FOOD AND DRUG ADMINISTRATION

NINETY-SEVENTH MEETING OF THE

CARDIOVASCULAR AND RENAL DRUG ADVISORY COMMITTEE

8:01 a.m.

Friday, July 19, 2002

Versailles Ballroom

Holiday Inn - Bethesda

8120 Wisconsin Avenue

Bethesda, Maryland

ATTENDEES

COMMITTEE MEMBERS:

JEFFREY BORER, M.D., Chairman

Director, Division of Pathophysiology

Weill Medical College

Cornell University

525 East 68th Street, Room F467

New York, New York 10021

JAYNE E. PETERSON, R.PH., J.D., Acting Executive Secretary

Advisors and Consultants Staff

Center for Drug Evaluation and Research

Food and Drug Administration

5630 Fishers Lane, Room 1093

Rockville, Maryland 20857

PAUL ARMSTRONG, M.D.

Professor, Department of Medicine

University of Alberta

251 Medical Science Building

Edmonton, Alberta, Canada T6C37

MICHAEL F. ARTMAN, M.D.

Professor of Pediatrics

Pediatric Cardiology

New York University Medical Center

530 First Avenue, FPO Suite 9-V

New York, New York 10016

BLASE A. CARABELLO, M.D.

Professor of Medicine

Houston Veterans Affairs Medical Center

Medical Service (111)

2002 Holcombe Boulevard

Houston, Texas 77030

SUSANNA L. CUNNINGHAM, PH.D.

Consumer Representative

Professor, Department of Biobehavioral Nursing

School of Nursing, Box 357266

Seattle, Washington 98195-7266

ATTENDEES (Continued)

COMMITTEE MEMBERS: (Continued)

THOMAS FLEMING, PH.D.

Professor and Chair

Department of Biostatistics

University of Washington

Box 357232

Seattle, Washington 98195-7232

STEVEN NISSEN, M.D., F.A.C.C.

Vice Chairman, Department of Cardiology

Professor of Medicine

Ohio State University

The Cleveland Clinic Foundation

9500 Euclid Avenue, F15

Cleveland, Ohio 44195

MEETING GUEST (NONVOTING):

THOMAS G. PICKERING, M.D., D.PHIL.

Professor of Medicine

Director, Integrative and Behavioral

Cardiovascular Health Program and

Hypertension Section

Michael and Zena A. Wiener Cardiovascular Institute

Mount Sinai School of Medicine

One Gustave L. Levy Place

New York, New York, 10029-6574

FOOD AND DRUG ADMINISTRATION STAFF:

ROBERT TEMPLE, M.D.

DOUGLAS THROCKMORTON, M.D.

ANNE TRONTELLE

BRISTOL-MYERS SQUIBB REPRESENTATIVES:

HENRY BLACK, M.D.

CHARLES H. HENNEKENS, M.D., PH.D.

ALLEN KAPLAN, M.D.

ELLIOTT LEVY, M.D.

MILTON PACKER, M.D.

MICHAEL WEBER, M.D.

ANTHONY WACLAWSKI, PH.D.

C O N T E N T S

NDA 21-188, Vanlev (omapatrilat)

Bristol-Myers Squibb Company,

Proposed for the Treatment of Hypertension

* * *

AGENDA ITEM PAGE

CONFLICT OF INTEREST STATEMENT

By Ms. Jayne Peterson 6

BRISTOL-MYERS SQUIBB PRESENTATION:

Introduction

By Dr. Anthony Waclawski 9

Clinical Efficacy Data

By Dr. Elliott Levy 15

Angioedema: Clinical Overview

By Dr. Allen Kaplan 71

Clinical Safety Data (Continued)

By Dr. Elliott Levy 85

OPEN PUBLIC HEARING 103

BRISTOL-MYERS SQUIBB PRESENTATION: (Continued)

OVERTURE Trial Review

By Dr. Milton Packer 141

Clinician's Perspective

By Dr. Henry Black 161

Conclusion

By Dr. Anthony Waclawski 170

COMMITTEE DISCUSSION AND REVIEW 171

P R O C E E D I N G S

(8:01 a.m.)

DR. BORER: It's not quite 8:01, so everybody has had some extra time. We'll begin this morning's session which is consideration of NDA 21-188, Vanlev, sponsored by Bristol-Myers Squibb.

The committee is slightly restructured today because of conflict of two members. So, we'll introduce the active members, including our nonvoting member guest. Before we do that, let me ask you please to turn off your cell phones, if they happen to be on.

Why don't we start on this side. Tom.

DR. PICKERING: I'm Tom Pickering from the Cardiovascular Institute at Mount Sinai Medical Center in New York.

DR. CUNNINGHAM: I'm Susanna Cunningham from the University of Washington in Seattle.

DR. CARABELLO: I'm Blase Carabello from the Houston VA and from the Baylor College of Medicine.

DR. NISSEN: Steve Nissen with the Cleveland Clinic School of Medicine.

DR. ARMSTRONG: Paul Armstrong from the University of Alberta.

DR. BORER: I'm Jeff Borer, Weill Medical College at Cornell University in New York City.

MS. PETERSON: I'm Jayne Peterson. I'm the

DR. FLEMING: Tom Fleming, University of acting Executive Secretary of the Advisory Committee. Washington, Seattle.

DR. THROCKMORTON: Doug Throckmorton. I'm the Director of the Cardio-Renal Division in the FDA.

DR. BORER: We'll have our additional member introduce himself when he comes in.

Jayne, will you please present the conflict of interest statement?

MS. PETERSON: Thank you.

