareporting.com And as I (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 98 1 development programs, the pediatric programs 2 are not as extensive because of the 3 following: 4 salmeterol and fluticasone have been 5 established in adults and adolescents as well 6 as children, and the efficacy of Advair in 7 adults and adolescents has been established. 8 9 The efficacy and safety of In this program, the trial included Advair Diskus and fluticasone. And you can 10 see that there's improvement in all the 11 endpoints compared to fluticasone, which 12 shows that salmeterol contributes to the 13 efficacy of Advair Diskus. 14 the robust adult data and the other trials in 15 the development program, the known efficacy 16 of salmeterol and fluticasone, these data 17 support the efficacy of Advair Diskus in 18 children. And taken with 19 The last product I'm going to show 20 you the data for is the Symbicort Inhalation 21 Aerosol. 22 one of two 12-week clinical trials taken from (202) 464-2400 This slide shows the results from Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 99 1 the FDA review for this product. 2 You can see that this program is 3 quite complex, and there's a couple things 4 I'd like to point out about this program. 5 addition to the factorial design including 6 each of the components, a fifth arm, with 7 budesonide and formoterol administered in 8 free combination or in separate devices, was 9 also included. 10 In The next thing I'd like to note is 11 the formoterol product used in this program 12 is the Oxis Turbohaler formoterol product. 13 And this is not approved in the United States 14 but is approved in the European Union. 15 the use of the Oxis Turbohaler product as a 16 comparator in this program is acceptable, 17 because the sponsor performed a study which 18 bridged the Oxis Turbohaler to the formoterol 19 and Symbicort. 20 think this will come up later in the 21 discussion today. 22 (202) 464-2400 And And I mention this because I This program also included Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 100 1 predefined criteria for asthma exacerbations. 2 And the data shown here is the percent of 3 patients withdrawn from predefined asthma 4 exacerbations. 5 the slide, there's a lot of information but I 6 only want to highlight a few points. 7 Actually going to the data on Compared to placebo, you can see 8 that the Symbicort Inhalation Aerosol group 9 showed an improvement in all the endpoints, 10 including the Asthma Quality of Life 11 Questionnaire. 12 of 0.5. 13 But it did not reach the MIB Two, compared to placebo, the 14 formoterol group showed an improvement in 15 most of the efficacy variables, particularly 16 post-dose FEV1, percent of rescue-free days, 17 and percent of awakening-free nights. 18 you can see that compared to placebo, all the 19 treatment groups, and especially the 20 combination, decreased the percent of 21 patients withdrawn due to the predefined 22 asthma events. (202) 464-2400 Beta Court Reporting www.betareporting.com And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 101 1 And finally, what I want to point 2 out is that the results for the Symbicort 3 versus the free combination treatment groups 4 are fairly similar, which emphasizes the 5 point that the combination products are 6 products of convenience, as there are similar 7 results when a long-acting beta agonist and 8 inhaled corticosteroids are administered in 9 the same device or in separate devices. 10 I've discussed a lot of different 11 endpoints. 12 across programs is the AQLQ. 13 is typically performed and evaluated in the 14 same manner. 15 this is cross-study comparison, so keep that 16 in mind. 17 the results of the AQLQ from several 18 different programs in adult and adolescent 19 patients with asthma. 20 And one endpoint to look at This endpoint But I want to remind you that But I have some data here showing I've already shown you the results 21 for Advair Diskus and Symbicort, and I did 22 note the data for the Advair Diskus program (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 102 1 crossed the minimum important difference of 2 0.5. 3 The results from montelukast or 4 Singulair, which is a leukotriene receptor 5 antagonist, is from a one-week clinical 6 trial, and shows a small improvement of the 7 AQLQ but does not reach the minimum important 8 difference. 9 another product, omalizumab, which is a Similar results are seen from 10 monoclonal antibody against IgE, used for 11 severe asthma. 12 that both long-acting beta agonist 13 combination products show a greater 14 improvement in the AQLQ compared to some 15 other asthma medications. 16 The point of this slide is So now that I've outlined the basis 17 of approval for the LABA products, let me 18 draw your attention to the current 19 indication. 20 the maintenance treatment of asthma and 21 prevention of bronchospasm in patients with 22 reversible obstructive airway disease, (202) 464-2400 These products are indicated for Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 103 1 including patients with symptoms of nocturnal 2 asthma. 3 patients not adequately controlled on other 4 asthma controller medications, or whose 5 disease severity clearly warrants initiation 6 of treatment with two maintenance therapies. 7 They should only be prescribed for These drugs are intended for 8 maintenance therapy and not as monotherapy. 9 And I'm going to get into more details on the 10 labeled use of these products later in my 11 presentation. 12 But now I'm going to begin a 13 chronological walk through the regulatory 14 history of these products that spans over a 15 decade. 16 and forth between different drugs, but I 17 think the chronological process is the best 18 way to present this data. 19 Bear with me as I kind of move back Serevent Inhalation Aerosol was 20 approved in February 1994, and this was the 21 first long-acting beta agonist approved and 22 was a meter dose inhaler. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 104 1 There was a Pulmonary Allergy Drug 2 Advisory Committee held prior to the approval 3 of Serevent in February 1993, and the results 4 of the Serevent nationwide surveillance study 5 were considered during the approval process. 6 And I'm going to briefly touch on this study. 7 The SNS study was a randomized 8 double-blind active controlled parallel group 9 16-week trial in the United Kingdom, and the 10 population were 25,000 patients 12 years of 11 age and older with asthma who were randomized 12 in a two-to-one fashion to salmeterol twice 13 daily or salbutamol 200 micrograms four times 14 daily. 15 agonist, known as albuterol in the United 16 States. 17 Salbutamol is a short-acting beta This was not a placebo controlled 18 study, but was an active controlled study in 19 which both arms were treated regularly with a 20 scheduled beta agonist. 21 were serious adverse events and reasons for 22 withdrawals. (202) 464-2400 The outcome measures This table displays the key Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 105 1 findings in the SNS study and remember that 2 the randomization was two-to-one. 3 I'd like you to note the following: 4 There was a numerical increase in the 5 respiratory- and asthma-related deaths in the 6 salmeterol, group with a relative risk of 7 three. 8 respiratory-related and asthma-related 9 hospitalizations or serious events, but there There was no difference in the 10 was a difference in respiratory- and 11 asthma-related withdrawals between treatment 12 groups, and that was statistically 13 significant, favoring the salmeterol group. 14 These results were known and 15 discussed during the 1993 Pulmonary Allergy 16 Drug Advisory Committee regarding Serevent 17 Inhalation Aerosol. 18 Just to spend a moment on this 19 meeting, the Serevent Inhalation Aerosol NDA 20 was the topic of discussion. 21 comments in voting from the summary minutes. 22 The Committee felt that the dose and dosing (202) 464-2400 Here are some Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 106 1 interval was supported, and that safety was 2 adequately demonstrated, but a few caveats 3 and concerns, as would be expected with any 4 drug in this particular class. 5 noted that the pediatric data was not 6 sufficient, and recommended additional 7 studies. 8 9 The Committee There was a unanimous recommendation to approve salmeterol for 10 chronic asthma in EIB in patients 12 years 11 and older, and a unanimous recommendation to 12 not approve salmeterol for chronic asthma in 13 patients less than 12 years of age. 14 upon the data submitted by the sponsor, and 15 consistent with the recommendations of the 16 Advisory Committee, Serevent Inhalation 17 Aerosol was approved in 1994 for patients 12 18 years of age and older. 19 Based Shortly after approval, reports of 20 life-threatening respiratory events and 21 fatalities with salmeterol were reported. 22 Some of these reports suggested that there (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 107 1 were possible inappropriate uses of 2 salmeterol such as as a rescue medication. 3 To address these events, the label 4 was revised in January 1995, and new warnings 5 were added to the label -- warnings that 6 serious acute respiratory events including 7 fatalities have been recorded with 8 salmeterol; salmeterol is not for acute 9 symptoms; is not a substitute for inhaled 10 corticosteroids; and should not be initiated 11 in worsening or acutely deteriorating asthma; 12 and that patients should have a short-acting 13 beta agonist for acute symptoms. 14 Following discussions with the FDA, 15 GSK developed a large simple trial to assess 16 the risk of asthma-related death and serious 17 asthma exacerbations with salmeterol. 18 trial is called the Salmeterol Multicenter 19 Asthma Research Trial, or SMART. 20 multicenter randomized double-blind 21 placebo-controlled trial of 28 weeks 22 duration. (202) 464-2400 Beta Court Reporting www.betareporting.com This SMART was a (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 108 1 The sample size was initially 2 planned to be 30,000, but then was increased 3 to 60,000 in 1999 because of fewer numbers of 4 events than expected. 5 of age, with a clinical diagnosis of asthma, 6 and were currently taking prescription asthma 7 medication. 8 salmeterol 50 micrograms twice daily or 9 placebo twice daily 28 for weeks, in addition 10 Subjects were 12 years Subjects were randomized to to usual asthma care. 11 Ideally, the endpoints for SMART 12 would have been asthma-related deaths and 13 serious asthma exacerbations. 14 historical data, the number of events was 15 expected to be low. 16 endpoint of the study was brought into the 17 combined respiratory-related deaths or 18 respiratory-related life-threatening 19 experiences, which meant intubation or 20 mechanical ventilation. 21 in June 1996. 22 (202) 464-2400 Based upon Thus, the primary SMART was initiated Now we're going to jump forward a Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 109 1 few years. Advair Diskus was approved in 2 2000, and prior to approval, there was 3 another Advisory Committee. 4 November 1999 to discuss the new drug 5 application for this product. 6 meeting, there was unanimous consensus that 7 the benefits of Advair outweighed the risks. This was held in At this 8 The following year, in 2001, 9 formoterol was approved as the Foradil 10 Aerolizer. 11 inhaler. 12 the higher dose of formoterol, 24 micrograms, 13 was not approved due to safety concerns of 14 serious asthma exacerbations. 15 Phase IV commitment was requested to further 16 evaluate the safety of these products. 17 This is a single-dose dry powder One very important point is that In addition, a The following year, the data safety 18 monitoring board for SMART noted an increase 19 in asthma events with salmeterol, 20 particularly in African-Americans. 21 also noted that enrollment was slow. 22 they recommended if the study could not be (202) 464-2400 Beta Court Reporting www.betareporting.com And they And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 110 1 completed in a timely fashion, that the study 2 be terminated and the company disseminate the 3 findings. 4 SMART was terminated in January 5 2003, and a summary of the interim analysis 6 was submitted to the data at that time. 7 released a communication to the public about 8 the interim findings of SMART, and later that 9 year, a box warning was added to the Serevent FDA 10 and Advair labels regarding a small but 11 significant risk increase in asthma-related 12 deaths, and that the risk may be greater in 13 African-Americans compared to Caucasians. 14 As I mentioned, this box warning 15 was applied to both the Serevent and Advair 16 labels. 17 the Agency made the decision to apply the box 18 warning to Advair, because Advair contained 19 salmeterol, and there was no convincing data 20 that inhaled corticosteroids mitigate the 21 risk of long-acting beta agonists. 22 (202) 464-2400 This is important to note because When the box warning was placed on Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 111 1 these products, FDA released another 2 communication that year to discuss the 3 labeling changes. 4 communication issued in 2003 alerting the 5 public about the risks of these products. 6 And this was the second Switching gears a bit, recall that 7 with the approval of the Foradil Aerolizer 8 product in 2001, there was a Phase IV 9 commitment for a safety study, and Novartis 10 submitted the results of this study in August 11 2004. 12 some updated results for SMART. 13 the FDA review in September 2004, the box 14 warnings and warnings for these products were 15 revised again, addressing the additional 16 findings for SMART, including appropriate 17 datasets and statistical analyses. 18 In that same year, GSK also submitted Based upon Because the Agency had the updated 19 results of SMART in the Phase IV study with 20 Foradil Aerolizer, we brought this 21 information to the Pulmonary-Allergy Drug 22 Advisory Committee in July 2005 to (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 112 1 specifically discuss the results of these 2 safety studies. 3 the labeling for the long-acting beta 4 agonists included a box warning on Serevent 5 Diskus and Advair, but there were no 6 medication guides and there was no box 7 warning on Foradil. 8 9 At the time of that meeting, On the next few slides, I'll show you the key results that were discussed at 10 that meeting. Here are some results from 11 SMART. 12 trial in 26,000 patients, and the results are 13 shown for the primary endpoint, the combined 14 respiratory deaths or respiratory-related 15 life-threatening experiences. Recall that this was a large safety 16 And you can see the relative risks 17 for the total population is 1.4, and that 18 African-Americans had a higher relative risk, 19 appearing to be a particular risk for this 20 endpoint. 21 22 A key secondary endpoint was asthma-related deaths, with a relative risk (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 113 1 of 4.37 for the total population. 2 similar finding to the results of the SNS 3 study. 4 This is a Here are the key results for the 5 Phase IV study with Foradil Aerolizer. And 6 this was a randomized blinded placebo 7 controlled 16-week trial in 2,300 patients 8 with asthma 12 years of age and older. 9 that the size was significantly smaller than Note 10 the SMART trial. 11 the higher dose of formoterol, 24 micrograms 12 twice daily, that was not approved. 13 This trial also included The results show that the rate of 14 events were too low to draw firm conclusions, 15 but the data trended towards an increase in 16 serious asthma exacerbations in the 17 formoterol groups compared to placebo. 18 After reviewing the SMART results 19 and the results of the Foradil Aerolizer 20 Phase IV study, the Preliminary Allergy Drug 21 Advisory Committee was posed the following 22 question: (202) 464-2400 Based on currently available Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 114 1 information, do you agree that salmeterol 2 should continue to be marketed in the U.S.? 3 And there was a unanimous recommendation to 4 keep salmeterol on the market. 5 A similar question regarding 6 formoterol: 7 information, do you agree that formoterol 8 should continue to be marketed in the U.S.? 9 And again, there was unanimous consensus to 10 recommend keeping formoterol on the market. 11 Based on currently available And finally, based on currently 12 available information, should the label of 13 the formoterol-containing product include 14 warnings similar to those in the salmeterol 15 label? 16 consensus to recommend class labeling 17 regarding the risks of these products. 18 And there was nearly a unanimous Some of the additional comments 19 from the 2005 Pulmonary-Allergy Drug Advisory 20 Committee included the following: 21 Recommendations to modify the box warning; to 22 discourage monotherapy; and encourage (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 115 1 co-administration with an inhaled 2 corticosteroid; to provide a medication guide 3 and direct patient information; and to 4 maintain a box warning on all the 5 salmeterol-contained products. 6 Following the July 2005 Advisory 7 Committee meeting, the FDA issued a Public 8 Health Advisory. 9 Advisory, FDA requested that the In this Public Health 10 manufacturers of these products update 11 product labels with new warnings for all the 12 LABA products, including the fact that 13 long-acting beta agonists should only be used 14 as additional therapy in patients who have 15 not adequately responded to other asthma 16 controller medications. 17 medication guides for each of these products. 18 FDA also requested At the same time, we issued health 19 care professional sheets regarding these 20 products. 21 health care professional sheets are tools FDA 22 can use to communicate important safety (202) 464-2400 And the Public Health Advisory and Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 116 1 information to the public and health care 2 providers. 3 the recommendations from the 2005 Advisory 4 Committee. 5 These actions are consistent with Because we've requested a 6 medication guide, I wanted to make sure that 7 you were familiar with this. 8 guide is FDA-required labeling that is 9 necessary for patients' safe and effective A medication 10 use of a drug. 11 optional, but per the regulations, a 12 medication guide can be required for one of 13 the following: 14 help prevent serious adverse event; or if the 15 drug product has serious risks of which 16 patient should be made aware of because the 17 information concerning the risk could affect 18 patient's decision to use or to continue to 19 use the product; or that the drug product is 20 important to health, and patient adherence to 21 direction is crucial to the drug's 22 effectiveness. (202) 464-2400 Normally, patient labeling is If the product labeling could Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 117 1 It is somewhat of a high hurdle to 2 require a medication guide, and the majority 3 of drugs do not have this. 4 long-acting beta agonists, a medication guide 5 was required so that patients are aware of 6 the risks of the medications. 7 But for the Following the FDA's request for 8 updated labeling and medication guides, all 9 the labels for the long-acting beta agonist 10 products, including Foradil Aerolizer, were 11 updated in 2006. 12 Inhalation Aerosol was also approved. 13 combination product would be budesonide and 14 formoterol. 15 included the class labeling, including a box 16 warning and medication guide. 17 In the same year, Symbicort The Importantly, the product label In the same year, pediatric 18 exclusivity was granted for salmeterol. 19 this is important because discussion of 20 products granted pediatric exclusivity are 21 required at a Pediatric Advisory Committee. 22 And that brings us to the Pediatric Advisory (202) 464-2400 Beta Court Reporting www.betareporting.com And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 118 1 Committee held in November 2007. 2 At this meeting, salmeterol was one 3 of many products discussed. Typically for 4 these meetings, the Office of Surveillance 5 and Epidemiology performs a review of the 6 post-marketing safety data. 7 safety profile of salmeterol, there was an 8 expanded discussion of this product, and 9 concern was raised regarding the risk/benefit Because of the 10 assessment in pediatric patients. The FDA 11 committed to discuss this issue at a future 12 Advisory Committee meeting, and that, of 13 course, is the meeting we're having today. 14 To be complete, I wanted to show 15 you the questions discussed at the 2007 16 Pediatric Advisory Committee. 17 voting questions to the Committee, but a very 18 long statement followed by a request for 19 discussion. 20 introductory statement; it is included in the 21 briefing package in the Pediatric Advisory 22 Committee summary minutes. (202) 464-2400 There were no And I will not read this Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 119 1 But I will paraphrase as follows: 2 The Office of Surveillance and Epidemiology 3 provided an analysis of available 4 observational studies and a subgroup analysis 5 of the pediatric population in clinical 6 trials. 7 a future Advisory Committee meeting. 8 9 FDA committed to bring this issue to The Committee was asked to discuss the following: Whether the current labeling 10 and medication guide adequately communicate 11 the potential risks in children, and whether 12 the labeling adequately addresses the 13 increase in pediatric hospitalizations; or 14 whether it's clear that salmeterol should 15 only be used as additional therapy for 16 patients not adequately controlled on other 17 asthma medications; and whether the labeling 18 is clear that there is no clear evidence that 19 the inhaled corticosteroids mitigate the risk 20 of asthma-related deaths. 21 22 This slide shows some of the Committee responses from the summary minutes. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 120 1 There were some members who discussed the 2 possibility of removal of salmeterol from the 3 market, but the Committee agreed more 4 extensive discussion was needed. 5 Committee requested that FDA report back 6 after additional review of the data, and some 7 expressed a sense of urgency from a public 8 health perspective. 9 update the labeling to identify pediatric The Some recommended the FDA 10 risks, including hospitalizations. 11 Committee generally agreed with the FDA's 12 recommendation to continue this assessment of 13 the risks of long-acting beta agonists, and 14 to seek advice from a future Advisory 15 Committee. 16 The Following the Pediatric Advisory 17 Committee, FDA requested that manufacturers 18 of these products provide data from the 19 controlled clinical trials to further 20 evaluate the safety of the long-acting beta 21 agonists when treating asthma. 22 2008, FDA updated its website, confirming (202) 464-2400 In March of Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 121 1 plans for a future Advisory Committee for 2 this topic. 3 have submitted the data FDA requested, which 4 was used with the meta-analysis in the FDA 5 briefing package. 6 Since that time, the sponsors And that brings us to today -- the 7 joint meeting held to discuss the 8 benefit/risk assessment of long-acting beta 9 agonists for the treatment of asthma in adult 10 and pediatric patients. 11 Recall that the discussion at the 12 Pediatric Advisory Committee was regarding 13 salmeterol use in pediatric patients, but the 14 Agency decided that the discussion should be 15 broadened to include all LABA products in 16 adults as well, because the majority of the 17 safety data with these products is in adults 18 and adolescents. 19 decision to label salmeterol and formoterol 20 similarly with class labeling; thus, the 21 discussion today is for both adult and 22 pediatric patients with asthma and for both (202) 464-2400 The Agency has made a Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 122 1 salmeterol and formoterol. 2 So now we're up-to-date on the 3 regulatory history. 4 presentation with the current labeling so 5 that everyone is clear on the language in the 6 labels. 7 And I'm going to end my The product labels for the 8 long-acting beta agonist products have 9 information regarding the risks in multiple 10 sections of the product label, including 11 those listed on this slide. 12 thought it was important to repeat the 13 message regarding the risks and appropriate 14 use throughout the label. 15 slides, I wanted to highlight some specific 16 sections of the label. 17 the box warning. The Agency Over the next few And I'll begin with 18 First, the box warning in all the 19 products is regarding asthma-related death. 20 The data regarding asthma-related deaths was 21 from SMART. 22 (202) 464-2400 There were other endpoints in Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 123 1 SMART, including respiratory-related deaths, 2 respiratory-related life-threatening 3 experiences, and hospitalizations. 4 the Agency did not want to label the 5 LABAs -- the Agency decided to label the 6 LABAs for asthma-related deaths, as this was 7 the most important endpoint, and we did not 8 want the message to be diluted with 9 discussion of all the other endpoints in However, 10 SMART. 11 data is with salmeterol, the Agency has 12 labeled formoterol with the same box warning. 13 It contains the same language. 14 And although the asthma-related death Second, although not explicitly 15 stated, the box warning applies to all age 16 groups. 17 asthma deaths seen in the pediatric 18 population, the Agency takes the conservative 19 approach and assumes that the risk of 20 asthma-related deaths applies to all ages. 21 22 Despite the fact that there's no Next, the box warning addresses the appropriate use of LABAs as an additional (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 124 1 therapy to a controller medication. 2 information was included not because inhaled 3 corticosteroids mitigate the risk, but 4 because patients with intermittent or mild 5 asthma are not appropriate for long-acting 6 beta agonist therapy. 7 allows for practitioners to add salmeterol or 8 formoterol to a variety of different 9 controller medications. 10 This And this language And finally, the box warning 11 contains information from SMART. 12 shows the current box warning for Serevent 13 Diskus. 14 These are in the briefing package. 15 box warning describes the risk of 16 asthma-related death, appropriate use because 17 of this risk, and it does in the salmeterol 18 product labels include some data from SMART. 19 This slide And I'm not going to read it to you. But the This slide shows some of the 20 figures from the clinical trials section 21 where SMART is discussed in the Serevent 22 Diskus product label. (202) 464-2400 The label includes the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 125 1 data regarding asthma-related deaths. Even 2 though this was not the primary endpoint, as 3 I mentioned earlier, the Agency felt this was 4 a most important endpoint. 5 to the relative risk, the data was also 6 expressed as the excess incidents. 7 10,000 patients treated for 28 weeks, there 8 were 8 additional asthma-related deaths in 9 patients treated with salmeterol in SMART. And in addition In per 10 The label also includes cumulative 11 incidence figures for asthma-related deaths. 12 So the SMART data is described in detail in 13 the product labels, and these types of 14 figures and tables provide prominence in the 15 label in this information. 16 This slide shows the current box 17 warning for Foradil Aerolizer, with 18 information regarding asthma-related death, 19 appropriate use because of this risk, and a 20 reference to SMART. 21 because it's in your briefing package. 22 (202) 464-2400 Again, I won't read it The Foradil Aerolizer does not Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 126 1 contain the same figure table from SMART, but 2 contains the result of the safety and 3 efficacy studies in the Phase IV safety study 4 with Foradil Aerolizer. 5 serious exacerbations are displayed in the 6 adverse reaction sections. 7 tables from the product label. 8 that these tables provide some prominence of 9 the safety information. 10 And the results for These are three You can see This is a repeat of an earlier 11 slide, because even though I'm only showing 12 you parts of the label, the warning regarding 13 the asthma-related death and appropriate use 14 is throughout many sections of the label. 15 Plus, I wanted to remind you that there's a 16 medication guide which has the information, 17 which is the required patient labeling. 18 And for the combination products, 19 the warning and appropriate use information 20 is very similar and is included in these 21 sections of the product label, so I'm not 22 going to show you details. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 127 1 2 But that brings us up-to-date on the current labeling. 3 And I'm going to close my 4 presentation with a brief summary. 5 long-acting beta agonists and the safety of 6 these products have a long regulatory 7 history, and this figure displays many of the 8 key milestones. 9 a different product, and diamonds represent 10 The Each colored line represents approval for different indications. 11 Each long-acting beta agonist 12 product was approved first for asthma in 13 adult and pediatric patients greater than 12 14 years of age and older. 15 Serevent Inhalation Aerosol was the first one 16 marketed, and as I mentioned, is no longer 17 marketed due to the phase-out of the 18 propellant chlorofluorocarbons. 19 As you can see, Other LABA products have been 20 approved for other indications during this 21 time period. 22 approved product that has not been marketed. (202) 464-2400 The Foradil Certihaler is an Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 128 1 This figure also shows, with the 2 red arrows, the Advisory Committees that have 3 been held regarding these products. 4 Pulmonary-Allergy Drug Advisory Committee, 5 dedicated to the discussion of the Serevent 6 NDA; the 1999 Pulmonary-Allergy Drug Advisory 7 Committee, dedicated to the discussion of the 8 Advair Diskus NDA; the 2005 Pulmonary-Allergy 9 Drug Advisory Committee, dedicated to the The 1993 10 discussion of the safety of these products, 11 including SMART and the Phase IV Foradil 12 Aerolizer study. 13 And the Pediatric Advisory 14 Committee, in which the one year 15 post-exclusivity of salmeterol was discussed. 16 And of course, we can add another red arrow 17 for this meeting we're having today. 18 I also want to explain this figure. 19 In the blue boxes are significant changes to 20 the product label throughout this time, 21 beginning, of course, following the approval 22 of Serevent Diskus in 1995, the additional (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 129 1 warnings to the label regarding these 2 products. 3 And then following the SMART study 4 ending in 2003, box warnings were added to 5 the labels and were revised again based upon 6 updated results in 2005, I believe. 7 following the Pulmonary-Allergy Drug Advisory 8 Committee meeting in 2005, medication guides 9 were added to these products, as well as And then 10 class labeling was extended to the formoterol 11 products. 12 the Agency's active participation in 13 addressing the risk of these products. 14 And I believe this timeline shows During my presentation today, I 15 described the basis of approval of the 16 long-acting beta agonists and the benefits of 17 long-acting beta agonists. 18 Dr. Lemanske also touched on this in his 19 presentation. 20 And I believe Long-acting beta agonists are 21 bronchodilators, approved based upon the 22 measurement of air flow, FEV1. (202) 464-2400 FEV1 is a Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 130 1 clinically meaningful endpoint and has been 2 used as the basis of approval of virtually 3 all asthma medications. 4 additional efficacy measures that supported 5 approval, such as rescue medication use, peak 6 expiratory flow, asthma symptoms, and 7 nocturnal awakenings. 8 9 There were The risks of LABA are known. There is a box warning regarding 10 asthma-related deaths, serious asthma 11 exacerbations, and the box warning regarding 12 asthma-related death applies to all ages, and 13 applies to salmeterol- and 14 formoterol-containing products. 15 provides appropriate use information, but 16 these products should only be used as 17 additional therapy in patients not adequately 18 controlled on other asthma controller 19 medications. It also 20 Today's discussion is a 21 continuation of the benefit/risk assessment 22 of these products for the treatment of (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 131 1 asthma. 2 Thank you. 3 DR. SWENSON: 4 At this juncture then, we have time Thank you, Dr. Seymour. 5 for questions both to Dr. Lemanske and to 6 Dr. Seymour. 7 of the several committees, if they would 8 simply put on their mic, we'll see the red 9 light and we'll try to take questions in 10 And I would ask if any members order. 11 So it's now open for questions. 12 MR. HENNESSY: Sean Hennessy. This is 13 a question for Dr. Seymour. I'm wondering what 14 the most plausible biological mechanism is for 15 the apparent increased risk in asthma-related 16 events and deaths associated with long-acting 17 beta agonists. 18 biological mechanisms? 19 DR. SEYMOUR: Has there been any discussion of I don't think we know 20 the answer to that for sure. 21 Dr. Lemanske touched a little bit on some of the 22 pharmacogenomics. (202) 464-2400 And I think And I think that the data is Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 132 1 not there to support the pharmacogenomics -- are 2 the cause of the difference in asthma-related 3 death with these products. 4 One theory is that these products 5 make symptoms -- make patients feel better, 6 and that maybe they mask an increase in 7 inflammation that's going on, and that the 8 patients build up to a catastrophic event. 9 But I think the bottom line is that we really 10 don't know the mechanism for this. 11 DR. SWENSON: Dr. Jones. 12 DR. LEMANSKE: If I could make one 13 other comment. 14 to consider here is the design of the trials on 15 which this data is based. 16 home the point that salmeterol should not be 17 used as monotherapy. 18 the SMART study, the patients were given the 19 medication and they could be using any other 20 medication that they wanted to but it wasn't 21 monitored. 22 these people -- because you feel better when you (202) 464-2400 I think the other thing you have I think I hammered And the trial, at least in So it's quite possible that many of Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 133 1 take long-acting beta agonists and other 2 medicines -- and they were getting this for 3 free, and other medicines, they'd have to pay 4 for -- that it's possible that they stopped 5 their other medications and continued their 6 long-acting beta agonists. 7 Now, this is pure speculation. We 8 have no data to say that's simply the case. 9 But I will tell you from a clinical 10 standpoint that it's harder to 11 prescribe -- to get somebody to take a 12 steroid because when they take it, they don't 13 feel any different acutely. 14 to use any form of a beta agonist because 15 they're getting symptom relief, and it's very 16 reinforcing in terms of what that inhaler can 17 or cannot do. 18 DR. SWENSON: 19 DR. JONES: It's much easier Dr. Jones? This is for Dr. Lemanske. 20 Could you talk a little bit about the biology of 21 inhaled corticosteroids and beta agonists, and 22 how they might biologically interact? (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 134 1 DR. LEMANSKE: Well, it's been 2 hypothesized and there is some in vitro data to 3 demonstrate this -- that corticosteroids will 4 increase the numbers of receptors for beta 5 agonists and perhaps have something to do with 6 their function as well in a good way. 7 And the other piece is the other 8 direction, and that is that beta agonists may 9 actually facilitate steroid metabolism, or 10 its effect in terms of its biologic effects. 11 So, it looks like there's an interaction 12 going both ways. Positive interaction. 13 DR. SWENSON: 14 DR. NEWMAN: Dr. Newman? Can you hear me now? 15 Okay. 16 PADAC meeting in 2005, we took a number of 17 actions that you described -- the FDA did -- in 18 terms of medical guidelines, revised box 19 warnings, the health care professional 20 materials. 21 the effectiveness of that in terms of modifying 22 practice? (202) 464-2400 For Dr. Seymour, on the heels of our What, if anything, do we know about Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 135 1 DR. SEYMOUR: I don't know that I can 2 answer the effectiveness of those different 3 changes in labeling in terms of modifying 4 practices. 5 pharmaceutical companies may be able to address 6 that because they can track their prescription 7 sales and see how their prescription sales are 8 compared with inhaled corticosteroids. 9 may be able to provide that information. 10 11 It's possible that some of the DR. NEWMAN: Thank you. And they I'll ask it again later then. 12 DR. SWENSON: 13 DR. WOLFE: Dr. Wolfe? Two very good 14 presentations. 15 the benefit side, you have clear, objective 16 evidence that FEV1 improves, which is what beta 17 agonists should be doing. 18 the spectrum, the risk side, you have clear, 19 objective evidence from asthma deaths, asthma 20 hospitalizations, and intubations. 21 between is this idea of severe exacerbations. 22 The question for I guess both of you is how (202) 464-2400 On one end of the spectrum, on On the other end of Beta Court Reporting www.betareporting.com Somewhere (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 136 1 precise is this definition? 2 severe exacerbation includes but is obviously 3 not limited to hospitalization. 4 various studies that you describe so nicely, how 5 specific and reproducible was the definition of 6 severe exacerbation? 7 DR. LEMANSKE: I'm sure that So in these That's a very, very 8 good question, and that's what we struggle with 9 in interpreting the literature. Exacerbations 10 were hopefully defined a priori in any study 11 design. 12 study to study. 13 taking all our trial data and getting DNA from 14 all the patients. 15 genotyping and looking at exacerbations as an 16 outcome, control as an outcome, et cetera, et 17 cetera. 18 an episode of asthma that requires Systemic 19 corticosteroids, whether it's oral, or 20 intravenously, or intramuscularly. 21 22 And the definition will differ from We're currently right now And we're going to look at And we're defining an exacerbation as But if you look at some of the trial data, some of the stuff that I showed (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 137 1 you this morning with maintenance and 2 reliever in the stay study, they had three 3 definitions of exacerbations. 4 definition. 5 corticosteroids as a definition. 6 also had oral steroids as a definition. 7 it really, really depends on how the 8 exacerbations are defined. 9 a clinician, whether you feel that what Peak flow They had increasing inhaled And they So And then you, as 10 they're telling you is clinically relevant or 11 not. 12 DR. WOLFE: Just a follow-up question 13 to Dr. Seymour, which is on your slide 12, which 14 is Regulatory History of Serevent Diskus 15 Efficacy Data, there was no difference -- I 16 mean, there was a difference in things such as 17 FEV1, obviously, but the last line is on page 6. 18 Asthmatic exacerbations -- there was really no 19 difference between the placebo and Serevent. 20 whatever -- again, the definition was there, it 21 did not show any difference. 22 (202) 464-2400 So And this is really sort of why I'm Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 138 1 asking this question. 2 I'm sure, is accurate that it varies all over 3 the lot. 4 described would be exacerbation, but not 5 necessarily severe exacerbation. 6 I think your answer, And some of the criteria you just So when you start looking at this 7 measure of the benefit, it seems far mushier, 8 obviously, than either FEV1 on one end or 9 asthma deaths, asthma hospitalizations, and 10 the other end. 11 That was the only point. DR. SEYMOUR: I think that's a good 12 point, and your observation on that slide is 13 correct. 14 really is no standardized definition for asthma 15 exacerbation. 16 study to study. 17 with comparing these cross-studies. I concur with Dr. Lemanske that there And they're going to vary from And that is one of the issues 18 DR. SWENSON: Dr. Krenzelock? 19 DR. KRENZELOCK: Thank you. The box 20 warning is rather intimidating, I think, when 21 you read it. 22 the high percentage of illiterate people we have (202) 464-2400 And I'm kind of wondering, given Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 139 1 in the U.S., if you took into account illiteracy 2 when you developed the box warning and maybe the 3 patient information -- and wondering if you took 4 into account again how a patient could make a 5 good informed decision on whether or not to use 6 the product based on the box warning and the 7 information that's available to them. 8 9 DR. SEYMOUR: Right. So the box warning that I showed you is from the 10 patient -- the product label, which is intended 11 for health care providers. 12 is at a level that they can understand. 13 And I hope that this But the medication guide and all 14 the patient labeling do go through our 15 patient labeling experts within the FDA. 16 they evaluate the literacy level for the 17 language in the medication guide. 18 not show you the medication guide language, 19 but it is adjusted to -- for patients, this 20 message. 21 is very difficult to write for patients, but 22 we did feel it was important to have a (202) 464-2400 And And I did But this is a message that I think Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 140 1 medication guide so at least they were aware 2 that there were some risks with these 3 medications, and that they need to talk with 4 their doctor to make sure that this product 5 is appropriate for them. 6 DR. SWENSON: Dr. Gardner? 7 DR. GARDNER: For Dr. Lemanske. 8 thinking ahead toward risk management 9 alternatives. 10 I'm Could you give us a 11 state-of-the-art of genetic testing? 12 of the research arena, how practical and 13 accessible are tests that would allow us to 14 distinguish between hyper-risk among 15 patients? 16 DR. LEMANSKE: Outside Well, to just answer 17 your question very quickly, we're not there yet. 18 But we're working on it. 19 together a Genome-wide Association Study with 20 all of the trials that the ACRN has done, CARE 21 has done, CAMP -- the Childhood Asthma 22 Management Program -- and that's in process. (202) 464-2400 The NIH is putting Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 141 1 We've defined our different phenotypes, and 2 we're now beginning to do genome-wide screening 3 to see if we can learn more about asthma genes, 4 or genes that are responsible for response to 5 therapy. 6 I think it's important to point out 7 that the data that I showed you with beta 8 adrenergic receptor polymorphisms is not the 9 reason why there's a black box label in the 10 product insert; okay? 11 disconnected. 12 be something to do with a person's genotype 13 that may confer increased risk, but we 14 haven't really established exactly what that 15 is at this point. 16 The two are It's possible that there may So to go one step further in terms 17 of risk management, there are companies now 18 who are making tests that you can look at. 19 Beta adrenergic receptor polymorphisms at 20 disposition 16 looking for the R RH (?) 21 genotype. 22 different areas, but we have no data to (202) 464-2400 And clinicians are using this in Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 142 1 support that that is going to be helpful in 2 terms of management or assessing risk in the 3 future. 4 that. So we have to be very careful with 5 DR. SWENSON: 6 DR. KNOELL: Dr. Knoell? Just a point of 7 confirmation or clarification regarding the 8 SMART trial. 9 were recruited initially, they were essentially Is it correct that when patients 10 handed seven containers of Serevent and then the 11 rest was 28 weeks over phone? 12 related to that, I understand that over the 13 course of the trial, the way the patients were 14 recruited into the study somewhat changed, and 15 it became more of a physician referral basis to 16 get patients into the study. 17 how the medications were given in bulk and 18 educated for use with the patients over that 19 time course? 20 DR. SEYMOUR: And then also Did that affect So I'll respond to this 21 question, and if I respond incorrectly, I'm sure 22 GSK will let me know. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 143 1 So I believe it is correct that at 2 the beginning of the trial, patients were 3 handed the medication for the entire 28-week 4 treatment period, and they were contacted by 5 telephone every four weeks. 6 remember that this is a huge trial. 7 simple safety trial, so this is -- in 30,000 8 patients, this was a way that the trial could 9 be conducted. You have to It's a And so it's certainly not 10 ideal trial process, I think, for the study 11 to get done. 12 There was a change in how patients 13 were recruited. I can't remember the exact 14 year cutoff, but I believe that the same 15 thing happened after they changed the 16 recruitment. 17 medications at the beginning and then 18 contacted every four weeks. They were still given all the 19 I would like to use this 20 opportunity to reiterate something 21 Dr. Lemanske said, and to remind everybody 22 that both the SNS and SMART study were done (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 144 1 in the 1990s. 2 guidelines had really become standard of care 3 with the use of inhaled corticosteroids. 4 And this was before all the So I think, you know, the timing of 5 these studies and what patients were 6 receiving at that time probably had some 7 impact on the outcomes of these studies. 8 if they were done today -- which they 9 wouldn't be done today; you wouldn't give And 10 salmeterol as monotherapy -- but there might 11 be different outcomes because patients would 12 be on inhaled corticosteroids. 13 DR. SWENSON: 14 DR. LEMANSKE: Dr. Goldstein? I think if I could just 15 make one more comment. 16 hard for us clinically is to try and determine 17 once and for all if these adverse effects that 18 we're seeing rarely are related to the drug 19 itself, or whether it's related to the fact that 20 the patients are not taking something else that 21 they need for their asthma, like inhaled 22 corticosteroids. (202) 464-2400 I think what's been very And that's a very, very Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 145 1 important point that you probably all recognize 2 from how we've gone over the data. 3 shown you a number of studies that clearly, 4 salmeterol should not be used as monotherapy. 5 And in the SMART study, there was really no way 6 of being able to monitor that. 7 DR. SWENSON: 8 DR. GOLDSTEIN: 9 Dr. Lemanske. I think I've Dr. Goldstein? Question for And I'm going to limit my 10 question to the pediatric population. In your 11 estimate, what is the percentage of pediatric 12 patients that are receiving long-acting beta 13 agonists? 14 anticipation of discussion later on -- if these 15 were removed from the market, what burden would 16 that place on the population? 17 follow-up question is, we've looked at data from 18 prospective studies and prospective data 19 gathering on deaths related to these 20 medications. The follow-up is that if -- in And then a 21 As a pediatric intensivist, I took 22 care of these children when they were in the (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 146 1 ICU after they had failed whatever treatment 2 they were on. 3 occurrence. 4 retrospectively and done an analysis on 5 pediatric deaths in terms of what types of 6 variables were associated with mortality? 7 Death was a very rare Has anybody looked DR. LEMANSKE: In answer to your 8 second part, the numbers of deaths are so small 9 that it's really hard to come up with a common 10 pathway that these patients are following to get 11 to that unfortunate end. 12 With regards to the percentage of 13 patients, I'm sure that my colleagues -- our 14 colleagues from GSK and Novartis, et cetera, 15 have data on what the uses in the pediatric 16 population. 17 say -- now, remember, I'm at a 18 university-based practice so I'm seeing kids 19 with more severe disease -- but I'm 20 prescribing it for about 25 to 50 percent of 21 my asthmatic children. 22 you data that clearly children respond very (202) 464-2400 In my own practice, I would That said, I've shown Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 147 1 well to monotherapy with inhaled 2 corticosteroids. 3 And many, many, many, many children 4 do not need combination therapy. It's this 5 step up from step II to step III, where 6 they're not being controlled on low dose 7 inhaled corticosteroids, and we as clinicians 8 have to make the decision should I push the 9 steroid dose or should I give them something 10 else? 11 address. 12 those kids? 13 over a dose of 200 micrograms of fluticasone 14 a day, you increase the risk of growth 15 suppression in the kids. 16 a cut point where we want to stay underneath 17 if we can. 18 And that's what BATGER is trying to What is the best way to approach We know that as soon as we get So that seems to be Now, obviously, that doesn't happen 19 in all kids, but it's a cut point where you 20 have to start thinking about other options. 21 22 DR. GOLDSTEIN: The follow-up to my other question -- I know there were three in (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 148 1 there -- if long-acting beta agonists were lost 2 as part of your pharmacologic armamentarium, 3 what burden would that pose for your patient 4 population? 5 DR. LEMANSKE: I think right now, in 6 terms of how I use these drugs clinically, if 7 they're not being controlled on low dose inhaled 8 corticosteroid, my next step would be to add 9 LABA as opposed to adding montelukast. So it 10 would be a burden for me because I wouldn't have 11 that option anymore. 12 something else, like give more steroid. 13 maybe that would be the best option. 14 we'll learn from BATGER which kids that would be 15 the best option for. 16 I'd have to choose And And again, As soon as you start limiting 17 choices -- the thing about asthma that's most 18 important and I didn't even say this is it's 19 so individualized. 20 is unique, I think. 21 Johnny is not going to work for Jane. 22 (202) 464-2400 Every patient with asthma And what works for And the challenge we have as Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 149 1 clinicians is to try to find which form of 2 therapy is going to work best for that 3 patient. 4 ability to do that gets more and more 5 limited. 6 And if we limit our options, our DR. SWENSON: At this point, our 7 schedule calls for a break, so we'll have to 8 hold on those questions. 9 at another point in the meeting. 10 I hope they'll come up And we'll reconvene at 11:00. 11 (Recess) 12 SPEAKER: Panel members, please select 13 your menu and pass it on over to me, or actually 14 our staff will be around to collect it. 15 Thank you. 16 DR. SWENSON: And now, Andy Mosholder 17 from the FDA will give us a presentation on the 18 background on the safety issues of LABAs. 19 20 DR. MOSHOLDER: slides up, please? 21 22 Can someone put my I'm Andy Mosholder from the Division of Epidemiology in the Office of (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 150 1 Surveillance and Epidemiology at FDA. 2 Pleased to be talking with you this morning. 3 I'll be giving some background on these 4 safety issues. 5 heard already earlier this morning, and also 6 for those of you who were participants in 7 last year's Pediatric Advisory Committee 8 meeting on Serevent, some of this will also 9 be familiar. 10 Some of this, we will have I want to start by acknowledging 11 the many collaborators who assisted me in 12 preparing this. 13 both within FDA and some people from outside 14 who were kind enough to supply information. 15 It clearly was a team effort This is just an outline of the 16 topics I'll be covering. 17 the history and some of the postmarking 18 surveillance data for long-acting beta 19 agonists; a brief discussion of some of the 20 observational pharmacoepidemiology studies 21 that have been done to look at safety issues 22 with these compounds. (202) 464-2400 First, I'll look at Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 151 1 We'll be looking at data from large 2 safety trials of salmeterol, some of which 3 we've already discussed this morning. 4 I'll be giving an overview of some of the 5 clinical trial meta-analyses that have been 6 done. 7 of LABAs in the United States, and there's 8 already been some interest expressed in 9 looking at that, so hopefully that will And We'll talk some about the current use 10 address some of those issues. And then 11 finally some concluding remarks. 12 So first let's look at 13 postmarketing surveillance data for the 14 LABAs. 15 note. 16 even before the era of the long-acting beta 17 agonists, there were concerns regarding beta 18 agonists used in the treatment of asthma. 19 I should add, too, that nothing And I want to start with a historical As has been alluded to previously, 20 that I'll be discussing today applies to use 21 of these products in COPD. 22 different subject. (202) 464-2400 That's a So this discussion is Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 152 1 limited to asthma. 2 But at any rate, in the 1960s, 3 there was an epidemic of asthma deaths in 4 countries -- in younger patients in countries 5 which had marketed a high dose isoproterenol 6 inhaler. 7 '90s in New Zealand, specifically, again 8 among younger patients, there was an epidemic 9 of asthma mortality which was linked to use 10 of a short-acting beta agonist, formoterol. 11 And when it was removed from the market, 12 asthma deaths declined. 13 long-acting beta agonists, there were 14 concerns about safety of this type of 15 compound. 16 And then similarly in the early So even before the So let's look now at postmarketing 17 surveillance data with the long-acting 18 compounds. 19 reported to FDA's spontaneous reporting 20 database, which is the Adverse Event 21 Reporting System, or AERS. 22 spontaneous cases reported by either (202) 464-2400 These are cases that have been And these are Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 153 1 consumers or health care professionals 2 involving use of one of the LABA products for 3 which the cause of death -- these are all 4 fatal events, and the cause of death appeared 5 to have been plausibly related to asthma. 6 We see here in the early 7 '90s -- actually, there was a peak coinciding 8 with the initial marketing of salmeterol with 9 Serevent, and this raised some concerns. And 10 there was a qualitative nature to some of 11 these reports, such that they seemed to be 12 describing overwhelming asthma attacks such 13 as descriptions of the patients found dead 14 clutching their inhalers and things of that 15 nature. So there were concerns then. 16 And then there is a decline. 17 then in recent years, some increase in 18 reports. 19 these are classified by age group. 20 specifically in 2008 -- this is a partial 21 year total -- since our discussion last year 22 at the Pediatric Advisory Committee reports (202) 464-2400 And Specifically -- I should add that Beta Court Reporting www.betareporting.com And then (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 154 1 of asthma deaths with these compounds, we 2 accrued a total of five additional cases here 3 all in children under 12 years of age. 4 So what conclusions can we draw 5 from the postmarketing surveillance data? 6 Well, despite the fact that there have been 7 these reports, it's actually very difficult 8 to assess causality from these types of data. 9 And the reason is that we have confounding by 10 indication. 11 events of interest being reported, which are 12 serious or fatal asthma exacerbations, are 13 actually related to the condition or which 14 the drug is being prescribed. 15 And by that, I mean that the Furthermore, it's not possible to 16 estimate an incidence from spontaneous 17 reporting data, as I'm sure many of you 18 realize, because we have an unknown degree of 19 under-reporting. 20 would receive all reports of all cases that 21 are occurring in the population. 22 of these limitations, we have to turn to more (202) 464-2400 We never assume that we Beta Court Reporting www.betareporting.com So because (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 155 1 systematic data sources. 2 So that brings us next to 3 observational pharmacoepidemiology studies of 4 LABA safety. 5 that have been performed. 6 touch on this. 7 there's interest, of course. 8 been a number of studies conducted in recent 9 years. And there have been a number I'm just going to We can go into it further if But there have I've listed seven of the more salient 10 ones here on this slide, and there are 11 probably others. 12 results of each one individually, for now, 13 I'll just make some summary observations. 14 But rather than go into the Basically, despite having the 15 advantage of large sample sizes, the problems 16 with these type of data include the 17 following: 18 inconsistent regarding the association or 19 lack thereof with serious asthma outcomes. 20 These studies have tended to have limited 21 data relative to the pediatric population. 22 There has been difficulty obtaining adequate (202) 464-2400 First, the findings have been Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 156 1 statistical power for outcomes of interest. 2 In fact, there was one study by GSK 3 that had to be abandoned because of lack of 4 statistical power. 5 difficulty methodologically accounting for 6 differences in the types of patients getting 7 the different types of drugs. 8 words, because these aren't randomized data, 9 the patient's treatment is tailored to their 10 And also, there's great In other condition. 11 And that is very hard to account 12 for in the analysis. 13 view these of less inferential value than 14 controlled clinical trial data. 15 So on balance, we would So let's turn next to the 16 controlled clinical trials. 17 already discussed two of these with you. 18 have the two large randomized safety trials 19 with salmeterol -- SNS and SMART. 20 be recapitulating some of the findings with 21 those. 22 comparable studies. (202) 464-2400 Dr. Seymour has We And I'll For formoterol, we have no such Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 157 1 And then I want to introduce the 2 concept of the meta-analyses, which is a 3 technique that's often used to enhance sample 4 size and statistical power by combining data 5 from different randomized trials into a 6 single analysis. 7 that throughout the day. And we'll be talking about 8 Before we look at the results from 9 these trials, there's a couple of concepts I 10 wanted to mention. 11 difference and the number needed to harm, 12 which are related, as I'll describe. 13 Basically, they're trying to ask the 14 question -- exposing how many patients will 15 produce one excess adverse event or adverse 16 reaction to the drug? 17 And these are the risk And they're related in that the 18 number needed to harm is actually the inverse 19 of the risk difference. 20 through this example. 21 22 And let me walk you Suppose one does a study and this particular side effect has an instance of (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 158 1 5 percent on drug and 3 percent on placebo 2 within that study. 3 The risk difference is going to be 4 .05 minus .03, which is a difference of .02. 5 So that's the risk difference. 6 say that that's equivalent to two excess 7 events for every 100 study subjects who 8 received the drug -- over the risk in 9 placebo. 10 Or you can Now, the number needed to harm is 11 calculated as the reciprocal of that. 12 it's 1 over .02, or 50. 13 is that for every 50 patients administered 14 the drug in the study, there was one excess 15 event over what would have occurred on 16 placebo. 17 So And what that means So with those concepts in mind, 18 let's take a look at the data from the 19 salmeterol large safety trials. 20 is the Serevent Nationwide Surveillance, the 21 so-called SNS study that was published in 22 1993. (202) 464-2400 First, this Just to reiterate, it was a randomized Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 159 1 16-week double-blind trial done in the U.K. 2 The purpose was to assess safety, and there 3 was no single a priori primary outcome 4 designated. 5 salmeterol 50 micrograms twice a day, 6 albuterol 14 micrograms four times a day, and 7 the randomization ratio is two-to-one. 8 when one looks at the results, one has to 9 keep that in mind. 10 The treatment groups were So The population was mostly adults, 11 although 6 percent were adolescents and were 12 considered to have mild asthma. 13 concomitant ICS use, which is of course a 14 critical issue, unfortunately was not 15 collected during the trial. 16 obtain complete data on patients who dropped 17 out, the investigators had to follow up 18 dropouts at the end of the study period to 19 ascertain whether they were still alive or 20 dead. 21 22 Data on In order to And so now let's look at the results. (202) 464-2400 Here, we see that there were close Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 160 1 to 17,000 patients received salmeterol and 2 just over 8,000 albuterol. 3 at asthma-related withdrawals, we see 4 actually that the relative risk is .8, which 5 that means it's favoring salmeterol. 6 then if we look at asthma-related deaths we 7 have 12 versus 2, and with the two-to-one 8 randomization, that works out to a relative 9 risk of 3. If first we look But And the p-value associated with 10 that of marginal statistical significance 11 .105. 12 And for all cause death, we see a 13 numerical imbalance again favoring albuterol, 14 which did not reach statistical significance. 15 Let's look next at the SMART trial, 16 which -- as you've heard, this was a U.S. 17 trial. 18 bit longer than SNS. 19 randomized treatment arms were salmeterol 42 20 micrograms twice a day, and placebo in a 21 one-to-one randomization ratio in this case. 22 Twelve percent of subjects were adolescents; (202) 464-2400 Actually, 28 weeks in length, so a And here, the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 161 1 the remainder being adults. And again, data 2 on concomitant ICS use was not collected 3 during the trial. 4 the National Death Index was used to complete 5 data on vital status for the participants, 6 including the ones who had dropped out of the 7 trial. 8 early and did not meet its targeted 9 enrollment. In order to be complete, This study was actually terminated 10 And so we'll look at the results. 11 Again, you saw this previously, but just to 12 review, there were roughly 13,000 in each 13 treatment arm. 14 been a priori designated was a combination of 15 respiratory-related death or life-threatening 16 experience. 17 imbalance favoring placebo. 18 statistical significance. 19 deaths, 13 versus 3, giving a relative risk 20 of 4.4, with a confidence interval that 21 excluded one -- that translates to a number 22 needed to harm of about 1,300. (202) 464-2400 The primary outcome that had And this was a numerical It did not reach However, asthma Meaning that Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 162 1 for every 1,300 patients who received 2 salmeterol in the trial, there was one excess 3 asthma death over what would have occurred on 4 placebo. 5 The other point here, with a 6 relative risk of this size, it means that in 7 actuality, one would attribute the majority 8 of the asthma deaths in the trial to the 9 action of salmeterol. There was another 10 outcome respiratory-related death which is a 11 somewhat larger category, as you see. 12 this was increased -- it was more than 13 doubled with salmeterol statistically 14 significant. 15 for this definition of about 1,000. 16 all cause death, again, a numerical imbalance 17 favoring placebo, which did not reach 18 statistical significance. 19 Again, And the number needed to harm And for Now, I wanted to put the results 20 from the two studies side by side to make a 21 few points here. 22 (202) 464-2400 I've labeled them LABA and no LABA Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 163 1 just for consistency with the later 2 meta-analyses. 3 the patients had randomized albuterol on a 4 scheduled basis. 5 use it as a reliever medication, so it's not 6 too much of a stretch to compare these two 7 trials. 8 look at withdrawals for lack of efficacy and 9 asthma-related withdrawals, we see that Now, it's true that in SNS, In SMART, they were able to And we see, first of all, if one 10 actually, the relative risk favors 11 salmeterol. 12 symptoms, as we've heard, are being -- are 13 responding to the salmeterol. 14 So the more ordinary types of But then as one goes up the 15 hierarchy we, see sort of the opposite 16 pattern. 17 hospitalizations, in SNS, it was a neutral 18 effect and slightly increased -- 1.2 or 19 20 percent higher with salmeterol in the 20 SMART trial. 21 more severe outcomes, either intubations or 22 asthma deaths, we see even higher relative (202) 464-2400 If you look next to asthma And then if one looks at even Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 164 1 risks. 2 paradoxical nature of the phenomenon here, 3 which is that the more ordinary routine 4 symptoms are being controlled, but at a price 5 of greatly increased asthma mortality and 6 life-threatening events. 7 So this I think illustrates the And then the other point to make as 8 we look later on at data from the 9 meta-analysis, it's important to note I think 10 that for asthma hospitalizations, which sort 11 of make up the majority of events in the 12 meta-analysis Dr. Levenson will present, 13 actually these trials, were close to being 14 neutral between treatment arms. 15 So when one lacks data on more 16 severe outcomes because of sample size or 17 whatever, I would argue that it's difficult 18 to draw too much reassurance from the finding 19 of neutrality on asthma hospitalizations, 20 because it's not necessarily predictive of 21 what will be the case for asthma deaths. 22 This is taken from the labels. (202) 464-2400 Beta Court Reporting www.betareporting.com I (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 165 1 think we've seen this already this morning, 2 but just to make the point, this is a 3 cumulative incidence plot for asthma deaths. 4 And this is the total sample at the top. 5 this is the African-American subgroup at the 6 bottom. 7 it appears the African-American subpopulation 8 out of the total sample was actually showing 9 a greater risk of asthma-related deaths 10 And And this is to make the point that relative to placebo. 11 So what can we conclude from these 12 trials? 13 European Respiratory Journal: "In view of the 14 results of the two studies, both of the 15 highest evidence class, the existence of 16 salmeterol-related excess mortality has to be 17 assumed with near certainty." 18 the Cochrane Review Group was able to 19 quantify that. 20 Here, I quote an editorial from the And actually, They calculated a combined odds 21 ration of combining the two trials for asthma 22 mortality and the result was 3.7, with a (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 166 1 p-value of .0074. 2 statistical probability testing on this 3 finding, that's consistent between the two 4 trials. 5 So putting some And the next, I hope Dr. Martinez 6 doesn't mind that I quote him here, but as he 7 wrote in the New England Journal, "One death 8 was attributable to salmeterol for every 700 9 patient years of treatment in SMART, a result 10 strikingly similar to that in the United 11 Kingdom study." 12 limitations of the trials preclude definitive 13 conclusions regarding the potential for 14 inhaled corticosteroids to limit or prevent 15 these adverse outcomes. 16 Unfortunately, the A point here -- an excess death 17 rate of 1 per 700 patients per year is not 18 going to be obvious to prescribers, 19 particularly if the milder symptoms are 20 responding at the same time. 21 illustrates part of the difficulty with 22 this -- handling these kinds of risks. (202) 464-2400 So that Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 167 1 To change gears now, look at 2 formoterol. 3 comparable large safety trials with 4 formoterol, but there was a pattern of 5 serious asthma events noted in the clinical 6 trials pre-approval, and in fact, it resulted 7 in the highest dose not being approved. 8 show you those data. 9 data from three of the key pivotal trials, 10 As I mentioned, we have no such I'll This is simply pooled and showing sort of the progression. 11 Here, we have placebo .4 percent; 12 albuterol .7 percent. And then formoterol 12 13 micrograms twice a day, 2 percent had serious 14 asthma events, and at 24 micrograms twice a 15 day, 4.5 percent. 16 included one asthma death and two patients 17 who required intubation. And that number actually 18 One can imagine if one had scaled 19 up a large trial with formoterol, one would 20 probably have seen many such events. 21 also, one could argue that the dose 22 relatedness of these phenomenon lends weight (202) 464-2400 Beta Court Reporting www.betareporting.com And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 168 1 to there being a causal relationship to 2 formoterol. 3 Let's change gears now and start to 4 talk about some of the meta-analyses that 5 have been done. 6 technique to improve the sample size and 7 statistical power. 8 meta-analysis was by Saul Peter and 9 Associates, (?) published in 2006. As I said, this is a And sort of the seminal So this 10 was not available at the time of the 2005 11 Advisory Committee meeting. 12 was to assess the risk for severe asthma 13 exacerbations with either of the compounds. 14 And the purpose They looked at 19 15 placebo-controlled trials at least three 16 months in duration with close to 34,000 17 subjects all told, and six of the trials were 18 pediatric. 19 and confidence intervals for outcomes. 20 did describe in the paper outcomes on asthma 21 deaths, but that's dominated by the SMART 22 trial. (202) 464-2400 They calculated PO odds ratios They So I didn't display that here but it Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 169 1 is in the paper. 2 But what this analysis did is it 3 extended the finding from asthma mortality to 4 asthma hospitalizations, for which there is 5 an odds ratio of 2.6, statistically 6 significant, and also exacerbations requiring 7 intubation for which there was an odds ratio 8 of 1.8, again, statistically significant. 9 And then next, this will be 10 familiar to those who participated in the 11 Pediatric Advisory Committee last year on 12 salmeterol. 13 data that Dr. Saul Peter was kind enough to 14 supply, and actually added data from the 15 SMART trial which we obtained pursuant to 16 last year's Advisory Committee for the 17 adolescent patients in SMART. 18 This was a subgroup of pediatric And one caveat I should mention, 19 some of these data include the unapproved 20 high dose of formoterol, in this trial at 21 least. 22 overall estimated odds ratio for asthma (202) 464-2400 But basically when one looks at the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 170 1 hospitalizations, they were close to tripled 2 over placebo with one of the two compounds. 3 So just a small digression. We've 4 talked about this already this morning a bit, 5 but the question came up of, given this sort 6 of paradoxical finding where the most severe 7 outcomes are increased even when sort of the 8 immediate symptoms may be reduced, what are 9 some possible explanations? 10 As we heard, one is masking a 11 progression of the underlying asthma by the 12 symptomatic relief provided by the 13 bronchodilation, as Dr. Seymour mentioned. 14 The second one is the genetic variant of the 15 beta adrenergic receptor, and Dr. Lemanske 16 already spoke to that. 17 on that, but that's one of the hypotheses. 18 The evidence is mixed And then third would be 19 desensitization or down-regulation of the 20 beta adrenergic receptors. 21 experimental evidence of tolerance, at least 22 in exercise-induced asthma. (202) 464-2400 There is some Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 171 1 I don't claim to be an expert in 2 these types of physiologic and pharmacologic 3 phenomenon, but this is just to introduce 4 that there are some plausible hypotheses to 5 explain the pattern seen in the data. 6 So we want to turn next to look at 7 how LABAs are being used currently in the 8 United States. 9 First, I present here the 10-year 10 prescription trends. 11 Vector I National Database. 12 this is salmeterol, the mono product. 13 red here is the combination product Advair 14 with fluticasone. 15 years, that has come to dominate the market. 16 And then salmeterol and the 17 formoterol-containing compounds making up a 18 much lower proportion. 19 This is from the STI And you see here And in You can see that in recent And something over 18 million 20 prescriptions annually for these products, 21 most of which being the Advair product. 22 (202) 464-2400 This is, again, from STI a Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 172 1 different database, the Total Patient 2 Tracker. 3 of prescriptions. 4 numbers of patients. 5 this as current as possible, so we went up to 6 the year ending September 2008. 7 that for Advair, there's about 5.5 million 8 patients receiving it, and then smaller 9 numbers of the remaining product. And the previous slide was numbers This is to show the And we wanted to make And we see So 10 altogether, about 6 million patients 11 receiving the long-acting beta agonists per 12 year. 13 And we also wanted to look at age 14 distribution of the use. 15 of the four marketed products, just breaking 16 down the age by 0 to 16 and 17 and over. 17 we see that -- I guess roughly speaking, 18 something like 10 percent or less in each 19 case are going to pediatric patients; the 20 remainder going to older patients. 21 varies, as you see, somewhat by specific 22 product. (202) 464-2400 This shows for each Beta Court Reporting www.betareporting.com And And it (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 173 1 And then finally, we're interested, 2 of course, in what the diagnoses associated 3 with their use is. 4 main -- as I said, this discussion is limited 5 to asthma, but of course, these products are 6 indicated for COPD as well. 7 taken from a physician survey known as the 8 Physician Drug and Diagnosis Audit. 9 see that for the four products, something As I said, the other So this is data And we 10 perhaps between 50 and 70 percent are being 11 prescribed for asthma. 12 Now, unfortunately, I don't have 13 this subgroup by age because this survey 14 isn't that granular, but roughly speaking, 15 something over half of the products going for 16 use in asthma. 17 figure of approximately 6 million patients 18 and takes about half of that, we would 19 estimate that there would be 3 million 20 patients receiving LABAs for asthma in the 21 U.S. annually. 22 perspective, that of course means we have to (202) 464-2400 So if one takes the earlier And from the public health Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 174 1 be extremely careful with judging the risks 2 of these products, because with an exposure 3 of 3 million per year, any risks are going to 4 be magnified in the population accordingly. 5 Then there's the interest, of 6 course, in the single entity products, which 7 are a minority of the use, but we wanted to 8 see if we could discern how many patients 9 receiving those were also receiving an 10 inhaled corticosteroid. 11 database that recently became available to us 12 where we are able to look at that. 13 a managed care organization -- a set of 14 managed care organizations, I should 15 say -- selecting patients who had been 16 enrolled for at least a year and who did not 17 have COPD. 18 diagnoses as well. 19 And we have a It's from We have access to information on So limiting it to asthma. We looked at prescriptions for 20 single ingredient LABAs over a three-year 21 period. 22 patients receiving single entity LABAs for (202) 464-2400 The sample was close to 12,000 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 175 1 asthma, and we found that almost half had 2 never received a concurrent prescription for 3 an inhaled corticosteroid. 4 Now, it's true that the majority of 5 these actually were single prescriptions for 6 LABA monoproducts, but still, I have to say 7 we were a bit surprised at this because it 8 was contrary to what's stated in the 9 guidelines, as we've heard. And the source, 10 I should mention -- this is IMS Health Plan 11 claims database. 12 So finally, we wanted to look at 13 what data we have to address the important 14 issue of whether concomitant ICS protects 15 against the catastrophic asthma events with 16 these drugs. 17 clinical trial meta-analyses completed 18 recently that have tried to address this. 19 There have been a couple of First, Bateman and Associates 20 published earlier this year, and that was 21 looking at salmeterol with ICS versus ICS 22 without salmeterol. (202) 464-2400 They found an odds ratio Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 176 1 of asthma hospitalizations of 1.1, not 2 statistically significant. 3 For more severe events, however, 4 there was only one asthma death and one 5 asthma intubation in their dataset. 6 happened to be with salmeterol. 7 there also was another meta-analysis 8 published just in September by Jaske and 9 Associates. Both And then And this looked at actually both 10 compounds, formoterol and salmeterol, with 11 ICS. 12 hospitalizations reduced, although it did not 13 reach statistical significance. 14 more severe events, all five deaths and 15 intubations were occurring in patients who 16 received the LABAs. 17 There, they found actually asthma However, for So although there's not much of an 18 indication of a problem with asthma 19 hospitalizations from these two 20 meta-analyses, we have sort of the 21 uncomfortable fact that all the deaths and 22 intubations were occurring in the patients (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 177 1 who received LABAs. 2 So for data on asthma mortality, we 3 have to turn again to the two large 4 salmeterol trials -- SNS and SMART. 5 problem, as has been mentioned, the data on 6 concomitant ICS use was not collected during 7 either trial. 8 entry into the study, and we can look at how 9 the data are subgrouped according to that. 10 Now, down at the bottom, this is And the What we do have is ICS use at 11 the SMART trial. 12 publication, and you see salmeterol with 13 baseline ICS and placebo with baseline ICS on 14 the left, and without ICS at baseline on the 15 right. 16 This is taken from the And then at the top is the SNS trial. This is data recently available 17 that the Cochrane Review Group obtained and 18 published earlier this year. 19 was -- some data was missing, so it's not 20 data on the complete sample in the trial. 21 Again, you can see how it's subgrouped 22 according to whether presence or absence of (202) 464-2400 And there Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 178 1 baseline ICS. And just sort of to the naked 2 eye, you can sense that there's a greater 3 imbalance in the group without baseline ICS. 4 And the Cochrane Group was able to 5 do some calculations which Dr. Cates was kind 6 enough to provide me. 7 down here, the group without ICS use at 8 baseline, we have a P odds ratio for asthma 9 mortality of 6.4, statistically significant. 10 Basically, if you look Actually, the lower band well away from 1. 11 So with ICS use at baseline, we 12 have a much more imprecise estimate. 13 ratio of 1.5, with a confidence interval of 14 .5 to 4.1. 15 just comparing these two groups. 16 marginal statistical significance of .06. 17 I would emphasize that that's only to say 18 that it seems probable that this group is 19 worse than this group. 20 this group has the risk and this group does 21 not have the risk. 22 to accept the odds ratio of 1.5 and were (202) 464-2400 Odds Additionally, they did a test It had So It's not to say that And in fact, if one were Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 179 1 willing to extrapolate that to the situation 2 of using salmeterol with fluticasone as per 3 Advair, one would still feel that the 4 3 million or so asthma patients receiving the 5 drugs with ICS would have a 50 percent 6 increased risk in asthma mortality. 7 So just to wrap up then, we looked 8 first at the post-marketing surveillance data 9 for which we have reports of asthma deaths 10 with LABAs, but as I explained, this is of 11 limited inferential value in terms of 12 causation. 13 observational safety studies that have been 14 done, but because of methodological issues 15 with non-randomized data, these are of 16 limited inferential value. 17 We touched on some of the We talked about the large 18 randomized safety trials with salmeterol, 19 which showed an increase in asthma mortality. 20 With formoterol, we're lacking comparable 21 trials, but there was definitely an imbalance 22 of serious asthma events seen even (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 180 1 premarketing. 2 clinical trial meta-analyses that have been 3 done, which, as I said, was to improve sample 4 size especially for the pediatric age group. 5 And then finally, use of LABAs with 6 concomitant ICS is currently very widespread, 7 but evidence that the ICS nullifies the 8 LABA-related risks is lacking. 9 And I'll stop there. 10 And we looked at some of the Thank you. (Applause.) 11 DR. SWENSON: Thank you. 12 Dr. Levenson now will take up the 13 story further here, with a bit more analysis 14 of these meta-analyses. 15 DR. LEVENSON: Hello. I'm Mark 16 Levenson. 17 the Quantitative Safety and Pharmacoepidemiology 18 Group in the Office of Biostatistics in CDER. 19 Today, I'm going to be talking about a 20 FDA-conducted meta-analysis on long-acting beta 21 agonists and serious asthma-related events. 22 (202) 464-2400 I'm a statistical safety reviewer in Here is the outline for my talk. Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 181 1 First, I'm going to present the goals and 2 objectives of the meta-analysis, then 3 describe the methods used in the 4 meta-analysis, including beta sources, 5 endpoints, conclusion criteria, comparisons 6 in subgroups, the statistical methods that we 7 used. 8 9 Then I'll present the results, starting with trial and patient summaries, 10 event summaries, the meta-analysis results, 11 results for subgroups. 12 the limitations of the meta-analysis and 13 compare it to other meta-analyses. 14 finally, I'll present some concluding 15 statements. 16 Then I'll talk about And The goal of this meta-analysis that 17 we conducted was to provide risk information 18 for a risk/benefit assessment of LABAs. 19 benefits in this risk/benefit assessment were 20 to come from action packages of approved 21 drugs. 22 in what drugs and doses we use in the (202) 464-2400 The So this has an important consequences Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 182 1 meta-analyses. 2 balance the risk against approved drugs, only 3 drugs and doses approved in the United States 4 were considered in this meta-analysis. 5 Because we're going to Here are the four LABAs that are in 6 the meta-analysis. These are the LABAs 7 approved for the treatment of asthma in the 8 United States. 9 about them, but I'll emphasize a few points You've already heard a lot 10 here. Advair and Serevent are 11 salmeterol-containing -- the LABA is 12 salmeterol, and Foradil and Symbicort, the 13 LABA is formoterol. 14 combination products. 15 gluticasone, and Symbicort has the ICS 16 budesonide. 17 Advair and Symbicort are Advair has ICS I should say throughout my 18 presentation when I refer to Advair or 19 Symbicort, I mean the combination product in 20 a single inhaler, not the two constituent 21 products that may be delivered in separate 22 inhalers. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 183 1 You also notice and you also heard 2 that the age which these drugs are approved 3 for differs. 4 Advair and Serevent are approved 5 from age 4 and up, and Foradil is approved 6 from age 5 and up. 7 from age 12 and up. 8 9 Symbicort is approved The first objective of the meta-analysis was to examine if LABAs are 10 associated with increased risk of the 11 following serious events: 12 death, asthma-related intubation, and 13 asthma-related hospitalization. 14 objective was to examine the risk by various 15 subgroups, including inhaled corticosteroid 16 use and demographics, particularly age, sex, 17 and race. 18 Asthma-related The second Now, I move on to the methods that 19 were used in the meta-analysis. 20 sources -- FDA requested patient-level and 21 trial-level data on LABAs from sponsors of 22 LABAs. (202) 464-2400 The data Only trials where the study Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 184 1 indication was asthma was requested. 2 only parallel placebo and/or active 3 controlled trials with or without ICS were 4 requested. 5 crossover trials was also requested. 6 And Actually, first period of The sponsors were instructed to 7 identify and adjudicate serious asthma 8 events, including the events I mentioned on 9 the last slide: the asthma death, intubation, 10 and hospitalization. 11 We defined these 12 endpoints -- again, asthma-related 13 death -- asthma-related death or intubation. 14 Because of the way the data was collected by 15 one of the sponsors, we couldn't consider 16 intubation separately from death. 17 asthma-related hospitalization and an 18 asthma-composite endpoint that we defined as 19 the death, intubation, or hospitalization. 20 And only events which occurred during the 21 protocol-blinded treatment period were 22 included in the analysis. (202) 464-2400 Beta Court Reporting www.betareporting.com We had (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 185 1 For the first comparison, the 2 primary comparison, we compared LABA 3 treatment versus no LABA treatment. 4 LABA includes ICS, short-acting beta 5 agonists, other non-LABA treatments, placebo, 6 or combinations of the above treatments. 7 subsets of that primary analysis, we looked 8 at two additional comparisons. 9 LABA with randomized defined ICS versus Here, no As The first is 10 randomized defined ICS. 11 the specific ICS and the dose must be the 12 same in both comparison arms within a trial. 13 And the final comparison is basically the 14 complement to that, the LABA without 15 randomized ICS versus no LABA. 16 In this comparison, Here are the inclusion criteria for 17 the analysis on the trial level. 18 only considered study indications of asthma, 19 and randomized blinded parallel trials or the 20 first period of a crossover trial. 21 trials -- this is an important inclusion 22 criteria -- the trial must have had both a (202) 464-2400 Beta Court Reporting www.betareporting.com Again, we The (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 186 1 LABA and a no LABA arm, which we refer to as 2 an internal control. 3 internal controls, different background rates 4 across trial may distort the overall findings 5 of the meta-analysis. 6 important criteria in our meta-analysis. 7 Without trials with So this is an The LABA arm may have had 8 additional therapy if the non-LABA arm has 9 the same therapy, including dosing. The 10 nominal blinded treatment duration had to be 11 at least seven days. 12 least 20 subjects in both the LABA and no 13 LABA arms, and subject level data had to be 14 available. 15 There had to be at Now, on the subject level, the 16 inclusion criteria was the assigned dose for 17 a subject had to be approved for the 18 treatment of asthma. 19 not part of the criteria. 20 subject had to be in the approved age range 21 for the drug. 22 (202) 464-2400 The delivery device was And the age of the Here are the subgroups we Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 187 1 considered for demographics. 2 four age groups. 3 and race of the subject and the location of 4 the subject's site, whether it was in the 5 United States or not in the United States. 6 We considered baseline ICS use, and whether 7 the nominal trial duration was less than 12 8 weeks of blinded treatment or greater than 9 equal to 12 weeks of blinded treatment. 10 We considered We considered the gender The primary analysis method was a 11 Manzel-Handel (?) risk difference. 12 stratified method. 13 control. 14 is a fixed-effect method in that it assumes 15 the common risk difference across trials. 16 The details of this were the uni-analysis was 17 the subject and the primary display was the 18 risk difference and 95 percent confidence 19 interval. 20 This is a It maintains within trial It uses trials with no events and We performed several sensitivity 21 analyses, including the influence of certain 22 trials. (202) 464-2400 First, the exclusion of SMART. Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 188 1 SMART was part of the primary analysis. 2 also examined the inclusion of SNS. 3 not part of the primary analysis because 4 subject-level data was not available. 5 considered the statistical properties of the 6 primary method using the exact method for an 7 odds ratio, and we examined trial 8 heterogeneity with a generalized linear mix 9 model with a random treatment effect. 10 We SNS was We Now I move on to the results. 11 First, trial and patient summaries. 12 were 110 trials that met the inclusion 13 criteria. 14 subjects were roughly divided between LABA 15 and the no LABA groups. 16 for 72 percent of the overall subjects. 17 Advair accounted for 22 percent. 18 Symbicort each supplied relatively few 19 subjects for the meta-analysis. 20 3,765 and Symbicort 1,270, which represents 21 only 2 percent. 22 duration -- the mean was 172 days. (202) 464-2400 Roughly 61,000 subjects. There Those Serevent accounted Foradil and Foradil The nominal trial Beta Court Reporting www.betareporting.com The (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 189 1 medium was 196 days. 2 to 476 days, and 50 percent of the subjects 3 were between 84 and 196 days. 4 It ranged from 7 days The majority of the subjects, 5 77 percent, the age was between 18 and 64. 6 Six percent were between age 4 and 11, which 7 represents 35 -- 115 subjects; 11 percent 8 between 12 and 17; and 7 percent greater than 9 or equal to 65. The median age was 37 and 10 the range was from 4 to 100. 11 the subjects' race were White Caucasian at 12 72 percent; 4 percent were Asian; 11 percent 13 were Black or African-American; and 14 14 percent were other/unknown. 15 The majority of Looking at other demographics, 16 50 percent of the subjects were female; 17 69 percent of the subjects had U.S. sites; 18 and 94 percent of the subjects were in trials 19 with a nominal treatment duration of at least 20 12 weeks. 21 characteristics, 20 percent of the subjects 22 were in trials that required ICS at baseline; (202) 464-2400 For other baseline Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 190 1 50 percent of the subjects reported that they 2 used ICS as baseline. 3 In terms of the actual treatment 4 duration, not the nominal treatment duration, 5 the mean was 154 days, and there was little 6 difference between the no LABA and the LABA 7 group. 8 and there was little difference in treatment 9 duration between the comparison groups. Likewise, the medium was 169 days, 10 Here are some selective summaries 11 by drugs. 12 display differences among the drugs. 13 characteristics were generally the same. 14 Serevent had a large or medium treatment 15 duration of 193 days versus Advair, Foradil, 16 and Symbicort. 17 percentage of subjects from U.S. sites, 18 84 percent versus 27, 48, and 38. 19 These summaries are chosen to Other Serevent also had a higher Both Serevent and Foradil had lower 20 percentages of subjects in trials that 21 required ICS at baseline. 22 Serevent, and 16 for Foradil, versus 49 for (202) 464-2400 Ten percent for Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 191 1 Advair and 100 percent for Symbicort. 2 Now I'll talk about some 3 descriptive summaries of the four endpoints 4 that we defined. 5 asthma-related deaths -- 16 in the LABA 6 group, 4 in the no LABA group; 71 deaths or 7 intubations -- 44 in the LABA group, 27 in 8 the no LABA group; 666 9 hospitalizations -- 369 in the LABA group and 10 There were 20 299 in the no LABA group. 11 As you recall, as I noted earlier, 12 the asthma composite is made up of the death, 13 intubation, and hospitalization endpoint. 14 And you can see that the hospitalization is 15 driving the asthma composite endpoint here. 16 Also note that the percentages of subjects 17 where each of these events in the LABA group 18 is higher than the corresponding percentage 19 in the no LABA group for each of these 20 events. 