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

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

Dr. Jeffrey Borer has been granted a waiver under 18 U.S.C. 208(b)(3) for his potential consulting for the sponsor on a competitor to Vanlev on unrelated matters. Potentially he could receive less than $10,001 a year.

Dr. Susanna Cunningham has been granted waivers under 18 U.S.C. 208(b)(3) and 21 U.S.C. 355(n)(4), amendment of section 505 of the Food and Drug Administration Modernization Act, for ownership of stock in a competitor to Vanlev. The stock is valued between $25,000 and $50,000.

Dr. Thomas Fleming has been granted a waiver under 18 U.S.C. 208(b)(3) for his participation on two data safety monitoring committees for a competitor and the parent of a competitor to Vanlev on unrelated matters. He receives less than $10,000 per year for each activity.

A copy of these waiver statements may be obtained by submitting a written request to the agency's Freedom of Information Office, room 12A-30 of the Parklawn Building.

We would also like to disclose for the record, because of her reported interest, Dr. Beverly Lorell, a committee member, is excluded from participating in all official matters concerning new drug application 21-188, Vanlev, omapatrilat, sponsored by Bristol-Myers Squibb proposed for the treatment of hypertension.

With respect to FDA's invited guest, Dr. Pickering has a reported interest that we believe should be made public to allow the participants to objectively evaluate his comments. Dr. Pickering is listed as a Vanlev consultant for Bristol-Myers Squibb and was paid in 2001. He has received a research grant from Bristol-Myers Squibb in 2001 for analyzing their data on 24-hour blood pressure. He has done nothing for the company in 2002.

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

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

Thank you. Dr. Borer.

DR. BORER: Mike, will you introduce yourself to the company?

DR. ARTMAN: I'm Mike Artman. I'm at New York University School of Medicine.

And I would just like the record to show that Dr. Borer's clock, according to the U.S. atomic clock, is about 3 minutes fast. Thank you.

(Laughter.)

DR. BORER: Well, it means we get through 3 minutes earlier.

Let's begin the sponsor's presentation then, if we can. Dr. Waclawski.

DR. WACLAWSKI: Thank you, Dr. Borer. Good morning to you, members of the advisory committee, FDA, ladies and gentlemen.

I'm Anthony Waclawski with the Regulatory Sciences Group at Bristol-Myers Squibb. It's my pleasure to take a few minutes today and introduce our presentation.

The purpose of our presentation today is to discuss the data that is relevant to the use of omapatrilat in hypertension, specifically in patients with hypertension that is difficult to control with other agents.

Omapatrilat is a vasopeptide ACE inhibitor. It is the first agent in this new class of antihypertensive agents to be discussed by this committee.

As background, I will briefly review the regulatory history of the application and then give you an overview of this morning's presentation.

The original NDA was filed in December of 1999. This NDA was based on an extensive preclinical and clinical development program. The clinical studies were mainly conducted as placebo-controlled or active-controlled, forced-titration studies.

In April of 2000, Bristol-Myers Squibb withdrew the NDA. This was in response to questions raised by the FDA regarding the comparative incidence and severity of angioedema with omapatrilat compared to existing agents.

In August of 2000, the 6-month, 25,000-patient OCTAVE study was initiated. OCTAVE stands for omapatrilat cardiovascular treatment assessment versus enalapril. This study was conducted to more clearly define the efficacy and safety of omapatrilat compared to the ACE inhibitor enalapril.

In December of 2001, based upon the review and analysis of the results of the OCTAVE study, the NDA for omapatrilat for the treatment of hypertension was resubmitted. The resubmitted NDA included data from approximately 19,000 subjects treated with omapatrilat, making it several times larger than recent NDAs submitted for hypertension. The size and scope of the omapatrilat NDA allowed for the characterization of the safety and efficacy of omapatrilat in a broad range of patients.

In addition, although not part of the NDA for hypertension, omapatrilat has been studied in an extensive heart failure program, including the recently completed OVERTURE study. There's a question today about OVERTURE and its implications for hypertension on the list of questions today.

With that background, I will now provide an overview of our presentation.

First, in terms of efficacy, data will be presented to demonstrate that omapatrilat is an effective antihypertensive agent, more effective as monotherapy than lisinopril, losartan, or amlodipine. In addition, data from the OCTAVE study will be presented. These data demonstrate that an omapatrilat-based regimen is more effective than an enalapril-based regimen in a broad range of patients under conditions that closely mimic clinical practice.

In terms of safety, data from the OCTAVE study will be presented that demonstrate that patients treated with omapatrilat experience angioedema about three times more frequently than those patients treated with enalapril. In OCTAVE, life-threatening angioedema occurred in patients treated with omapatrilat at a rate of approximately 2 per 12,000 patients. In OCTAVE, no patients treated with enalapril experienced life-threatening angioedema.

In terms of benefit and risk, these data, taken together, present difficult and complex questions about benefit and risk. How should one evaluate a compound that may offer superior benefit when it also carries an increased risk of a potentially life-threatening adverse event? How should the expected benefit be estimated? What level of risk is acceptable? And in what patients is perhaps the benefit-to-risk favorable? Data will be presented today to help address these issues.

Let me tell you about the approach that we have taken.

Since the filing of our NDA in December of last year, we have performed numerous additional statistical analyses of the OCTAVE data and have had extensive consultations with medical and regulatory experts and the FDA aimed at helping us to answer these questions.

In light of the risk of angioedema, we have looked for ways to maximize the benefit and minimize the risk. Maximizing the benefit means to target the use of omapatrilat to those patients that are most likely to benefit from therapy. These patients would have an increased cardiovascular risk and would have hypertension that is difficult to control with available therapies. Data will be presented today which demonstrate that omapatrilat provides substantial blood pressure reductions in these patients.

Regarding the management of risk, we have initiated discussions with the FDA about how to manage the risk of angioedema. We have thus far focused on the identification of the risk factors of angioedema and on the use of patient education about angioedema to help minimize the risk of severe outcomes.

You have in your briefing book an FDA review of our proposed risk management plan. The review points out that risk management will not likely reduce the risk of angioedema with omapatrilat to that of an ACE inhibitor. We agree with this, and this is not the objective of the plan. Rather, the objective is to minimize the risk of life-threatening angioedema using education. The review acknowledges that this might be possible, and we are continuing to work with FDA on this plan. We are confident that if omapatrilat is approved on the basis of the clinical data, that we can find a mutually acceptable plan with the FDA.

I will now come back to the target population and be a little bit more specific since our presentation today is focused on these patients.

We'll present data that supports the use of omapatrilat in patients that can be described with two broad criteria. These patients will have comorbid conditions or characteristics associated with high cardiovascular risk, such as a history of cardiovascular disease, patients with target organ damage, those with three or more cardiac risk factors, or patients with diabetes or renal disease. They would also have hypertension that is difficult to control with existing agents.

As you'll see from our presentation today, black patients and patients who smoke are at a higher risk of angioedema. Use of omapatrilat in these patients must be accompanied by particular caution.

This is the target population. We will present data today that supports the use of omapatrilat in these patients. When evaluating these data, we recognize that the advisory committee and the FDA will rely upon their scientific judgment when considering how these data may support a recommendation for the approval of omapatrilat. We've been working through these issues for some time and are looking forward to your deliberations.

Bristol-Myers Squibb has invited several consultants to the meeting today. They are Drs. Black, Hennekens, Kaplan, Packer, Neaton, and Weber. These experts are here to facilitate the advisory committee discussions and deliberations.

Finally, the agenda for the presentation is as follows. Dr. Levy, who leads the clinical development program for omapatrilat at Bristol-Myers Squibb, will present the clinical efficacy data. Dr. Kaplan, from the University of South Carolina, an expert in angioedema and a member of the OCTAVE angioedema endpoint adjudication committee, will provide a short background on this event. Dr. Levy will then return to present the safety data and the benefit-risk summary.

I should note that although Dr. Hennekens was listed on the agenda that may be in your briefing package, he will not make a presentation today on risk-benefit, but he is here to answer any questions.

There is also a question about OVERTURE. We have asked Dr. Packer to come and make a short presentation about OVERTURE. This is also a small change from your agenda.

Dr. Black will follow Dr. Packer and he will provide a clinician's perspective. I will then return and conclude our presentation.

That ends the introduction. I would now like to introduce Dr. Elliott Levy who will present the clinical efficacy data.

DR. LEVY: Dr. Borer and members of the committee, thank you for your attention. My name is Elliott Levy, and I lead the omapatrilat clinical development team.

Before discussing the efficacy of omapatrilat, I'd like to reemphasize a point made by Dr. Waclawski in his introduction. Bristol-Myers Squibb is asking the advisory committee to consider omapatrilat for use in patients who have established cardiovascular disease or other characteristics associated with similarly high cardiovascular risk and whose blood pressure is difficult to control with existing therapies. In this population, the benefit of omapatrilat treatment strongly outweighs the risk of angioedema.

I'll present efficacy data this morning in the following order. In four placebo-controlled trials, including approximately 2,400 subjects, omapatrilat was shown to reduce systolic and diastolic blood pressure in dose-dependent fashion.

In six active-controlled trials involving approximately 2,700 subjects, the maximum intended dose of omapatrilat, 80 milligrams, was shown to reduce blood pressure more effectively than the maximum labeled dose of the widely used antihypertensives lisinopril, amlodipine, and losartan.

In OCTAVE, which included about 2,500 subjects, an omapatrilat-based regimen was shown to reduce blood pressure more effectively than one based on enalapril. Omapatrilat was also shown to reduce blood pressure effectively in the proposed target population: patients with high cardiovascular risk and difficult-to-control hypertension.

In four placebo-controlled, randomized, double-blind, dose-ranging studies, omapatrilat at doses of 10 to 80 milligrams was shown to reduce systolic and diastolic blood pressure in dose-dependent fashion. At the proposed starting dose of 10 milligrams, omapatrilat produced statistically significant reductions in blood pressure relative to placebo. At the maximum intended dose of 80 milligrams, omapatrilat reduced systolic blood pressure by about 16 millimeters of mercury relative to placebo and 19 millimeters of mercury overall.

These changes in blood pressure were substantially larger than those historically reported with existing agents, and based on these findings a series of six active-controlled, randomized, double-blind trials were performed in which omapatrilat 80 milligrams was directly compared to the maximal recommended dose for the widely used antihypertensive agents amlodipine, lisinopril, and losartan. For clarity, I'll present the systolic blood pressure results in these studies. The results for diastolic blood pressure were similar.

In three of these studies presented here, efficacy was assessed by measurement of seated blood pressure in the physician's office using standard cuff methodology at the time of trough blood levels, so about 24 hours after administration of the previous dose. Omapatrilat produced statistically significant reductions in blood pressure relative to amlodipine, lisinopril, and losartan, ranging from 3 millimeters of mercury systolic relative to amlodipine on the left-hand side, and moving right, 5 millimeters of mercury relative to lisinopril, and 7 millimeters of mercury relative to losartan.

You may have noted that in one of these studies conducted versus lisinopril, reductions in blood pressure were smaller than observed elsewhere. This study was performed in African Americans in whom the response to drugs that inhibit the renin-angiotensin system is known to be diminished. As expected, the response to both omapatrilat and lisinopril was reduced in this study, but systolic blood pressure was reduced about 5 millimeters of mercury more with omapatrilat than with lisinopril.