21 22 Now I'm going to look at these event summaries for the four individual (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 192 1 drugs. First, Advair. There were no deaths 2 or intubations in Advair trials. 3 21 hospitalizations in the LABA group, and 20 4 in the no LABA group. 5 is responsible for all the deaths and 6 intubations in the meta-analysis, and the 7 percentages in the LABA group for all the 8 events are higher than the no LABA group. 9 Again, Foradil had no deaths or There were For Serevent, Serevent 10 intubations. 11 hospitalizations in the LABA group and 14 in 12 the no LABA group. 13 there was no deaths or intubations. 14 were six hospitalizations in the LABA group 15 and one in the no LABA group. 16 There were 18 asthma For Symbicort, again, There Now the results of the 17 meta-analysis. 18 endpoints -- asthma death, intubation, 19 hospitalization, composite. 20 risk difference estimates. 21 greater than zero implies that the LABAs are 22 associated with a greater rate of events (202) 464-2400 These are the four asthma I'm displaying A risk difference Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 193 1 relative to the no LABA group. 2 You can see from the estimates for 3 all four asthma endpoints, the risk 4 differences are greater than zero, implying 5 that the LABAs are associated with an 6 increased rate of events relative to the no 7 LABA group. 8 95 percent confidence intervals, none of them 9 contain zero. Also, when you look at the So the null value of a no 10 association is not contained in the 11 confidence intervals. 12 than zero on the lower side. 13 They're all greater This is a forest plot which I'm 14 going to show several of, so let me go 15 through the first one slowly. 16 remainder of the presentation, I'm going to 17 be concentrating on the asthma composite 18 endpoints, and I'm going to be showing the 19 asthma composite here by the four drugs. 20 For the The square plot symbol represents 21 the risk difference estimate, and the line 22 that extends beyond it is the confidence (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 194 1 interval, also displayed numerically over 2 here. 3 Here is the overall result across 4 all the drugs. For three of the four 5 drugs -- Serevent, Foradil, and 6 Symbicort -- the risk difference estimate was 7 greater than zero. 8 confidence interval did not contain zero. 9 And Advair, the risk difference was slightly For Serevent, the 10 negative. 11 differences between these four drugs was not 12 statistically significant, and that's that 13 p-value .307. 14 Based on a post hoc analysis, Now, it's a forest plot, but now 15 broken down by the comparison groups. 16 first comparison group is LABA without 17 randomized ICS versus no LABA. 18 The The second comparison group is LABA 19 with randomized ICS versus randomized ICS. 20 In both cases, the risk difference estimate 21 was greater than zero. 22 randomized ICS, the confidence interval did (202) 464-2400 For the LABA without Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 195 1 not contain the value of zero. 2 hoc analysis showed that the difference 3 between these two comparisons was not 4 statistically significant. 5 p-value .339. 6 Again, a post And that's that We examined whether the increased 7 association of events occurred -- at what 8 time period did that occur -- early in the 9 trial, late in the trial. And here, we 10 looked at that two ways. 11 table approach and one through a Kaplan-Meyer 12 approach. First I'll show the life table 13 estimate. Here, you have four periods of 14 time starting from randomization or starting 15 from initiation of treatment. 16 periods of 90 days from 0 to 89, 90 to 179, 17 up to a year. 18 longer -- six days longer -- to include the 19 remaining events that were observed. 20 One through a life Roughly four The last one is slightly For each of these periods -- well, 21 for three of these four periods, the hazard 22 of the asthma composites endpoint was higher (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 196 1 in the LABA group than the no LABA group. 2 Here, .11 versus .08. 3 time periods -- the first, second, and 4 fourth -- the hazard was higher in the LABA 5 group than the no LABA group. 6 So for three of the You could look at this in terms of 7 a Kaplan-Meyer incident curve. 8 is the LABA group. 9 no LABA group. 10 The top curve The bottom curve is the And it's over a period of about a 11 year. 12 the two curves are generally spreading apart, 13 so the increased rate of events in the LABA 14 group seem to occur throughout a year, which 15 is what we have data on. 16 And you can see throughout the year Some sensitivity analyses. This is 17 the asthma composite by drug risk difference 18 estimates, but excluding SMART. 19 was a Serevent trial, so excluding SMART only 20 affects the Serevent estimates. 21 SMART actually increased the risk difference 22 estimate for Serevent and increased the risk (202) 464-2400 Beta Court Reporting www.betareporting.com Now, SMART Excluding (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 197 1 difference estimate overall. 2 had a lower differential rate of 3 hospitalization than the other Serevent 4 trials. 5 SMART actually Other sensitivity analyses, 6 including the exact odds ratio which examines 7 statistical properties, and the generalized 8 linear mix model, which examined trial 9 heterogeneity, did not show any deviations 10 from the overall result. 11 Now I'll talk about subgroups. 12 First, age subgroups. Again, here's a forest 13 plot. 14 age subgroups -- 4 to 11, 12 to 17, 18 to 64, 15 and 65 and up. 16 forest plot is that there's a clear trend 17 that the lower age groups are at higher 18 relative risk compared to the older age 19 groups. The different rows represent the four What you can see by this There's a clear trend here. 20 And again, a post hoc analysis 21 actually did demonstrate that there is a 22 linear effect in this effect across age (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 198 1 groups. 2 And the p-value there was .081. I'm going to look at this effect by 3 the four individual drugs. 4 seen in three of the four drugs. 5 Advair. 6 Advair had the reverse situation, where the 7 oldest age group, 65 and up, had the largest 8 risk difference estimate. 9 It was actually First, It was not seen in Advair. In fact, For Serevent, the trend is very 10 similar to the overall results, which is not 11 surprising, because Serevent makes up a large 12 majority of the data. 13 the dataset is quite a bit smaller and the 14 estimates quite a bit more variable, I think 15 the trend is relatively clear, with the 16 youngest subjects at higher risk -- higher 17 relative risk compared to the older subjects. 18 For Foradil, although Symbicort, because it's not 19 approved for age 4 to 11, there's no subjects 20 in this first age group. 21 have subjects, but there were no events, so 22 we couldn't calculate a confidence interval, (202) 464-2400 The 65 and up did Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 199 1 but the estimate falls on zero. And for the 2 limited amount of data available for 3 Symbicort, the age trend might exist as well. 4 Here's the asthma composite by race 5 subgroups. The Black African-American 6 subgroup had the highest risk difference 7 estimate. 8 represents a significant amount of the 9 subjects in this meta-analysis, so perhaps This was seen in SMART, and SMART 10 it's not surprising. 11 the results without SMART -- excluding SMART. 12 And although the estimate for the Black 13 African-American subgroup is lower than with 14 SMART, it's still the highest among the 15 racial subgroups considered. 16 This next plot shows Here is the asthma composite by 17 gender subgroups. 18 somewhat larger risk difference estimate than 19 the male subgroup. 20 this is this was seen in each of the four 21 drugs -- this difference between genders. 22 For other subgroups, there were no noticeable (202) 464-2400 The female subgroup had a What's interesting about Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 200 1 differences among subgroups based on 2 location, nominal treatment, duration, and 3 baseline ICS use. 4 I'll now discuss the limitations of 5 this meta-analysis, and some comparison with 6 other meta-analyses. 7 trials using this meta-analyses were not 8 generally designed for the present purposes. 9 The endpoints that we're analyzing were not First of all, the 10 systematically collected. 11 identified and adjudicated post hoc. 12 Information on subjects in trials are limited 13 in the meta-analysis. 14 The events were We did not collect information on 15 dropout, so it's possible a differential 16 dropout rate might explain some differences. 17 However, as I showed you earlier, the 18 treatment durations between the LABA and the 19 no LABA group were very similar, so we don't 20 expect there's a large difference there. 21 22 We collected information on concomitant ICS use, whether it was used or (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 201 1 not, but we weren't available -- information 2 on the amount of ICS use or the specifics of 3 the ICS use was not available. 4 The SMART results were known prior 5 to the design of this meta-analysis, so it's 6 not surprising a lot of the results we see in 7 SMART are confirmed -- are seen in this 8 meta-analysis. 9 analyses that excluded SMART did not show any However, the sensitivity 10 major deviations without SMART. 11 there were many comparisons performed on this 12 meta-analysis, so there's exploratory and 13 multiple testing issues. 14 And again, Now I'm going to discuss three 15 other meta-analyses, mainly not to summarize 16 these meta-analyses, but to compare how they 17 compare to the FDA meta-analyses so you can 18 understand the similarities and differences. 19 In parentheses are the drug that 20 the meta-analyses considered. 21 correction to the slides that you were 22 distributed. (202) 464-2400 There's a In those slides, I had Advair Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 202 1 under Bateman, but it's really salmeterol 2 plus ICS. 3 product and separate products -- separate 4 inhalers. 5 Both the Advair combination So the Bateman considers salmeterol 6 plus ICS. Saul Peter discussed salmeterol 7 and formoterol, and Sears (?) was based on 8 formoterol. 9 Bateman meta-analysis. First I'm going to discuss the Again, the drugs 10 under evaluation were salmeterol plus ICS. 11 The comparison was salmeterol plus ICS versus 12 ICS. 13 since SMART was a placebo-controlled trial. 14 So SMART is not in the Bateman analysis. 15 This comparison does not include SMART The ICS did not have to be the same 16 drug or dose within a trial in this 17 comparison. 18 dose, but I don't know about the ICS. 19 the results for asthma hospitalization, the 20 risk difference was .2, with a confidence 21 interval from -.1 to 2.3 per 1,000 subjects. 22 When you compare that to the FDA (202) 464-2400 Salmeterol had to be an approved Beta Court Reporting www.betareporting.com And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 203 1 meta-analysis for asthma hospitalization, 2 only Advair, which is slightly more narrow, 3 the risk difference was -.15 minus 2 to 1.7. 4 The confidence intervals were actually very 5 qualitatively similar. 6 good agreement between the Bateman and the 7 FDA meta-analysis, at least when it comes to 8 Advair. 9 So there's actually a Now, the Saul Peter meta-analysis. 10 They considered the drug formoterol and 11 salmeterol. 12 placebo. 13 Non-approved dosages were included. 14 The comparison was LABA versus This would include SMART. The results for asthma 15 hospitalization of LABA versus placebo in 16 terms of an odds ratio was 2.6, with a 17 confidence interval from 1.6 to 4.3. 18 believe SMART is actually included in this 19 outcome. 20 I don't And this compares -- it's 21 relatively similar to the comparable FDA 22 results for asthma hospitalization of the (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 204 1 LABA versus no LABA, which is 1.3, with a 2 confidence interval from 1.1 to 1.5. 3 slightly lower here, but it's actually 4 relatively qualitatively similar. 5 It's Finally, the Sears 6 meta-analysis -- this is based on the 7 AstraZeneca safety data -- considered the 8 drug formoterol. 9 formoterol with and without ICS versus ICS The comparison was 10 and other no LABA treatments. 11 pulled together, where with and without ICS 12 was combined. 13 non-LABA control, and non-approved dosages 14 were included in the Sears analysis. 15 All this was Trials may not have had a Let me -- since the Sears analysis 16 differs so much in terms of number of 17 subjects than the FDA analysis, I'm going to 18 explain how you get from roughly 60,000 19 subjects in the Sears analysis to the 1,270 20 for the FDA analysis of Symbicort. 21 all, we did not compare trials -- specific 22 trials -- so the details are not known. (202) 464-2400 Beta Court Reporting www.betareporting.com First of I (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 205 1 just know that 34,000 subjects were submitted 2 to the FDA based on our inclusion criteria. 3 Thirty thousand of the Sears 4 analysis probably came from trials with only 5 LABA arms -- no non-LABA comparisons. 6 Sears paper, these are described chiefly as 7 Symbicort versus Symbicort trials. 8 9 In the So that brings us mainly down to the 34,000 that were submitted to FDA. From 10 those 34,000, 31,000 were removed because 11 they were unapproved dose or product. 12 that brings us down almost all the way now. 13 256 subjects had no comparison in the 14 comparisons that we defined. 15 had an age less than 12, so they're not 16 approved for the drug and were excluded based 17 on the FDA inclusion criteria. 18 subjects had to be removed because of data 19 limitation in how we asked for the data. 20 They came from a single trial and there were 21 no events in that trial. 22 (202) 464-2400 So 807 subjects And 362 Finally, the conclusions. Beta Court Reporting www.betareporting.com Overall, (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 206 1 there was an effect of LABAs on each of the 2 asthma-related endpoints -- asthma-related 3 death, asthma-related death or intubation, 4 asthma-related hospitalization, and asthma 5 composite, which is made up of the death, 6 intubation, and hospitalization. 7 For the asthma composites endpoint, 8 the risk difference was 2.8, with a 9 confidence interval from 1.1 to 4.5 per 1,000 10 subjects. 11 adverse signals, but the sample size varied 12 substantially among the drugs. 13 Three of the four drugs had Differences among the drugs were 14 not statistically demonstrated. 15 differences may be related to trial or 16 subject differences. 17 And observed LABA with defined ICS appears to 18 have less adverse effect; however, the 19 difference was not statistically 20 demonstrated. 21 22 The effect was dominated by Advair, which supplied most of the data for this (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 207 1 comparison. 2 with the youngest at greatest risk. 3 talked about the limitations of the 4 meta-analysis, including the ascertainment of 5 events, the limitations on information 6 available, and its exploratory nature. 7 There is an apparent age effect And I And finally, there was an agreement 8 or at least resolution among the three 9 meta-analyses that I discussed. 10 Thank you. 11 DR. SWENSON: Thank you, Dr. Levenson. 12 It's now 12:15, and so we're right on schedule. 13 I know there's more questions and they'll just 14 have to be retained, and we'll have chances 15 later today and tomorrow. 16 So we'll meet back again at 1:15. 17 And just a reminder to the Committee members 18 that discussions of the subject matter at 19 hand should not occur during lunch. 20 should be within this forum. It 21 (Whereupon, at approximately 22 12:15 p.m., a luncheon recess was (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 208 1 taken.) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 209 1 A F T E R N O O N S E S S I O N 2 (1:15 p.m.) 3 DR. SWENSON: We'll resume, but 4 slightly departing from the schedule, because 5 there's still a number of questions that have 6 been percolating. 7 minutes with questions directed to our last two 8 speakers. 9 then we'll go ahead and get started. We'll spend about the next 10 So if you'll identify yourselves, 10 I think I see Dr. Goldstein there. 11 We'll start with you. 12 DR. GOLDSTEIN: 13 Is Dr. Levenson back yet? 14 15 Is Dr. Levenson back? DR. SWENSON: Dr. Levenson? 16 I don't know. He's coming. DR. GOLDSTEIN: Thank you. I have two 17 quick questions. 18 meta-analysis study design, was the study design 19 done prospectively? 20 inclusion and exclusion criteria and the 21 statistical analysis plan, et cetera, proscribed 22 prior to requesting data from the various (202) 464-2400 One is that in the In other words, were the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 210 1 companies? 2 And then my second question was 3 that you mentioned an effect both in age, 4 gender, and race. 5 effect as well? 6 a 4-year-old black female, was that analyzed 7 to see if that was the ultimate risk? 8 9 Was that a composite In other words, if you were DR. LEVENSON: The first question about the prespecified plan, the data requests 10 were initially made in March, which specified 11 some inclusion criteria. 12 analysis plan was not done more until the June 13 time frame, before we actually got the data. 14 the request was out there before the analysis 15 plan was finalized -- but before we got the 16 data, the plan was finalized. 17 deviations from the plan that are noted in the 18 briefing package. 19 everything we wanted to look at because of 20 limitations in data. 21 22 The final statistical So There were some We weren't able to look at You know, there were a lot of missing values for certain variables. (202) 464-2400 Beta Court Reporting www.betareporting.com We (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 211 1 wanted to look more at baseline asthma 2 control. 3 The second question about perhaps 4 an interactive effect between the risk 5 factors, no, we did not look at that. 6 didn't look at interactions between the 7 demographic variables. 8 DR. SWENSON: 9 DR. WOLFE: We Dr. Wolfe? In your presentation, you 10 showed a slide on the median duration of the 11 treatment. 12 median duration for Advair was barely more than 13 half of the median duration for Serevent. 14 then on the Kaplan-Meyer curve, you show very 15 nicely that as a function of time out there, the 16 risk gets progressively more. 17 may be about twice as much at a couple hundred 18 days than it is at 100 days. 19 And it was interesting that the And It looks like it So the question is, was it possible 20 to do any kind of analysis with Advair since 21 one explanation could be that since the 22 median duration of treatment was shorter, (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 212 1 that that in and of itself independent of the 2 drug could have made it look better? 3 DR. LEVENSON: Yes. I actually did do 4 these curves by drugs, but I do not have them 5 available. 6 that some of this data towards the longer time 7 points may be more dominated by Serevent is 8 probably a correct observation. 9 the information handy, and I don't recall the 10 DR. WOLFE: it. 13 14 I'd like to see it. I should be able to get those to you. DR. WOLFE: It's certainly a confounder based on your Kaplan-Meyer analysis. 17 The time. 18 DR. LEVENSON: 19 DR. WOLFE: 20 DR. SWENSON: 21 DR. HENNESSY: 22 I do not have If you could maybe send DR. LEVENSON: 15 16 The observation results. 11 12 But you're right. Yes. Thank you. Dr. Hennessy? I have a question for Dr. Mosholder but I don't see him in the room. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 213 1 Is he here? 2 SPEAKER: No, he -- 3 DR. HENNESSY: Okay. 4 back, I have a question for him. 5 DR. SWENSON: 6 DR. NEWMAN: When he comes Dr. Newman is next. Yes, I had one for 7 Dr. Mosholder as well, but I do have a follow-on 8 question for Dr. Levenson. 9 know you had written in what we received that In terms of -- I 10 you did not have access to the baseline control 11 data and the baseline severity data. 12 wondering if you can just elaborate on how 13 that -- how you consider that in terms of 14 limitations of the work that you're presented. 15 And I'm also curious why you wouldn't have that 16 information, because I would think that 17 information exists. 18 DR. LEVENSON: I'm We had some of that 19 information, but when we do these meta-analyses, 20 we have to come up with a somewhat unified data 21 request that will cover hundreds of trials. 22 it's hard to find characteristics that are (202) 464-2400 Beta Court Reporting www.betareporting.com So (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 214 1 uniformly collected across all these trials. 2 we tried to collect some, but it turned out 3 there were a lot of missing information. 4 So In terms of baseline asthma 5 control, I can only comment on what I read. 6 It could be -- I mean, the editorial on SMART 7 could be a factor why results differ across 8 different trials. 9 SMART unique in that it's different. And I think that does make It's 10 real world. 11 relevant, but it's unique. 12 asthma control is something I think that 13 would be of interest, and is a limitation 14 that we did not have more information on it 15 and were able to analyze it. 16 It may be more relevant or less DR. SWENSON: I see Dr. Mosholder is 17 back up. 18 you have a question for him? 19 So the baseline I think we had -- Dr. Hennessy, did DR. HENNESSY: I do. Dr. Mosholder, 20 on your slide -- 33 I think it is -- you show 21 that only 11 percent of long-acting beta use is 22 consistently together with an inhaled (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 215 1 corticosteroid, but your slide 27 appears to 2 show that of all the long-acting beta agonist 3 use, the vast majority of it is the combination 4 product Advair. 5 things can both be true. 6 DR. MOSHOLDER: And I'm wondering how those two Yes, the data from the 7 managed care organization was looking at the 8 question of single entity LABA products, mainly 9 Serevent and Foradil, and how many were 10 co-prescribed -- or how many patients got 11 concomitant prescriptions for an ICS as a 12 separate device. 13 answer your -- 14 So I think that's -- does that DR. HENNESSY: Sure. So if I'm 15 interpreting correctly, the very low lines at 16 the bottom of the single agent use, relatively 17 little of that is in combination with an inhaled 18 corticosteroid, but if you look at all the 19 long-acting beta use, it looks like the majority 20 of it is in combination with an inhaled 21 corticosteroid and the combination product. 22 (202) 464-2400 Is that right? Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 216 1 DR. MOSHOLDER: Yes, by virtue of the 2 fact that the inhaler itself is a combination; 3 therefore, of the total universe of LABA use, 4 most is coming with an inhaled corticosteroid. 5 But if one drills down on use of the single 6 entity inhalers, we found only about half were 7 being prescribed with a concomitant separate ICS 8 inhaler. 9 DR. HENNESSY: 10 11 DR. SWENSON: Thank you. We have time for just a few more. 12 Dr. Zito? 13 DR. ZITO: Yes, I have a question for 14 Dr. Mosholder. 15 around the limitations of observational research 16 studies. 17 that we are not taking full advantage of 18 community-based information on practice 19 patterns, for the reason that trial data often 20 do not generalize to the populations that we are 21 seeking to bring to treatment. 22 do a reasonable job of benefit to risk (202) 464-2400 I appreciate your presentation But at the same time, I would wonder And if we are to Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 217 1 assessment for future use, I would think that we 2 would like to take advantage of the racial 3 ethnic information that's available in the 4 Medicaid population, which represents 40 percent 5 of the children now -- the pediatric 6 population -- if we include the S-CHIP kids, for 7 example. 8 9 And in addition, that we are now able methodologically to look across time so 10 that we can see the length of duration of the 11 use of monotherapy versus the combination, 12 which is a big question here. 13 finally, that we could begin to think about 14 looking at subgroups of kids who have other 15 problems that would suggest that they are 16 using -- that will tell us that they're using 17 medications that increase the risk burden for 18 severe adverse events. 19 DR. MOSHOLDER: And then Well, I guess by way 20 of reply, I'd just say when I looked at the 21 observational studies, I was really focusing on 22 their value in determining the risk of -- the (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 218 1 intrinsic risk of the product. 2 the methodological issues, for example, it 3 appeared that LABAs were going to a somewhat 4 different patient population in practice than 5 comparative drugs. 6 studies to account for that. 7 And because of And it was very hard in the One way that people have attempted 8 to adjust, say, for asthma severity is to 9 look at asthma hospitalization. Well, the 10 problem is, as we've seen, if asthma 11 hospitalization is itself associated with 12 LABA use, then the jargon is you're adjusting 13 for something that's in the causal pathway, 14 and that you'll get maybe a bias. 15 So because of all those 16 limitations, just for the purpose of looking 17 at the intrinsic risk of the products 18 relative to comparative groups, that was the 19 nature of the limitations. 20 21 DR. SWENSON: We have Dr. Martinez next. 22 (202) 464-2400 DR. MARTINEZ: Thanks. Beta Court Reporting www.betareporting.com This is a (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 219 1 question for Dr. Levenson based on 2 Dr. Mosholder's presentation. 3 he proposed to us some interesting 4 interpretation of both SMART and SNS, which is 5 that the effects seen for asthma death and 6 intubation is much stronger than the effects 7 seen for asthma hospitalizations. 8 understood correctly what he was implying was 9 that perhaps there's something specific about In his slide 17, If I 10 the individuals who die of asthma or need to be 11 intubated. 12 attacks with respect to those who have less 13 severe manifestations. In other words, have very severe 14 And I think also what he was trying 15 to imply was that if the explanation that has 16 been given for this effect, which is that 17 individuals were taking separate canisters of 18 these two medicines, stopped taking the 19 inhaled corticosteroids when they're taking 20 the canister that they feel is effective for 21 their disease, which in this case would be 22 LABAs, then why would the effect be only seen (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 220 1 for asthma deaths and not for asthma 2 hospitalizations? 3 increase in asthma hospitalizations. 4 that respect, my question for Dr. Levenson 5 is, did you see the same trend when you 6 analyzed the data as a whole? 7 words, for the effect to be stronger with 8 asthma deaths and intubations than when they 9 were with hospitalizations? You would also see an So in In other And therefore, 10 if that is true, would the compound outcome 11 that you have selected underestimated some of 12 the most severe effects? 13 DR. LEVENSON: In terms of the risk 14 difference, the largest effect was seen in 15 hospitalization. 16 of odds ratio, the story may be different. 17 I don't have that information available now. 18 can get you that. 19 20 DR. MARTINEZ: And I Because also, the number of events is much larger. 21 22 But that, of course, in terms DR. LEVENSON: Right. Yes. So I'll have to look at some notes to see what the trend (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 221 1 is in terms of like a relative measure. 2 DR. SWENSON: Dr. Schoenfeld? 3 DR. SCHOENFELD: I just wanted to just 4 sort of quantify this. 5 calculations of this that the number needed to 6 harm for patients taking the combination 7 of -- you know, taking both an ICS and a LABA 8 is -- the number needed to harm is about 4,000. 9 Is that right? 10 My sort of brief That's what I think your estimate was. 11 SPEAKER: Yes. 12 DR. SCHOENFELD: But that's per year. 13 And the other estimate seemed like about the 14 same. Am I close there? 15 16 DR. LEMANSKE: Or did I-Which presentation are you referring to? 17 DR. SCHOENFELD: Well, both 18 presentations. 19 I think that was the number. 20 then you would multiply -- it would be roughly 1 21 in 4,000. 22 estimates but -- and you have to multiply by two (202) 464-2400 Your -- the first presentation, It was .25 and In your presentation, you had several Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 222 1 because most of the patients only were treated a 2 half a year. 3 4,000. 4 5 So it would be something like 1 in Maybe you didn't calculate that. DR. LEMANSKE: I didn't calculate number needed to harm in my analysis. 6 DR. SCHOENFELD: Am I right in the 7 first presentation, that that would be about 1 8 in 4,000 for people taking concomitant ICS? 9 DR. LEMANSKE: People taking -- oh, 10 actually, I don't think I did the number needed 11 to harm for that. 12 1.5. 13 from that odds ratio. 14 about whether the conquering group combined the 15 two trials. 16 That was the odds ratio of I didn't derive the number needed to harm I think you're talking DR. SCHOENFELD: I'm just trying to 17 get a notion of what the risk estimate is for 18 that subgroup. 19 DR. LEMANSKE: If you take the 1.5 20 point estimate, it's 50 percent increase, which 21 you want to derive the number needed to harm 22 from that, but I haven't done that. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 223 1 DR. SWENSON: Well, I regret to say 2 that we have to keep on schedule here. 3 apologize for the number of questions I know are 4 out there. 5 6 I But we'll move on now. Dr. McMahon from the FDA will speak. 7 DR. McMAHON: Good afternoon. I would 8 like to thank the Committee for their 9 participation and input on this important public 10 health matter: the benefits and risks of 11 long-acting beta-2 agonists for the asthma 12 indication. 13 My name is Ann McMahon, and I will 14 be giving an overview of benefit/risk issues 15 before the Committee on the subject of LABAs 16 from a perspective in CDER's Office of 17 Surveillance and Epidemiology. 18 acting director of the Division of 19 Pharmacovigilance II and have been working on 20 post-marketing surveillance at the FDA since 21 2002. 22 (202) 464-2400 I am the I'm a pediatrician. My purpose in this brief Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 224 1 presentation is to discuss the risks of 2 long-acting beta-2 agonists for the asthma 3 indication; to discuss the benefits of LABA 4 use for the asthma indication; and to pose 5 questions of benefit versus risk. 6 I will start by giving a brief 7 background of LABAs for asthma indication, 8 followed by a presentation of the highlights 9 of the FDA meta-analysis aimed at summarizing 10 the risks. 11 by the FDA in drug approval will be 12 discussed, highlighting documented benefits 13 of the drugs. 14 Then the pivotal studies reviewed Thereafter, there will be a 15 discussion of weighing the risk and benefits 16 of LABA use for the asthma indication as a 17 whole and in population subsets. 18 be followed by conclusions, recommendations, 19 and a brief presentation of the questions for 20 the Advisory Committee, although those will 21 be in your packet and won't be read at this 22 presentation. (202) 464-2400 Beta Court Reporting www.betareporting.com This will (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 225 1 This slide briefly outlines the 2 approved long-acting beta-2 agonists for the 3 asthma indication in the United States. 4 information was reviewed by Dr. Seymour this 5 morning, and I will refer you to her slides. 6 7 I think that's probably my Blackberry. 8 9 This Try not to forget it; okay? The Serevent Nationwide Surveillance study, or SNS from the United 10 Kingdom, and the Serevent Multicenter Asthma 11 Research Trial, or SMART from the United 12 States, both were randomized controlled 13 trials and were discussed at some length this 14 morning. 15 This slide simply is a reminder of 16 some of the key outcomes. 17 column farthest to the right, in both trials, 18 there was an elevated relative risk in the 19 three to four range of deaths from asthma in 20 salmeterol-treated patients as compared to 21 controls. 22 (202) 464-2400 Notice in the OSE reviewed a number of Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 226 1 meta-analyses, summarizing a variety of 2 outcomes with LABA use for the asthma 3 indication. 4 those trials designated a mean or median 5 trial duration, and all of those had a trial 6 duration of at least three months. 7 three are shown in this table. 8 9 We notice that only a subset of Those There are many others shown in the similar table in the OSE review which you 10 have in your background package. Two of the 11 three in this table included children and 12 adults and compared formoterol and/or 13 salmeterol to placebo and/or salbutamol. 14 Those are the first two, by Kates and Saul 15 Peter as primary authors. 16 Bateman et al. compared Advair or Salmeterol 17 plus ICS with ICS. The third by 18 Note that where available in the 19 righthand column, the odds ratio of asthma 20 exacerbation was higher, though not 21 statistically significantly so, in children 22 than in the overall population. (202) 464-2400 Beta Court Reporting www.betareporting.com Also, the (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 227 1 odds ratio for asthma exacerbation with 2 Advair overlapped one, as we've been hearing 3 about this morning, but the upper end of the 4 confidence interval approached two. 5 Therefore, this subset of meta-analyses 6 underlined the risk of salmeterol and 7 formoterol as single agents, and bring out 8 the following questions. 9 higher risk than adults for severe Are children at 10 asthma-related events with LABA use? And 11 two, what role, if any, in risk mitigation is 12 played by inhaled corticosteroids? 13 I will now review brief highlights 14 of the FDA meta-analysis of the data related 15 to risks of LABA use, performed and presented 16 by Dr. Mark Levenson this 17 morning -- presented this morning, that is. 18 As was outlined by Dr. Levenson, 19 the meta-analysis of data submitted by 20 sponsors was performed by the FDA. 21 were approximately 60,000 patients included 22 in the meta-analysis, and approximately half (202) 464-2400 Beta Court Reporting www.betareporting.com There (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 228 1 of those received a LABA. 2 the patients received Serevent; many of them 3 from the SMART study. 4 included were adults, though there were 5 approximately 10,000 children and 6 adolescents, and approximately 4,000 older 7 individuals included. 8 The majority of Most of the patients The risk difference, as you heard 9 in Dr. Levenson's talk, is essentially the 10 attributable risk, and was measured by the 11 incidence of the adverse event in the group 12 that received a LABA minus the incidence of 13 the adverse event in the comparison group. 14 risk difference of zero, shown by the dotted 15 line in the middle -- a risk difference of 16 zero, indicated by the dotted line in the 17 middle, shows risk neutrality. 18 A Overall, combining data for all the 19 drugs, the youngest patients showed the 20 highest risk difference, and this risk 21 difference decreased with increasing age. 22 fact, the FDA meta-analysis showed a (202) 464-2400 Beta Court Reporting www.betareporting.com In (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 229 1 significant age trend in the endpoint of 2 asthma composite index, which included, as 3 Dr. Levenson said, asthma-related deaths, 4 asthma-related intubations, and 5 asthma-related hospitalizations, but was 6 driven by asthma-related hospitalizations. 7 It showed a significant age trend in the 8 asthma composite index, with a p-value of 9 0.018, which was mostly driven, as I said, by 10 asthma hospitalization. 11 This slide shows the risk 12 differences for the asthma composite index 13 for the drugs that include salmeterol as 14 their active LABA component. 15 Advair, which has an inhaled corticosteroid 16 as the other active ingredient, has a risk 17 difference which was essentially zero. 18 differs from Serevent, which was 19 significantly elevated -- had a significantly 20 elevated risk difference. 21 pediatric data for Serevent do show an 22 elevated risk difference compared to the data (202) 464-2400 In this case, This Note also that the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 230 1 in older individuals; whereas for Advair in a 2 small subset of 400 or so individuals over 65 3 years of age, there was actually an increase 4 in the risk difference. 5 This slide shows the two drugs with 6 formoterol LABA component. Note that there 7 are no data for children less than 12 years 8 of age for Symbicort in this analysis. 9 the other hand, there were less than 500 On 10 children and adolescents in the meta-analysis 11 that received Foradil. 12 there was an elevated risk difference for 13 patients receiving both Foradil and 14 Symbicort, although the confidence intervals 15 did not -- the confidence intervals both did 16 overlap zero. 17 Notice that overall, I will now briefly review some 18 highlights of the efficacy endpoints in 19 pivotal trials of U.S.-approved LABAs for the 20 asthma indication. 21 about each of the four main LABAs approved 22 for the asthma indication, and will divide my (202) 464-2400 I will talk separately Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 231 1 comments into those related to pediatric 2 versus adult data. 3 meant for reference. 4 from the pivotal trials highlighting the 5 efficacy of Serevent. 6 a significant increase in peak expiratory 7 flow rate, and the secondary endpoints show 8 some clinical improvement with Serevent 9 compared to albuterol, depending on baseline 10 This slide is busy and is It is a sample of data Notice that there was and other details. 11 Notice that, for example, 0.6 less 12 puffs of rescue medication per day might or 13 might not have a large clinical impact. 14 that in these studies, the percent of asthma 15 exacerbations were no different in the 16 Serevent versus albuterol group. 17 Note Again, this table has many figures 18 and is for your reference, but please note 19 that although the spirometry measures were 20 significantly improved for Serevent compared 21 to placebo in 4- to 11-year-old children, the 22 secondary endpoints in children show little (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 232 1 difference in the treatment group compared to 2 placebo. 3 post-treatment, there was a 10 percent net 4 increase in asthma exacerbations in the 5 Serevent group compared to placebo. 6 Note that during two weeks In summary, in patients over 11 7 years of age comparing salmeterol with 8 albuterol, the primary spirometric endpoints 9 were met. The change in percent nights with 10 no symptoms and change in percent days with 11 no symptoms favored salmeterol by 10 to 12 20 percent, but all other symptoms measured 13 in secondary endpoints showed little 14 difference between salmeterol and albuterol, 15 or favored albuterol in the case of change in 16 percent with asthma exacerbations. 17 In children 4- to 11-years of age 18 receiving salmeterol compared to placebo, the 19 primary spirometric endpoints were met, and 20 there was little difference in the asthma 21 symptoms score, but other secondary endpoints 22 measured showed slightly less asthma (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 233 1 symptomatology in salmeterol-treated 2 patients. 3 So on to Advair. This slide again 4 is busy and is meant for reference. It is a 5 sample of data from a pivotal trial 6 highlighting the efficacy of Advair Diskus in 7 individuals over 11 years of age. 8 that there was a significant increase in 9 FEV1, and the secondary endpoints show some Notice 10 clinical improvement with Advair compared to 11 fluticasone, depending on baseline and other 12 details. 13 awakening-free nights most likely does not 14 have a large clinical impact from a baseline 15 of 90 percent. 16 However, 4 percent more These are summary points regarding 17 Advair efficacy. 18 11 years of age comparing Advair and 19 fluticasone, all three primary experimetric 20 endpoints were met. 21 12-hour under the curve FEV1 was higher at 22 one week than at 12 weeks. (202) 464-2400 In individuals greater than Notably, the change in Notably, for the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 234 1 secondary endpoints, both change in percent 2 symptom-free days and change in percent days 3 without rescue medication showed 4 approximately 20 percent increase, which, 5 depending on baseline and other clinical 6 parameters, may have a large or less large 7 impact. 8 data on Advair collected for children that 9 were presented by Dr. Seymour this morning, In children, there were clinical 10 and data for approval in children were 11 extrapolated from adults. 12 This is another slide to give you a 13 sense of the clinical data in some of the 14 pivotal trials for -- in this case, for 15 formoterol compared to albuterol. 16 endpoint was met, but the secondary endpoints 17 do not appear to correlate with the FEV1 18 endpoint and extent of improvement. 19 The FEV1 This is an example of a pivotal 20 trial of a LABA -- in this case 21 Foradil -- that was performed -- the trial 22 performed in children 5- to 12-years of age. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 235 1 Here, Foradil is compared to albuterol, and 2 showed spirometric significant improvement in 3 subjects receiving Foradil compared to 4 albuterol. 5 the baseline data presented, most if not all 6 of the secondary outcome measures amount to 7 arguably minimal clinical improvement in 8 children. 9 This slide shows that in light of Review of the pivotal trials for 10 the Foradil products in patients greater than 11 12 years of age showed that the two FEV1 12 endpoints were met, and FEV1 was 13 significantly greater in patients receiving 14 Foradil compared to albuterol. 15 endpoints of percent nights awakened and 16 percent nights using rescue medication were 17 both lower in Foradil-treated patients than 18 albuterol-treated patients, though the 19 decrease was less than 20 percent. The secondary 20 In all other secondary endpoints, 21 there was a small difference in the Foradil 22 recipients compared to albuterol. (202) 464-2400 Beta Court Reporting www.betareporting.com In (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 236 1 children 5 to 12 years of age, the two 2 primary spirometric endpoints again were met, 3 and none of the secondary endpoints, however, 4 showed large difference in the Foradil 5 compared to the albuterol group or placebo. 6 Another busy slide again for 7 reference, this time showing an example of 8 data from a pivotal trial for Symbicort in 9 individuals over 11 years of age. Here, 10 comparing Symbicort with Budesonide. 11 again that this LABA is an effective 12 bronchodilator. 13 values, the improvements in the various 14 secondary endpoints are likely not to be very 15 clinically significant. 16 Notice But given the baseline For example, a difference of 17 10 percent or a total of 40 percent rescue 18 medication-free days compared to a baseline 19 of 30 percent may or may not be terribly 20 clinically significant. 21 22 Here, I will summarize the pivotal trials comparing Symbicort and Budesonide. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 237 1 In patients greater than 11 years of age, the 2 primary spirometric endpoints were met. 3 There was very little, if any, difference 4 between Symbicort and Budesonide in the 5 secondary endpoints, including change in 6 percent nocturnal awakening, but there was 7 some improvement over Budesonide in change in 8 percent rescue-free days, a net of 10 percent 9 increase, and change in percent with 10 exacerbations, a decrease of 14 percent. 11 There was no data reviewed on individuals 12 less than 12 years of age. 13 This slide summarizes overall 14 comments from efficacy endpoints for pivotal 15 trials of U.S.-approved LABAs for asthma. 16 Some measure of forced expiratory volume 1, 17 or FEV1, was the primary endpoint in all of 18 the U.S. studies. 19 reviewed, there was a statistically 20 significant increase in some spirometric 21 measure compared to either placebo or 22 albuterol or an individual component product. (202) 464-2400 In all of the studies Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 238 1 Some degree of response to a 2 short-acting beta agonist was required for 3 enrollment. 4 And in adults, the largest clinical 5 differences between LABA and comparator were 6 approximately 1.5 -- sometimes a little more, 7 sometimes a little less -- of rescue medicine 8 per day, or 15 to 20 percent more 9 symptom-free days in the LABA versus 10 That's a parenthetical comment. comparator group. 11 Although the pediatric studies met 12 the spirometric endpoints, there was little 13 improvement over comparator in the secondary 14 endpoints for children less than 12 years of 15 age. 16 separately; thus, an analysis of this group 17 of patients with respect to LABA benefits is 18 not possible here. 19 Adolescents were not assessed So some considerations for risk 20 versus benefit. 21 is spirometric endpoints were consistently 22 met in the LABA pivotal trials. (202) 464-2400 The first comment to be made Beta Court Reporting www.betareporting.com Data (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 239 1 reviewed did not suggest large benefits on 2 secondary endpoints, especially in children 3 less than 12 years of age. 4 not powered to study were not powered, as has 5 been mentioned this morning, to study more 6 quality of life endpoints. 7 difficult to assess age-specific benefits 8 because the numbers were small in the 9 pediatric and elderly populations, although I 10 Efficacy trials And it was haven't mentioned elderly much. 11 With respect to risk, the age of 12 the recipient may have had a large impact on 13 the risk/benefit ratio. 14 was a trend of decreasing age -- decreasing 15 risk with increasing age -- that was present 16 overall in the FDA meta-analysis. 17 reviewed do not suggest large secondary 18 benefits, especially in children less than 12 19 years of age. 20 distinctions between LABA drugs was not 21 possible, at least in part due to small 22 sample sizes. (202) 464-2400 Specifically, there Data Additionally, drawing Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 240 1 So our conclusions based on these 2 observations are that these data substantiate 3 and quantify an increased risk overall in 4 serious asthma-related events with the use of 5 LABAs compared with non-LABA therapy for 6 asthma. 7 risk of the adverse effects of LABAs than 8 adults. 9 improving spirometric outcomes. Children appear to be at greater LABAs have a consistent effect of And data 10 reviewed do not suggest large secondary 11 benefits, especially in children less than 12 12 years of age. 13 Correlation of spirometric 14 endpoints with clinical benefits was not 15 clear. 16 The LABA Safety Review Team within 17 the Office of Surveillance and Epidemiology 18 shared most conclusions and recommendations, 19 and agreed that there is a signal for 20 asthma-associated morbidity and mortality in 21 adults; that that signal is largely from the 22 SMART and SNS studies, which as you know were (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 241 1 performed using a single entity LABA 2 salmeterol. 3 It is unclear what role inhaled 4 corticosteroids play in mitigating 5 LABA-associated risk; that there is an age 6 effect, such that the risk of 7 age-attributable asthma-associated 8 hospitalization anyway is increased in 9 children; and the extent of increased risk of 10 LABA use for other demographic subsets among 11 adults is less clear. 12 Therefore, the team unanimously 13 recommends withdrawing the asthma indication 14 from LABAs for individuals less than 18 years 15 of age, and removing the asthma indication 16 contraindicating the use of single entity 17 LABAs Serevent and Foradil for all ages. 18 There was a point of disagreement 19 among OSE reviewers on recommendations, which 20 was whether to recommend removing the asthma 21 indication for Advair and Symbicort for 22 adults. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 242 1 I'd like to thank the individuals 2 who have helped with this presentation and 3 analysis. 4 through the questions, but they are in your 5 handouts for your reference and we'll be 6 talking about them tomorrow. As I mentioned, I will not read 7 That concludes my comments. 8 DR. SWENSON: 9 10 Thank you, Dr. McMahon. Next, Dr. David Graham will continue this discussion. 11 DR. GRAHAM: Good afternoon. I'm 12 David Graham with the Office of Surveillance and 13 Epidemiology, and during the next half hour or 14 so, I would like to talk about public health 15 considerations in weighing benefits and risks, 16 and particularly as it pertains to LABAs. 17 This is important because FDA is a 18 public health agency. 19 Public Health Service. 20 makes are at a population level. 21 be directed at the entire population and not 22 at the level of the individual patient. (202) 464-2400 We are part of the Decisions that FDA Beta Court Reporting www.betareporting.com They are to And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 243 1 so that's why it's necessary to now move from 2 the level of clinical trials and individual 3 patients up to a little higher altitude, 4 where we're going to look at the population 5 and what is it we're doing with benefits and 6 risks. 7 This is a rough outline of the 8 talk. What I'd like to talk about first is 9 that in the approval process and in the 10 assessment of safety post-marketing, that 11 there is an asymmetric handling of benefit 12 and risk, and that that needs to be 13 considered when weighing population effects. 14 I also will discuss how the uncertainty 15 created by low power to exclude high levels 16 of asthma mortality risk with the 17 non-Serevent LABAs is critically important to 18 the consideration of this committee. 19 I'll use Foradil pivotal trials as 20 an example to illustrate the pitfalls of low 21 statistical power. 22 evidence addressing the presence or absence (202) 464-2400 I'll then summarize the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 244 1 of important clinical benefits with LABAs, 2 and then contrast the decision-making 3 paradigms in controlled trials at the 4 population level in the public health arena. 5 We'll then move to focus on Advair, 6 its health benefits and its mortality risks. 7 Why? 8 presentation earlier, Advair dominates the 9 market -- the LABA market now. 10 Well, as you saw from Dr. Mosholder's It has the largest population exposure. 11 It has the most data to work with. 12 I'll show that the problems of asymmetric 13 power exist, and raise the question for 14 consideration by this committee, which is, is 15 the absence of proof proof of absence? 16 the absence of proof of harm -- the 17 definitive proof of harm -- proof that there 18 isn't harm? 19 be made as the basis of that? Is And what kind of decision should 20 The current drug approval paradigm, 21 and the way safety is handled and efficacy is 22 handled post-marketing, is that pivotal (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 245 1 trials are powered to show efficacy. 2 a prespecified value of what qualifies a 3 success. 4 states the drug doesn't work. 5 that can be tested and rejected. 6 Type I error, which is the probability of 7 finding a difference if there isn't 8 one -- given that there isn't one -- at 9 5 percent. 10 positives. 11 There's There's a null hypothesis that It's something We set the And we do that to minimize false This is sometimes referred to as 12 the regulator's risk, because the regulator 13 doesn't want to falsely approve a drug that 14 doesn't work. 15 protective of the public, because there's an 16 opportunity cost if we approve a drug that 17 doesn't work and there's another drug out 18 there that does and I happen to take the drug 19 that doesn't work. 20 opportunity cost. 21 22 And that's maximally Well, you can see the Now, there's also a prespecified Type II error, the beta, which is the (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 246 1 probability of failing to find a difference 2 given there is one. 3 to minimize false negatives. 4 sometimes referred to as the company's risk, 5 because the company wants to be sure to find 6 an effect if there is one, because that'll be 7 the basis for approval. 8 power to find the clinical effect they're 9 looking for is in the range of 80 to 10 The purpose of this is And this is So typically, the 90 percent. 11 Now, statistical tests are then 12 used to reject the null hypothesis, either 13 p-values or the lower bound of the confidence 14 interval, excluding the null. 15 illustrated in this slide, where here we have 16 a p-value that's below .05. 17 interval does not include the null effect 18 level. 19 efficacy is demonstrated, and FDA-approved. 20 If the confidence interval or the p-value 21 don't show significance, the null hypothesis 22 is not rejected. (202) 464-2400 And that's The confidence The null hypothesis is rejected, So then the conclusion Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 247 1 really is the drug doesn't have efficacy, at 2 least operationally, because the drug isn't 3 approved. 4 Now, under the current, paradigm we 5 have asymmetric handling of safety and 6 efficacy. 7 you know, are not powered to demonstrate 8 safety or exclude harm. 9 that in a little more detail. The pivotal trials, as I'm sure Well, let's examine There's no 10 prespecified value for safety, so what is a 11 safe drug? 12 that I have a safe drug? 13 serious harm is acceptable in exchange for 14 the prespecified value of efficacy? 15 the pivotal trials exclude this prespecified 16 level of risk -- of acceptable harm, if you 17 will. 18 do that, and so those things aren't tested. 19 The null hypothesis is the drug is safe 20 rather than the drug is harmful. 21 there an a priori presumption of safety? 22 think -- I don't have an answer for that, but (202) 464-2400 What qualifies as success to show What level of And do Well, the studies aren't designed to Beta Court Reporting www.betareporting.com Why is I (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 248 1 it is an inherent bias in the way the drug 2 approval process works. 3 Now, a question I would ask is why 4 aren't drugs presumed a priori to carry some 5 specified level of unacceptable risk which I 6 can then test? 7 level of unacceptable risk, then at least I 8 sort of have capped the level of risk that 9 I'm dealing with, and I can say as far as And if I can reject that 10 we're able to say, the benefits that we know 11 about exclude the risks that might exist 12 because we capped the risk. 13 greater than a particular level. 14 not done. 15 drugs; it hasn't been done with any drug that 16 FDA has ever approved. 17 We know it's no But that's It hasn't been done with these Now, further, when we come to the 18 post-marketing setting and we're talking 19 about harm, FDA has insisted on a standard of 20 definitive proof in order to reject its 21 presumption of safety. 22 that was used in the past recently with (202) 464-2400 This was the standard Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 249 1 Avandia and Actos. 2 minute, it was also the standard used in the 3 original approval of SNS -- of Serevent in 4 light of the SNS study. 5 And as I'll show you in a Now, in looking at safety, there's 6 no attention to the Type II error to minimize 7 false negative conclusions of safety. 8 showed before, the typical beta to show harm 9 is very high, which means that the power in a As we 10 typical clinical trial to even cap the risk 11 at a particular level is below 50 percent; 12 whereas the power to show the effect they 13 want is 80, or 90, or 95 percent. 14 So again, then there is a focus on 15 p-values and the confidence intervals and we 16 get the same situation. 17 but now, it's applied to safety rather than 18 efficacy. 19 looking at a harmful effect. 20 p-value is less than .05 and the confidence 21 interval doesn't include the null effect, we 22 will say that we have definitive proof of (202) 464-2400 It's the same slide, And in this situation here, we're And if the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 250 1 harm and the risk is real. 2 hand the p-value is greater than .05 and that 3 confidence interval happens to go down below 4 the null effect level, the null hypothesis of 5 safety isn't rejected. 6 If, on the other Harm hasn't been proven 7 definitively, and so then the risk 8 operationally isn't real because the drug 9 continues more or less to be used the way it 10 was before. 11 So we end up then with a situation 12 that's a little bit difficult to deal with. 13 We have a double standard. 14 of approval, but especially post-marketing, 15 it becomes more important to address this in 16 a realistic way, especially where we know 17 that one of the drugs we're talking about is 18 beyond doubt lethal to some patients -- that 19 it carries an increased mortality risk. 20 in a population level as I'll show you in a 21 few minutes, that could be quite substantial. 22 (202) 464-2400 And at the time And So this slide summarizes then when Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 251 1 we compare efficacy and safety, are we 2 evidence-based or is the deck really stacked 3 in favor of the drug and against finding 4 harm? 5 the drug doesn't work and it can be tested; 6 an a priori presumption that the drug is safe 7 and that can't be tested. 8 to the randomized trials, and you'll find out 9 that where we're talking about the We have the a priori presumption that You go down here 10 probabilities of false positive and false 11 negative conclusion, that they're the 12 opposite. 13 I now want to focus your attention 14 on the SNS study, which as Dr. Seymour told 15 us earlier, was available to FDA at the time 16 that Serevent was approved. 17 original BJM article. 18 Relative risk of 3. 19 is what the confidence interval was on 20 that: .7 to 28. 21 and said it's not statistically significant. 22 And so if you look at the label when Serevent (202) 464-2400 This is from the I've blown it up here. The p-value was.1. This FDA looked at the p-value Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 252 1 was approved, there's no mention whatsoever 2 of asthma mortality. 3 They didn't look at a potential for 4 a 28-fold increase in asthma mortality, which 5 these data are consistent with. 6 would argue is that for an evidence-based 7 rational benefit/risk weighing, one has to 8 have a benefit that you know about that would 9 exceed the risks that you can be certain And what I 10 about. 11 certain that the risks are not as high as 28. 12 In this situation, we can't be So a few more things about SNS and 13 SMART. 14 speakers before. 15 hasn't been mentioned was that although 16 inhaled corticosteroid therapy wasn't 17 monitored during the studies, it was captured 18 at the start of the study. 19 trial, about 70 percent of patients were 20 taking inhaled corticosteroids at the time of 21 randomization. 22 it was about 30 or 40 percent. (202) 464-2400 They've been talked about by several One point I think that And in the SNS In the SMART study, I think Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 253 1 What I want people to see in this 2 slide is that although these studies were 3 done 10 years apart, opposite ends of the 4 Atlantic Ocean, that they found almost 5 identical attributable risks. 6 10,000 per year -- and this is for asthma 7 mortality -- with a combined relative risk of 8 about 4, with an upper bound of about 13. 9 About 17 per Now, if you take that attributable 10 risk, you can use the drug use data that we 11 have and you can estimate what the number of 12 asthma deaths could be based on that 13 attributable risk in the population. 14 Dr. Mosholder earlier showed us the 15 prescription use of Serevent and Advair and 16 all the other LABAs. 17 1994 to 2004, if we just focus on Serevent, 18 where we have the SMS and the SMART data, 19 there's over 5,000 deaths that are likely to 20 have occurred in the United States as a 21 result of that. 22 10 percent of all asthma deaths during that (202) 464-2400 And so for the years That would account for about Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 254 1 11-year period. 2 Now, if that same risk applies to 3 all LABAs used in that time period, then 4 we're talking about 9,000 total deaths. 5 for the period 2005 to 2007, the last three 6 years where we have data on drug use, the 7 Serevent number of deaths that are 8 attributable to it are low. 9 in the drug use data, the use of Serevent as Now, But as you saw 10 a single entity product has fallen quite 11 dramatically, but it's been replaced by the 12 use of Advair, which is three times as high 13 as what the highest use of Serevent ever was. 14 And so if that same level of risk applies to 15 Advair, for example, then in three years, we 16 have half as many deaths as occurred in 11 17 years with all the other LABAs. 18 public health perspective, this is an 19 important point to deal with and to be 20 cognizant of. So from a 21 Now, I said I would go to Foradil 22 to show what is the blind side -- the blind (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 255 1 spots that result from low statistical power. 2 This summarizes all the studies that were in 3 the pivotal trials for the approval of 4 Foradil in adults and children. 5 And what you can see is that the 6 number of patients, the number combined, the 7 number of cases here -- now, this is serious 8 asthma exacerbations -- the number of person 9 years -- look at that. Two hundred person 10 years of use for a drug that we expect to be 11 used potentially by hundreds of thousands of 12 children. 13 a large population exposure. 14 That's a small amount of data for The attributable risk that we get 15 is 66 events per 10,000. 16 confidence interval on that risk. 17 be as high as 570, which means the 18 attributable risk could be -- the number 19 needed to harm could be as high as 17 per 20 year. 21 per year to produce one extra serious asthma 22 exacerbation. (202) 464-2400 But look at the It could What that means is treating 17 people The data in the entire pivotal Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 256 1 trial -- the entire program that FDA bases 2 assumption on couldn't exclude a number 3 needed to harm of 17. 4 was made that the benefits exceed the risks. 5 And yet the decision So now within the Foradil NDA for 6 the dose that's marketed, there was no asthma 7 mortality. 8 ago, this was pointed out by other colleagues 9 in the Office of New Drugs, that the fact At an internal meeting two years 10 that there were no deaths was proof that the 11 drug was safe. 12 Well, I want to describe to you how 13 in such a small database of 235 person years, 14 zero cases of death really is not reassuring 15 at all. 16 mortality rate as high as 1 per 6 per 100 per 17 year. 18 We have the inability to exclude a So that isn't anything. Now, how 19 does Foradil -- the approval of Foradil 20 compare to what we know about Serevent 21 salmeterol? 22 bound for salmeterol could be 34 per 10,000 (202) 464-2400 Well, for incidence, the upper Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 257 1 per year. Because of the small database for 2 Foradil, it could be as high as 160. 3 down here to number needed to harm, it's 4 about -- the upper bound is about 370. 5 it could be as low as 70. 6 that there's really a lot that we don't know 7 about the drug. 8 it's safe or to even conclude that it's safe 9 really I think is premature because there's Going Here, So we could see And for us to presume that 10 no actual evidence to conclude that the drug 11 is safe. 12 There's absence of evidence. Okay, now what I'd like to do is 13 look at the single entity products -- the 14 single ingredient products and then the 15 combination products, and summarize a couple 16 of measures abstracted from the action 17 packages. 18 presented by Dr. Seymour earlier, and some of 19 it was presented by Dr. McMahon. 20 Some of this information was When Dr. Seymour presented these 21 data, she presented them in comparison -- the 22 active drug in comparison to placebo. (202) 464-2400 Beta Court Reporting www.betareporting.com What I (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 258 1 will be presenting here is, wherever 2 possible, to the active therapy that was also 3 included in that treatment, because a natural 4 question would be if LABAs are -- if that 5 indication is removed from these drugs for 6 any of the age groups or all of them, well, 7 what's the consequence of that? 8 is the only place where we really have 9 head-to-head information. 10 And so this Dr. Lemanske gave a really 11 excellent presentation this morning, but one 12 thing I'd like to point out to people there 13 is that in none of the studies was there 14 actually a comparison with the LABAs and say 15 scheduled albuterol were there in the same 16 trial. 17 various combinations of do I add steroid? 18 I not add steroid? 19 steroids with the TREXA study at the end 20 where he's going to do all these different 21 combinations of albuterol and steroids. But 22 there isn't one where they're together. And (202) 464-2400 You've got all these LABA studies and Do I've got these albuterol Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 259 1 so this is the only place where we have that 2 information. 3 So now it's indisputable that the 4 LABAs are bronchodilators. 5 dispute that. 6 And so we don't What we do dispute is whether that 7 bronchodilation translates into meaningful 8 clinical benefits that justify the mortality 9 risk that we believe these drugs confer at a 10 population level. 11 pivotal trials for Serevent for the two 12 formulations -- the one that's not marketed 13 anymore and the one that is because of the 14 inhalant -- looking at asthma quality of 15 life, asthma score, and rescue medication 16 use. 17 studies, at that point maybe asthma quality 18 of life wasn't being used, but asthma score 19 was. 20 the medical officer's review for these drugs 21 changes in the asthma score of up to like 22 about .4 going down were characterized as (202) 464-2400 So now this takes all four And you can see for the Serevent And we had these minor differences. Beta Court Reporting www.betareporting.com In (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 260 1 being small. 2 here -- this would be a small at best to 3 really not much of anything. 4 So we're not talking about For asthma rescue therapy, one 5 place in one of the reviews the medical 6 officer said .6 rescue therapies a day change 7 isn't clinically meaningful. 8 a statement about what does 1.7 mean. 9 if we look at Foradil, where we have asthma I couldn't find Now, 10 quality of life, which is a validated 11 instrument and where as previously described 12 you need to have a change of .5 or more to 13 have an effect that's actually clinically 14 meaningful -- that actually means you've done 15 something that the patient actually is 16 experiencing what they would call a clinical 17 benefit. 18 don't -- they don't make it. 19 the .5. 20 Here, they were just cited by the medical 21 officer as not being different. 22 (202) 464-2400 You see, for Foradil these just They don't make The asthma scores were minimal. I'd like to point out that in one Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 261 1 of -- in this 2303, where there were 80 2 patients exposed to Foradil, one of them was 3 hospitalized with life-threatening 4 exacerbation required intubation. 5 anecdote in a way, but the fact is that in a 6 study as small as this, you had this happen. 7 And remember, when these clinical trials are 8 done, you're not selecting patients who are 9 sort of at the worst end. You don't expect 10 this. 11 intubation in clinical trials. 12 So it's an You don't expect patients to require Now, if we look at the combination 13 products, Advair and Symbicort, the set-up of 14 the slide is the same. 15 it doesn't -- neither Advair nor Symbicort 16 make the asthma quality of life. 17 is comparing the combination products to 18 inhaled corticosteroid alone. 19 Symbicort, there's no change in rescue 20 therapy; no change in asthma score. 21 Advair, the data were presented as 18 percent 22 increase in symptom-free days here, (202) 464-2400 And you can see that Now, this Now, for Beta Court Reporting www.betareporting.com For (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 262 1 22 percent increase in rescue-free days here. 2 And the question is well, what does that 3 mean? 4 In this slide, what I've tried to 5 do is show like the correlation between 6 asthma quality of life, asthma score, rescue 7 use, and then what do these things actually 8 mean. 9 of life. And you can look across asthma quality None of them are meaningful from 10 the clinical perspective, so they're all 11 pretty much the same. 12 here, we don't have them for Advair but 13 they're not meaningful there. 14 use isn't important. 15 The asthma scores And the rescue If we look here, these are the 16 numbers from the previous slide. 17 over here for one of the Symbicort studies. 18 19 percent rescue-free days. 19 much the same as 21 or 22. 20 But look That's pretty 11 percent symptom-free days. 21 Well, that's not very far different from 18 22 or 16 percent. (202) 464-2400 So the question in my mind Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 263 1 then is what do these things mean that isn't 2 being captured here? 3 validated instrument and it examines like 4 four different axes that relate to asthma 5 control, sense of well-being, functioning, 6 environmental stimuli, need for rescue 7 control, sleep disturbances at night. 8 captures a whole array of symptomatology, and 9 it puts it down in a single value that we can 10 Because this is a It look at. 11 Now we'll look at children. The 12 evidence in children -- from these data, it 13 doesn't look like the LABAs do a whole lot in 14 adults. 15 think Dr. Lemanske pointed that out as well 16 in one of his slides, where he pointed up and 17 he says it really raises a question why you'd 18 add a LABA to the ICS in the first place. 19 Look for Foradil first. 20 microscopic. 21 any difference. 22 difference. (202) 464-2400 They do even less in children. I I mean, these are And Advair, those don't make These don't make any There's no evidence of an actual Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 264 1 clinical benefit to the children who were 2 getting these drugs. 3 mortality effects that we have to talk about 4 as well. 5 But we have the Now, what's the level of certainty 6 about asthma mortality risk in combination 7 products? 8 to mention is that the questions that you'll 9 have tomorrow, you're asked to vote And at this point, what I'd like 10 separately on children up to the age of -- I 11 guess it's 11 or 12; and then adolescents, 12 which are like 12 to 17; and then older than 13 that. 14 it was important to carve out the adolescents 15 is because in the clinical trials that we've 16 talked about up until now, the adolescents 17 are lumped together with adults. 18 look back at Dr. Levenson's slides where he 19 summarizes the ends in these studies, you'll 20 see that the number of adolescents across all 21 of the LABAs is pretty small. 22 (202) 464-2400 And the reason why our office thought But if you And I'd just like to give an Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 265 1 illustration of one study that formed part of 2 the basis of approval for Foradil for adults 3 and adolescents. 4 adolescent. 5 arm, a 13-year-old. 6 seek approval for adults and adolescents. 7 that's why we want people to focus on 8 adolescents as well. 9 It was a study. It had one One adolescent in the active And the indication is we So Now, getting to this notion of 10 power to detect. 11 entire LABA database from Dr. Levenson's 12 review. 13 ratio. 14 have a background rate of risk of death. 15 how many times above that -- how many 16 multiples of that risk do I have evidence to 17 say I can exclude it? 18 is it's not until I get up to a 43-fold 19 increase in asthma mortality risk do I 20 actually have enough power -- comparable 21 power to the power the company had to show 22 that it met its FEV1 measure -- to say that (202) 464-2400 This is looking at the And what I have here is a hazard That's a relative risk. That's -- I And And what you can see Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 266 1 it's not that high. 2 The risk could be easily 25-fold or 3 15-fold or 10-fold increased. 4 shows you what the person years you would 5 need to exclude the risk. 6 remember from the previous slide, we only had 7 about 235 person years of Foradil use in 8 children. 9 And this just And if you So I think you get the picture. We have an absence of evidence, and 10 there should be no reassurance or assurance 11 on the part of anyone in this room that we 12 know anything about how bad or how high the 13 mortality risk could be. 14 evidence-based, my argument at least is then 15 we need to say what is it we know? 16 it we can reject? 17 what is it we can reject? 18 reject an acceptable level of risk, then 19 we've got a real problem, because the 20 benefits and the risks don't balance. 21 22 And if we're What is What is it we know and And if we can't Okay, now I want to talk about decision-making paradigms and public health. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 267 1 And this is a log scale, and it goes from 2 like 5 percent up to 100 percent. 3 these adjectives are sort of the terminology 4 that people might use in everyday life to 5 interpret that level of statistical 6 certainty. 7 weatherman when he says it's likelihood it's 8 going to rain today; do you bring an 9 umbrella? And then And it's kind of like the You know, if he says there's an 10 80 percent chance of rain, you bring an 11 umbrella. 12 Well, FDA and industry use this 13 level of evidence to establish efficacy and 14 harm. 15 problem they're dealing with, there's a 16 threshold for clinical and public health 17 significance that's important to consider and 18 upon which decisions actually have to be 19 made. 20 more balanced. 21 had I believe would be inversely related to 22 the magnitude, severity, or intensity of the (202) 464-2400 But in public health, depending on the And when you do this, then things are And the action level that you Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 268 1 benefit. 2 So if I'm talking about a very 3 large harm, I might accept that it's possibly 4 a very large harm if the benefits aren't 5 demonstratively very large. 6 a sliding scale. 7 So it's kind of Now, this stylized slide shows the 8 Y axis of the certainty going from very 9 likely to very unlikely, and the X axis being 10 basically the size or the magnitude of 11 whatever the effect is. 12 beneficial effect; it can be a harmful 13 effect. 14 It can be a And if we were to design the ideal 15 drug, we would have high certainty of a large 16 benefit, and equally high certainty that we 17 don't have substantial harm -- that harms are 18 small and we know that with high certainty, 19 and that the benefits are great and we know 20 that with high certainty. 21 ideal drug. 22 (202) 464-2400 That would be the In the next several slides, I'm Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 269 1 going to try to graphically describe, 2 integrating the data that I presented in 3 previous tables, summarizing the asthma 4 quality of life and the asthma scores for 5 these various LABAs into a diagram that tries 6 to at the same time show the benefits and 7 risks in a common place at a common time. 8 9 What I've done here is it's a similar X and Y axis to before, but what I've 10 done is I've reversed the direction of the 11 log. 12 And I did that -- in a normal log 13 scale as you go higher up, the marks get 14 closer and closer together. 15 What I did was flip it because I 16 wanted to magnify the places where we either 17 have the greatest certainty or we have the 18 greatest either positive or negative effect. 19 And so if you look back at the 20 previous tables for all of the LABAs compared 21 to existing therapies, the asthma quality of 22 life score and the asthma scores were all in (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 270 1 the small to trivial range for all of those 2 studies. 3 that FDA reviewers or that people who have 4 standardized instruments for asthma quality 5 of life thought were important or meaningful. 6 They did not meet the definitions The reason why the certainty goes 7 all over the board is because different drugs 8 within the LABA class had different amounts 9 of data that provide us with different levels 10 of certainty about what that small -- what 11 are the probabilities associated with that 12 very small benefit. 13 With the excess harms, on the other 14 hand, we know that from SMART and SNS that 15 we're talking about a relative risk of about 16 four and that that could translate into many, 17 many deaths. 18 This slide, I've added one 19 additional feature, which is this dotted line 20 which I've labeled the threshold of public 21 health importance. 22 curvilinear because of the log scale. (202) 464-2400 The line is not -- it's Beta Court Reporting www.betareporting.com And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 271 1 what I'm trying to convey here is that 2 there's some threshold -- this is just -- I 3 just put it there. 4 could be down there. 5 that would have to be informed by the 6 situation, by the benefits you're dealing 7 with, by the public health problem. 8 for teaching, heuristic purposes -- is that 9 there is some line beyond which if an effect It could be up here; it It's really something But it's 10 is a beneficial effect, it would be 11 considered important even if the certainty 12 about it was low. 13 So if I knew that there's a drug 14 that -- you know, cures asthma but I don't 15 have a lot of certainty yet because there 16 isn't a lot of data for it, well, that might 17 go over here. 18 threshold and be something that would be 19 important to consider if some other harms 20 came up that we needed to balance it against; 21 likewise, with the harms. 22 suggesting here is that the evidence we have (202) 464-2400 And it would cross the And what I'm Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 272 1 on harms for the LABAs and the evidence we 2 have on benefits are really on opposites 3 sides of that decision threshold for public 4 health importance. 5 Now, FDA, by approving the LABAs, 6 and the company by marketing the LABAs, have 7 implicitly stated -- or explicitly on the 8 part of FDA -- that the benefits exceed the 9 risks. So what they're saying basically in 10 terms of diagrams is that here's where the 11 benefits are for the LABAs, and they exceed 12 the harms in some type of calculus. 13 problem is, from anything I've seen today, we 14 don't have any proof that these are where the 15 benefits are, and certainly not from large 16 RCTs. 17 speculative at best. 18 And the So I would submit that that's Now, looking at our data, this 19 should have a minus sign there -- that should 20 be a -3, not a +3. 21 22 The question came up earlier that Advair within Dr. Levenson's review seemed to (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 273 1 have a lower point estimate for the 2 composite. 3 there was any risk for the composite. 4 one possibility is that the population that 5 was studied in the adverse studies is 6 different than the populations that maybe 7 were studied in other studies. 8 done over here is look at the control groups 9 for each of the studies, and we can see what 10 the incidence rate for the composite outcome 11 was per 10,000 person years. 12 is that the Advair control group was very 13 different than the groups for the other 14 products. 15 it real or is it Memorex? 16 Advair looked good in Dr. Levenson's analysis 17 an actual property of Advair, or is it a 18 property of the controls that were used in 19 the study? 20 In fact, there didn't look like Well, So what I've And what we see So that raises the question of is Is the fact that Now, what's important to note also 21 as well, and Dr. Mosholder pointed this point 22 in Dr. Levenson's review, is that there are (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 274 1 only 20 outcome events in the Advair group. 2 Well, if you look at the data for Serevent in 3 Dr. Levenson's review, there's about 15 or 20 4 hospitalizations per outcome event. 5 fact that we don't see one -- that we don't 6 see a death in the adverse study should 7 provide no reassurance either. 8 address that a little more next. 9 So the And I'll We talked about the asymmetry of 10 handling benefits and risks in pivotal trials 11 before. 12 that the benefits in the Advair trial to 13 identify a clinically important change in 14 asthma quality of life was 88 percent. 15 was right out of the biostatistics part of 16 the submission that the company sent in. 17 Symbicort, for their FEV1, they had 95 power. 18 This is the power to exclude a 10-fold 19 increase in mortality. 20 Advair, 10 percent for Symbicort. 21 increase in mortality. 22 40 percent, even 100-fold. (202) 464-2400 And what I want to summarize here is That For Eight percent for Fifty-fold 31 percent, So statistical Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 275 1 power. 2 craters. 3 has no craters? 4 asthma deaths with Advair because I have 5 8 percent power to show it, does that mean 6 there isn't? 7 conclude that. 8 presumption; that's not evidence-based. 9 Without a telescope, the moon has no Does that mean that the moon really The fact that I don't see I don't think that we can I think that that's a Now, given such low power, how 10 could we objectively conclude that either of 11 these drugs are safe, or that their clinical 12 benefits exceed their risks, assuming that 13 risks haven't been adequately or accurately 14 measured -- that we haven't even capped the 15 risk -- unless you're willing to say that 16 100-fold increase in asthma mortality is an 17 acceptable level of risk -- the benefits that 18 Advair confers. 19 have an easy solution to your dilemma. 20 If that's the case, then you Now, the level of certainty 21 regarding asthma mortality -- this is from 22 the entire Advair controlled database that (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 276 1 Dr. Levenson presented earlier. Okay, so not 2 just -- the previous slide was clinical 3 trials from the approval. 4 know, seven or eight years of additional 5 data, we have a larger database. 6 the power to detect the hazard ratio. 7 got enough power now to detect a hazard ratio 8 of 10. 9 Advair below 10. Now with, you And this is We've We can't exclude a hazard ratio for 10 Serevent confers something between 11 3 and 4. 12 30 and 40 percent power to deal with that. 13 We basically were flying blind. 14 You have somewhere probably between So now -- I said I would focus on 15 Advair specifically because it's the leading 16 LABA that's marketed -- and I'll show you now 17 the same diagram of the certainty and the 18 intensity of the effect. 19 benefits. 20 mortality risk. 21 the ADA power to conclude that Advair health 22 benefit is small and not clinically (202) 464-2400 Blue is clinical The dotted red lines are the This box here is based on Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 277 1 important, because that's the power that you 2 had for asthma quality of life. 3 Now, mortality risk, well, you 4 could place the box in one of two places 5 depending on how you want to view the 6 problem. 7 had low power to document the level of Advair 8 mortality because the studies were small. 9 But if you do that, then you're rewarding You could place it here because you 10 companies for not doing studies, because the 11 solution to this problem is do the smallest 12 studies you can get away with and then you 13 don't have an answer and you're guaranteed a 14 null result. 15 because there's no proof that Advair's 16 mortality risk is different than Serevent's. 17 You could put it up here because ICS hasn't 18 been proved to reduce LABA mortality -- at 19 least not in a very large prespecified RCT. 20 So this is the characteristics of You could put it up here 21 an ideal drug that I showed you earlier. 22 this is where I think we're dealing with (202) 464-2400 Beta Court Reporting www.betareporting.com And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 278 1 Advair. 2 opposite. 3 drug. 4 high certainty. 5 Serevent from Advair. 6 for Serevent with high certainty is large and 7 it's high. 8 9 It is the antithesis. It is the It i the mirror image of an ideal There's a small clinical benefit with We can't distinguish We know that the risk And there's no proof this risk isn't very large. So I would say that from 10 an evidence-based perspective, that the 11 evidence of benefit -- we know with high 12 certainty that at least the things we can 13 measure aren't good. 14 Maybe we need better measures in the clinical 15 trials. 16 presented with don't suggest that we're 17 getting a whole lot. 18 bullet. 19 It's a bronchodilator. But the things that we've been It's not the magic So the conclusions that our group 20 came to -- I'll go over those. 21 some of them as well. 22 entity LABA products should be withdrawn from (202) 464-2400 Ann went over We believe that single Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 279 1 the market. 2 that the clinical benefits are generally 3 small compared to albuterol. 4 the benefit is even less. 5 of the single ingredient salmeterol Serevent 6 or Foradil are used without ICS. 7 And the reasons for that are In children, Nearly 50 percent So we know with certainty that 8 Serevent confers a high mortality risk. 9 this is an engineering solution. And If we use a 10 combination product, then nobody who uses a 11 LABA could get it without an ICS. 12 risk mitigation strategy, if you believe that 13 ICS cures the problem, this solves that 14 problem. 15 So from a But here's the contradiction. ICS 16 hasn't been shown to reduce asthma 17 mortality -- LABA mortality. 18 rationale for concluding that ICS should be 19 used with LABAs? 20 our entire group agreed with -- the four of 21 us on the LABA OSC team. 22 that we also agreed with, which was that the (202) 464-2400 So what's the Now, that first conclusion The second is one Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 280 1 asthma indication for use of all combination 2 LABA products in children and adolescents 3 should be withdrawn. 4 it's been inadequately studied, the safety is 5 unproven, and there's no evidence -- there's 6 no proof that the benefits exceed the risk. 7 From my perspective, There's no basis to conclude that. 8 The benefits, if any, seem to be small 9 compared to placebo, because most of these 10 studies in children were done against 11 placebo. 12 substantially increased mortality risk. 13 we get back to this question. 14 proof of mortality is not the same as the 15 proof of absence of such risk. 16 exclude a 43-fold increase in asthma 17 mortality for these LABAs in children -- LABA 18 ICS products in children. 19 acceptable level of risk for the small amount 20 of benefit we've had, then leave the 21 indication there. 22 it isn't, we believe the indication should be (202) 464-2400 There's a high potential for And The absence of We can If that's an If it's not, and we think Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 281 1 pulled. 2 Now, the asthma indication for all 3 combination LABA products in adults should be 4 withdrawn. 5 this recommendation. 6 Dr. Busaber didn't. 7 have for it are that while it's clear that 8 LABA ICS is superior to ICS alone as a 9 bronchodilator, I don't know what that means Dr. Mosholder and I subscribe to Dr. McMahon and And the reasons that I 10 clinically. 11 translate into something that's been measured 12 that I can say is a clinical benefit that I 13 can measure against the harm as measurable as 14 mortality. 15 favorable data is available, there's no 16 evidence that asthma hospitalizations are 17 prevented. 18 As far as I can see, it doesn't Even for Advair, where the most The LABA component of these 19 products increases mortality, and we know 20 that. 21 evidence to say the risk isn't the same as it 22 is for Serevent. (202) 464-2400 With Advair, basically we have no Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 282 1 And then finally, additional 2 reasons why -- to support the reason why we 3 think it should be withdrawn in adults as 4 well. 5 approved indication. 6 an FDA endorsement or recommendation that a 7 product is safe and effective for use in a 8 specified disorder. 9 presented previously where we have small I point your attention here to the This is equivalent to Given what I've 10 proven clinical benefits, how can we justify 11 exposing millions to what we must conclude is 12 an extremely high risk of death? 13 We haven't rejected -- we haven't 14 even capped the risk at a level that's 15 probably reasonable to accept. 16 concluding that the benefits exceed the 17 risks. 18 And yet we're To me, this question could have 19 been answered many, many years ago if FDA had 20 insisted that the manufacturers of Advair go 21 out and do a really large SMART-like study 22 comparing the combination product to ICS (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 283 1 alone, to scheduled albuterol plus ICS, and 2 we'd have an answer today. 3 wasn't asked for by FDA. 4 apparently didn't volunteer to do it, and so 5 we don't have those data. 6 But that study The company So the question then is how can we 7 conclude these products are safe and 8 effective if the efficacy measures don't 9 predict clinical benefit. So at best, I 10 would say -- I would conclude that the LABA 11 ICS products have been inadequately and 12 insufficiently studied to be marketed, and 13 that their continued marketing is really in a 14 sense a natural experiment where we don't 15 have good data collection. 16 And then a couple of final 17 thoughts. 18 and the therapies that we have are not 19 adequate. 20 therapies, better therapies that work. 21 our desire to have better therapies shouldn't 22 trump making evidence-based decisions as a (202) 464-2400 Yes, asthma is a serious disease And we would love to have new Beta Court Reporting www.betareporting.com But (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 284 1 population -- and a population in a public 2 health perspective. 3 And what we have right now is we 4 don't have evidence, so it's all based on 5 presumption. 6 safe. 7 risks, but when you test it, it's like the 8 emperor with no clothes. 9 substance there. 10 We assume that the drug is We assume that the benefits exceed the There isn't So we have diamonds, we have zircons, and we have glass. 11 And I would say that zircon at LABA 12 ICS are closer to glass than diamonds in 13 terms of clinical benefits, but they command 14 diamond prices in terms of mortality risk. 15 Okay, thank you. 16 DR. SWENSON: 17 Well, thank you, Dr. Graham. 18 Well, at this point, we've heard 19 some -- we've heard two serious talks here, 20 and I think that we'll just break a little 21 bit from the schedule because I'm certain 22 there are a number of questions, and we'll (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 285 1 take a bit of time here to have questions for 2 these two recent speakers. 3 4 And so let's start with Dr. Brantly. 5 I see him first. DR. BRANTLY: So this is a question 6 for Dr. Graham. And again, I'd like to focus in 7 and get into public health and the epidemiology 8 of asthma, and I'd like to try to rectify that 9 with the data that you've presented. And that 10 is that, number one, you predict that there's 11 more asthma deaths based on the LABAs than there 12 actually are, number one. 13 And number two is, even with the 14 introduction of LABAs, there's actually been 15 a continued decrease in the deaths associated 16 with asthma. 17 So the numbers don't quite add up. 18 DR. GRAHAM: Well, a couple things. 19 One, the total deaths that I listed are the 20 total deaths over all the years. 21 about 4- or 5,000 asthma deaths a year, so my 22 numbers don't exceed the number of asthma deaths (202) 464-2400 Beta Court Reporting www.betareporting.com There are (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 286 1 that there are, A. 2 B, it is true that trend for 3 reported asthma death has declined over time, 4 but it's an ecologic association. 5 shown that that decline has anything to do 6 with LABA use. It's not done in a clinical 7 trial setting. And as an epidemiologist, you 8 know that's ecological. 