In three other studies displayed here, efficacy was assessed by ambulatory blood pressure monitoring. Ambulatory blood pressure has been shown to correlate more closely with target organ damage than does office blood pressure. And ambulatory blood pressure also captures the effect of drug on blood pressure over 24 hours during normal daily activities, rather than at a single time point in the physician's office.

In these studies, omapatrilat was also shown to reduce blood pressure more effectively than maximal recommended doses of amlodipine, lisinopril, or losartan. Here the differences ranged from about 5 to 6 millimeters of mercury relative to amlodipine to about 7 millimeters of mercury relative to lisinopril and 8 to 9 millimeters of mercury relative to losartan. These differences between omapatrilat and comparator were somewhat greater than observed in the office blood pressure studies previously presented, which is the opposite of what one might expect since ambulatory pressures tend to be lower than office blood pressures and to vary over a smaller range.

The course of blood pressure reduction over 24 hours is illustrated in this representative tracing from the amlodipine comparison study. At every time point over 24 hours, omapatrilat reduced blood pressure more than amlodipine, as illustrated by the bottom curves. Similar results were observed in ambulatory blood pressure trials conducted versus lisinopril and losartan.

In sum, in these active-controlled trials, omapatrilat at 80 milligrams produced greater reductions in blood pressure than the maximum recommended doses of amlodipine, lisinopril, and losartan. A major objective of OCTAVE was to determined whether omapatrilat would be superior to another agent in conditions similar to those encountered in clinical practice where an antihypertensive therapy is titrated electively to reach blood pressure target and supplemented by other agents as needed.

OCTAVE used a simple protocol of a large sample size and few exclusion criteria so that the efficacy and safety of omapatrilat could be assessed in a variety of demographic and clinical subgroups.

In OCTAVE, 25,000 hypertensive patients were randomized in equal number to treatment with omapatrilat beginning at 10 milligrams or enalapril beginning at 5 milligrams. After an initial fourth titration step at week 2, physicians were instructed to titrate patients as needed to reach blood pressure target at weeks 4 and 6. At week 8, the end of the study drug titration phase, the dose of study medication was fixed, and investigators were instructed to add other antihypertensive agents as needed in order to reach blood pressure target at weeks 8 and 16. The dose range selected for omapatrilat reflected the intended clinical dose range, while the enalapril dose regimen was selected in accordance with the label and customary clinical practice.

For assessment of efficacy, subjects were assigned at randomization to one of three prespecified study groups, each representing a potential manner of use of omapatrilat. Patients not receiving antihypertensive therapy at enrollment, about 9,000 patients, were assigned to study group 1 and received omapatrilat or enalapril as initial therapy for hypertension.

Patients receiving antihypertensive therapy at enrollment but not controlled were assigned to study groups 2 or 3. Those with mildly elevated blood pressure, systolic blood pressures of 140 to 159, or diastolics of 90 to 99, were assigned to study group 2 and received omapatrilat or enalapril as replacement for existing therapies, all of which were discontinued at randomization. About 11,000 patients were assigned to this group.

Study group 3 patients included those with more markedly uncontrolled blood pressure at randomization, systolic blood pressure of 160 to 179 or diastolic pressure of 100 to 109, and whose baseline regimen did not include an ACE inhibitor. These patients received omapatrilat or enalapril in addition to existing therapies which were continued beyond randomization. About 5,000 patients were assigned to this study group.

Two efficacy objectives were specified as co-primary study endpoints. The first, change in systolic blood pressure from baseline to week 8, reflected the effect of study drug on blood pressure, titrated electively as needed to reach target. The second co-primary efficacy objective, the use of new adjunctive antihypertensive therapy between weeks 8 and 24, reflected the extent to which a more effective monotherapy might reduce the need for additional antihypertensive therapy.

Important safety objectives included the assessment of the incidence of adverse events, as well as the incidence and severity of angioedema. These will be discussed in more detail in the safety portion of the talk.

The study results at week 8, the end of the study drug titration period, are summarized here. If omapatrilat had greater inherent efficacy, then one might expect that subjects randomized to enalapril would be more likely to be titrated upward in order to reach blood pressure target than subjects randomized to omapatrilat, and this in fact was observed. As shown on the right-hand panel of this slide, subjects randomized to enalapril were more likely to be titrated to top dose of study drug than subjects randomized to omapatrilat, and this was true whether study drug was used as initial therapy for hypertension in study group 1, as replacement for existing therapy in study group 2, or in addition to existing therapy as in study group 3.

Between 33 and 52 percent of patients randomized to enalapril were titrated to 40 milligrams, the maximal dose. This pattern of therapy with robust doses of enalapril is considerably more aggressive than that encountered in clinical practice. Despite greater use of maximal study therapy in patients randomized to enalapril, those randomized to omapatrilat had greater reductions in systolic blood pressure at week 8, as shown in the left-hand panel. The difference in systolic blood pressure reduction of about 3 to 4 millimeters of mercury was highly consistent whether patients received study drug as initial therapy, as replacement, or add-on therapy.

You might note that the blood pressure reductions with both study drugs were smaller in group 2 than in groups 1 and 3 because group 2 subjects discontinued all prior antihypertensive therapy at enrollment. Their blood pressure changes reflect both the antihypertensive effect of study drug and the effect of withdrawal of other active therapies.

The results at week 24, the end of the study, are summarized here. It was hypothesized that if omapatrilat reduced systolic blood pressure more than enalapril at week 8, it would also reduce the use of other antihypertensive agents from weeks 9 through 24, and this was observed. As summarized on the right-hand panel, subjects randomized to omapatrilat were significantly less likely to receive additional antihypertensive therapy than subjects randomized to enalapril. Despite receiving less top-dose study drug and less adjunctive therapy, subjects randomized to omapatrilat had consistently greater reductions in systolic blood pressure at week 24 as shown in the right-hand panel, about 3 millimeters of mercury more than subjects randomized to enalapril.

Now, these study findings were highly consistent across patient subgroups. OCTAVE included about 7,000 patients over the age of 65, 2,000 over the age of 75, and 2,500 black patients. Omapatrilat reduced systolic blood pressure about 3 millimeters of mercury more than enalapril at study's end in each major demographic subgroup as shown on the right-hand column of this slide. Not surprisingly, reductions in blood pressure with both omapatrilat and enalapril were smaller in black patients than in others, but nevertheless blood pressure was reduced about 4 millimeters of mercury more with omapatrilat than with enalapril in these subjects.

OCTAVE also included a large number of patients with comorbid characteristics or other features associated with increased risk of cardiovascular disease. About 3,300 patients with diabetes and 2,300 patients with established cardiovascular disease were studied in OCTAVE. Omapatrilat produced consistently greater reductions in systolic blood pressure than enalapril, on the order of 3 to 5 millimeters of mercury, as shown on the right-hand side of this chart, in patients with severe hypertension, those with diabetes, atherosclerotic disease, isolated systolic hypertension, renal disease, or heart failure.

In summary, OCTAVE demonstrated greater blood pressure reduction with an omapatrilat-based regimen than with an enalapril-based regimen despite more use of top-dose enalapril and more use of adjunctive antihypertensive therapy with enalapril.

The results of OCTAVE were highly consistent, regardless of patient demographics or comorbidity and regardless of the manner in which study drug was used.

Lastly, the greater blood pressure reduction observed with omapatrilat at week 8, the end of the study drug titration period, was preserved to the end of the trial despite the use of adjunctive therapy in order to reach a common blood pressure target in all patients.

The advisory committee has been asked to consider why the efficacy advantage observed at week 8 in OCTAVE was preserved at week 24 and whether this suggests that an omapatrilat-based regimen provides a reduction in blood pressure that cannot be achieved with a regimen based on enalapril or existing therapies.

OCTAVE provides a unique data set with which to answer this question. While we acknowledge that in many patients hypertension can be readily controlled with enalapril or other existing treatments, OCTAVE suggests -- and other clinical trials confirm -- that hypertension is difficult to control in many patients, even with multi-drug regimens. Therefore, for many patients the question is not whether omapatrilat can be used in place of a combination regimen, but whether omapatrilat should be used as part of a combination regimen. The results of OCTAVE strongly confirm that a combination regimen which includes omapatrilat reduces blood pressure to a greater extent than a combination regimen containing enalapril because omapatrilat is a more efficacious antihypertensive agent.

The effect that greater drug efficacy can have on regimen efficacy can be most clearly appreciated in those most likely to require a multi-drug antihypertensive regimen, namely those whose blood pressure is difficult to control with single agents. In this presentation, the blood pressure changes at week 24 are summarized for study group 1 subjects stratified according to their baseline severity of hypertension; that is, from left to right, mild or JNC VI stage I, moderate or JNC VI stage II, severe or JNC VI stage III.

The difference between omapatrilat and enalapril at week 24 is present in all three groups, but it is most apparent in those with most severe hypertension at baseline in whom, as shown on the right-hand panel of the slide, the rate of use of adjunctive therapy was also the greatest. This suggests that the benefit of a more efficacious antihypertensive agent might be greatest in those most likely to require combination therapy, those with hypertension that is difficult to control.

Another representative group of patients with difficult-to-control hypertension is those who remained significantly above blood pressure goal in spite of treatment with existing therapies. In data from patients randomized to OCTAVE study group 3 who continued to have JNC VI stage II hypertension in spite of treatment with two or more antihypertensives or three or more antihypertensives at baseline are shown on this slide. In these patients, the addition of omapatrilat provided significantly greater blood pressure reduction compared to the addition of enalapril, demonstrating the benefit of adding a more effect agent.

The FDA review has raised the question that the efficacy difference between omapatrilat and enalapril may be easily overcome with greater use of adjunctive therapy. I would like to make two important points here.

First, many of these difficult-to-control patients are already on multiple treatments and have limited options for additional therapy.

Second, many of these patients remain significantly above goal even after adding enalapril or omapatrilat, as illustrated here in these 700 patients in whom the rate of control with enalapril on top of three baseline meds is only 28 percent at the end of the study, and even with omapatrilat only 42 percent reached target. If there is opportunity to add more treatment, if there are options, it would occur in both patients treated with omapatrilat and those treated with enalapril. And while the use of adjunctive therapy would increase with both drugs, the blood pressure reduction in the regimen with more effective components would still be grater.

Hence, in patients with difficult-to-control hypertension, a regimen containing omapatrilat would be expected to provide persistent benefit compared to a regimen using less effective agents due to the greater antihypertensive efficacy of omapatrilat.

To maximize the benefit this drug has to offer, we are focusing on patients with high cardiovascular risk and hypertension that is difficult to control with existing agents, and I'd like to provide you with some more data in patients not achieving blood pressure goal on current therapies.

I'll now present data collected in another group of patients with hypertension that's difficult to control with existing agents, namely those who are resistant to ACE inhibitor therapy. This group of patients is of particular interest since ACE inhibitors are widely used to treat hypertension and since omapatrilat acts in part through ACE inhibition. I'll review data from two sources, a study conducted specifically in ACE inhibitor resistant patients, study -73, and the large number of such patients from OCTAVE.