9 Nobody has The other thing is that 50 percent 10 of that decline is due to the change in 11 classification coding from ICD-9 to ICD-10. 12 And so I think that ecologic data -- I mean, 13 it's in the first chapter of epidemiology 14 textbooks that you can be led astray. 15 16 DR. BRANTLY: Can I just follow that up? 17 I just want to say we have at least 18 one historical example of increase in asthma 19 deaths associated with the introduction of a 20 drug onto the market for treatment, and 21 that's isoproterenol, where there was 22 actually an increase in the number of asthma (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 287 1 deaths associated, clearly identifying that 2 drug as the -- as one of the -- something 3 that was clearly just cause and effect. 4 I mean -- 5 DR. GRAHAM: But here's my question 6 back to you. Then Serevent by itself doesn't 7 increase mortality? 8 are wrong and Serevent does not increase asthma 9 mortality, then yes, you could have your death Because if SNS and SMART 10 rates going down over time or you could fail to 11 see a spike. 12 very crude measure of what's going on. 13 But what you're talking about is a The coding for asthma mortality, as 14 you well know, is very inefficient, and very 15 imprecise and prone to lots of error. 16 unless you're prepared to say that Serevent, 17 that SNS and SMART, that those controlled 18 randomized trials are wrong and that Serevent 19 doesn't increase asthma mortality, I think 20 that your argument is lacking in strength, 21 and I disagree. 22 (202) 464-2400 DR. SWENSON: And so Dr. Goldstein. Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 288 1 DR. GOLDSTEIN: This is for Dr. Graham 2 as well. 3 that the current paradigm for drug development 4 is asymmetric, and that the approval process is 5 different for efficacy -- that is different in 6 terms of efficacy and safety. 7 the questions -- and I agree with your analysis 8 and the numbers that you showed. 9 I certainly don't disagree with you I think one of I think one of the -- the main 10 question that I have is that the current bar 11 that's set for efficacy for this drug 12 population, my interpretation is that it's 13 FEV1, and it's a hard number and a 14 physiologic process that can be measured. 15 DR. GRAHAM: Right. 16 DR. GOLDSTEIN: For other drugs in 17 other areas, including asthma, scoring systems 18 or surrogate markers have not been accepted. 19 The hard markers are needed. 20 blood pressure, survival, those things are 21 required. 22 designs a trial for a primary outcome variable, (202) 464-2400 FEV1, change in And therefore, when a drug company Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 289 1 they use these hard primary outcome variables to 2 power their study design. 3 To then go back and say that 4 efficacy was really not demonstrated because 5 even though they met the primary outcome 6 variable, the secondary outcome variables 7 were really not all that significant and they 8 are now more important than the primary 9 outcome variable, A, doesn't seem fair 10 because you can't change the rules after the 11 game is over. 12 And B, the study wasn't designed 13 with those secondary outcome variables as the 14 primary, so they really weren't powered to 15 show differences in those. 16 DR. GRAHAM: 17 DR. GOLDSTEIN: 18 Right. I just wanted your comment on that. 19 DR. GRAHAM: Well, everything that 20 you've said, I agree with, but what I would 21 point is the following: 22 difference between efficacy and effectiveness, (202) 464-2400 That there's a big Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 290 1 between efficacy and benefit. 2 other drugs. 3 We see that with If I had a statin -- I mean, you 4 know, we're pushing to look at clinical 5 outcomes. 6 but it doesn't translate into improved 7 cardiovascular survival, have I accomplished 8 anything? 9 efficacy, which is the standard that FDA If I lower your cholesterol level So there's a difference between 10 uses, and benefit, which when you're talking 11 about whether a drug should remain on the 12 market or not, there you're talking about 13 benefits and risks. 14 the most simplistic level, how many lives do 15 I save, how many lives do I take. 16 You're talking about, at And here, it's not changing the 17 rules. 18 situation has arisen. 19 certainty that LABAs increase asthma 20 mortality. 21 that? 22 to pay $10,000 for a piece of glass or (202) 464-2400 It's saying an extraordinary We now know with high Do the health benefits justify Is it justifiable? Does it make sense Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 291 1 $10,000 for a diamond? 2 well, free choice, and everybody should have 3 options and do what they can, but there are 4 certain exchanges that just aren't 5 reasonable. 6 regulatory agencies to set a bar about what's 7 acceptable. 8 9 Now, you could say, And I believe it's the role of But if people disagree, then what I would say is you have to be very explicit 10 about what you're doing. 11 data we have here, that would mean saying 12 that for children, an increase in asthma 13 mortality of 43-fold is dominated by the 14 health benefits that that drug confers to 15 children; and that a morality increase of 16 tenfold with Advair justifies what appears to 17 be small clinical benefits. 18 And based on the And if you're prepared to say that 19 and to enunciate it so that the world can see 20 and scratch their head and say does that make 21 sense to them, then I can sit down because 22 then I don't have an argument anymore. (202) 464-2400 Beta Court Reporting www.betareporting.com But (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 292 1 where I would place the balance and the 2 exchange is much lower than 43-fold increased 3 death in children or tenfold increase with 4 Advair. 5 DR. GOLDSTEIN: I appreciate that. 6 Quickly, I just would comment that because I 7 think the studies that we would ideally love to 8 see have not been done, that's why I asked 9 Dr. Lemanske earlier what would be the burden on 10 the patients if these drugs were removed? 11 specifically referring to children. 12 DR. GRAHAM: 13 DR. GOLDSTEIN: 14 DR. GRAHAM: And Right. It was just a comment. Right. Well, if I can 15 respond to that question, I think for children, 16 since they're not deriving much, if any, 17 clinical benefit, it's hard to imagine that 18 there would be a harm that would occur as a 19 result of withdrawing the indication. 20 anything, we're going to have fewer asthma 21 deaths in children, which I would think is a 22 good thing. (202) 464-2400 Beta Court Reporting www.betareporting.com If (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 293 1 In adults, I would argue the same 2 thing. 3 end of Dr. Lemanske's presentation, where he 4 was presenting data on use of albuterol and 5 ICS, and suggesting that albuterol and ICS 6 may actually -- even if it's used as a rescue 7 therapy, might actually be beneficial, that 8 it might have promise. 9 nobody's compared those therapies. 10 I would also argue that at the very The problem is We know that albuterol, scheduled 11 albuterol, is safer than Serevent. 12 us that. 13 is that the fact that we don't have those 14 studies would mean that I would go to the 15 thing that, based on the comparisons, looks 16 to be safer, which would be albuterol and 17 inhaled corticosteroids. 18 SNS tells And so the question in my mind is, And yes, there are going to be 19 individual patients that are in control, that 20 are controlled by LABAs, but the question is 21 do you expose the entire population to that 22 so that you can get the occasional patient (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 294 1 that responds that way? 2 way of identifying who's going to have that 3 miraculous response to Advair, just as we 4 have no way or predicting who's going to end 5 up in the cemetery because of Advair. 6 DR. SWENSON: 7 DR. KRAMER: Because we have no Dr. Kramer. This is not specifically 8 for David Graham, but for the FDA in general. 9 I'd actually like a response from both the 10 Pulmonary and Drugs Division as well as the 11 Office of Surveillance and Epidemiology. 12 I'd like to step up at a higher 13 level about what these three different 14 advisory committees are being asked to do 15 today. 16 concerned when I reviewed the material in 17 advance, and as the presentations have gone 18 on, I've gotten increasingly concerned. 19 And I've been growing -- I was My understanding -- first of all, 20 we've got a disease that in itself is a very 21 serious disease. 22 these two days, we've got patients and (202) 464-2400 And I think at the end of Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 295 1 practitioners waiting with bated breath for 2 what this group is going to do, which is why 3 I'm asking this question this way. 4 As I understand it, from as far 5 back as 2005, both the label and increasingly 6 guidelines have advised against monotherapy. 7 And in 2007, as I understand it, the 8 Pediatric Advisory Committee asked for a 9 follow-up update on the safety in children 10 and continued update on safety in general. 11 And yet my understanding of what 12 we've been presented and what's continuing to 13 be presented here is, first, a meta-analysis 14 that is heavily dominated by the old data on 15 monotherapy, and the quantitative estimates 16 heavily dominated by the monotherapy arms of 17 these trials. 18 And now, in the last presentation, 19 we're having a philosophical debate about the 20 basis for approving drugs in this country in 21 the context of three advisory committees 22 trying to decide on the proper treatment for (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 296 1 patients with asthma. And I think that this 2 is an impossible task that you're asking us 3 to do. 4 us some guidance. And I just wonder how you could give 5 I'd like FDA to be clear about what 6 new data are we being asked to look at that 7 wasn't present in 2007, that we're reviewing 8 now. 9 going to opine and decide different criteria And what are -- you know, are we really 10 for basis of approval of drugs at this 11 meeting? 12 (Applause) 13 DR. JENKINS: Okay. I'll be brave and 14 try to answer that. 15 fundamentally want this committee to advise us 16 on today and tomorrow is the question of do the 17 benefits of these products exceed their risk, 18 such that they should remain on the market? 19 if so, how should those risks be managed to try 20 to maximize benefit and minimize risk? 21 22 The issue that we And You say you have an impossible task. (202) 464-2400 We have the same task, so we're asking Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 297 1 you to give us advice. 2 data that you have in front of you is the FDA 3 meta-analysis that Dr. Levenson presented 4 earlier. 5 you to address. 6 I think the main new That's the main question we want We're not here to address comments 7 or concerns about the approval standard for 8 drugs. 9 by the way. 10 It's not only in the United States, That's the way drugs are approved around the world. 11 This should really focus on benefit 12 versus risk for LABAs and the LABA 13 combinations. 14 population benefits and the individual 15 patient benefits warrant keeping these 16 available given the known risk of serious 17 adverse reactions with these drugs? 18 Do you think that the DR. KRAMER: Then why didn't the FDA 19 present us with data to come to bear with the 20 current guidelines on how asthma's treated? 21 22 DR. JENKINS: Well, I think you're going to see some of that in Dr. Chowdhury's (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 298 1 presentation, which is coming up next. 2 haven't see all of the presentations and you 3 obviously haven't seen the sponsors' 4 presentations either, because they probably have 5 some viewpoints on this issue as well. 6 So you You know, one thing to keep in mind 7 is that the analysis that you just saw is 8 probably true for any chronic symptomatic 9 therapy that has a known serious risk. 10 probably true for NSAIDs, where you're 11 treating -- using them for symptom relief, 12 not disease modification. 13 expecting to cure osteoarthritis by using an 14 NSAID chronically, but you know they carry a 15 risk of -- serious risk of GI bleeding and 16 even potential death. 17 construct for any of those. 18 It's You're not You can make this Tylenol, you could make the same 19 construct for acetaminophen about the 20 benefits, and yet it's the number one cause 21 of drug-induced liver failure in the United 22 States every year. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 299 1 So you've got to look at what are 2 the benefits in your mind for the population 3 as a whole and individual patients? 4 the risks? 5 manageable? And if so, how should we do that 6 management? That's the fundamental question 7 that we're asking the committee to opine on. 8 9 What are Are those risks acceptable and I mean, we don't have a lot of disagreement, I don't think, within the FDA 10 that the risk is real. 11 you know, does the risk mean that these drugs 12 are unacceptable for marketing, or is that 13 risk acceptable in your mind given the 14 benefits that are accrued to patients 15 individually and collectively? 16 what we'd like to hear your opinion on. 17 The real question is, And that's You're hearing different opinions 18 from within the FDA. 19 that you're hearing different opinions. 20 a very complex data set. 21 is a philosophical question, as you said. 22 hopefully, that helps you understand what (202) 464-2400 It's not surprising It's To some degree, it Beta Court Reporting www.betareporting.com So (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 300 1 we're trying to achieve. 2 DR. SWENSON: Well, at this point, I 3 think it would be appropriate that we move on 4 with Dr. Chowdhury's comments. 5 quite a bit more time tomorrow to discuss these 6 very important aspects. 7 DR. CHOWDHURY: And we will have Good afternoon. 8 will be speaking about the risk-benefit 9 assessment of long-acting beta agonist 10 I bronchodilators in the treatment of asthma. 11 Before I start, I would like to 12 thank members of my division for the work 13 that I'm presenting. 14 our statistical team leader, Dr. Chen Lee 15 (?). 16 Dr. Seymour and Dr. Kalimisha, Dr. Michel, 17 Dr. Boskin, Peter Stowick (?), and others. 18 also thank Dr. Rosebrough and Dr. Jenkins of 19 the office for their oversight and support. 20 Particular thanks to Thanks to the medical reviewers, I Here is the outline of the 21 presentation. 22 introductory comments, then talk about risks (202) 464-2400 I'll initially make some Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 301 1 of LABAs in adults and children, benefits of 2 these drugs in adults and children, 3 risk-benefit assessment, and then I'll 4 reflect on some of the comments that you 5 heard from our OSE colleagues. 6 make some concluding remarks. 7 And finally, First, let me go through some 8 introductory comments and then talk about the 9 risks and then benefits. As you have heard 10 in the discussion just preceding my talk, got 11 into that, that there were differing views on 12 managing risks from inhaled LABAs in the 13 treatment of asthma. 14 expressed by our OSE colleagues and us in the 15 Division of Pulmonary and Allergy Drugs are 16 somewhat different on how to manage these 17 risks. 18 And the views that were You have heard unanimous 19 recommendation from our colleagues in the OSE 20 which goes for withdrawing medication for all 21 long-acting beta agonists for patients below 22 18 years of age and removal of asthma (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 302 1 indication and contraindicate its use as 2 single-ingredient agents for all ages. 3 other words, what will remain are combination 4 products only for managing asthma above the 5 age of 18. 6 In The rest will all go. From our viewpoint, we think the 7 products containing long-acting beta agonists 8 should continue to be in the market, and the 9 safety risk should be managed through 10 labeling. 11 other ways of managing risk has been 12 expounded by Dr. Seymour, which includes 13 various aspects such as: 14 which you have seen; medication guides; 15 directions for appropriate use; particularly 16 that long-acting beta agonists should be used 17 only as additional therapy for patients not 18 adequately controlled on other asthma control 19 medications such as inhaled corticosteroids. 20 And some of this labeling and Labeling changes, And I'd like to add that opposition 21 is consistent with major asthma treatment 22 guidelines, such as those of the NAPP of NIH (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 303 1 and also the GENA (?), which is the Global 2 Asthma Treatment Guidelines. 3 Now you have heard what asthma is. 4 In the top bullet, I give some features of 5 asthma which is generally accepted, which 6 includes shortness of breath, wheezing, chest 7 tightness, cough, airflow obstruction, and, 8 of course, bronchial hyper-responsiveness and 9 underlying inflammation. As you can see, 10 many of these are symptoms because of airflow 11 limitations that these patients have. 12 long-acting beta agonist dilates airways and 13 relieving their symptoms. 14 A It was heard in the early 15 presentation that asthma is classified based 16 on level of symptoms, awakenings, 17 short-acting beta agonist use, and 18 interference with normal activity and lung 19 function. 20 intermittent or persistent, and the 21 persistent being mild, moderate, severe. 22 (202) 464-2400 And asthma is classified as And as you have also heard in the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 304 1 morning, asthma is not a fixed disease and 2 not exactly the same disease in everybody. 3 And patients with asthma can move down in 4 severity or move up in severity. 5 words, the disease is very variable in 6 patients. 7 In other Now, what are the current treatment 8 choices that we have for treating asthma? 9 Well, asthma treatments can be classified in 10 two categories: 11 medications, which include short-acting 12 bronchodilators for all practical purpose in 13 the U.S. -- this means inhaled 14 albuterol -- and (inaudible) corticosteroids. 15 For the very mild asthma, short-acting beta 16 agonists or inhaled albuterol is all the 17 patient may need. 18 The quick relief The long-term control medications 19 as classes are listed here. 20 through these classes one by one so that you 21 understand what the choices are. 22 listed alphabetically. (202) 464-2400 And let me go Beta Court Reporting www.betareporting.com These are (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 305 1 First, there are or soon will be 2 (inaudible) inhaled cromones. 3 are propelled by CFCs and are to be phased 4 out because of CFC and these have not been 5 reformulated with alternate propellants. 6 drugs in this class were nedocromil and 7 cromilin. 8 9 These drugs The Then we have immunomodulator that we heard before. The trade name is Xolair. 10 This drug is specifically for patients who 11 have got asthma with an allergic component, 12 specific levels of IGE, more directed for 13 more severe patients, and have serious risk. 14 The two warnings: 15 anaphylaxis. 16 Malignancy and Third, we have heard about inhaled 17 corticosteroids and heard a lot about it. 18 Then there are liquid modifiers. 19 practically two in the market: 20 we have heard in the morning how effective 21 these are; and also zileuton. 22 cause liver injury. (202) 464-2400 There are Montelukast, Zileuton can And the product label Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 306 1 asks for periodic monitoring of blood for its 2 safe use. 3 Then, of course, we have 4 long-acting beta agonists, which we're 5 discussing now. 6 or theophylline, which we all know what the 7 safety profiles of these drugs are. 8 theophylline one needs to do periodic blood 9 monitoring to ensure safe use. And then the methylxanthines For Theophyllines 10 are quite narrow drug. And of course, we 11 have systemic corticosteroids coming under 12 both. 13 well-known, well-characterized safety 14 profile, including metabolic effects. 15 keep in mind, inhaled corticosteroids, even 16 at high doses, can have systemic effects. And systemic corticosteroids has their 17 And So these are the drugs that we 18 have. 19 main drugs that currently are used by 20 patients with asthma are albuterol, inhaled 21 corticosteroids, and long-acting beta 22 agonists. (202) 464-2400 For really all practical purpose, the These forms the main banquet for Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 307 1 asthma treatment today. 2 what we're talking about is removal of this. 3 What particularly The removal of long-acting beta 4 agonists as a group, as recommended by our 5 colleagues in the OSE, will force patients 6 who are not adequately controlled on ICS to 7 use other medications. 8 symptom relief medications effects that are 9 acutely noticeable by patients, patients are Since LABA are 10 likely use -- are likely to use inhaled 11 short-acting beta agonists or albuterol 12 chronically and perhaps at high doses, and 13 also, burst of oral (inaudible) steroids. 14 As I'll discuss later, you will 15 appreciate that chronic high-dose 16 short-acting beta agonist also has risk of 17 its own, which goes back to what the risk 18 with long-acting beta agonist is, which is 19 asthma-related death. 20 steroids have known risks that you're aware 21 of. 22 (202) 464-2400 Oral and (inaudible) Now, asthma-related death with Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 308 1 short- acting beta agonists is not new. 2 There were some discussions on this a couple 3 minutes ago. 4 And here is a slide that I borrowed from a 5 publication, Lancet, showing similar 6 well-described population-based risk, which 7 we heard about a minute ago, with some of 8 these drugs. 9 It goes back over 50 years. The first that we saw was with 10 epinephrine in 1940s and '50s in some 11 countries. 12 New Zealand, England, and Wales, and 13 Netherlands -- the U.S. is not here -- and 14 there was an increase with asthma-related 15 death that was thought to be linked to 16 epinephrine. 17 The countries in the slide are Second, in 1950s, we heard just a 18 comment before about isoproterenol resulting 19 in increased death. 20 there going back about 50 years, which 21 describes this death. 22 high-dose formulation of isoproterenol (202) 464-2400 The publications out Then there was a Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 309 1 introduced to the market in some countries, 2 resulting in quite high increase and, again, 3 population-based increase in death. 4 Then finally, in these -- I would 5 point out the well-described cases with 6 fenoterol and other short-acting beta 7 agonists. 8 New Zealand. 9 characterized in population-based studies Initially, the risk was seen in And subsequently, the risk was 10 conducted in Saskatchewan, Canada. And in 11 this analysis the authors actually looked at 12 albuterol and concluded in the articles in 13 different journals on this that, based on 14 exposure, it is possible, but albuterol at 15 high doses can carry the same risk. 16 led to the use of albuterol, which currently 17 we use today, which is more of as-needed use 18 and not chronic high-dose use. And that 19 Now, going back to a question that 20 was posed early in the morning, what are the 21 mechanisms of asthma-related death? 22 short answer is we do not know. (202) 464-2400 Beta Court Reporting www.betareporting.com The (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 310 1 The couple of factors that we 2 should consider or think about what we know. 3 First, there are some contributing factors. 4 Use of high-dose beta agonist drugs can cause 5 problem. 6 previous slide, with isoproterenol. 7 also (inaudible) with a case with fenoterol. 8 And also we have data from recently the case 9 with the long-acting beta agonist formoterol. 10 And we saw that, as I showed in the It was Second, use of less selective drug 11 can contribute, and we have seen data with 12 isoproterenol and, again, fenoterol. 13 Now, there are also some 14 hypothesized mechanisms that can contribute 15 to this. 16 earlier, is reduction of protection against 17 bronchoconstrictor stimuli. 18 if patients are on these drugs, then the 19 responsiveness to a bronchoconstrictive agent 20 may be blunted. 21 with various agents and shown to be the case. 22 And agents included, the published articles (202) 464-2400 First, which we heard about In other ways, And this has been tested Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 311 1 on these, using methylcholine, histamine, 2 adenosine, exercise, cold air, allergen. 3 The other thing which we've also 4 heard is masking symptoms of worsening 5 asthma. 6 long-acting beta agonists can have asthma 7 worsening, which they would not really 8 respond to and seek treatment because with 9 the long-acting beta agonist on board, the In other ways, the patients taking 10 symptoms are controlled. 11 they seek help, perhaps inflammation is too 12 advanced. 13 And by the time Now, we have heard about 14 phenogenetic studies and things for the 15 future, but at this time, we do not have any 16 certain length of phenogenetic marker or 17 phenotypic marker (inaudible) patients who 18 are at risk. 19 for the briefing package I've listed some of 20 the studies which has tested for the 21 possibility of looking for a genetic link for 22 this asthma risk, and we do not have one. (202) 464-2400 And in my summary documentation Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 312 1 Now, mechanistically one can think 2 the mechanism by which inhaled short-acting 3 beta agonists and long-acting beta agonists 4 cause asthma-related death are not known, but 5 maybe they're likely to be similar because 6 the basic action of both these drug classes 7 are the same. 8 bronchodilators and really do not have any 9 market effect of inflammation -- on 10 They're basically both inflammation. 11 Now, what are the risks of 12 long-acting beta agonists in adults? 13 over the next couple of slides talking about 14 adults and then come back and comment about 15 pediatrics. 16 I'll go Of course, the main risk is 17 asthma-related death, which we have been 18 talking about, and serious exacerbation. 19 addition, there are other effects with these 20 drugs that are well-known and they're listed 21 here. 22 on the slides are rare with inhaled drugs (202) 464-2400 In The systemic effects which I'm listing Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 313 1 because the inhaled doses are lower than the 2 doses known to cause the systemic effects. 3 And local irritation of airways are also seen 4 with this drug. 5 practical purpose, what we're talking about 6 are asthma-related death and serious asthma 7 exacerbation. 8 9 So for today, for a more Now I'll go over some of the safety data and safety findings with the two 10 long-acting beta agonists: 11 formoterol. 12 through my slides pretty quickly because most 13 of this has been described earlier, and I'll 14 just make some comments which I may want to 15 make on these studies. 16 Salmeterol and For this section, I'll go As we have heard, for salmeterol 17 there are two studies: 18 study. 19 SNS study showing the death risk with the 20 risk of 3. 21 study design, which was a study comparing 22 salmeterol versus placebo. (202) 464-2400 SNS and the SMART And we have seen the results of the And we have heard about the SMART I'll not go into Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 314 1 the design in this slide any further. 2 And we have heard about the conduct 3 of the study, the powering of the study. The 4 main point is that it was powered to rule out 5 three times increase in asthma death. 6 number three comes essentially from the SNS 7 study. 8 important endpoint of death is shown in this 9 slide. The And the primary endpoint and the You also have this slide in my 10 briefing document and has been presented 11 earlier, so I'll not go through the details 12 here. 13 there were 13 deaths with Serevent versus 3 14 with placebo. 15 African-Americans. 16 the same risk with death. 17 The main point, which we all know, And it seemed to be more in All the Caucasians had Two of the slides quickly, 18 secondary endpoints on the death. 19 one was a question that we have addressed 20 already, whether inhaled corticosteroids 21 protects or not. 22 (202) 464-2400 The first And here is the data broken down by Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 315 1 use of inhaled corticosteroid at baseline 2 versus no, and the numbers are here. 3 we really look at the African-Americans, it 4 doesn't seem to be much different, three 5 versus one, four versus zero. And if 6 Another question came up in the 7 morning regarding the phases of the study 8 and, yes, it was done in two phases. 9 first phase, which was in these four years, And the 10 patients were recruited primarily through 11 advertisements from their essentially home or 12 are patients and not necessarily through 13 physician contact. 14 In the second phase, to encourage 15 enrollment, patients were actually 16 (inaudible) through physician's office. 17 Patients coming in Phase II were actually 18 seeing physicians for the asthma control in 19 most cases. 20 And as expected, the number of 21 events were high in the Phase I, where 22 patients would be coming in, if you would (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 316 1 call it, from the wild. 2 different. 3 yellow is placebo. 4 deaths. 5 The numbers were The green is salmeterol and the There were 10 versus 3 The second phase, which the 6 patients were under physicians' control, also 7 have events; much less, and three events were 8 with salmeterol. 9 Now, let me go over again very 10 quickly with the formoterol studies. 11 seen the studies presented before. 12 through them pretty quickly. 13 You've I'll go As you've heard, there were two 14 Phase III studies in patients 12 and above, 15 one Phase III study in patients 5 to 12, and 16 one Phase IV study, 16-week study duration, 17 in patients 12 years of age and older. 18 show all those studies in one slide. 19 And here is the result. I'll You have 20 the same slide in my briefing document. 21 I'll not go through the numbers here because 22 the numbers have already been presented. (202) 464-2400 Beta Court Reporting www.betareporting.com And The (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 317 1 main point is there was serious asthma 2 exacerbations that were seen with formoterol. 3 The numbers were higher than albuterol and 4 the placebo. 5 ordering. 6 And there seemed to be a dose Now, I want to touch on, briefly, 7 about the combination product, Symbicort 8 Phase III NDA study, and bring out one safety 9 finding from this -- these Phase III studies. 10 There are two Phase III studies that 11 supported approval of Symbicort in the U.S.: 12 Study 1 and 2 done in patient with moderate 13 to severe asthma and mild to moderate asthma. 14 The two co-primary endpoints, as you've 15 heard, are (inaudible) present, which was 16 based on FEV1. 17 predefined asthma worsening criteria were 18 required to be withdrawn. 19 these patients were withdrawn and make some 20 comment here. 21 22 And patients who satisfied And I'll show Here are the number of patients and percentages who were withdrawn from these (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 318 1 studies because of asthma worsening. Again, 2 you have the slide in the briefing document. 3 But I want to highlight here that these, of 4 course, are efficacy parameters and also they 5 can be used as safety parameters. 6 On the safety side, if you look at 7 the studies -- for example, study 1 -- there 8 are more patients on formoterol alone who 9 have withdrawn due to decrease in FEV1, peak 10 flow, and clinical exacerbation. 11 compare to placebo. 12 these are very effective bronchodilator 13 drugs, but also a small percentage of 14 patients have worsening of their asthma with 15 this drug, and this is what we're talking 16 about. 17 including in these studies. 18 The numbers The point here is that It comes out in multiple studies, Now, let me just briefly comment on 19 the inhaled LABAs risk in children. 20 12 and above, it's the same as older children 21 and adults because the safety studies were 22 conducted in this age group. (202) 464-2400 For ages For children 4 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 319 1 to 11, as we heard before, there are no large 2 safety studies and Phase III studies did not 3 show any asthma death. 4 not relevant because the drug is not approved 5 at this age group. 6 Ages below four is Now, keep in mind for these drugs 7 there actually were studies done in 4- to 8 11-year-old children, for example, with 9 Serevent, Advair, Foradil. Each have studies 10 with children in this age group, ranging in 11 number between 200 to approximately 500 12 patients. 13 Another comment about the risk in 14 children: 15 for now carry the same labeling warning 16 irrespective of age with the conclusion that 17 asthma-related death seen in studies 18 conducted in adults and adolescents, in 19 patients 12 years of age and older, apply to 20 patients below 4 years of age. 21 have taken a conservative view. 22 (202) 464-2400 Products containing inhaled LABA Therefore, we A couple of points here. Beta Court Reporting www.betareporting.com Based on (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 320 1 the control trials, the available data do not 2 suggest that the safety risk with long-acting 3 beta agonists is higher in children four 4 years of age and older. 5 consider that patients four years of age and 6 older as they compare to adults, as far as 7 the disease is concerned, the disease is the 8 same. 9 the same. 10 And we also need to Asthma, the pathophysiology also is The target of the beta agonists, 11 the beta receptors, function similarly. 12 Therefore, the response to long-acting beta 13 agonists is expected to be the same. 14 Now, our OSE colleagues concluded 15 the safety risk is higher in children. 16 conclusion is not based on death, but on 17 hospitalization. 18 that conclusion later in my presentation. 19 This And I'll comment further on You have, again, seen this slide 20 before, which is a risk interpretation. 21 the point that I want to make here is the 22 risk in a population basis, based on the (202) 464-2400 Beta Court Reporting www.betareporting.com And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 321 1 SMART study, is 8 per 10,000 over 28 weeks. 2 And we are to keep in mind that for the SMART 3 this perhaps is the worst-case scenario. 4 LABA in the study was used, in many cases, 5 without controller medications. 6 moved. 7 of asthma has moved. 8 9 Management has moved. The Time has The treatment In real life, long-acting beta agonists mostly, as we'll see in slides, are 10 used by controller medications such as 11 inhaled corticosteroids. 12 came up how does this risk really translate 13 into asthma mortality in the U.S. on a 14 population basis? 15 presenting the population data with some 16 other short-acting beta agonists drugs. 17 And the question And you've seen before by For the long-acting beta agonist 18 drug, as we see now, it doesn't seem to have 19 translated to increased mortality. 20 Here's the mortality shown over 21 age, over years, rate, and number. 22 look at -- approximately late '90s, it has (202) 464-2400 Beta Court Reporting www.betareporting.com And if we (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 322 1 changed. 2 10 code change. 3 decrease here. 4 seems to continue the gradual decline. 5 this is in spite introductions of the drugs 6 that we're talking about here today. 7 And this is the ICD-9 to code -- to And of course, there was a But all the ICD-10 code it And Inhalation aerosol in 1994. 8 years later, the Diskus formulation. 9 couple of years later, Advair Diskus, Two A 10 inhalation aerosol, and lately Symbicort. 11 And the use of these drugs also per volume is 12 increasing as you've seen in some slides 13 before. 14 Now, what are the benefits of LABA 15 in adults? 16 quickly, and you've heard this before. 17 were approved for use in asthma, EIB or 18 exercise-induced bronchospasm, and COPD. 19 some of them, as you heard before, of course 20 increases quality of life, specifically the 21 ones which are combination products 22 containing LABA and inhaled corticosteroids. (202) 464-2400 I'll go through this pretty Beta Court Reporting www.betareporting.com LABAs And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 323 1 The reason I bring up quality of life, 2 because we've heard a lot of discussions 3 around it and to keep in mind the AQLQ for an 4 instrument to be used in trials was really 5 not validated up until late '90s and old 6 trials do not have much AQLQ data. 7 to keep in mind that AQLQ itself was 8 validated based on FEV1 measure and 9 (inaudible) challenge, which is also based on 10 FEV1. And also So some was circular. 11 The benefits of LABA in children 12 for ages 12 and above we concluded it is the 13 same as older children and not necessarily 14 any worse. 15 showing efficacy, and ages below 4 is not 16 relevant. 17 And there are studies 4 to 11 And again, for -- same as on the 18 safety side, on the benefit aspect of it, the 19 same logic that I applied for the safety 20 applies here. 21 the target being the same, the response being 22 the same. (202) 464-2400 The disease being the same, So there shouldn't be really much Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 324 1 reason to believe that efficacy will be 2 substantially different. 3 children studies were done and shown 4 efficacy. 5 Nevertheless, in As far as combination product goes, 6 we have heard before the combinations are 7 combinations of convenience. 8 necessarily require the same level of study 9 to show efficacy of the combination product. So they did not 10 The point here is does the combination give 11 the same two active ingredients? 12 true for Advair the number of studies are 13 limited. 14 multiple studies done with the 15 single-ingredient products in children. 16 And it is But again, they're based on For example, in Advair, in children 17 there was one study primarily for safety, 18 which also showed efficacy, but this was 19 built upon multiple studies. 20 with fluticasone, there are three studies 21 with Flovent Diskus in patients 4 to 11, 22 involving approximately a thousand patients. (202) 464-2400 For example, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 325 1 There was one study with Flovent (inaudible) 2 involving about 250 children. 3 another study involving over 200 patients 4 with Flovent inhalation aerosols. 5 also similar studies with salmeterol. 6 therefore, the point I'm making is the 7 combination Advair has multiple studies with 8 single ingredients, which supports the 9 approval and the combination is built on 10 There was also There are So, those studies. 11 Now, on the benefits, just to 12 summarize then, LABAs are effective for 13 asthma in terms of improvement in FEV1, peak 14 flow, rescue albuterol use, symptom control, 15 nocturnal awakenings, and others. 16 have seen a summary of the studies presented 17 by Dr. Sally Seymour and others. 18 addition, you have seen many of the studies 19 which have been done in the public domain by 20 either public funding or private funding that 21 has been summarized Dr. Lemanske before. 22 I'm naming some of the major studies here (202) 464-2400 Beta Court Reporting www.betareporting.com And you In And (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 326 1 that includes combination products, including 2 formoterol and budesonide, salmeterol and 3 (inaudible), salmeterol and fluticasone. 4 these all show the benefit of these 5 combination products over single ingredients. 6 And And I would like to use one study 7 which is a widely quoted study conducted over 8 multiple years. 9 And this is a GSK-funded study 10 which showed the contribution of combination 11 product over inhaled corticosteroid. 12 next three slides I'll walk through the study 13 to make the point. 14 And the The goal study involved 15 approximately 3,400 patients, ages 12 through 16 80, group, multinational. 17 step-wise increase of doses of fluticasone or 18 fluticasone plus salmeterol in achieving 19 asthma control. 20 which I show based on daytime symptoms, 21 rescue drug use, morning peak flow, nocturnal 22 awakenings, exacerbations, emergency visits, (202) 464-2400 And it compared And the asthma control here, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 327 1 and treatment of adverse event, these come 2 out from the (inaudible) document. 3 are combinations of asthma control. 4 So these The design was somewhat 5 complicated. The patient were broken up into 6 stratus 1, 2, and 3, based on the severity, 3 7 being more severe. 8 phases of trials, which is called Phase I and 9 Phase II. And patients entered into And the aim for these phases was 10 to stabilize patients on a dose of 11 fluticasone or a combination product and 12 achieve control. 13 maintain that. 14 the steroid complement and go up on that. 15 It's basically the way asthma is managed and 16 this allows comparison of fluticasone and 17 salmeterol versus fluticasone in different 18 levels of asthma severity. 19 If control was achieved, If not, increase the dose of Here is a summary slide. On the 20 left-hand side, I'm showing patients 21 achieving well-controlled; on the right-hand 22 side, patients achieving totally controlled. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 328 1 Without going into the details you can see 2 that there was more patients or there was 3 more control achieved by the combination 4 product over fluticasone. 5 based on FEV1, the base on asthma control, 6 which I defined earlier. 7 right here across the strata. 8 the patients, there's more improvement. 9 And this is not And the numbers are And the sicker And also note that totally 10 controlled asthma, which we tried to achieve 11 was not really achieved in about half of the 12 patients. 13 necessarily that get us there. 14 talking about removing salmeterol and 15 formoterol from these choices. 16 So we have good drugs, but not And now we're Now I'll talk about the 17 risk/benefit assessment and then come back 18 and comment on some of the OSE comments here. 19 On the risk side, the main risk, of 20 course, is asthma-related death. 21 benefit side, it is that most patients derive 22 symptomatic benefit in the form of various (202) 464-2400 Beta Court Reporting www.betareporting.com On the (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 329 1 control measures that we have talked about 2 this morning. 3 balance this risk against the benefit. 4 The main thing is how to Patients, health care providers, 5 and society have accepted serious adverse 6 reactions and even death in a small number of 7 patients for symptom control in a large 8 number of patients. 9 minutes ago, we have heard two examples: 10 And just a couple of Acetaminophen and NSAIDs. 11 These control -- these drugs 12 provide symptom control, minor aches, and 13 ailments, but have serious adverse reactions. 14 So our view from the division side is that 15 the safety risk of long-acting beta agonists 16 can be managed through labeling to inform 17 health care providers and patients of the 18 risk and thereby directing use of these drugs 19 to the appropriate patient population. 20 our position is consistent with the 21 recommendation from the July 2005 Advisory 22 Committee meeting. (202) 464-2400 And And as I said earlier, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 330 1 also consistent with the U.S. asthma 2 guidelines from NIH and also the global 3 guidelines, the GENA document. 4 Now, let's think through some of 5 the consequences of removal of asthma 6 indication for long-acting beta agonists. 7 will reduce choice (inaudible) unable to 8 control patients' asthma on ICS alone. 9 drugs are listed here, and I've mentioned the 10 It The limitations of these drugs earlier. 11 Since LABAs are symptom relief 12 medications, effects that are acute and 13 noticeable by patients, patients will most 14 likely use short-acting beta agonists 15 chronically and at high doses and oral 16 steroids with or without theophylline. 17 will shift and push back the asthma treatment 18 and asthma care to where it was about 20 19 years ago. 20 mortality, but may increase it. 21 short-acting beta agonists itself may have 22 risk of asthma-related death. (202) 464-2400 This This shift will not reduce As we know, And oral and Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 331 1 (inaudible) steroids have known safety risks. 2 Theophylline is a narrow therapeutic index 3 drug. 4 Now, another concern will be 5 inappropriate use of long-acting beta 6 agonists in patients with asthma. 7 to keep in mind long-acting beta agonists are 8 also proved for (inaudible) to the 9 bronchospasm and for COPD. You need Therefore, health 10 care providers and patients will have access 11 to the medications and these can continue, 12 although off-label use. 13 OSE's presentation commenting about the 14 labeling and how effective it can be. 15 also keep in mind with the removal of the 16 asthma indication specific recommendation 17 regarding appropriate safe use of LABAs in 18 asthma will also be removed from the label. 19 And we have seen And We have often heard about the 20 possibility of removal of asthma indication 21 for single-ingredient LABAs in favor of the 22 combination products. (202) 464-2400 Again, it would reduce Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 332 1 choice, choice for inhaled corticosteroids. 2 Because what we'll have now are the two 3 corticosteroids: 4 which are available as combination products. 5 And also, it will preclude combination use of 6 inhaled LABAs with other long-term control 7 medications other than corticosteroids. 8 of course, we have heard about that there's 9 no data that such combination can mitigate 10 Fluticasone and budesonide, And, the risk. 11 Now let me move on and reflect on 12 some of the comments that we have heard from 13 our colleagues earlier in the session. 14 a couple of brief points about limitations of 15 meta-analysis, and what I'm commenting here 16 is for any meta-analysis be it for safety, be 17 it for efficacy. 18 Now, First of all, when meta-analyses 19 are done they're very, very useful to 20 overcome the problem of reduced surgical 21 power in studies with some small sample sizes 22 by pooling similar studies in appropriate, (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 333 1 orderly way. A weakness in meta- analysis is 2 that any source of bias in the various 3 original studies is carried forward. 4 has to be very careful about the critical 5 issues and lack of control of these can lead 6 to bias. 7 were selected, their heterogeneity in the 8 results, conduct design of the studies, 9 patient populations are all information that And one For example, how and what studies 10 one needs is available and was an appropriate 11 analysis performed. 12 And generally, a large, controlled, 13 randomized clinical study is the gold 14 standard of obtaining information of a 15 specific question. 16 agonists, we have the luxury of having two 17 such studies. 18 And with long-acting beta But for rare events, the same 19 points that I made earlier applies, and even 20 can get magnified because they're even so 21 small. 22 actually become important because of lower (202) 464-2400 And for rare events meta-analysis can Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 334 1 signals and lower sample sizes. 2 these rare events, conclusions should be 3 based on endpoints, and these require a good 4 understanding of all potential bias 5 introduced in meta-analysis. 6 But for Now, let me talk about a few points 7 about the FDA meta-analysis. I'm just using 8 the term OSE up here just to point out, but 9 this is actually FDA meta-analysis conducted 10 by our Office of Biostatistics. 11 that patient dropout is important and can 12 confirm results. 13 placebo arm or less effected treatment arms 14 compared to more effected treatment arms may 15 result in patients staying longer in the 16 active treatment arm or more effected 17 treatment arm and, therefore, provide more 18 adverse events of interest. 19 One point is Differential dropout in An example, if you look at the 20 SMART trial, the number of dropouts in the 21 placebo arm was 535 patients compared to 434 22 in the salmeterol arm, a difference of (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 335 1 approximately 100. 2 As for the comparators, no LABA 3 which has been used as a comparator in the 4 FDA meta-analysis is informative, but, again, 5 has some limitations. 6 question one is answering -- is asking. 7 example, to understand if LABA risk is 8 mitigated by inhaled corticosteroid, the 9 right comparator arm is ICS plus LABA versus 10 It depends what For LABA. 11 The fact that many studies were 12 designed to address objective meta-analysis 13 is not an issue of meta-analysis in general. 14 The important point is whether data is ready 15 for analysis. 16 Two of the points to keep in mind 17 is whether the risk difference is constant or 18 variable over time, and thus -- this -- and 19 how this effects the selection of studies 20 based on 30-day ration as a cutoff. 21 seen the cutoff used in the FDA 22 meta-analysis. (202) 464-2400 Beta Court Reporting www.betareporting.com We have (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 336 1 Another point to keep in mind, is 2 it reasonable to exclude doses of drugs 3 higher than labeled dose, which has been done 4 in some meta-analysis? 5 knowledge that safety risk is generally 6 dose-related with higher dose giving more 7 safety findings. 8 9 It's general Now I'll comment on meta-analysis for asthma-related death. And the question 10 is really, what is the role of meta-analysis 11 when large control studies are available? 12 And they are. 13 inclusion of studies with limited duration of 14 exposure at a low, even rate. 15 of patients, such as milder patients or 16 patients on adequately controller drugs on 17 board, we have low risk and can dilute the 18 findings. 19 comparator would have the same effect. 20 Risk can be diluted by Heterogeneity And, of course, the non-LABA as a Now, hospitalization is a very 21 important consideration for asthma and is a 22 major cause of asthma morbidity. (202) 464-2400 Beta Court Reporting www.betareporting.com The FDA (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 337 1 meta-analysis did a remarkable job to get 2 data on this important endpoint. 3 some points that you need to keep in mind as 4 you review the hospitalization data in the 5 FDA meta-analysis and other meta-analysis and 6 see how informative these are. 7 There are The general expectation is that the 8 asthma-related hospitalization will track 9 with asthma-related death and intubation, if 10 one is to believe that these are spectrum of 11 LABA risk. 12 did not track. 13 should be controlled, and we talked about the 14 biases earlier. 15 In the FDA meta-analysis, this And all potential biases And one important point is that the 16 underlying cause of the hospitalization, as 17 much as possible, should be obtained because 18 it is very informative. 19 of hospitalization events without accounting 20 for the underlying cause of hospitalization 21 does not provide a complete picture of the 22 risk. (202) 464-2400 Because the number Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 338 1 Now, you have seen part of this 2 slide before at least twice. And what I'm 3 showing here is the FDA meta-analysis, the 4 primary endpoint stratified by age, the risk 5 difference, and (inaudible) per thousand 6 subjects, LABA versus no LABA. 7 this before, which is a composite endpoint 8 showing that the risk is high in the younger 9 patients. You have seen You've also see and heard before 10 that the risk is driven primarily by 11 hospitalization. 12 limitations in the hospitalization data which 13 has commented -- which I commented to 14 earlier. 15 And there are some For asthma death or asthma death 16 and intubation, the FDA meta-analysis indeed 17 does dilute the death signal seen in SMART. 18 The asthma death risk difference in SMART is 19 about .8 per 1,000. 20 difference in this meta-analysis is 21 approximately half of that. 22 death difference of -- in patients were (202) 464-2400 Overall risk death The overall Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 339 1 largely seen in patients 18 to 65 years of 2 age and virtually not seen in younger 3 patients. 4 In the same FDA meta-analysis 5 Symbicort seemed to have the worst risk of 6 all products, even worse than the single 7 ingredient formoterol or Foradil here. 8 finding goes against other observation that 9 the risk of asthma death and other related This 10 outcomes are seen more in patients with 11 single-ingredient LABA products compared to 12 patients on LABA plus ICS. 13 The two other points I would like 14 to briefly make is in this slide. 15 needs to consider what patients are included 16 under combination products Advair and 17 Symbicort. 18 assessment, patients treated with fixed-dose 19 combination products, as well as patients 20 receiving both individual components 21 contained in those fixed-dose products, 22 should be included under the combination (202) 464-2400 One, one To get a complete risk Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 340 1 product groups. 2 is the same, whether they receive Advair 3 Diskus as a single product or they receive 4 Advair Diskus plus the steroid formulation. 5 Because the risk to patients Second point, the number of 6 patients in this analysis and some of the 7 published meta-analysis are different and we 8 have seen some explanation before. 9 need to understand to make sure that there's 10 no potential bias introduced in selection or 11 exclusion of studies. 12 And you In the FDA meta-analysis the number 13 of patients in the two formoterol-containing 14 products are very low compared to some 15 published meta-analysis. 16 complete risk assessment study using all 17 relevant products, either marketed in the 18 U.S. or internationally, should ideally be 19 included. 20 Turbuhaler, which is a (inaudible) 21 formulation of formoterol, which contributed 22 about 8,000 patients, were not included. (202) 464-2400 To (inaudible) And I believe the product Oxis, a Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 341 1 The final couple of slides here, 2 the unanimous OSE recommendation that you 3 have heard before needs to be considered 4 carefully in light of all existing data. 5 I discussed before, it is sort of 6 questionable whether there's conclusive data 7 that shows long-acting beta agonists below 8 patients 18 years of age are more riskier. 9 Withdrawal of drugs or contraindication of As 10 drugs are serious recommendations. 11 withdrawal, these drugs will not be available 12 for patients. 13 mean that these drugs should not be used 14 because the risk of use clearly outweigh any 15 possible benefit. 16 the drug, there's no benefit that can 17 overcome the risk. 18 removal you've heard before. 19 With And contraindication would It simply means do not use And the consequences of Briefly, some reflections on some 20 OSE comments which you heard before and also 21 in the briefing documents. 22 which benefit might outweigh the risk? (202) 464-2400 One question is Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 342 1 Again, this is a judgment call and we are 2 asking you to opine on that. 3 benefit is in larger number of patients who 4 take these drugs. 5 numbers you have seen before. Again, the The risk is real and the 6 And you have seen comments that 7 LABA are more risky and less effective in 8 patients below 18 years of age compared to 9 older patients. And this is mainly driven by 10 hospitalization. And conclusions and 11 recommendations based on hospitalization from 12 meta-analysis has many problems, as I 13 discussed earlier. 14 On the efficacy side, I would like 15 to comment that paucity of evidence does not 16 mean that the drug is less effective. 17 in fact, they have data down to four years of 18 age that shows similar efficacy profile 19 across ages. 20 And, And one point is that long-acting 21 beta agonists and short-acting beta agonists 22 are not necessarily fully interchangeable. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 343 1 These are drugs that belongs to different 2 classes. 3 paradigm falls into categories: 4 pain reliever drugs for as-needed use; 5 long-acting pain controller drugs, although 6 the science in the future going forward may 7 change. 8 come from the discussion here. 9 And in the current asthma treatment Short-acting And perhaps some of the changes may I'd like to reflect on the two 10 large remaining questions that our colleagues 11 in the OSE poses in the briefing document. 12 And the questions are, which we have heard 13 before: 14 short-acting beta agonists plus ICS compare 15 with safety and efficacy of long-acting beta 16 agonists plus ICS? 17 How does the safety and efficacy of Again, as I said earlier, this goes 18 somewhat different than the current paradigm 19 of asthma treatment and APB and GENA, which 20 recommends albuterol to be (inaudible) 21 as-needed basis. 22 (202) 464-2400 And the second question which I Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 344 1 think is quite important is, does concomitant 2 ICS use mitigate the LABA risk in other 3 patients? 4 safety studies of Advair and Symbicort to 5 answer these two questions. 6 comparators are short-acting beta agonist 7 plus ICS as a control arm, ICS as a control 8 arm. 9 Our colleagues have proposed large And the proposed As one looks into this proposed 10 study, there are a couple of thoughts that 11 come to mind. 12 ICS is not necessarily the standard of care. 13 And it raises the question of using 14 short-acting beta agonists chronically, 15 perhaps on top of it using short-acting beta 16 agonists on an as-needed basis for patients 17 who are not adequately controlled. 18 Short-acting beta agonist plus Second, if short-acting beta 19 agonist is found to be safer than low-acting 20 beta agonist it will be very hard to imagine 21 combination products containing albuterol 22 with formoterol -- with, sorry, fluticasone (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 345 1 or budesonide, which are the steroids in 2 Advair and Symbicort. 3 frequency of albuterol is four times a day, 4 whereas the steroids are four times a day. 5 So we'll not really have (inaudible) with 6 these specific steroids. 7 Because the dose and And another point is that the 8 second question, which is a rather important 9 question, is not going to be addressed by 10 this design because of appropriate control in 11 the LABA arm is missing. 12 GENA guidelines and product labels for these 13 long-acting beta agonists, it is unlikely 14 that a large study with single-ingredient, 15 long-acting beta agonists with mortality or 16 other serious safety endpoint is actually 17 even visible. 18 Based on NIPP and Some concluding remarks with the 19 last two slides. 20 agonists is real. 21 risk. 22 one can call it small, but the numbers are (202) 464-2400 Risk with long-acting beta Their mortality, serious It happens in a number of patients, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 346 1 all there. 2 the SMART study and SNS study. 3 The worst-case scenario is from The benefit, which is, again, real, 4 is in most patients. 5 accept the risk and manage the risk by 6 appropriate labeling, including dose warning, 7 medication guide, inform patients and 8 practitioners how to safely use this drug, 9 and to label the drug for appropriate and 10 One way to vanish is to safe use. 11 And finally, for our position has 12 been that for patients four years of age and 13 older LABAs should be used in patients not 14 adequately controlled on other 15 asthma-controller medications; e.g., low to 16 medium dose ICS, or whose disease severity 17 clearly warrants initiation of treatment with 18 two maintenance therapies. 19 agonists should not be used in patients whose 20 asthma can be managed by the ICS alone with 21 occasional use of inhaled short-acting beta 22 agonists. (202) 464-2400 Long-acting beta For patients less than four years Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 347 1 of age, this is not approved. 2 Thank you very much. 3 DR. SWENSON: At this time, since we 4 are behind schedule, we're going to shorten the 5 scheduled break to ten minutes, and I'd like 6 everyone back then to resume in ten minutes. 7 That will be 3:50. 8 (Recess) 9 DR. SWENSON: We'll begin this last 10 portion of the meeting today with a first 11 overview of Asthma Guidelines in the Diagnosis 12 and Management, by Stuart Stoloff. 13 Dr. Stoloff? 14 DR. STOLOFF: Good afternoon. I'm Dr. 15 Stuart Stoloff. 16 Carson City, Nevada. 17 family physician, as a solo physician, for over 18 30 years. 19 care, and continue to provide care, daily for 20 infants, children, adults, pregnant women, and 21 people that I put in the "older age group"; 22 other people put them in the geriatric age (202) 464-2400 I'm a family physician from I have practiced as a During that time, I have provided Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 348 1 group. 2 For the past 15 years, I have also 3 been a member of the expert medical panel for 4 the Guidelines for the Diagnosis and 5 Management of Asthma, of which some of my 6 other colleagues are in the room at the 7 present time. 8 9 Because of this special interest, it has allowed me to increase my practice in 10 this field, and I presently have over 2,000 11 infants, adults, pregnant women and older 12 individuals in my practice with that disease. 13 So it's my pleasure to have an opportunity 14 today to extend the comments that 15 Dr. Lemanske started with, provide some other 16 information. 17 opportunity. And I appreciate the 18 What this next slide shows you is 19 the United States, and what's going on on a 20 daily basis in the United States. 21 not only common, it's a chronic disease, and 22 when you look at it, about 5,000 people every (202) 464-2400 Beta Court Reporting www.betareporting.com Asthma is (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 349 1 day end up in the emergency room in this 2 country. 3 Additionally, 63,000 people miss 4 work or miss school. And in fact, asthma is 5 the number one reason why children miss 6 school in the United States. 7 ten people a day die in the United States. 8 So the death rate is not 5,000, the death 9 rate is not 4,000, the death rate is Additionally, 10 somewhere between 3,400 and 3,600 in the 11 United States. 12 So what is this disease? Well, 13 this is a picture diagram you can look at, 14 and the number one thing that you see is it 15 is a bronchoconstriction. 16 you not in medicine, what I'd like you to 17 think of is a hose, and the outside of the 18 hose has got a smooth area but you don't have 19 a nozzle to turn off that hose, so what you 20 do is you clamp it down with your hand. 21 is bronchoconstriction. 22 (202) 464-2400 Now, for those of That At the same time, unfortunately, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 350 1 one of your kids got dirt and pebbles inside 2 that hose, and that's inflammation. 3 these two components, this smooth muscle 4 dysfunction, combined with the inflammation, 5 can and do occur concurrently in asthma. 6 that's asthma. 7 And And That's the pathophysiology. So we have the guidelines. Well, 8 the guidelines are actually a developmental 9 process. The first guideline was in 1991, 10 and in August 2007, as you heard, we released 11 the third expert panel report, which utilized 12 a strictly evidenced-based approach to care. 13 It is a group of which I'm part of, and other 14 members here are of -- 18 experts, and we 15 worked for 3-1/2 years, and what we reviewed 16 when we started was over 15,000 articles with 17 over 2,100 full text reviews and 18 approximately 80 percent -- somewhere between 19 50 and 80 percent looked at specifically 20 therapeutics, pharmaco-therapy. 21 22 Now, we have evolved from that first version in 1991 that was a (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 351 1 consensus-based, to an evidence-based 2 process. 3 disease is heterogeneous. 4 We have recognized that this It involves, as I stated, both 5 airway inflammation and a dysfunction of the 6 smooth muscle that surrounds the outer 7 airway. 8 based upon how this airway inflammation and 9 smooth muscle dysfunction play out and affect We do make our therapeutic decisions 10 the patient. 11 driven by severity; it is a disease for which 12 it is driven by control. 13 It is no longer a disease And really, that should make sense 14 to everyone in this room. 15 about high blood pressure, when we think 16 about diabetes, when we think about 17 hyperlipidemia, what are we really thinking 18 about is what does it take to gain and 19 maintain control. 20 diseases, and asthma is a chronic disease. 21 That's how we should look at this disease of 22 asthma. (202) 464-2400 When we think And these are chronic Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 352 1 Now, as part of that, as for other 2 chronic diseases, education by the clinician, 3 and engagement of the patient with this 4 education, is imperative. 5 relationship of the patient, the physician, 6 and the family is a key component, and it is 7 a strong component in our guideline. 8 So where are we? 9 This therapeutic Well, this is a picture that was just shown to you, but I've 10 added some to this. 11 mortality data in the United States, and as 12 you can see, the mortality incidence is 13 diminishing. 14 approximately 1995, 1996. 15 introduction of long-acting bronchodilators 16 in approximately 1994, we have seen a change. 17 This is the latest The height of it was With the Is this change entirely due to a 18 drug? 19 education; two, a better understanding of the 20 disease; and additionally, the improvements 21 in the medications that we utilize. 22 Educational efforts of the National Heart, (202) 464-2400 No. The improvement is due to, one, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 353 1 Lung, and Blood Institute, coordinated by the 2 National Asthma Education and Prevention 3 Program, have facilitated both the 4 development of the guidelines as well as the 5 dissemination of these guidelines. 6 Now, despite all these efforts, we 7 still have approximately 10 people a day in 8 this country dying from asthma. 9 So here are the definitions, and 10 Dr. Lemanske put them up for you. We think 11 of severity as the intrinsic state of the 12 disease, and we define severity actually like 13 a runner who's in the starting block at the 14 start of a race. Before the gun goes off, 15 that's severity. The minute he takes or she 16 takes that first step, all we are interested 17 in is gaining and maintaining control to meet 18 our goal, the goal devised by the patient and 19 the clinician, the physician. 20 We define severity most easily in 21 patients who have a diagnosis of persistent 22 asthma but are not on a controller therapy. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 354 1 Control is defined by looking at the two 2 domains: that is impairment and risk. 3 that is the key, looking at the present 4 impairment and risk for the future. 5 decisions that we make, both 6 pharmacologically as well as other 7 therapeutic decisions, are based upon gaining 8 and maintaining control, the same way we do 9 in the other chronic diseases, as I 10 And The discussed. 11 So what is the goal of therapy? 12 Well, the goal is really gaining and 13 maintaining control of the disease. 14 established during a patient/physician 15 dialogue. 16 severity of the disease, we are always 17 looking at both decreasing impairment and 18 risk. 19 It's For any patient, regardless of the What's impairment? Well, 20 impairment is the daytime symptoms, the 21 nighttime symptoms, the use of quick relief 22 inhalers, and most importantly, quality of (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 355 1 life for the patient. 2 We want, and our patients want, to 3 be able to sleep, to work, and to play to 4 their full capability. 5 patients to have near normal or preferably 6 normal lung function. 7 look at risk. 8 is an asthma exacerbation. 9 exacerbation in our document as an emergency In addition, we want At the same time, we We want to reduce risk. Risk We define an 10 room visit, a hospitalization, and use of 11 oral corticosteroids. 12 And regardless of the severity, we are always 13 stating that the goals of therapy are the 14 same -- a well-controlled individual. 15 That's exacerbation. So to help you with this, I'm going 16 to tell you about a patient of mine. 17 CJ. 18 some time. 19 given a diagnosis of asthma at approximately 20 age two in another practice. 21 during the viral season, which is actually 22 right now. (202) 464-2400 This is I've been taking care of him for quite He's nine years old and he was He has symptoms When he plays, and we'll talk Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 356 1 about this further, he has some difficulties 2 especially when he's tried to play soccer. 3 In addition, he is atopic with multiple 4 allergies. 5 He came to me on a low dose inhaled 6 corticosteroid, which is and was the 7 appropriate therapy. 8 quick relief medications. 9 however, on further questioning, was that he He had also received What I found out, 10 was having to use his quick relief 11 medications unrelated to exercise three to 12 four times a week. 13 bursts of oral corticosteroids each for seven 14 days in the past year, and he had to miss 15 five days of school because of asthma 16 attacks, as he called them. 17 He had received two He loved to play soccer; he could 18 not play soccer anymore. 19 play on the soccer team, but in fact when he 20 could play, the only position he could 21 play -- goalie. 22 normal lung function, as you can see in the (202) 464-2400 He really wanted to Now, he had relatively Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 357 1 bottom, an FEV1 of 88 percent predicted. 2 So CJ came in on this therapy, and 3 once you're on a controller therapy, we 4 assess control rather than severity to 5 determine treatment, and what you see here 6 are the two domains -- the domains of 7 impairment and risk in his age group for 8 assessing control. 9 was using a quick relief medication, what has 10 already been called appropriately SABA, short 11 acting bronchodilator, and he was using it 12 regularly three to four times a week. 13 And if you remember, CJ Looking at this figure, what you 14 will see is, he falls in this category of not 15 well-controlled. 16 severe when you look at controlled is where 17 you define the level of control. 18 according to the guideline grid right here, 19 he is not well-controlled despite being on 20 what many would perceive appropriate therapy, 21 a low dose inhaled corticosteroid. 22 (202) 464-2400 The category which is most So I did a questionnaire on him, a Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 358 1 validated questionnaire. 2 childhood asthma control test. 3 That is consistent with not well-controlled, 4 so that went along here. 5 he got oral corticosteroids, two bursts in 6 the past calendar year. 7 not well-controlled. 8 9 I used the He scored 18. In addition, I said Right here, again, So what did I do for CJ? Well, I looked at our algorithms for care -- and this 10 is exactly what I did, following the 11 guidelines that we all wrote. 12 longer Step II. 13 We had to step up. 14 someone who was not well-controlled 15 regardless of age, we need to step up 16 therapy. 17 He is no He is in Step III therapy. We talked about this as And for stepping up, I went to 18 combination therapy. 19 concurrent of a low dose inhaled 20 corticosteroid with a long-acting 21 bronchodilator. 22 some letters here, and these letters are (202) 464-2400 CJ got a combination And what you also see are Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 359 1 associated from the alphabet with the 2 recommended therapy. 3 the quality of the evidence to support our 4 therapeutic recommendations. 5 These letters relate to A, low dose inhaled corticosteroids 6 are double blind, randomized, placebo 7 controlled trials in very large populations 8 with a lot of wealth of data; B are double 9 blind, randomized, placebo controlled trials 10 but with much smaller populations with a 11 limited body of data; C are observational and 12 non- randomized studies. 13 D, this is called panel consensus. 14 our group of 18, based upon an extensive 15 review of the literature, we determined what 16 the preferred therapy was for each level of 17 severity and control. 18 And D, when we see So for The preferred therapies were 19 distinguished from alternatives. 20 corticosteroids plus a long-acting 21 bronchodilator are preferred therapies at 22 Step III, and in fact are the only preferred (202) 464-2400 Beta Court Reporting www.betareporting.com Inhaled (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 360 1 therapies at Step IV through Step VI in 2 children five to eleven. 3 What about for 12 and above? For 4 ages 12 and above, the data is more robust. 5 As you can see, evidence A here, for 6 concurrent combination, low dose, inhaled 7 corticosteroids with a long-acting 8 bronchodilator, and these are based upon 9 large, randomized, placebo controlled trials 10 with a wealth of data. 11 What about the alternatives, down 12 here? 13 reviewed all the literature, to tell us which 14 of the alternatives should be a preferred 15 alternative. 16 alternative therapies in alphabetical order, 17 as shown in the figure. 18 There is insufficient data, when we Therefore, we listed all the Once you treat the patient and make 19 a change, you need to have them controlled a 20 minimum of three months. 21 the data in the world and found out that the 22 key was three months. (202) 464-2400 We looked at all If you step down Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 361 1 therapy at an interval of less than three 2 months, the risk of exacerbation increases. 3 So I put him -- I put CJ -- on combination 4 therapy. 5 Now, I previously discussed that we 6 looked at a lot of robust data. Among the 7 data we looked at was the Cochrane analysis, 8 meta-analysis, written in 2005, and it 9 provided us insight into how to make our 10 decisions for the populations who were not 11 well-controlled on monotherapy with a low to 12 high dose inhaled corticosteroid. 13 And this looked at all the English 14 language studies to the time. 15 combination of an ICS and a long-acting 16 bronchodilator versus monotherapy with an ICS 17 resulted in statistically and clinically less 18 exacerbations requiring oral corticosteroids, 19 as the first portion of the upper portion of 20 the slide. 21 shows you that the withdrawal rate is much 22 less in the population who gets combination (202) 464-2400 The The bottom portion of the slide Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 362 1 therapy versus low to increasing to medium or 2 high dose inhaled corticosteroids as 3 monotherapy. 4 So what's happened to CJ? Well, CJ 5 now is 10 years of age. I saw him this past 6 month. 7 on the road. 8 of a low dose inhaled corticosteroid with a 9 long-acting bronchodilator for over 10 I see patients every day when I'm not He's been on concurrent therapy 10 months. 11 put him on this therapy? 12 changed about the middle of last year in the 13 school year. 14 exacerbations. 15 a school day. 16 important thing to CJ is, he's on a soccer 17 team, he's on the traveling soccer team, he 18 does not play goalie. 19 the field. 20 So what's happened to him since I And that was Number one, he has no Number two, he has not missed Number three, the most He runs up and down So where are we? Well, we're with 21 a disease that has 20 million individuals who 22 have the disease at the present time. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 363 1 Additionally, approximately 9.2 million 2 children under the age of 18 have this 3 diagnosis. 4 children in the United States have a 5 diagnosis of asthma. 6 mortality rate has decreased 35 percent. 7 That is one out of every eight Why? In the past decade, the Better education, better 8 understanding of the pathophysiology of the 9 disease, and better therapy. We have a much 10 better understanding of what appropriate 11 therapy, including all medications, should be 12 for the population. 13 So where are we? 14 CJ, the concurrent use of an inhaled 15 corticosteroid, combined with a long-acting 16 bronchodilator, is an effective and safe 17 treatment option for patients who are not 18 well-controlled on monotherapy with low dose 19 ICS, or if they come to me or any other 20 clinician physician with a severity, a 21 classification of moderate persistent asthma, 22 not on controller therapy or on a controller (202) 464-2400 Well, just as in Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 364 1 therapy but not well-controlled, our 2 guideline recommendations are that therapy 3 should be instituted with a low dose inhaled 4 corticosteroid combined with a long-acting 5 bronchodilator. 6 Today, CJ is well, but my concern 7 for CJ and hundreds of my other patients like 8 him is what options will be available to him 9 and other children and adults if there is any 10 restriction to access for long-acting 11 bronchodilators. 12 Thank you. 13 DR. SWENSON: Thank you, Dr. Stoloff. 14 We'll go now to GlaxoSmithKline, and Dr. Jones, 15 who will start off. 16 DR. JONES: Good afternoon. My name 17 is Elaine Jones, and I'm vice president of 18 Regulatory Affairs at GlaxoSmithKline. 19 behalf of GlaxoSmithKline, I would like to thank 20 the Agency and the Advisory Committees for this 21 opportunity to participate in the review of the 22 efficacy and safety data for salmeterol. (202) 464-2400 Beta Court Reporting www.betareporting.com On (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 365 1 Our review of the benefits and 2 risks of salmeterol will encompass data from 3 all approved salmeterol-containing products. 4 Current products available are Serevent 5 Diskus, which contains only salmeterol, and 6 Advair Diskus and Advair HFA, both of which 7 contain salmeterol with the inhaled steroid 8 fluticasone propionate in a single device. 9 As we have heard, after the first 10 approval in 1990 in the United Kingdom, two 11 large surveillance studies of Serevent were 12 conducted. 13 Surveillance Study, or SNS, and the 14 Salmeterol Multicenter Asthma Research Trial, 15 or SMART. 16 increase in serious asthma-related outcomes 17 of Serevent, which helped contribute to the 18 lingering concerns around the safety of 19 salmeterol. They were the Serevent Nationwide In these two studies, there was an 20 These studies did not mandate the 21 concomitant use of inhaled corticosteroids, 22 and their use was not monitored. (202) 464-2400 Beta Court Reporting www.betareporting.com However, (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 366 1 the results suggested that when Serevent was 2 used in patients reporting inhaled 3 corticosteroids at baseline, the risk of 4 serious asthma-related outcomes was 5 mitigated. 6 Immediately following termination 7 of SMART in January 2003, an ADIA health care 8 practitioner letter was distributed informing 9 of the findings. A box warning was 10 subsequently incorporated into the labeling 11 for Serevent and Advair describing the 12 asthma-related deaths in SMART, and 13 suggesting that African-Americans may be at 14 higher risk. 15 As Dr. Seymour presented earlier, 16 in July 2005, the Pulmonary-Allergy Drugs 17 Advisory Committee reviewed salmeterol data 18 including SMART, and comprehensively 19 evaluated the benefit/risk profile of 20 salmeterol in the treatment of asthma. 21 result was a unanimous vote confirming the 22 positive benefit/risk profile of long-acting (202) 464-2400 Beta Court Reporting www.betareporting.com The (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 367 1 agonists for the treatment of asthma in 2 patients four years of age and older. 3 As a result of recommendations of 4 the Committee, additional labeling revisions 5 were made for both Serevent and Advair 6 advising that long-acting beta agonists 7 should only be used in patients when a single 8 controller agent is not sufficient. 9 addition, a medication guide was approved for 10 In distribution to patients. 11 In November 2007, the Pediatric 12 Advisory Committee evaluated the spontaneous 13 adverse event data obtained in the one-year 14 period following the granting of pediatric 15 exclusivity for salmeterol and to the Best 16 Pharmaceuticals for Children's Act. 17 Committee concluded that no new safety 18 signals were identified in the year following 19 the granting of exclusivity. 20 evaluation of pediatric hospitalizations from 21 SMART, and studies with formoterol, resulted 22 in the recommendation that a further (202) 464-2400 The However, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 368 1 evaluation of the benefit/risk profile of 2 long-acting beta agonist-containing products 3 was appropriate. 4 In preparation for this Advisory 5 Committee, GSK has extensively evaluated the 6 benefits and risks of salmeterol-containing 7 products. 8 salmeterol to an inhaled corticosteroid in 9 the treatment of asthma in adults and To explore the benefit of adding 10 children, all clinical studies that were 11 specifically designed to measure efficacy 12 between salmeterol plus inhaled 13 corticosteroids compared with alternative 14 treatment options were analyzed. 15 In July this year, we submitted to 16 FDA a safety database of GSK randomized 17 control trials, including over 200 studies 18 with a salmeterol-containing treatment arm, 19 representing over 100,000 patients. 20 reviewed other databases to fully elucidate 21 the benefit to risk profile of 22 salmeterol-containing products. (202) 464-2400 Beta Court Reporting www.betareporting.com GSK (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 369 1 Dr. Kate Knobil, head of 2 Respiratory Clinical Development and a 3 pulmonologist, will now present the 4 evaluation of the benefit to risk of adding 5 salmeterol to an inhaled corticosteroid in 6 the treatment of asthma in adults and 7 children. 8 9 DR. KNOBIL: Thank you, Dr. Jones. Today, I will discuss the safety and efficacy of 10 salmeterol-containing medicines for the 11 treatment of asthma, which includes both 12 Serevent and Advair. 13 20 years of clinical experience, 14 comprising nearly 60 million patient years of 15 exposure, has demonstrated substantial 16 efficacy and no significant increased risk of 17 asthma-related events when salmeterol is used 18 appropriately with an inhaled corticosteroid. 19 First, I will review the efficacy 20 of salmeterol-containing products when used 21 appropriately with an inhaled corticosteroid, 22 which will include both Serevent and Advair. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 370 1 Then, I will review the methodology for the 2 main safety analysis, followed by the data 3 for Serevent and then for Advair. 4 After the review of the data, 5 Dr. Jones will return to discuss GSK's 6 recommendations. 7 Here is a representative study that 8 shows the benefit in lung function of adding 9 salmeterol to an inhaled corticosteroid. 10 This study compared Advair, shown in yellow, 11 to fluticasone propionate, in green, Serevent 12 in blue, and placebo in white. 13 show the serial FV1 response to treatment 14 over time, with Day One on the left and Week 15 Twelve on the right. 16 These graphs As shown on the left, the onset of 17 action after the first dose of salmeterol was 18 within 30 minutes, and the duration of action 19 was at least 12 hours. 20 lung function improved after chronic dosing 21 with salmeterol, with no diminution of 22 effect. (202) 464-2400 Shown on the right, At week 12, the FEV1 response for Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 371 1 Advair was over 600ml, which is a substantial 2 improvement in lung function. 3 that the effective FP never obtained the same 4 benefit as Advair. 5 Also, note While not shown, symptoms in 6 Albuterol use were also significantly 7 improved over FP alone. 8 single study, but all of the studies that 9 have compared the efficacy of salmeterol plus This is just a 10 an inhaled corticosteroid have consistently 11 shown a benefit over ICS alone. 12 This slide includes all studies in 13 adults and adolescents that were specifically 14 designed to measure efficacy between 15 salmeterol plus ICS compared with alternative 16 treatment options. 17 treatment difference in an individual study. 18 For mean change and morning peak flow, the 19 additional improvement provided by salmeterol 20 added to ICS over that seen with ICS at the 21 same dose is shown in yellow, over higher 22 dose ICS in blue, and over ICS plus add-on (202) 464-2400 Each bar represents Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 372 1 therapy with either Theofolin or Montalucast 2 in white and purple. 3 In all of these studies, there was 4 a statistically significant improvement in 5 lung function for patients receiving 6 salmeterol plus ICS over the alternative 7 treatment options. 8 salmeterol is a very effective 9 bronchodilator, but that is only one measure 10 These data show that of asthma control. 11 All patients with asthma should be 12 provided a short-acting bronchodilator to 13 relieve acute symptoms. 14 percentage of days that patients receiving 15 salmeterol plus ICS did not require rescue 16 Albuterol to treat acute symptoms compared to 17 patients receiving other treatments. 18 of these studies, there was a statistically 19 significant improvement in rescue-free days 20 for patients receiving salmeterol plus ICS 21 over alternative treatment options. 22 (202) 464-2400 Shown here as the In all Extrapolating these data, adding Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 373 1 salmeterol to inhaled corticosteroids would 2 result in two fewer months per year in which 3 patients require rescue medication to treat 4 asthma symptoms. 5 Finally, asthma exacerbations are 6 an indicator of loss of asthma control, and 7 as such, a reduction in exacerbations is one 8 of the most important goals of asthma 9 therapy. Shown here is a fourth plot of a 10 meta-analysis of risk differences, which 11 included 24 studies assessing severe asthma 12 exacerbations requiring oral corticosteroids. 13 Each study is listed on the left with the 14 corresponding risk difference and confidence 15 intervals plotted to the right. 16 An effect that favored the addition 17 of salmeterol is to the left of zero, and an 18 effect that favored ICS alone is to the 19 right. 20 at the bottom, indicated that salmeterol 21 added to inhaled corticosteroids, 22 significantly decreased the risk of severe (202) 464-2400 The combined result, shown in yellow Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 374 1 exacerbations. 2 250 per 10,000 patients. 3 The risk difference was minus This means the 10,000 patients 4 receiving salmeterol plus ICS would 5 experience 250 fewer exacerbations than 6 10,000 patients treated with ICS alone. 7 The authors also reported the odds 8 ratio for this outcome, which was .65, 9 representing a 35 percent reduction in asthma 10 exacerbations. 11 For the efficacy review in 12 children, we included studies compared 13 salmeterol plus ICS compared with a higher 14 dose of ICS. 15 Advair label, which states that children must 16 be symptomatic on an ICS before stepping up 17 to alternatives like Advair. 18 representative study in children aged four to 19 eleven years compared Advair in yellow to 20 twice the approved dose of FP, shown in blue. 21 22 This is consistent with the This Treatment with Advair resulted in significantly improved lung function, and the (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 375 1 results for asthma symptoms and rescue-free 2 days showed similar trends. 3 higher doses of inhaled corticosteroids may 4 not provide better asthma control. 5 Therefore, Exacerbations of asthma are the 6 leading reason for missed school days in 7 children, and that is one reason why 8 preventing exacerbations is so important in 9 this age group. Shown here are the 10 exacerbation rates for all pediatric trials 11 comparing Advair with the same dose or higher 12 dose of inhaled corticosteroid. 13 exception of one trial, treatment with Advair 14 resulted in a similar or lower rate of asthma 15 exacerbations. 16 bottom of the table compared Advair with a 17 higher dose of ICS. 18 With the The three trials at the Because of the risk of dose-related 19 adverse events with ICS such as growth 20 suppression, guidelines recommend that after 21 achieving asthma control, ICS should be 22 titrated to the lowest effective dose. (202) 464-2400 Beta Court Reporting www.betareporting.com The (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 376 1 addition of salmeterol to ICS allows for 2 better overall asthma control and a lower 3 dose of ICS. 4 In summary, results from randomized 5 control trials consistently showed that 6 Advair, or the combination of salmeterol plus 7 ICS, provided better overall asthma control 8 compared with alternative treatment options. 9 As you heard in CJ's case from Dr. Stoloff, 10 improvements in lung function, a decrease in 11 the need for rescue medications, and the 12 prevention of serious asthma exacerbations 13 allows patients to achieve and maintain 14 normal daily functioning, and reduces the 15 number of missed school and work days. 16 these reasons, evidence-based asthma 17 treatment guidelines place long-acting 18 agonists plus ICS as preferred medications 19 for treating patients with persistent asthma. 20 For Before reviewing the safety data 21 for Serevent, I will examine how salmeterol 22 is used in clinical practice today. (202) 464-2400 Beta Court Reporting www.betareporting.com I will (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 377 1 then describe the methods of the 2 meta-analysis of GSK clinical trials. 3 When Serevent was introduced in 4 1994, the role of inflammation and the 5 pathophysiology of asthma was not as widely 6 appreciated, and bronchodilators were often 7 prescribed alone. 8 1994 to 1996, only about one third of 9 Serevent was dispensed with a concurrent As shown on the left in 10 inhaled corticosteroid. 11 dispensed with no other controller medication 12 at all. 13 Another third was The current understanding of asthma 14 has changed since the 1980s and '90s, and 15 now, asthma treatment guidelines and 16 clinicians recognize that all patients with 17 persistent asthma should be treated with an 18 anti-inflammatory agent, preferably an 19 inhaled corticosteroid. 20 as shown on the right, salmeterol is 21 dispensed with an ICS over 98 percent of the 22 time, with the majority of this in the form (202) 464-2400 Today, for asthma, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 378 1 of Advair. 2 salmeterol use is in Advair. 3 Currently, 97 percent of The safety data considered for 4 analysis included all GSK-sponsored clinical 5 studies of salmeterol which were randomized, 6 controlled, double blind, and chronic dosing 7 and design. 8 adjudicated by independent, external 9 physicians based on blinded case narratives. The outcomes of interest were 10 If, in the clinical judgment of the 11 physicians, the outcome could reasonably be 12 considered asthma-related, then it was 13 adjudicated as asthma-related. 14 Today, my review will focus on 15 asthma-related death and hospitalization. 16 Although I will not discuss asthma-related 17 intubation and all-cause death in detail, the 18 results are included in your briefing 19 document. 20 The analysis populations 21 constructed from the studies in the database 22 allow for appropriate comparison of safety (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 379 1 outcomes between the treatments of interest. 2 The comparison of salmeterol-containing 3 treatments versus non-LABA treatments contain 4 215 studies, representing over 106,000 5 patients. 6 LABA treatment, and includes all other 7 treatments, such as ICS, Theofolin, and 8 placebo. 9 heterogeneous. 10 The non-LABA group excluded any This population is the most To aid in the interpretation of a 11 potential treatment effect of salmeterol, 12 more homogeneous populations were 13 constructed. 14 left evaluated salmeterol in the absence of 15 ICS, compared with placebo, also with no ICS. 16 To evaluate the effect of salmeterol in the 17 presence of ICS, the populations were grouped 18 by the way they received ICS in the study, 19 corresponding to the degree of control of the 20 use of ICS; therefore, the level of 21 confidence and adherence to ICS increases as 22 you move from left to right. (202) 464-2400 The groups shown on the far Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 380 1 In the first, salmeterol was added 2 to background inhaled corticosteroids. It is 3 important to understand the difference 4 between ICS given as a background medication 5 and ICS administered as a study drug. 6 Background ICS refers to patients who 7 reported taking an ICS prior to the screening 8 visit and who were instructed to continue 9 that ICS throughout the study. However, 10 there was no systematic reinforcement or any 11 measure of continued adherence to the ICS. 12 The confidence in the adherence ICS 13 as a background medication is lower than in 14 the other treatment comparisons. 