This slide summarizes the design of study -73, conducted in patients who remained above blood pressure target despite aggressive ACE inhibitor therapy. Subjects with systolic blood pressure of at least 140 millimeters of mercury or diastolic pressure of at least 90 millimeters of mercury despite therapy with an ACE inhibitor at maximal or near maximal dose for at least a month were enrolled and, after a 2-week stabilization period, randomized to treatment with either omapatrilat starting at 20 milligrams and up-titrated to 80 milligrams, or lisinopril starting with 10 milligrams and up-titrated to 40 milligrams.

The lisinopril arm was intended to reproduce under blinded conditions the potential effects of continued therapy with maximal ACE inhibitor. All patients were treated with top doses of omapatrilat or lisinopril for 4 weeks prior to the final evaluation. Ambulatory blood pressure was used as the primary method for the assessment of treatment effect.

At study's end, 24-hour ambulatory systolic blood pressure was reduced 8.8 millimeters of mercury more with omapatrilat than with lisinopril. Blood pressure was also reduced more with omapatrilat than with lisinopril at each time point during 24-hour ambulatory blood pressure monitoring. While the differences between omapatrilat and lisinopril were greatest during the daytime hours, a difference of 7 millimeters of mercury persisted at trough, 24 hours post dose administration.

The results of this study indicate that patients resistant to ACE inhibition are not equally resistant to treatment with omapatrilat and suggest that omapatrilat can be used as an alternative to ACE inhibitors to provide substantial additional blood pressure reduction in patients failing to reach target with an ACE inhibitor.

Of course, subjects treated with an ACE inhibitor alone could achieve additional blood pressure reduction through addition of a second or third agent. This study evaluated not only patients uncontrolled on ACE inhibitor monotherapy, but also those uncontrolled on ACE inhibitor as part of a combination antihypertensive regimen. In such subjects, the ACE inhibitor was discontinued at randomization while other antihypertensive medications were continued without alteration in dose.

As shown here, reductions in blood pressure were highly consistent whether subjects entered the study on ACE inhibitor monotherapy, as shown in the left-hand bars, or on an ACE inhibitor as part of a combination antihypertensive regimen, as shown in the right-hand bars. Numerically the reductions in systolic blood pressure relative to enalapril were greater in those who entered the study on an ACE inhibitor-containing regimen than those who entered the study on an ACE inhibitor monotherapy, 11.5 versus 7.6 millimeters of mercury.

Now, OCTAVE also included over 4,000 subjects whose enrollment blood pressure remained above target despite therapy with an ACE inhibitor or ACE inhibitor-containing regimens. In these patients, prior treatments were discontinued at study entry and patients were randomized to either omapatrilat or enalapril.

Blood pressure was reduced consistently more with omapatrilat than with enalapril whether patients were receiving an ACE inhibitor alone at randomization or as part of a regimen containing one or more additional antihypertensives. Numerically the greatest reductions relative to enalapril of about 6 millimeters of mercury were observed in those receiving an ACE inhibitor plus two or more antihypertensive medications at randomization.

Now, the proposed target indication also includes patients with difficult-to-control hypertension who have comorbid conditions and other characteristics that put them at increased risk of cardiovascular events. As representative data for this population, the results from OCTAVE in subjects with diabetes and blood pressure above target at enrollment despite ACE inhibitor therapy are summarized here. About 1,000 patients are included in this analysis.

Omapatrilat reduced blood pressure significantly more than enalapril in these subjects whether they had been treated with an ACE inhibitor alone at randomization as shown in the left-hand bars or with an ACE inhibitor-containing antihypertensive regimen in the middle and right-hand bars. Reductions in blood pressure with omapatrilat relative to enalapril ranged from about 5 up to 9 millimeters of mercury, and the greatest reduction was again observed in those receiving the most intensive antihypertensive regimen at baseline.

In summary, a large clinical development program has demonstrated that an omapatrilat-based regimen reduced blood pressure more than the regimens containing enalapril. This blood pressure advantage was consistent across patient subgroups regardless of the manner of the use of the study drug. And OCTAVE further suggested that the blood pressure advantage observed with omapatrilat in clinical trials can be maintained under clinical use conditions.

Lastly, data from OCTAVE, as well as data from other trials, indicate that in patients that cannot readily achieve blood pressure target with existing drugs, omapatrilat provides further blood pressure reduction that's not otherwise available.

Now, let's go on to safety. In a few minutes, I'm going to ask Dr. Kaplan to come to the podium to present an overview of angioedema.

DR. BORER: I'm sorry. Just before you do that, because these are a lot of data and there will be a lot of questions, maybe if it's okay we can stop here and ask questions to clarify the efficacy data, and then we'll move on to the safety and do the same thing.

Does anybody on the committee have substantive questions about the data? Tom.

DR. PICKERING: Yes. I have one general question. It's well known that ACE inhibitors' effectiveness is increased by sodium depletion or diuretic treatment, and I don't think in any part of your presentation you specifically referred to the use of concomitant diuretics. I don't think I've seen any head-to-head comparison between omapatrilat and an ACE inhibitor-diuretic combination, which many of us use in clinical practice. Do you have such data?

DR. LEVY: If I could refer to my backup deck for a moment. Thank you. Could I have slide HP-8?