15 ICS administered as a study 16 medication with salmeterol in separate 17 blinded inhalers is the next treatment 18 comparison. 19 systematic reinforcement of adherence. 20 However, patients could selectively 21 discontinue either one of the inhalers. 22 (202) 464-2400 In this comparison, there was Finally, salmeterol and ICS could Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 381 1 be administered as a blinded study medication 2 in a single inhaler as Advair. 3 last analysis population evaluated outcomes 4 when the concurrent use of salmeterol and ICS 5 was assured. 6 Only this For safety, I will first review the 7 data for Serevent in the overall population. 8 For our meta-analysis, risk differences will 9 be expressed per 10,000 patients, to be 10 consistent with the way results from SMART 11 have been incorporated into the Serevent and 12 Advair labels. 13 differences per 1,000 patients, and in 14 general, the results reported by FDA and GSK 15 were consistent. 16 FDA reported results as risk The first outcome that I will 17 discuss is asthma-related death. 18 the risk difference comparisons I will show 19 you, the analysis populations and 20 corresponding outcomes will be listed on the 21 left, and the risk differences per 10,000 22 patients will be on the right. (202) 464-2400 Beta Court Reporting www.betareporting.com For all of (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 382 1 For salmeterol versus 2 non-LABA-containing treatments, there were 28 3 asthma-related deaths in patients receiving 4 salmeterol and seven asthma-related deaths in 5 patients receiving non-LABA. 6 but five of the 35 total asthma-related dates 7 occurred in SNS and SMART. 8 were initiated in the early to mid-1990s, 9 which was a time when inhaled corticosteroids 10 were not as routinely used as they are today. 11 Since this top population is the most 12 heterogeneous, the treatment comparisons 13 below this one are more informative for 14 specific treatment regimens with Serevent. 15 Of note, all These studies It is now well-accepted that 16 Serervent should not be used with without an 17 anti-inflammatory medication in persistent 18 asthma. 19 used without an inhaled corticosteroid, had a 20 risk difference of nearly nine per 10,000, 21 and this increase in asthma-related death is 22 already described in the product label. (202) 464-2400 The data showed that Serevent, when Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 383 1 As we move down the next three 2 analysis populations, the confidence and 3 adherence to ICS increases. 4 was added to background ICS, which was not 5 dispensed as part of the study, there were 6 five asthma-related deaths on Serevent and 7 three on ICS. 8 in SMART. 9 When Serevent All of these events occurred When Serevent and ICS were both 10 given the study medications in separate 11 inhalers, there was one asthma-related death 12 with Serevent, and none with inhaled 13 corticosteroid. 14 As a point of reference, Advair 15 data are included on the slide, and I will 16 discuss these data in more detail later in my 17 presentation. 18 assured, there were no asthma-related deaths 19 for patients receiving Advair. 20 When concurrent use of ICS was There were more patients with an 21 asthma-related hospitalization, so this 22 outcome was more amenable to analyzing the (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 384 1 more homogeneous treatment comparisons. 2 Serevent compared with placebo, each without 3 ICS, which again is not acceptable management 4 of persistent asthma because it does not 5 control inflammation, the risk was elevated. 6 For Why would the risk be elevated 7 versus placebo? Serevent is an effective 8 bronchodilator, and chronic use may blunt the 9 patient's awareness of worsening underlying 10 airway inflammation, and could lead to a 11 delay in seeking appropriate medical 12 attention. 13 risk factor for untoward outcomes in asthma. 14 This behavior is an established Similarly, when comparing the use 15 of Serevent added to background ICS, with 16 background ICS alone, the risk difference was 17 also elevated. 18 finding, and a possible explanation for this 19 increased risk may be that patients, because 20 they feel better, selectively discontinued 21 non-study ICS during the clinical trials. 22 (202) 464-2400 This was an unexpected Why would patients taking Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 385 1 salmeterol with background ICS have worse 2 outcomes than background ICS alone? 3 data from table two in your briefing document 4 show that even in the absence of Serevent, 5 the risk of hospitalization was five times 6 greater in patients reporting background use 7 of ICS compared with patients receiving ICS 8 as a study medication. 9 increase in asthma-related death in this These There was also an 10 population. 11 discontinued their background ICS during the 12 study which resulted in severe asthma 13 outcomes. 14 like Serevent improves symptoms, and patients 15 may be more likely to discontinue their 16 inhaled corticosteroid. 17 These data suggest that patients Using an effective bronchodilator This leaves inflammation untreated 18 which is associated with an increase in 19 serious asthma morbidity and mortality. 20 Thus, the inappropriate treatment of asthma, 21 specifically under-use of inhaled 22 corticosteroids, can lead to an increase in (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 386 1 asthma-related outcomes. 2 If we now examine the population 3 where Serevent and inhaled corticosteroid 4 were both administered as study drug in 5 separate devices, there were 16 6 hospitalizations in patients receiving 7 Serevent plus ICS, and 14 hospitalizations 8 with an ICS and study drug. 9 difference of two events in nearly 6,000 10 This was a patients. 11 When salmeterol and inhaled 12 corticosteroid were administered in the same 13 device as Advair, there was no increase in 14 risk. 15 I will now review similar safety 16 information in the pediatric population. 17 the majority of studies, these patients were 18 between the ages of four and eleven. 19 review of Serevent data included 37 studies 20 in over 7,400 children. 21 asthma-related death, and this occurred in a 22 patient that was receiving Albuterol four (202) 464-2400 For The There was one Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 387 1 times daily. 2 Although not discussed in detail 3 for the overall population, we felt it also 4 important to mention the occurrence of 5 intubations in children, since it may be 6 considered a surrogate for risk of 7 asthma-related death. 8 asthma-related intubations in children. 9 occurred in a patient that was receiving There were two One 10 Albuterol four times daily in a different 11 patient than in an asthma-related death, and 12 then the other occurred in a patient 13 receiving Serevent. 14 experiencing these events were receiving 15 concurrent inhaled corticosteroids. 16 None of the children An examination of asthma-related 17 hospitalization found similar results in 18 children as in the overall population. 19 look at the specific treatment comparisons 20 for Serevent compared with placebo, the 21 number of events was low, but resulted in an 22 elevated risk difference. (202) 464-2400 If we When comparing the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 388 1 use of Serevent added to non-study background 2 ICS with background ICS alone, the risk 3 difference was also elevated in this 4 population. 5 The most plausible explanation, as 6 discussed for the overall population, is that 7 more patients on Serevent selectively 8 discontinued their background inhaled 9 corticosteroid during this study, thus 10 leaving airway inflammation untreated. 11 If we now examine the population 12 where Serevent and ICS were both administered 13 as a study drug in separate devices, there 14 was one hospitalization in children receiving 15 Serevent plus ICS, and two in the children 16 receiving ICS alone. 17 one hospitalization in children receiving 18 Advair compared with two receiving ICS alone. 19 Similarly, there was Concurrent use of Serevent and 20 inhaled corticosteroids was shown to be 21 highly effective for the treatment of asthma 22 in adults and children by demonstrating (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 389 1 substantial improvements in lung function, 2 decreased symptoms, and a decrease in asthma 3 exacerbations. 4 not meet the needs of all patients. 5 are patients who need a different ICS than 6 fluticasone propionate, or need a different 7 dose of ICS than is available in Advair. 8 also allows for easier titration of ICS for 9 patients who require frequent changes in A fixed dose combination may There It 10 their regimen. 11 relatively small number, it is important that 12 those individuals who cannot use a 13 combination inhaler have effective options 14 available to them. 15 risk of serious asthma outcomes with Serevent 16 when ICS use was not controlled. 17 there was no safety signal when Serevent was 18 used concurrently with inhaled 19 corticosteroids. 20 Although it may be a There was an increased However, GSK acknowledges that there is a 21 question as to whether Serevent should 22 continue to be indicated for the treatment of (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 390 1 asthma. 2 thus restricting the use of salmeterol and 3 asthma to the combination product Advair 4 would ensure the appropriate use of Serevent 5 with ICS. 6 for physicians and patients and could have a 7 negative impact on patient care. 8 Removing the asthma indication and However, this would limit choice Balancing these considerations, we 9 favor continued availability, and have 10 proposed the labeling changes and will 11 discuss a plan of action to reinforce that 12 Serevent should only be used concurrently 13 with inhaled corticosteroids. 14 Patient safety is our top priority, 15 and we believe that the benefits demonstrated 16 by Serevent, used concurrently with an 17 inhaled corticosteroid in asthma, outweigh 18 the potential risks in the population of 19 patients who need broader treatment options. 20 You've already seen some of the 21 data with Advair, and now I will provide a 22 more comprehensive review. (202) 464-2400 Since the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 391 1 majority of current salmeterol use is in 2 Advair, the meta-analysis of safety data from 3 the Advair clinical trials is more reflective 4 of how salmeterol is used today. 5 of the meta-analysis of the Advair data for 6 the overall population are shown here. 7 The results There were no asthma-related deaths 8 in over 22,000 patients receiving Advair or 9 ICS. Further, there was no increased risk of 10 asthma-related hospitalization in this 11 population. 12 there were no asthma-related intubations in 13 patients receiving Advair. 14 Although not shown on the slide, I will now review the meta-analysis 15 results for Advair in children. 16 the overall population as shown on the 17 previous slide, the results for Advair in 18 children were positive. 19 asthma-related deaths in over 2,400 children 20 receiving Advair or ICS. 21 no increase in risk of asthma-related 22 hospitalization in children. (202) 464-2400 Similar to There were no Further, there was Although not Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 392 1 shown on the slide, there were no 2 asthma-related intubations in children 3 receiving Advair. 4 African-Americans experience 5 comparatively greater asthma morbidity and 6 mortality than Caucasians and other ethnic 7 groups. 8 serious asthma-related outcomes for 9 African-Americans receiving salmeterol, and SMART found an increased risk of 10 nearly all of these events occurred in those 11 not reporting use of ICS at baseline. 12 In response to SMART, GSK conducted 13 a yearlong study in African-Americans to 14 examine the rate of asthma exacerbations 15 between Advair and FP. 16 significant difference in the rate of asthma 17 exacerbations in patients receiving Advair 18 compared with FP, demonstrating that 19 appropriate use of salmeterol with an inhaled 20 corticosteroid was not associated with an 21 increased risk of exacerbations in 22 African-Americans. (202) 464-2400 The results showed no Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 393 1 We also examined the number of 2 asthma-related hospitalizations in all 3 African-American patients in all studies 4 comparing Advair and ICS. 5 increase in asthma-related hospitalizations 6 with the addition of salmeterol to ICS. 7 These results were similar to the overall 8 population. 9 that African-Americans are at increased risk There was no There was no evidence to suggest 10 for serious asthma-related outcomes when 11 using Serevent or salmeterol concurrently 12 with ICS. 13 I will now move to the results of 14 studies of use of Advair in clinical 15 practice. 16 Advair, odds ratios were reported, so they'll 17 be presented here in the same manner. For observational studies of 18 In order to assess the overall 19 effect of the use of Advair in adults in 20 clinical practice within the U.S., a 21 meta-analysis was undertaken of all 22 observational cohort studies that compared (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 394 1 Advair to ICS. 2 comprising nearly 83,000 patients met the a 3 priori criteria for study inclusion, with 4 59,000 receiving Advair. 5 Advair resulted in a combined odds ratio of 6 .84, which corresponds to a 16 percent 7 decrease in patients with an asthma-related 8 emergency department visit compared to 9 patients treated with ICS. 10 A total of four studies Treatment with For asthma-related 11 hospitalizations, the combined odds ratio 12 shown in yellow on the lower half of the 13 screen was .85, corresponding to a 15 percent 14 decrease in patients with asthma-related 15 hospitalizations with Advair compared to 16 patients treated with ICS. 17 that Advair significantly reduced severe 18 events over and above that seen with ICS 19 alone. 20 These data show In pediatrics, a similar 21 meta-analysis was performed, including over 22 43,000 children and adolescents, with over (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 395 1 16,000 receiving Advair. 2 included diverse populations ranging from 3 Medicaid patients to patients with mild 4 asthma. 5 asthma-related hospitalizations observed in 6 these studies, only the combined endpoint of 7 asthma-related emergency department visits 8 and hospitalizations is shown. 9 The analysis Due to the low frequency of The studies above the dotted line 10 compared Advair to ICS alone. 11 odds ratio was .91, which corresponds to a 12 9 percent decrease in the number of patients 13 with an emergency department visit or 14 hospitalization. 15 Advair was significantly more effective in 16 reducing this outcome in children than ICS 17 alone. 18 The combined This analysis showed that The studies below the dotted line 19 compared Advair to ICS plus Montelukast in 20 children. 21 which corresponds to a 54 percent decrease in 22 the number of patients with an emergency (202) 464-2400 The combined odds ratio was .46, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 396 1 department visit or hospitalization, again 2 showing that Advair was significantly more 3 effective than the combination of ICS plus 4 Montelukast in reducing this outcome in 5 children. 6 For Advair, the benefit to risk 7 profile is positive for adults and children. 8 Advair is highly effective in the treatment 9 of asthma, and has demonstrated substantial 10 efficacy in improvement in lung function, 11 decreased symptoms and a decrease in 12 exacerbations. 13 Use of Advair was also associated 14 with significant reductions in emergency 15 department visits and hospitalizations versus 16 ICS alone in adults and children, and also a 17 reduction in a combined emergency department 18 visit and hospitalization versus ICS plus 19 Montelukast in children. 20 alternative treatment options may not be more 21 efficacious. 22 (202) 464-2400 Therefore, The combination of salmeterol and Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 397 1 FP in a single device ensures concurrent use 2 of the ICS, and does not allow selective 3 discontinuation of ICS. 4 that Advair is associated with the risks 5 identified in the early studies of Serevent. 6 This comprehensive review of Advair, which 7 included nearly 18,000 patients from 8 randomized clinical trials, as well as over 9 75,000 patients from observational studies, There is no evidence 10 did not identify any increased risk 11 associated with Advair. 12 asthma-related deaths reported in children or 13 adults taking Advair, and there was no 14 increase in asthma-related hospitalizations. 15 While I did not show the data, there were no 16 asthma-related intubations and no increase in 17 all-cause death. 18 There were no The case for Advair is clear. 19 Substantial efficacy has been demonstrated, 20 and there is no increased risk of untoward 21 asthma outcomes. 22 more complex. (202) 464-2400 The case for Serevent is We know that it is Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 398 1 inappropriate to use long-acting beta 2 agonists alone to treat asthma, and the data 3 show that the use of Serevent alone is 4 associated with increased risk. 5 However, there was no significant 6 increase in risk when Serevent was used 7 concurrently with inhaled corticosteroids. 8 For the relatively small number of patients 9 who deserve access to guideline preferred 10 therapy, but whose needs cannot be met with 11 Advair, it is important that Serevent remains 12 available. 13 I would now like to turn the podium 14 back over to Dr. Jones, who will discuss the 15 actions that GSK is recommending to ensure 16 appropriate use of this important medication. 17 DR. JONES: As Dr. Knobil has shown, 18 there is a positive benefit to risk profile when 19 salmeterol is used concurrently with an inhaled 20 corticosteroid. 21 ensure the concurrent use of inhaled 22 corticosteroids is by the fixed-dose combination (202) 464-2400 The most effective way to Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 399 1 products. 2 While Advair meets the medical 3 needs of most patients, there are some that 4 require separate inhalers, and for these 5 individuals, it is important that Serevent 6 remains available. 7 the following actions to prevent 8 inappropriate use of Serevent in the 9 treatment of asthma. 10 Therefore, GSK recommends The first action, already 11 undertaken, was submission of a labeling 12 supplement. 13 September and is currently under review. 14 indications section was revised to restrict 15 the use of Serevent in patients with asthma 16 to concomitant therapy with an inhaled 17 corticosteroid. 18 This action was taken in The Also, information concerning the 19 risk of asthma-related hospitalizations was 20 added to the box warning that already 21 included a warning for asthma-related death. 22 (202) 464-2400 Finally, for patients, the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 400 1 medication guide was strengthened to 2 reinforce that if using Serevent Diskus to 3 treat asthma, an inhaled corticosteroid must 4 be used every day. 5 patients must continue taking an inhaled 6 corticosteroid every day, and must not stop 7 or decrease the dose even if they feel 8 better. It also emphasizes that 9 The next step is the enhancement of 10 our risk management approach, which builds on 11 the important changes to the label that I've 12 just reviewed. 13 targets health care providers, pharmacies, 14 and patients. 15 addition to the labeling revisions, GSK is 16 proposing to communicate with all health care 17 providers who prescribed or dispensed 18 Serevent in the last year, emphasizing 19 concurrent use of Serevent with inhaled 20 corticosteroids. 21 initiate educational programs emphasizing 22 that Serevent should only be used with an (202) 464-2400 The comprehensive plan For health care providers, in In addition, we will Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 401 1 inhaled corticosteroid. 2 Next, GSK will work with managed 3 care and pharmacy benefit management 4 organizations and retail pharmacies. 5 initiative will change fulminary (?) 6 Algorithms, update pharmacy computer systems, 7 and inform physicians of these changes. 8 would result in the pharmacist being alerted 9 that Serevent is being prescribed without an 10 inhaled corticosteroid and should have 11 contact with a physician. 12 This This Finally, to communicate to patients 13 directly, in addition to revisions to the 14 medication guide, GSK proposes changes to the 15 packaging for Serevent Diskus, alerting 16 patients that in the treatment of asthma, 17 Serevent must only be used as concomitant 18 therapy with an inhaled corticosteroid. 19 This risk management plan will 20 enable the safe and appropriate use of 21 Serevent in the treatment of asthma. 22 (202) 464-2400 In conclusion, we have presented Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 402 1 data showing Advair and Serevent plus inhaled 2 corticosteroids provide substantial benefits 3 to patients with asthma, and these medicines 4 have significantly advanced the care and well 5 being of both adults and children. 6 medications remain a preferred treatment 7 option in the NHLBI evidence-based asthma 8 treatment guidelines. 9 these life changing medicines would be 10 extremely detrimental to patient care. 11 critical that these medicines continue to be 12 available to maintain the high standard of 13 care that is currently available to patients 14 with asthma. 15 Thank you. 16 DR. SWENSON: These Denying access to It is Now we'll turn just with 17 a slight departure from the schedule, to the 18 representatives from Novartis, who I think we'll 19 begin with Mathias Hukkelhoven to begin the 20 presentation. 21 22 DR. HUKKELHOVEN: Dr. Swenson, Dr. Rappley, members of the three Advisory (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 403 1 Committee meetings, members of the FDA and 2 guests, good afternoon. 3 Hukkelhoven. 4 Regulatory Affairs at Novartis. 5 to present today on behalf of my 6 colleagues -- and Linda Armstrong will present 7 as well -- the safety and efficacy data for 8 Foradil. 9 My name is Mat I'm the Global Head of Drug It's a pleasure To begin with, I will present the 10 regulatory history of Foradil, part of which 11 has actually already been very well-explained 12 by Dr. Seymour. 13 Foradil was first approved in 1990 14 in France, and is currently approved in over 15 80 countries. 16 approved in 2001 under the tradename Foradil 17 Aerolizer for the maintenance treatment of 18 asthma in patients five years of age and 19 older, for the acute prevention of 20 exercise-induced bronchospasm, and for the 21 maintenance treatment of Chronic Obstructive 22 Pulmonary Disease (COPD). (202) 464-2400 In the U.S., Formoterol was Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 404 1 Foradil Aerolizer consists of 2 capsules for oral inhalation containing 12mcg 3 of Formoterol and a delivery device, a 4 single-dose dry powder inhaler, the 5 Aerolizer. 6 daily, which corresponds to a total daily 7 dose of 24mcg. 8 NDA holder for Foradil Aerolizer, and 9 Schering-Plough Corporation markets the The approved dose is 12mcg twice In the U.S., Novartis is the 10 product. 11 multi-dose dry powder inhaler which contains 12 10mcg of Formoterol per activation, is 13 approved at the dose of 10mcg BID, 14 corresponding to a total daily dose of 20mcg. 15 In addition, Foradil Certihaler, a It was approved in 2006 for the 16 maintenance treatment of asthma five years of 17 age and older, but it is not marketed. 18 mention the total daily dose, TDD, as this is 19 how the results of the clinical data analysis 20 will be presented. 21 22 I Outside of the U.S., Foradil Aerolizer is approved in many countries at 12 (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 405 1 to 24mcg BID, which corresponds to a TTD of 2 24 to 48mcg for the prophylaxis and treatment 3 of asthma in patients five years of age and 4 older, for the prophylaxis of bronchospasm 5 induced by inhaled allergens, cold air, or 6 exercise, and for COPD. 7 Note that the 24mcg BID dose which 8 is twice the approved dose in the U.S. is 9 only recommended in adults. 10 In some countries, Formoterol is 11 also approved as an inhalation aerosol in the 12 MDI, and as Foradil Certihaler which is also 13 not marketed ex U.S. 14 approved for multiple indications, the focus 15 of today's meeting is of course asthma, and 16 the analyses to be presented are based on the 17 studies in asthma. 18 Although Formoterol is Events leading up to the meeting 19 today following approval of Foradil Aerolizer 20 in 2001 include the Pulmonary Advisory 21 Committee meeting in 2005 to discuss the 22 safety of LABAs, and the subsequent labeling (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 406 1 revisions for Foradil in 2006. In 2007, the 2 Pediatric Advisory Committee meeting raised, 3 as has been discussed before, concerns about 4 safety of LABAs in pediatric patients with 5 asthma, and recommended that this be further 6 discussed at a subsequent meeting. 7 result, early this year, FDA requested 8 information regarding controlled clinical 9 trials conducted with LABA for manufacturers As a 10 of LABA containing products. 11 will be presenting this data, specifically 12 the events of interest identified by FDA 13 which we have termed "serious asthma 14 exacerbations." 15 Dr. Armstrong I would like to focus for a moment 16 on the 2006 labeling revisions. 17 time, a box warning was added to the Foradil 18 label. 19 possible increased risk of asthma-related 20 death for LABA as a class based on the signal 21 seen with salmeterol in the SMART study. 22 also clearly advocates the concomitant use of (202) 464-2400 At that The box warning addresses the Beta Court Reporting www.betareporting.com It (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 407 1 inhaled corticosteroids with Formoterol. 2 The label revisions of 2006 also 3 included the addition of data on the 4 incidence of serious asthma exacerbations 5 from the three pivotal Foradil registration 6 trials, including data from the one-year 7 pivotal pediatric study which you see here, 8 and where we did observe a higher incidence 9 of serious asthma exacerbations in the 10 Formoterol groups compared to placebo. 11 Note that all three registration 12 studies were conducted in the 1990s before 13 inhaled corticosteroids became the standard 14 of care. 15 incidences of serious asthma exacerbations 16 outlined in the label differ from those 17 outlined in the briefing document provided by 18 Novartis in preparation for this meeting. 19 Those in the label include all events up to 20 30 days following the end of treatment, and 21 those in the briefing book included events to 22 the end of blinded treatment, as requested by (202) 464-2400 Please also note that the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 408 1 FDA. 2 The label update also included the 3 results of two pivotal registration studies 4 in adolescents and adults, and in these two 5 trials, we saw no difference in serious 6 asthma exacerbations in the 12mcg BID group 7 compared to placebo. 8 was a small increase in serious asthma 9 exacerbations in the highest dose, the 24mcg However, since there 10 BID, FDA asked for a post-marketing study 11 specifically investigating serious asthma 12 exacerbations at the approved dose and twice 13 the approved dose. 14 In this slide, you will see the 15 labeling resulting from this post-marketing 16 study which was completed in 2004. 17 can see, there was no dose response observed, 18 and the incidence of serious asthma 19 exacerbations was actually very low. 20 As you Since approval of Foradil 21 Aerolizer, a number of activities have been 22 completed or are ongoing to evaluate and (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 409 1 minimize the risk of serious asthma 2 exacerbations. 3 post-marketing commitment study that I just 4 mentioned, which is reflected now in the 5 label and will be discussed in greater detail 6 by Dr. Armstrong in the next presentation. 7 These include a Also, we did pharmacogenetic 8 analysis looking at polymorphesence (?) in 9 the receptor from two Phase III Formoterol 10 studies which was submitted to the FDA 11 through a voluntary genomic data submission; 12 epidemiological studies to investigate the 13 incidence of serious asthma exacerbations, 14 asthma related emergency room visits, 15 hospitalizations, and intubations; global 16 pharmaco-vigilance to review and assess 17 events, with a focus on asthma-related 18 serious adverse events, and that for 19 Formoterol, revised labeling as explained as 20 well as the addition of a medication guide in 21 2006 to reflect the potential risks and to 22 describe the appropriate use of Formoterol. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 410 1 And finally, education and 2 communication of risks through a website that 3 educates prescribers and consumers on the 4 appropriate use of Formoterol. 5 Our clinical presentation by 6 Dr. Armstrong will show that based on a 7 review of clinical data from Novartis and 8 publications for Formoterol, in the context 9 of the current treatment guidelines and the 10 approved label, and based on our 11 post-marketing surveillance data both from 12 the Novartis ARGUS database as well as the 13 FDA AERS database, we conclude that Foradil 14 continues to exhibit a favorable benefit/risk 15 ratio. 16 intubations were observed in children in any 17 of our controlled studies with Formoterol, so 18 all serious asthma exacerbations were 19 hospitalizations. 20 Please note that no deaths or Furthermore, Formoterol remains an 21 important therapeutic option in the treatment 22 of patients with asthma. (202) 464-2400 Further, that the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 411 1 continued availability of non-combined LABAs 2 provide flexibility in choosing the type and 3 dose of inhaled corticosteroids, and that the 4 appropriate use of Foradil is adequately 5 outlined in the current labeling. 6 Based on the outcome of this 7 Advisory Committee meeting today and 8 tomorrow, we will work with FDA to update the 9 label and medication guide as appropriate. 10 Dr. Linda Armstrong from our 11 Clinical Development and Medical Affairs 12 Department at Novartis will now present our 13 clinical data on Foradil. 14 Gary Cook from the University of North 15 Carolina and Dr. James Kemp, clinical 16 professor of pediatric oncology, are also 17 attending on behalf of Novartis and are 18 available to answer any questions. 19 Dr. Armstrong? 20 DR. ARMSTRONG: 21 As Dr. Hukkelhoven has just 22 In addition, Dr. Thank you. mentioned, our data support that Foradil, (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 412 1 when used according to the product label and 2 treatment guidelines, continues to have a 3 positive benefit/risk profile, and should 4 remain available for the treatment of asthma. 5 I will start with a review of the 6 efficacy and safety data from the three 7 pivotal trials that were the basis for 8 Foradil's approval in 2001, and a safety 9 study that was conducted as a post-marketing 10 commitment. 11 data that Novartis provided to the FDA. 12 will review the use of the drug in the United 13 States and reports of spontaneous 14 asthma-related events. 15 address the overall benefit/risk profile of 16 long-acting beta agonists, including Foradil, 17 in the treatment of asthma. 18 I will then describe the pool I And finally, I will The asthma treatment paradigm has 19 evolved, and the treatment guidelines have 20 been updated over the 17 years in which 21 Novartis has conducted Foradil trials. 22 the pivotal trials were designed and (202) 464-2400 Beta Court Reporting www.betareporting.com When (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 413 1 conducted between 1995 and 1998, inhaled 2 corticosteroids were not yet considered the 3 standard of care for asthma, and in fact, 4 regularly dosed Albuterol was widely used to 5 treat patients with persistent disease and 6 was used as an active comparator in many of 7 our trials. 8 9 As Dr. Lemanske stated, regularly dosed Albuterol in the Baggs and Barts Study 10 was found to be no more effective than PRN 11 dosing, and might be harmful in some 12 patients. 13 These are the three pivotal studies 14 that supported the registration of Foradil. 15 Remember that given the time frame in which 16 these studies were conducted, only about half 17 the patients reported inhaled corticosteroid 18 use at any time during this treatment period. 19 In these trials, approximately 500 20 adolescent and adult patients, randomized to 21 receive Foradil either 12mcg twice daily, the 22 approved dose, 24mcg twice daily, Albuterol, (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 414 1 dosed regularly four times a day, or placebo. 2 Patients in all of the treatment groups were 3 to remain on their stable controller 4 medications, including inhaled 5 corticosteroids, and were permitted to use 6 Albuterol as needed. 7 As you can see from the results of 8 Trial 40, Foradil showed superior efficacy 9 compared to placebo and Albuterol after 12 10 weeks of treatment. 11 measured by lung function, improved symptom 12 control, and decreased use of rescue 13 medication. 14 This improvement was After 12 weeks, even before 15 treatments, patients treated with Foradil 16 12mcg, shown here in green, demonstrated a 17 significantly higher baseline FEV1. 18 dosing, FEV1 remained significantly higher 19 than placebo, which is shown in white. 20 contrast, patients treated with Albuterol, 21 shown in blue, had an initial improvement 22 after dosing which returns to baseline within (202) 464-2400 Beta Court Reporting www.betareporting.com After In (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 415 1 four hours. 2 These may be associated with increased 3 symptoms during the course of the day or 4 overnight. 5 Note the troughs with Albuterol. On average, patients treated with 6 Foradil demonstrated a 22 percent improvement 7 in FEV1 from baseline. 8 showed that within 60 minutes of dosing, 9 89 percent of patients demonstrated an A responder analysis 10 improvement in FEV1 that was clinically 11 significant. 12 This improvement in lung function 13 was also associated with improvements in 14 other measures of asthma control, including 15 asthma symptoms and rescue medication use. 16 The Y axis shows the median symptom score 17 which was measured on a five point scale from 18 zero to four. 19 was one. 20 30 percent reduction in symptoms. At baseline, the median score Patients treated with Foradil had a 21 This improvement was associated 22 with a decrease in rescue medication use. (202) 464-2400 Beta Court Reporting www.betareporting.com In (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 416 1 contrast, patients treated with Albuterol had 2 a smaller reduction in symptoms, and as 3 expected, patients on placebo remained 4 symptomatic. 5 In this graph, the Y axis is median 6 inhalations of rescue medication. At 7 baseline, most patients use between one to 8 two inhalations during the day. 9 of the trial, rescue medication use was By the end 10 60 percent lower in the Foradil group 11 compared to placebo. 12 A similar trend was seen for 13 nighttime symptoms and rescue medication use. 14 Because many patients with asthma are 15 troubled by symptoms at night, the impact of 16 a medication on nocturnal symptoms and rescue 17 medication use is of particular clinical 18 interest. 19 symptom score for nocturnal symptoms. 20 baseline, the median score was .7. 21 treated with Foradil had a 70 percent 22 reduction in symptoms. (202) 464-2400 Here, the Y axis shows the median At Patients This was associated Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 417 1 with fewer nocturnal awakenings and a 2 decrease in rescue medication use. 3 In the Y axis of the next graph, 4 the median inhalations of rescue medication 5 during the night are recorded. 6 again, most patients use one to two 7 inhalations per night. 8 trial, rescue medication use at night was 9 70 percent lower in the Foradil group At baseline, By the end of the 10 compared to placebo. Although not measured 11 in this study, patients treated with Foradil 12 had demonstrated significant improvements in 13 quality of life compared with placebo in 14 other clinical trials. 15 Turning now to safety, there were 16 no asthma-related serious adverse events in 17 patients receiving Foradil 12mcg twice daily. 18 There were four asthma-related serious 19 adverse events in the 48mcg group, two in the 20 Albuterol group, and no events in the placebo 21 group. 22 one event in the approved dose group, five (202) 464-2400 In the other pivotal trial, there was Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 418 1 events in the higher dose group, none in the 2 Albuterol group, and two in the placebo 3 group. 4 These are the results of the trials 5 that were the basis for the adult and 6 adolescent approval in 2001. 7 pediatric data look like? 8 9 What does the Novartis has conducted seven trials in children to support approvals. Most of 10 the placebo controlled data comes from 11 Study 49 which was a one-year study conducted 12 in children five to twelve. 13 were randomized to four 24mcg twice daily, 14 12mcg twice daily, or placebo. 15 required to use concomitant anti-inflammatory 16 medications, although compliance with these 17 medications was not monitored. 18 518 children Patients were 75 percent reported using inhaled 19 corticosteroids at any time during the 20 treatment period and mostly they were on low 21 dose becla-methasone, triamcinolone, and 22 budesonide. (202) 464-2400 As with the overall population Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 419 1 in our studies, pediatric patients treated 2 with Foradil in Study 49 demonstrated 3 significant improvements in average FEV1. 4 Patients treated with 12mcg BID demonstrated 5 a 150ml improvement over placebo, which is a 6 10 percent improvement in lung function. 7 This improvement was shown after 12 weeks of 8 treatment, which is the primary endpoint, and 9 maintained throughout the one-year trial. 10 12 micrograms twice daily was 11 selected as the lowest effective dose with 12 the greatest benefit/risk profile. 13 higher doses are approved for adults in some 14 countries, 12mcg twice a day is the maximum 15 daily dose for children worldwide. 16 Although The overall improvement in 17 pediatric patients was confirmed in analysis 18 of daytime symptom reduction. 19 one-year trial, Foradil decreased daytime 20 symptom scores by more than 50 percent. 21 improvement in symptoms was accompanied by a 22 similar reduction in rescue medication use. (202) 464-2400 Throughout the Beta Court Reporting www.betareporting.com This (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 420 1 Patients did not have to be 2 symptomatic for inclusion in this trial, 3 which explains the relatively mild symptoms 4 at baseline, but in those patients with 5 baseline symptoms, the reduction in symptom 6 and rescue medication use is clinically 7 relevant. 8 9 Deterioration days is a composite endpoint that was used in this study to 10 describe days in which patients use more than 11 four puffs of rescue medication, had a 12 symptom score greater than one, or used 13 additional controller medications. 14 treated with Foradil had a 45 percent 15 reduction in deterioration days, confirming 16 the benefit in this population. 17 Patients Once again, Foradil also decreased 18 nighttime symptom scores by almost 19 50 percent, and additionally decreased the 20 use of nighttime rescue medication. 21 improvements persisted throughout the 22 one-year trial. (202) 464-2400 Beta Court Reporting www.betareporting.com These (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 421 1 In the safety analysis of 49, there 2 were more asthma-related SAEs among patients 3 receiving Foradil compared with placebo. 4 Note that there were no asthma-related SAEs 5 among patients randomized to placebo. 6 is unusual, as other long-term studies in 7 children have demonstrated asthma-related 8 hospitalization in the 1 to 4 percent range 9 for placebo-treated patients. This Eight subjects 10 randomized to 12mcg twice daily, and 11 11 randomized to 24 experienced asthma-related 12 SAEs throughout the course of the one-year 13 trial. 14 during this trial. 15 There were no deaths or intubations Although there were no SAEs among 16 the placebo patients, the premature 17 discontinuation rate was actually higher 18 among placebo patients than those who were 19 treated with Foradil. 20 quite symptomatic, often describing severe 21 symptoms with increased rescue medication use 22 and significant reductions in peak flow, who (202) 464-2400 These patients were Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 422 1 required multiple courses of oral 2 corticosteroids. 3 Patients who discontinued were not 4 followed after dropping out of the trial. A 5 disproportionate dropout in patients may have 6 lead to a healthy survivor effect in the 7 placebo arm, meaning that better-controlled 8 patients remained in the study and were less 9 likely to experience serious adverse events. 10 These limitations do not fully account for 11 the imbalance that we saw in Study 49; 12 however, they do make it more difficult to 13 fully assess the magnitude of risk associated 14 with Foradil in this study. 15 the basis for approval for Foradil in 16 patients five to twelve, and is reflected in 17 the current label. 18 This study was Studies 40, 41, and 49 were 19 conducted 15 years ago. 20 Novartis conducted Study 2307 as a 21 post-marketing commitment. 22 to further assess whether an imbalance in (202) 464-2400 After approval, It was designed Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 423 1 asthma-related hospitalizations, seen between 2 the 12 and 24mcg doses, would be confirmed in 3 a larger safety study. 4 Four times the number of patients 5 in the pivotal trials were enrolled in this 6 study. 7 older were randomized to receive Foradil 8 12mcg twice daily, 12mcg twice daily plus PRN 9 Foradil, Foradil 24mcg twice daily, or 2085 patients 12 years of age and 10 placebo. 11 criteria were matched to approximate the 12 populations enrolled in the earlier studies. 13 The inclusion and exclusion Patients were to continue their 14 anti-inflammatory medications. 15 were taking inhaled corticosteroids, compared 16 to 50 percent in the earlier trials. 17 number of events was lower than anticipated, 18 and the differences between the treatment 19 groups were similar. 20 in the 12mcg group, one in the 12mcg plus PRN 21 group, two in the 24mcg group, and one in the 22 placebo group. (202) 464-2400 70 percent The There were three events Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 424 1 These four studies are the largest 2 studies included in the pooled data set 3 requested by the FDA. 4 the FDA's pre-specified criteria. 5 randomized, blinded asthma studies in which 6 Foradil was delivered by the Aerolizer or the 7 Certihaler devices. 8 were randomized in these trials. 9 45 clinical trials met They were More than 8,000 subjects Approximately 5,300 received 10 Foradil, 2,000 received placebo, and almost 11 1,000 received Albuterol, delivered primarily 12 as a 180mcg four times a day. 13 In all groups, patients were 14 permitted to use Albuterol PRN and continue 15 their controller medications. 16 studied at three dose strengths, 12mcg total 17 daily dose, which was studied as either 6mcg 18 twice daily or 12mcg as a single daily dose; 19 24mcg total daily dose, the approved dose in 20 the United States studied as 12mcg twice 21 daily or 24 as a single daily dose; 48mcg, 22 studied as 24mcg BID. (202) 464-2400 Foradil was We'll focus on Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 425 1 comparisons between the marketed dose of 2 Foradil with placebo and also regularly dosed 3 Albuterol. 4 Although the pooled safety database 5 provides important information, it doesn't 6 address the question of whether or not any 7 inhaled corticosteroids mitigate the risk of 8 serious exacerbations. 9 most of the Novartis studies occurred prior 10 to the time when ICS became the standard of 11 care, and in our database, inhaled 12 corticosteroid use was permitted in 41 of the 13 45 trials, but were required in only four 14 studies, including only two multiple dose 15 studies. 16 not provided by the sponsor, nor was ICS 17 compliance monitored. 18 As you may recall, In all the other studies, ICS were The focus of the pooled analysis is 19 the asthma composite endpoint, which was 20 defined by the FDA as asthma-related deaths, 21 intubation, or hospitalization. 22 from events described in our label in that (202) 464-2400 Beta Court Reporting www.betareporting.com This differs (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 426 1 asthma-related SAEs may also include other 2 medically important events such as ER visits. 3 Events occurring between 4 randomization and the final day of double 5 blind treatment for parallel group studies, 6 or the end of the first period for crossover 7 studies, were identified by blinded physician 8 review of the clinical trial database. 9 In contrast, our label includes 10 SAEs reported up to 30 days after 11 discontinuation of the study medication. 12 our analyses, we've compared Foradil to 13 placebo and regularly dosed Albuterol, while 14 the FDA has compared LABAs to non-LABAs. 15 In In the 45 trials, there were no 16 pediatric deaths or intubations. 17 one asthma-related adult death. 18 66-year-old woman had a fatal asthma 19 exacerbation 19 days after randomization to 20 Foradil 48mcg, the higher dose, which is not 21 approved in the United States. 22 (202) 464-2400 There was A So what is the effect of Foradil on Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 427 1 non-fatal asthma exacerbations? This slide 2 shows the incidence and rate of the asthma 3 composite endpoint for patients overall and 4 for different age groups. 