What we've done here is summarize the blood pressure reductions at study end in patients who received a variety of additional therapies after week 8. On the left-hand panel are displayed the findings in those who received hydrochlorothiazide in addition to either omapatrilat or enalapril, as well as those who received a variety of other antihypertensive agents. As you can see, both omapatrilat and enalapril have additional efficacy when supplemented by hydrochlorothiazide, but the blood pressure reduction with omapatrilat remains greater. And the same is true really regardless of the antihypertensive agent or class which is added on top of omapatrilat or enalapril.

DR. PICKERING: You are saying that no study has been done with a randomized direct comparison between omapatrilat and ACE inhibitor-diuretic combination. Is that correct?

DR. LEVY: Yes. We're not proposing that the drug be used in patients who can readily be controlled with an ACE inhibitor-diuretic combination. And the patients I've shown you are patients who are typically already treated with combination therapy in whom the option of adding a diuretic to an ACE inhibitor is no longer available.

DR. BORER: Are there other substantive issues? Bob.

DR. TEMPLE: This is to some extent the same question. But on slide 25 where you're looking at ACE inhibitor plus two or more antihypertensive meds, what would those antihypertensive meds have been? I ask because it matters. For example, if they're all on beta blockers, you don't really expect too much more. The effectiveness overlaps. Were they all on diuretics, as they presumably should have been? What were they on?

DR. LEVY: The majority of these patients were on diuretics, and then, of course, the third med was a variety of medications, in some cases a calcium channel blocker, in some cases a beta blocker.

DR. TEMPLE: Okay, but again I ask because the question that I'm sure will come up repeatedly always is this extra 3 millimeters or 10 millimeters or whatever it is -- could you have done it just as easily by adding amlodipine? So, all of these things raise that question. I'm just trying to direct it there early because I think that's going to come up repeatedly.

So, those people would have mostly been -- it's not that many, but 169 of them -- on at least a diuretic, do you think? Do you know exactly?

DR. LEVY: Can we go back a slide? Again, this is a cut of a cut, but the previous slide, slide 24, is of a larger number of patients who were on an ACE inhibitor at randomization and failed to reach target. As you can see, there was almost 600 in the group on two or more antihypertensive meds. And yes, these patients are in general receiving an ACE inhibitor, in most cases plus a thiazide diuretic and then a calcium channel blocker or a beta blocker.

DR. TEMPLE: You don't ave a precise breakdown of that.

DR. LEVY: We can provide you with that information later, if you'd like.

DR. TEMPLE: Okay.

DR. BORER: Paul.

DR. ARMSTRONG: My question is in the same area. Just to pursue this, if we're going to get more data to see later, not only would I be interested in the types of adjunctive therapies that were added in instances where one or other choice of therapy in OCTAVE was perceived to be inadequate, but the doses of those agents. In other words, were the doses of those agents pushed to equal intensity in the instance where it was perceived that the primary therapy had failed?

DR. LEVY: Certainly there's a wealth of data on that question. Let me provide you with the one patient subgroup where the data is most clearly defined. As mentioned, this was a simple trial. The case report form was simple, and the amount of information about study drug dosing is therefore limited. But for a few certain drugs, we do have specific dosing information, and perhaps I can show you some information there that will illustrate what happens when omapatrilat is used rather than enalapril in patients who were receiving high-dose, aggressive antihypertensive regimens. I think they're representative of the whole study, but for this particular subgroup, we have very detailed information about dosing.

These would be our slides comparing the efficacy of omapatrilat and enalapril at week 24 in study group 3 subjects who entered the trial on a two-drug regimen, including amlodipine and hydrochlorothiazide.

DR. BORER: While you're pulling that up and looking for slides, Steve is our committee reviewer, and he'll have a number of questions I'm sure. But I had a specific question on the same issue and that was from your slide number 23 where the addition of lisinopril actually caused no change, an average .6 millimeter of mercury increase in blood pressure, when added on to other therapy. And I too wanted to know what the other drugs were, what their doses were, how you would explain that, what the population was. Were there some vagaries there that could explain the absolute lack of any activity of the ACE inhibitor in that population? So, while you're looking all this up, go back to your slide 23 also, if you would.

DR. LEVY: All right. Let me answer this question first and then I'll return to your question.

Again, in this group we have the most specific information about study drug dosing and the doses of adjunctive therapy. Now, these are patients who entered the trial uncontrolled on two or more antihypertensive agents, which included amlodipine or hydrochlorothiazide, and for those two drugs we have the dosing information. These patients at baseline had blood pressures that remained at JNC VI stage II, systolic pressure of 160 to 169 or diastolic of 100 to 109. And their mean systolic pressures were about 166.

These patients were receiving a minimum of hydrochlorothiazide and amlodipine. The mean dose of amlodipine was 7 milligrams. The mean dose of hydrochlorothiazide was 20 milligrams. So, the patients were about split between amlodipine 5 and 10 and hydrochlorothiazide 12.5 and 25. About 40 percent of them were also receiving a beta blocker, and 10 percent were also receiving an angiotensin receptor blocker. So, over half of these patients were actually receiving three drugs at randomization.

They then received, in addition to their existing therapy, omapatrilat or enalapril. As I mentioned, these drugs were used very aggressively in the course of the trial. Over 60 percent of these patients were titrated to enalapril 40 milligrams, which is a dose that's considerably higher than that generally used in clinical practice, and then some number of these patients received a fifth or a sixth antihypertensive agent in the course of the trial.

So, these represent really an extraordinarily aggressively treated group of patients in terms of the number of drugs and the dosing of those drugs. And there's still an advantage in both systolic and diastolic blood pressure reduction at the end of the study. So, I present these as representative. I happen to have the most detailed dosing information for these patients, but they're presented to you simply because you asked a question about dosing.