5 presented here as the incidence or percent of 6 patients with events, and the rate corrected 7 for exposure presented as events per 100 8 patient years. 9 the 24mcg total daily dose, the approved U.S. Events are Among patients randomized to 10 dose,.7 percent met the asthma composite 11 endpoint, compared to .4 percent of 12 placebo-treated patients. 13 regularly dosed Albuterol-treated patients 14 experienced serious asthma exacerbations. 15 When corrected for exposure, there were 2.7 16 events per 100 patient years in the 17 Formoterol group, compared to 1.4 in the 18 placebo group, and 4.5 in the regularly dosed 19 Albuterol group. 20 One percent of If we assess asthma composite 21 endpoints by age, among adults greater than 22 18, there were similar event rates. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 428 1 0.4 percent of patients in the 2 approved dose met the endpoint with 1.9 3 events per 100 patient years, compared to 4 .3 percent of placebo patients with 1.4 5 events per 100 patient years, and .7 percent 6 of Albuterol treated patients with three 7 events per 100 patient years. 8 9 10 Adult studies generally enrolled subjects greater than 12. We analyzed adolescents separately. 11 There were relatively few patients 12 in this age group, but the number of events 13 is similar across all treatment arms. 14 Among children five to twelve, the 15 number of patients who met the asthma 16 composite endpoint was increased among 17 patients receiving Formoterol and regularly 18 dosed Albuterol compared to placebo, 19 2.4 percent of patients randomized to 20 Formoterol experienced events, with 5.2 21 events per 100 patient years, and 3 percent 22 of patients randomized to Albuterol, with 16 (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 429 1 events per 100 patient years, compared to 2 only .3 percent or .5 events per 100 patient 3 years in placebo. 4 This is a forced plot of the crude 5 odds ratio versus placebo for all subjects by 6 age. 7 An odds ratio favoring Foradil is on the left 8 and placebo is on the right. 9 Formoterol, the odds ratio versus placebo was The scale on the X axis is logarithmic. Overall for 10 1.7. 11 than 18 years. 12 For children 13 to 18, relatively small 13 numbers and low number of events make 14 confidence intervals quite wide. 15 The same is true for patients older The odds ratio is about 1.3. In the five to twelve year age 16 group, the odds ratio was 8.4. 17 the odds ratio of regularly dosed Albuterol 18 versus placebo was greater than Foradil. 19 For all ages, To better align our data with the 20 data presented from the pooled analysis with 21 that shown in our label, we also performed an 22 analysis including events that occurred (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 430 1 during the observation period 30 days after 2 the last dose. 3 observation period occurred within five days 4 of the last dose of the double blind 5 treatment period. 6 the Formoterol group occurred in the 7 observation period, with two in children five 8 to twelve. 9 Most events occurred in the Six additional events in With the addition of these events, 10 the point estimate for the overall analysis 11 and the analysis by age are 2.2, 1.7, .3 and 12 10.5 in the five to twelve year age group. 13 These numbers are consistent with the 14 analyses presented in the Novartis and FDA 15 briefing box. 16 Next, we will review how Foradil is 17 used in the United States, and spontaneous 18 reports of asthma-related serious adverse 19 events. 20 regarding the use of single LABA agents based 21 on a physician's survey. 22 with you some more specific data on Foradil (202) 464-2400 This morning, you heard data I will now share Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 431 1 prescribing tied directly to patient records. 2 80 percent of Foradil prescriptions 3 are written for the treatment of COPD. 4 20 percent are written for the treatment of 5 asthma, and of these, 8 percent are written 6 for children less than 18. 7 prescriptions are evenly split between 8 children five to twelve and 13 to 18. 9 Pediatric To determine the extent of Foradil 10 concomitant use with inhaled corticosteroids, 11 we analyze data from a large U.S. pharmacy 12 benefit manager from 2006 to 2007. 13 three-quarters of the members under the age 14 of 40 who filled a prescription for Foradil 15 also filled a prescription for an inhaled 16 corticosteroid within one year. 17 percentage of concurrent prescriptions in 18 children five to twelve was 84 percent. 19 majority of the patients filled an ICS 20 prescription within 30 days of the Foradil 21 prescription. 22 (202) 464-2400 More than The The The FDA adverse event reporting Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 432 1 system, or the AERS database, includes all 2 spontaneous reports from U.S. sources for all 3 marketed drugs. 4 physicians and patients and mandatory for 5 manufacturers. 6 for cases in which Foradil was identified as 7 a medication and an asthma-related serious 8 adverse event was reported. 9 Reporting is voluntary for We searched the AERS database Approximately 1,000 events of any 10 kind were reported from January 1, 2001 until 11 March 31, 2008. 12 shown here is the number of asthma-related 13 serious adverse events per 100 adverse events 14 of any kind. 15 been a slow decline in the number of 16 spontaneously reported serious asthma 17 exacerbations. 18 until the first quarter of this year, there 19 were seven reports of asthma-related serious 20 adverse events, and no asthma-related death 21 in children. 22 safety data do not show new findings that (202) 464-2400 The reporting proportion Since a peak in 2003, there's Since Foradil was approved Overall, the post-marketing Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 433 1 heighten concern regarding the safety of 2 Foradil in patients of any age. 3 With this data in mind, Novartis 4 remains committed to the ongoing assessment 5 and mitigation of risk for patients using 6 Foradil. 7 study of seven state Medicaid databases to 8 better understand drug utilization and risk 9 associated with the use of Foradil with We are fielding an epidemiological 10 various treatment regimens. 11 medication guide accurately and clearly 12 reflect our current understanding of the 13 class risks of long-acting beta agonists, and 14 the safety data from the Foradil trials. 15 We're committed to working with the Agency to 16 address any questions or concerns that arise 17 from these proceedings. 18 Our labeling and We highlight the appropriate use 19 and risk in all asthma promotional and 20 educational materials, and continue to 21 monitor spontaneous and post-marketing 22 surveillance data for these events. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 434 1 We've reviewed the efficacy and 2 safety data from Foradil, discussed the 3 pooled clinical trial data, and shared data 4 from the FDA AERS database. 5 know about the efficacy and safety of LABAs 6 including Foradil, what's their benefit/risk 7 profile and how do they compare to other 8 available treatments? 9 Given what we These are the updated treatment 10 guidelines that Dr. Stoloff presented earlier 11 and that you've seen a couple times today. 12 As you know, inhaled corticosteroids are 13 foundation for asthma treatment. 14 therapy, LABAs are recommended in Step III, 15 with alternatives including leukotriene 16 modifiers, theophylline and increased ICS. 17 It's recommended that physicians As add-on 18 add LABAs to ICS in patients whose symptoms 19 are not controlled on ICS alone. 20 symptoms are controlled and lung function is 21 near normal, guidelines recommend that the 22 patients step down to less therapy. (202) 464-2400 Beta Court Reporting www.betareporting.com Once the Of note, (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 435 1 the use of non-combined LABAs and inhaled 2 corticosteroids allow for greater choice for 3 physicians, and facilitates the step-up and 4 step-down approach that's recommended by the 5 NHLBI. 6 If the indication for LABAs were 7 withdrawn, what alternatives would be left? 8 In short, as discussed earlier, the 9 alternatives provide a less favorable 10 benefit/risk profile than LABAs. 11 superior efficacy compared to the other 12 therapies in Step III. 13 effective add-on than leukotriene modifiers, 14 more effective than increasing the dose of 15 inhaled corticosteroids, or the addition of 16 theophylline. 17 LABAs have They're a more Several alternatives to LABAs have 18 significant adverse events -- theophylline 19 has toxicities that include cardiac 20 arrhythmia, seizures, and death. 21 monitoring is required for theophylline and 22 also for zileuton, which is associated (202) 464-2400 Beta Court Reporting www.betareporting.com Systemic (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 436 1 hepatotoxicity. 2 there's less data overall. 3 pediatric studies, doubling the dose of 4 inhaled corticosteroids has not been shown to 5 improve lung function or symptoms. 6 very little data on the efficacy and safety 7 of theophylline in children, and little data 8 on montelukast as an add-on to inhaled 9 corticosteroids. 10 In pediatric studies, In several There's Among LABAs, Foradil provides the 11 only non-combined Formoterol option for 12 asthma treatments in the U.S. today. 13 gives physicians the flexibility of adding 14 Formoterol to any of the approved inhaled 15 corticosteroids delivered by an array of 16 devices across a range of doses. 17 ICS choice, not all patients have the same 18 treatment response to a given ICS, and as a 19 result, there are eight inhaled 20 corticosteroids available for the treatment 21 of asthma. 22 (202) 464-2400 It In terms of If Foradil were no longer Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 437 1 available, patients currently taking any of 2 the other improved ICS other than Fluticasone 3 and budesonide would be required to alter 4 their inhaled corticosteroid regimen when 5 stepping up therapy with LABAs, and then 6 would be required to do so again when 7 stepping back down. 8 9 Children five to twelve would have even fewer options, since Advair Diskus is 10 the only combination of LABA and ICS that's 11 approved for this age group. 12 Many patients have difficulty using 13 metered dose inhalers or have a device 14 preference that makes Foradil's single dose 15 dry powder inhaler an attractive option. 16 In conclusion, Foradil improves 17 lung function, reduces symptoms, and reduces 18 a need for rescue medication. 19 directed, in addition to inhaled 20 corticosteroids, it remains an important and 21 effective option that should remain in the 22 therapeutic armamentarium with other (202) 464-2400 When used as Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 438 1 recommended alternatives. 2 Foradil, like all LABAs, may be 3 associated with an increased risk of 4 asthma-related serious adverse events. 5 data from more than 3,000 patients across 45 6 studies with the approved dose shows this 7 potential signal. 8 medication guide address the risk of 9 asthma-related serious adverse events in Our The current label and 10 adults and children -- the increased rate of 11 asthma hospitalizations observed among 12 children five to twelve in the pooled 13 analyses. 14 by Study 49, in which a differential dropout 15 rate may have contributed to the imbalance. This finding was primarily driven 16 Beyond the clinical trial data, 17 spontaneous reports resulting from almost 18 five million patient years of exposure do not 19 suggest a risk of hospitalization in the 20 pediatric age group that's greater than the 21 other age groups. 22 (202) 464-2400 It's the opinion of Novartis that Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 439 1 the totality of evidence supports the 2 appropriate use of Foradil as an effective 3 bronchodilator, and provides benefits for 4 asthma patients of all ages. 5 an important option for physicians and 6 patients. Foradil remains 7 Thank you for your attention. 8 DR. SWENSON: 9 Now I'll ask the team from AstraZeneca to be first seated. 10 This must be Catherine Bonuccelli. 11 DR. BONUCCELLI: Thank you and good 12 afternoon to everybody here in the home stretch 13 of a long day. 14 I'm the Global Product Vice President for 15 Symbicort at AstraZeneca. 16 Agency and the Committees for the opportunity to 17 be here today to present the benefit/risk of 18 AstraZeneca's Formoteral-containing products, 19 with a particular focus on the combination 20 product available in the U.S., the Symbicort 21 (pMDI). 22 (202) 464-2400 I'm Dr. Cathy Bonuccelli, and I want to thank the I will start by introducing the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 440 1 expert advisors who are here with AstraZeneca 2 today: Gary Anderson, professor of 3 Respiratory Pharmacology and Immunology from 4 Melbourne, Australia; Gary Cook, professor of 5 Biostatistics from Chapel Hill, North 6 Carolina; Craig LaForce, Clinical Professor 7 of Pediatrics from Raleigh, North Carolina; 8 Harold Nelson, professor of medicine from 9 Denver, Colorado; and Malcolm Sears, 10 professor of medicine from Hamilton, Ontario. 11 Multiple speakers today have 12 reviewed the history leading up to today's 13 meeting, so I'm not going to go through that 14 background again. 15 that both GSK and Novartis have shown you 16 data illustrating the significant clinical 17 benefits of LABAs, and the importance of 18 using LABAs in conjunction with ICS. What I will point out is 19 We will now present a substantial 20 amount of data on Formoterol used either in 21 free or in fixed combination with budesonide, 22 reaffirming the significant benefits of LABA (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 441 1 ICS combination therapy, and providing 2 reassurances that risks are minimized in this 3 setting of concomitant use. 4 Dr. Kramer asked, what new data 5 will you show us today? 6 data for salmeterol-containing products and 7 Novartis' data for Foradil was considered in 8 2005, the data that we will now show you has 9 not been previously considered. 10 While much of GSK's This is what we will take you 11 through this afternoon. 12 introduction, Dr. Thomas Anderson will review 13 the AstraZeneca products of relevance, and 14 the scientific evidence demonstrating the 15 benefits that Symbicort provides for patients 16 with asthma, both by reducing future 17 asthma-related risks and by improving current 18 asthma control. 19 After my Then, Dr. Kevin Carroll will review 20 AstraZeneca's analysis of the dataset 21 requested by the FDA, which demonstrates that 22 there is no signal of an increased risk for (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 442 1 Symicort for the severe asthma-related events 2 of specific interest today. 3 Dr. Carroll will also try to put 4 the differences between the AstraZeneca 5 analysis that you saw in your briefing 6 documents and the FDA analysis into context 7 with each other. 8 9 Following Dr. Carroll's presentation, I will come back and review our 10 conclusions from this review of benefit and 11 risk, that Symbicort is an important 12 treatment option for patients with persistent 13 asthma that has a positive benefit/risk when 14 used as currently indicated, and that the 15 current prescribing information adequately 16 covers any potential risks of the class. 17 I will now introduce Dr. Thomas 18 Anderson, the medical science director for 19 Symbicort at AstraZeneca, to review the 20 AstraZeneca products of relevance and the 21 benefit data. 22 (202) 464-2400 DR. ANDERSON: Thank you, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 443 1 Dr. Bonuccelli. 2 are three AstraZeneca products that are relevant 3 for today's discussion. 4 recent is Symbicort (pMDI), a fixed combination 5 of the inhaled corticosteroid Budesonide, and 6 the long and rapid acting agonist Formoterol, 7 which was approved by FDA in 2006. 8 9 Just to set some context, there The first and most Symbicort is indicated for long-term maintenance treatment of asthma in 10 patients 12 years of age and older. 11 Symbicort is not for all asthma patients. 12 You heard it earlier today, and as clearly 13 stated in the label, Symbicort should only be 14 used for patients not adequately controlled 15 on other asthma controller medications, or 16 whose disease severity clearly warrants 17 initiation of treatment with two maintenance 18 therapies. 19 However, There are two doses of Symbicort 20 available, with a maximum daily dose of 21 640mcg of Budesonide and 18mcg of Formoterol. 22 It should be noted that the dose of Fomoterol (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 444 1 is the same for both Symbicort strengths. 2 There are two other products that 3 are highly relevant for today's discussion. 4 Symbicort is available outside the U.S. as 5 Symbicort Turbuhaler. 6 inhaler. 7 Symbicort Turbuhaler is now registered for 8 asthma in more than 100 countries, and it has 9 an estimate patient exposure of more than 10 It's a dry powder And first available in 2000, 4,400 million treatment days. 11 The approved Formoterol doses in 12 Symbicort Turbuhaler are 9, 18, or 36mcg 13 daily, expressed as delivered doses. 14 Therefore, outside the U.S., Formoterol is 15 approved at a higher daily dose than the 16 18mcg daily approved in the U.S. 17 OXIS turbuhaler is a fomoterol 18 monoproduct first approved in 1996 and now 19 registered in 79 countries outside of the 20 U.S. 21 doses, and has an estimate patient exposure 22 of more than 1,400 million treatment days. (202) 464-2400 It's approved also for a range of Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 445 1 Both OXIS and Symbicort turbuhaler 2 are approved for children with asthma from 3 the age of four in some countries, but mostly 4 the age of six and upwards. 5 In specific studies, these products 6 have been proven to be therapeutically 7 equivalent with regards to the Formoterol 8 component. 9 to include data from all three products when Therefore, it's highly relevant 10 determining the overall benefit/risk of 11 Formoterol. 12 I will now turn to the benefits 13 review of Symbicort. 14 clinically relevant setting -- that is when 15 Formoterol is being added to Budesonide. 16 will show you data on the benefits of adding 17 Formoterol to Budesonide on reducing the 18 future risk of asthma, such as asthma 19 exacerbations and asthma worsenings. 20 we will show you benefits on important 21 measures of current control, the impairment 22 of asthma, as Dr. Stoloff told you about (202) 464-2400 We will focus on the Beta Court Reporting www.betareporting.com We Also, (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 446 1 before -- namely asthma symptoms, lung 2 function measured by FV1, asthma stability 3 measured as morning peak flow, and 4 asthma-related quality of life. 5 The asthma-related endpoints that 6 measure the future risk of the disease, such 7 as asthma exacerbations, are being assessed 8 today here as possible adverse effects of 9 therapy. Therefore, I would like to start by 10 presenting data that shows that adding 11 Formoterol to Budesonide actually lowers 12 asthma future risk. 13 This is the Facet study. It's a 14 clinical trial designed specifically to study 15 severe asthma exacerbations. 16 included 852 subjects with moderate to severe 17 asthma, and it compared a treatment with a 18 low dose Budesonide to the same dose with the 19 addition of Fomoterol. 20 The study It also compared a moderate dose of 21 Budesonide to the same dose plus Fomoterol 22 over the one-year treatment period. (202) 464-2400 Beta Court Reporting www.betareporting.com You saw (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 447 1 it briefly earlier in Dr. Lemanske's 2 presentation. 3 England Journal of Medicine -- Medical 4 Journal in 1997, and it's a very important 5 and groundbreaking study supporting the 6 development of ICS-LABA combinations, and it 7 was also the first major study to have asthma 8 exacerbations as a primary endpoint. 9 Facet was published in the New Adding Formoterol to either a low 10 or a moderate daily dose of Budesonide 11 significantly reduced the risk of a severe 12 asthma exacerbation compared to treatment 13 with Budesonide alone. 14 A severe asthma exacerbation was 15 defined as the need for oral steroid 16 treatment, or a more than 30 percent decrease 17 in morning peak flow from baseline. 18 only look at oral steroid treatments, the 19 results are the same -- numbers creep down of 20 course. 21 not part of the definition of asthma 22 exacerbations in this trial. (202) 464-2400 If we Asthma-related hospitalizations were The number of Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 448 1 asthma-related hospitalizations are low but 2 go in the same direction as the main result. 3 In the low dose Budesonide 4 comparison, there were three versus one 5 asthma-related hospitalizations. 6 dose Budesonide comparison, there were four 7 versus one, both favoring the groups treated 8 with Formoterol. 9 In the high Again, these results suggest that 10 Symbicort reduces future asthma risk, even 11 for the more severe events such as 12 hospitalizations. 13 that the high dose combination treatment in 14 this figure is equivalent to the higher 15 approved U.S. dose of Symbicort. 16 I would like to point out Another key endpoint of Facet, mild 17 asthma exacerbations, is a measure of asthma 18 worsenings with increased reliever use, 19 decrease in lung function, and awakenings due 20 to asthma. 21 with uncontrolled asthma. 22 (202) 464-2400 For a patient, this means days As you can see from this data, Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 449 1 similar to the severe asthma exacerbations 2 adding Formoterol to Budesonide significantly 3 reduced the rate of asthma worsenings. 4 What you are seeing here are the 5 results from Facet from asthma symptoms. A 6 higher score means worse symptoms, and thus, 7 lower is better. 8 adding Formoterol to either dose of 9 Budesonide resulted in significant reduction The results confirm that 10 in asthma symptoms that was maintained 11 without diminution over the 12 months of the 12 study. 13 need for reliever medication, and improved 14 lung function. In addition, Formoterol reduced the 15 Four years after Facet, the Optima 16 study confirmed the Facet result, once again 17 showing that adding Formoterol to Budesonide 18 reduces severe asthma exacerbations as well 19 as asthma worsenings. 20 hospitalizations went in favor of those 21 treated with Budesonide and Formoterol. 22 Therefore, when rigorously studied in studies (202) 464-2400 Once again, asthma Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 450 1 that are designed for the purpose, adding 2 Formoterol to Budesonide reduces rather than 3 increases the risk of asthma exacerbations. 4 Study 717 is one of the pivotal 5 trials for Symbicort (pMDI) in moderate to 6 severe asthma in the U.S. 7 curve shows asthma worsenings, and lower 8 numbers indicate more patients with events. 9 Asthma worsenings were predefined as This Kaplan-Meier 10 increases in asthma symptoms, increases in 11 reliever use, reduction in lung function, or 12 clinical worsenings. 13 In terms of benefit, you can see 14 that fewer patients on Symbicort, shown in 15 green, had a worsening, and it took longer 16 for the worsening to occur. 17 shows the combination of separate components 18 which had equal efficacy to Symbicort. 19 The purple line In terms of risk, it should be 20 noted that this study mandated the removal of 21 patients with worsening asthma who were then 22 not followed further. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 451 1 This makes this particular study by 2 design less suitable for assessing more 3 serious asthma events such as 4 hospitalizations. 5 the four trials included in the analysis 6 provided by the FDA today. 7 was not. 8 However, this was one of Optima and Facet This study, combined with Facet and 9 Optima, provides strong evidence that adding 10 Formoterol to Budesonide -- that is treating 11 patients with Symbicort -- reduces rather 12 than increases future risk of asthma events. 13 Continuing with other data from 14 717, I will now discuss the benefits of 15 Symbicort on measures of current control of 16 asthma patients. 17 constricted airways and poor lung function 18 are key burdens to asthma patients, and 19 improving lung function is a key goal of 20 treatments. 21 22 It's important to note that To the left, in green, we see the rapid and long lasting bronchodilation of (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 452 1 Symbicort over 12 hours, provided on the 2 first day of treatment. 3 point out three things on the right. 4 12 weeks of treatment with Symbicort, shown 5 in green, improves pre-dose FEV1 more than 6 Budesonide alone, which is shown in blue. 7 This is a measure of the baseline lung 8 function and asthma stability. 9 I would like to First, Secondly, the effect of LABAs used 10 as monotherapy after withdrawal of the 11 steroid can be seen here. 12 shown in orange, improves lung function 13 better than placebo, but it's less effective 14 during chronic dosing than the combination. 15 This also reaffirms that LABAs should not be 16 used without ICS. Formoterol alone, 17 Finally, after chronic treatment 18 for 12 weeks, with Symbicort in green, and 19 the separate components in purple, the 20 magnitude, the speed of onset, and the 21 duration of bronchodilatation are fully 22 maintained. (202) 464-2400 It also illustrates, when you Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 453 1 look at the difference between the orange 2 curve and the green and purple ones, how 3 concomitant use of Budesonide maintains the 4 responsiveness to Formoterol. 5 I will take the next couple of 6 slides to put the efficacy seen in children 7 into context with what's seen in older 8 patients. 9 of the seven Symbicort studies in the U.S. This figure provides an overview 10 that included morning pre-dose peak flow, a 11 measure of asthma stability, and the lung 12 function, before they take their dose. 13 studies are arranged by age group, and 14 includes comparison of Symbicort to 15 Budesonide. 16 patients age six to less than twelve. 17 The Five of these studies included Point estimates to the left favor 18 Budesonide, while dose to the right favor 19 Symbicort. 20 Symbicort is greater than what is achieved 21 with Budesonide, in itself a very active and 22 effective treatment. (202) 464-2400 As shown here, the benefits of These improvements are Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 454 1 seen across all age groups from children age 2 six and older. 3 Quality of life, as we heard 4 earlier today in the presentations of 5 Dr. Seymour and Dr. Chowdhury, it's an 6 important overall measure of asthma control 7 for both children and adults. 8 in the FDA Pulmonary Divisions briefing 9 document, ICS-LABA combination therapies are As pointed out 10 the only asthma treatments that have shown 11 meaningful improvements in this endpoint. 12 both of the U.S. Pivotal trials, Symbicort 13 demonstrated clinically relevant improvement 14 in quality of life versus placebo, using the 15 Asthma Quality of Life Questionnaire, AQLQ, a 16 validated instrument. 17 In This morning, you saw one of the 18 studies presented by Dr. Seymour, and here is 19 the other one, Study 716 in patients with 20 mild to moderate asthma. 21 22 Quality of life was improved in all active treatment groups, with the greatest (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 455 1 improvement from baseline observed for 2 Symbicort. 3 quality of life has been assessed using the 4 Pediatric Asthma Quality of Life 5 Questionnaire, PAQLQ, in patients aged twelve 6 to eleven, (?) as shown in the data to the 7 right from a six-month study in children. 8 9 In Symbicort pediatric studies, Although this study did not have a placebo comparator, the treatment differences 10 for Symbicort versus Budesonide were similar 11 to that seen in adults, and the improvement 12 from baseline met the MID of 0.5. 13 these data demonstrate improved quality of 14 life with combination therapy, and suggest 15 there there's similar benefits in children 16 and adults. 17 Overall, So as demonstrated today, Symbicort 18 has a positive benefit profile. 19 of Fomoterol to Budesonide as provided with 20 Symbicort reduces the future risk to patients 21 of asthma worsenings and exacerbations. 22 Furthermore, Symbicort results in significant (202) 464-2400 Beta Court Reporting www.betareporting.com The addition (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 456 1 and meaningful improvements of lung function, 2 symptoms, reliever use, and health-related 3 quality of life. 4 For patients, this means 5 significantly more days without impairment 6 from asthma, and for any one of you who 7 treats asthma or who has asthma yourself, you 8 know that this is meaningful to patients. 9 Finally, there's evidence that 10 children also experience benefits similar to 11 those seen in adults for validated measures 12 of asthma control and quality of life. 13 I will now hand over to Dr. Kevin 14 Carroll, who will present our analysis of 15 risk with focus on the events of particular 16 interest. 17 Thank you. 18 DR. CARROLL: 19 Over the next 15 minutes, I will Thank you, Thomas. 20 present an overall assessment to risk to 21 address the key question here today, being, 22 does Formoterol increase the risk of (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 457 1 asthma-related events. 2 be based primarily on trials and trial data 3 that met FDA's pre-specified criteria for 4 inclusion as outlined in the AstraZeneca 5 briefing document, and I will also briefly 6 refer to some epidemiological data. 7 The assessment will During the discussion, I will 8 clarify why and how the analyses presented in 9 the AstraZeneca briefing document differ from 10 those presented in FDA's statistical briefing 11 package. 12 When looking at a potential issue 13 of drug safety, the commonly accepted 14 practice is to consider all trial data at all 15 doses, both approved and unapproved, in order 16 to maximize the available information on the 17 potential risk. 18 assess the potential risks associated with 19 LABAs, FDA asked sponsors to provide data 20 from all blinded parallel group, randomized 21 control trials conducted with a LABA in the 22 treatment of asthma. (202) 464-2400 Thus, in order to best Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 458 1 The data from these trials then 2 underwent a pre-specified adjudication 3 process for serious asthma-related events 4 occurring on treatment. 5 slide, a total of 42 trials in just over 6 23,500 patients met FDA's predefined 7 criteria. 8 9 As shown in this As previously mentioned, the analysis provided by FDA in their materials 10 includes only 1,270 patients from four 11 AstraZeneca trials, meaning that 38 trials 12 and around 95 percent of the data in over 13 22,000 patients have been excluded. 14 analysis is therefore based on just 15 5 percent -- 5 percent of the data originally 16 supplied by AstraZeneca. 17 FDA's This dramatic falloff in patient 18 numbers is because after receiving the trial 19 data from AstraZeneca, the Qualitative Safety 20 and Pharmaco-epidemiology group at FDA 21 applied additional exclusion criteria to the 22 data, as shown on the slide. (202) 464-2400 These criteria Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 459 1 resulted in the exclusion of highly relevant 2 data such as study arms with Formoterol and 3 Budesonide received in free combination at 4 the same ratio as contained in Symbicort, in 5 data in children less than 12, and data at 6 doses higher than the currently approved 7 total daily Formoterol dose. 8 Now, in line with comments that we 9 heard this morning from Dr. Seymour in which 10 the therapeutic equivalents of free and fixed 11 combinations of Budesonide and Formoterol 12 have been established, and also bearing in 13 mind the comments from Dr. Chowdhury relating 14 to the importance of looking at higher doses, 15 be that free or fixed in combination or in 16 terms of OXIS turbuhaler, AstraZeneca 17 strongly believes it is essential to consider 18 all the control trial data that were 19 originally requested by FDA to ensure a 20 comprehensive and thorough assessment of 21 risk. 22 about this later. (202) 464-2400 And I shall say a little bit more Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 460 1 Hence, the data I'm about to show 2 you is based on all 42 trials and 23,500 3 patients that met FDA's predefined inclusion 4 criteria. 5 again, as requested by FDA, are as listed on 6 the slide. 7 Formoterol versus non-LABA containing 8 products, but I will also show you more 9 refined comparisons, looking specifically at The events I'll be looking at, The main comparison will be for 10 Formoterol in combination with ICS, and 11 Formoterol at doses equal to or greater than 12 the U.S. approved dose of 18mcg per day. 13 Finally, I will also show you data 14 in the pediatric subpopulation of patients 15 aged under 12 years. 16 So now I'd like to turn to the data 17 themselves. 18 overall trial data set, there were seven 19 deaths, splitting three versus four for 20 Formoterol and non-Formoterol groups 21 respectively. 22 evidence for an increased risk of death with (202) 464-2400 Here, you can see that in the Consequently, there is little Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 461 1 Formoterol-containing products based on these 2 adjudicated data. 3 asthma-related intubation in the Formoterol 4 group, and importantly, there were no 5 asthma-related deaths. 6 Also, there was one In terms of asthma-related 7 hospitalizations, this is the one event type 8 occurring frequently enough to allow a 9 detailed assessment of risk. Here, you can 10 see that in the third line, there were a 11 total of 161 events, giving event rates in 12 the four line of approximately 12 and 16 per 13 1,000 patients per year for Formoterol and 14 non-Formoterol groups respectively. 15 both the relative risk and rate difference 16 analyses show numerically lower risk of 17 asthma-related hospitalization for 18 Formoterol-containing products compared to 19 non-Formoterol-containing products. 20 Hence, And as outlined in our briefing 21 materials, we also looked at the data in key 22 subgroups such as gender, race, baseline (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 462 1 asthma severity, and we also looked at the 2 data in terms of longer term trials of at 3 least six months' duration. 4 these analyses were consistent with the 5 overall results, showing a numerically lower 6 risk of asthma-related hospitalization for 7 the Formoterol group relative to the 8 non-Formoterol group. The results of 9 In this slide, the hospitalization 10 data are redisplayed as a Kaplan-Meier plot. 11 As you would expect from the previous slide, 12 time to first asthma-related hospitalization 13 is longer for the Formoterol group, with the 14 curves diverging smoothly, and no evidence of 15 an increasing risk with time. 16 Now, in the previous two slides, I 17 showed you the overall result across the 42 18 trials, and here, I take the opportunity to 19 display the trial data individually. 20 horizontal bar represents a separate trial, 21 in the middle of which is shown the point 22 estimate for that trial and at the foot of (202) 464-2400 Beta Court Reporting www.betareporting.com Each (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 463 1 the slide, shown in orange, is the overall 2 relative risk and confidence interval 3 estimate. 4 Symbols to the left favor 5 Formoterol and to the right, non-LABA. 6 you can see, there is some variability but 7 the confidence interval tend to include the 8 line of unity. 9 risk and rate difference analyses show a 10 numerically lower risk of asthma-related 11 hospitalization for Formoterol-containing 12 products compared to 13 non-Formoterol-containing products. 14 As As before, both the relative At this point, I'd like to pause 15 for a moment and turn back to reflect on 16 FDA's analysis of the Formoterol data. 17 mentioned before, additional exclusion 18 criteria were applied to the AstraZeneca 19 data, and this resulted in FDA's analysis 20 being based on just four partial trials in 21 1,270 patients. 22 data originally requested by FDA. (202) 464-2400 As I That's only 5 percent of the Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 464 1 Here, you can see the four trials 2 included in FDA's analysis, together with the 3 corresponding relative risk result and 4 confidence interval shown in green at the 5 foot of the slide. 6 relative risk result based on all 42 trials 7 for comparison. 8 width of the confidence interval, indicating 9 the loss of precision in FDA's restricted Also shown in orange is a Note the difference in the 10 analysis, which is based on just seven 11 events. 12 This shows the importance of 13 evaluating all the data that were originally 14 requested by FDA to allow for a more complete 15 and personal assessment of the safety of 16 Formoterol as used in the U.S. 17 Moving on, and as I promised 18 earlier, I'd now like to show you a more 19 refined comparison looking at Formoterol in 20 combination with ICS, which is very pertinent 21 to the treatment of patients as outlined both 22 in the label and in current treatment (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 465 1 guidelines. 2 A total of 29 trials and just over 3 18,000 patients contributed to this analysis. 4 As you can see in our data for Formoterol, 5 the patent for LABA plus ICS versus ICS alone 6 is essentially the same as that seen for LABA 7 versus non-LABA containing products. 8 9 Now, we can even further refine the analysis of Formoterol plus ICS by looking 10 only at trials where the total daily dose of 11 Formoterol was equal to or greater than the 12 U.S. approved dose of 18mcg a day. 13 analysis, a total of 17 trials and 7,213 14 patients contributed. 15 slide, there is no evidence for an increased 16 risk of asthma-related hospitalization for 17 Formoterol in combination with ICS when 18 administered at a total daily dose of 18mcg 19 or greater. 20 For this As in the previous Now, it is important to note that 21 this particular analysis includes all the 22 patient data included in FDA's analysis, but (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 466 1 is more complete, as it also includes 2 Formoterol in combination with ICS at doses 3 at least as high as the approved U.S. dose. 4 Hence, if there was truly an excess 5 risk of hospitalization for Formoterol in 6 combination with ICS, this analysis, with 7 Formoterol at doses of at least 18mcg per 8 day, is one where we might reasonably expect 9 to see a signal. 10 And as you can see, no signal is evident. 11 To summarize, I have shown you 12 asthma-related hospitalization data analyzed 13 in terms of first, the LABA versus non-LABA; 14 secondly, in terms of ICS plus LABA versus 15 ICS alone; and finally, in terms of ICS plus 16 LABA versus ICS alone, where the total daily 17 dose of Formoterol is at least 18mcg per day. 18 The overall results of these 19 analyses are shown on this slide for 20 comparison. 21 22 As you can see, the data are very consistent across analyses, showing no (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 467 1 increase in the risk of asthma-related 2 hospitalization for LABA-containing products. 3 It's also a value to consider the other 4 analyses of interest such as those provided 5 by GSK, where data on the free and fixed 6 combination of an ICS and a LABA are 7 considered separately. 8 9 When we do this for Formoterol, we say there is no evidence of an increased risk 10 of hospitalization either when Formoterol is 11 given in free combination with an ICS or when 12 Formoterol is given in fixed combination as 13 in Symbicort. 14 On this next slide, we've looked at 15 the number of hospitalizations as a function 16 of accumulating Formoterol dose. 17 see in the column to the right, there's no 18 indication of an increasing risk of 19 hospitalization with increasing dose. 20 I've added the corresponding data for the 21 non-LABA-treated patients. 22 addition to there being no dose response for (202) 464-2400 As you can Now You can see, in Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 468 1 hospitalizations with increasing Formoterol 2 dose, the hospitalization rate was 3 consistently lower than that for 4 non-LABA-treated patients. 5 Now, this all strongly argues 6 against a pharmacological dose-related risk 7 with Formoterol when used appropriately 8 together with an ICS. 9 I would now like to move on to look 10 at the pediatric patient data. 11 overall trial data, there were 3,423 children 12 aged between four and twelve years, and 13 around 2,150 of whom were treated with 14 Formoterol. 15 in the FDA's analysis of risk. 16 deaths or intubations amongst these patients. 17 In the None of these data were included There were no Here, you can see there were a 18 total of 39 hospitalization events in 19 pediatric patients, given event rates of 20 approximately 35 and 29 per 1,000 patients 21 per year for Formoterol and non-Formoterol 22 groups respectively. (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 469 1 Here, the data are displayed by 2 individual trial. While the overall relative 3 risk and risk difference analyses show point 4 estimates that tend to favor non-LABA-treated 5 patients, the confidence intervals are wide 6 and consistent with no increase in risk, 7 reflecting the individual trial data where 8 the confidence intervals all include the line 9 of unity. 10 As we pointed out in our briefing 11 materials, the result here is influenced by 12 the poor performance of a sub-therapeutic 13 dose of Budesonide and Formoterol in Trial 14 673. 15 were observed in this trial due to a lack of 16 efficacy. 17 reconsidered without this sub-therapeutic 18 dose, it is interesting to note that the 19 point estimate reduces to .8. 20 point out that as outlined in our briefing 21 document, our data in adolescents is 22 consistent with the overall trial data. (202) 464-2400 Consequently, seven hospitalizations If the pediatric data are And I should Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 470 1 To summarize then, in roughly 3,400 2 pediatric patients aged between four and 3 twelve years, there was little difference in 4 the hospitalization rate between LABA and 5 non-LABA-treated patients. 6 While the data did not favor 7 Formoterol numerically, confidence intervals 8 were wide both for the overall analysis and 9 for individual trials straddling the line of 10 unity. 11 Moving on now briefly to look at 12 epidemiological data, while asthma prevalence 13 has been increasing right into the early 14 2000s, there is no signal of an increase in 15 asthma mortality after the introduction of 16 LABAs, as seen in this slide of U.S. data 17 previously shown by Dr. Stoloff. 18 As a complement to U.S. data, here 19 you can see the development of asthma 20 mortality of the past 10 years in Sweden 21 based on data from the National Death 22 Registry. (202) 464-2400 Sweden is the first country where Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 471 1 Symbicort was approved and where Formoterol 2 is the predominant LABA. 3 reassuring to see that during the period with 4 a marked increase in the use of LABAs, and of 5 Formoterol in particular, there is no 6 evidence of any negative trend in asthma 7 mortality. It's therefore 8 On the contrary, decline in asthma 9 mortality seen is continued over this period 10 of time strongly support the current 11 treatment practice guidelines for the 12 management of asthma. 13 I will now turn back to 14 Dr. Bonuccelli, who will conclude today's 15 presentation. 16 17 DR. BONUCCELLI: Thank you, Dr. Carroll. 18 In summary, AstraZeneca believes 19 that the benefits of Symbicort have been 20 shown unequivocally in controlled clinical 21 trials in adults, adolescents and children. 22 In asthma patients not well-controlled on ICS (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 472 1 alone, Symbicort improves current control of 2 asthma, and also reduces the risk of future 3 asthma worsenings and exacerbations. 4 In addition, in the FDA-requested 5 and adjudicated data set of 23,510 patients 6 across 42 controlled, randomized, blinded, 7 parallel group trials, there was no evidence 8 of increased risk of asthma-related deaths, 9 intubations, or hospitalizations for 10 Formoterol when used with an inhaled 11 steroid -- that is for Symbicort. 12 Specifically, in this data set, there were no 13 asthma-related deaths, and asthma-related 14 hospitalizations occurred less frequently in 15 Formoterol-treated patients. 16 are comprehensive and more precise, with 17 narrower confidence intervals than the 18 analysis done by the FDA, which included only 19 5 percent of the FDA requested data. 20 These results In conclusion then, Symbicort 21 (pMDI), when used as indicated, is a safe, 22 effective, and important medicine for (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 473 1 patients whose asthma is not controlled on 2 inhaled steroids alone. 3 Symbicort is appropriate, and the box warning 4 and safety precautions in the current label 5 adequately describe any potential risks. 6 The indication for Given the compelling scientific and 7 clinical evidence supporting the positive 8 benefit/risks of Symbicort (pMDI), we feel 9 strongly it should remain available to 10 patients and prescribers. 11 And as always, AstraZeneca is happy 12 to discuss any additional steps that we could 13 take to ensure the safe and effective use of 14 Symbicort. 15 Thank you for your attention. 16 DR. SWENSON: We've heard a lot, and 17 we're now 30 minutes past schedule. I'm sure 18 there's still a number of questions. 19 And I'd like to just possibly get a 20 sense from the panels here if they wish to at 21 least have a few questions before we end the 22 day here. (202) 464-2400 There will be time tomorrow, but Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca 474 1 of course, memories are fresh now from these 2 presentations. 3 here, but could I get a show of hands of 4 whether there's any sentiment to continued 5 questioning here or to save it for tomorrow? 6 7 Those in favor of any further question period? 8 9 We may not have everyone Okay. Then we meet again here tomorrow morning at 8:30. 10 (Whereupon, at approximately 5:57 11 p.m., the PROCEEDINGS were 12 adjourned.) 13 * * * * * 14 15 16 17 18 19 20 21 22 (202) 464-2400 Beta Court Reporting www.betareporting.com (800) 522-2382 800530c0-80cb-45e8-97e2-4623877f3dca