DR. BORER: Do you have any idea why 7 milligrams and 20 milligrams was the average? It's certainly not the maximum labeled dose of amlodipine, and the thiazide dose, though, one can go way up the scale. One might choose not to because of safety issues, but 20 milligrams is kind of low. So, why is it that those adjunctive therapies or those initial therapies were limited in those patients? Do we have any idea at all?

DR. LEVY: Well, you know, in practice physicians tend to prefer the use of low-dose therapies, particularly for drugs that do have dose-related toxicity. Amlodipine has a much higher incidence of peripheral edema at 10 milligrams than at 5 milligrams. That may have been a factor in physicians' choice of 5 milligrams in some half of these patients, and thiazide diuretics also have a dose-related adverse effects that may have influenced the selection of study dose. But these are actually relatively high doses compared to those encountered in usual practice.

DR. BORER: Right, but that's not really the point. What's encountered in usual practice may be reasonable or it may be unreasonable, and I know that there are many reasons people may do things. That doesn't make them rational or right. The question is, do we actually have information about why people weren't given the higher doses? Maybe you don't, and I'm not suggesting you had to have such information but I'm just asking if you do.

DR. LEVY: No, we don't.

DR. BORER: Before you go on with Paul's issue, can you go back to your slide 23? Do you know anything about that group that received lisinopril on top of something else?

DR. LEVY: Again, this isn't OCTAVE. This is a trial that was specifically conducted in patients who were resistant to ACE inhibitor therapy. In this study patients who were on combination regimens discontinued the ACE inhibitor at randomization but continued all other medications. They were already on maximal or near maximal ACE inhibitor therapy at randomization. So, that meant lisinopril 20 milligrams, enalapril 20 milligrams. What you see there reflects the replacement of their prior ACE inhibitor with study ACE inhibitor, which was titrated to 40 milligrams.

DR. BORER: Do you have any information about the characteristics of that population that would explain their relative resistance just for our edification?

DR. LEVY: Well, there were more black patients represented in this study and a slightly higher incidence of diabetics, characteristics which might be associated with diminished response to drugs which inhibit the renin-angiotensin system. But it was actually a quite representative hypertensive population.

DR. BORER: Paul, have you completed?

DR. ARMSTRONG: Yes.

DR. BORER: Steve.

DR. NISSEN: First of all, I really want to compliment BMS on one of the most extraordinary development programs for a hypertensive drug. The number of patients studied, the robustness of the data is really quite extraordinary. I think there a lot of insights, obviously, to gain from a 25,000-patient study.

I also wanted to say I really appreciated the review from Drs. Lawrence, Stockbridge, and Throckmorton. I think we had a really comprehensive package. So, we've got a lot of information and I want to go through a little bit of it.

I wanted to begin by asking something about mode of action, and the question I want to get at is why does this agent have greater antihypertensive efficacy. I'll offer you a hypothesis, and I want to know whether there's any data to support it.

The hypothesis is that by increasing levels of natriuretic peptides, that there's a weak diuretic effect from the drug. So, what we're looking at here is something that looks like the combination of an ACE inhibitor with a very weak diuretic. As we all know, when you add a little bit of diuretic to an ACE inhibitor, you get a lot of bang for the buck. You get a lot of blood pressure reduction, even 6.25 milligrams of hydrochlorothiazide will add a few millimeters of blood pressure reduction to ACE.

Is that really what we're seeing here that we have in a single compound a drug that's combining a little bit of diuretic effect with an inhibition of the renin-angiotensin system? And any of your consultants, if you could shed some light on this, I would be appreciative.

DR. LEVY: If I can just make a few comments. That's an excellent question. Certainly when we began developing the drug, it was a major question. In our clinical pharmacology program, in which subjects were actually studied in clinical research units and their intakes and outputs could be carefully measured, there was no evidence of a natriuretic or diuretic effect with omapatrilat at doses well above and below those studied.

In our hypertension development program, as I've shown you, patients who received a diuretic in addition to omapatrilat, experienced the same incremental reductions in blood pressure that one sees when adding a diuretic to an ACE inhibitor, suggesting that its additional antihypertensive effect is not mediated through diuresis. It appears to be the vasodilator effect of the natriuretic peptide that contributes to the antihypertensive effects of this drug. In particular, the drug may have a unique central vasodilatory effect on the large conduit vessels.

DR. NISSEN: You did formal salt balance studies and that sort of thing. Are those available for us? Because I think that would be very interesting to see is, in that first week after you start the drug, what happens to salt balance, not later on, but as I understand diuretics, what you see is an initial fall in sodium and then it returns to normal again. I'd be very interested in seeing any salt balance studies that you have.

The reason it's relevant I guess is let's suppose that that's right, that this is a drug that has weak diuretic properties. Then it still might be true that adding additional, say, hydrochlorothiazide to the regimen would produce additional blood pressure reductions. I mean, if you go from 6.25 to 12.5 to 25 milligrams of hydrochlorothiazide, you see additional efficacy. It's highly relevant in my view because it speaks to Tom Pickering's question, which is, is the real comparator for omapatrilat ACE plus a little bit of diuretic? Is that really what we're talking about as a comparator?

DR. LEVY: Well, of course, we're not recommending that the drug be used in patients who can be controlled with ACE plus a little bit of diuretic. We're proposing it be used in patients who can't be controlled with an ACE-diuretic combination, and there's evidence that it provides substantial incremental benefit in those patients.

So, with regard to the mechanistic considerations, with a thiazide-diuretic, one would see a brisk diuresis within hours of administration of the drug, an excretion of 200 to 300 millimoles of salt. We don't see anything like that with early dose administration.

DR. NISSEN: So, there are salt balance studies that you can provide us to take a look at?

DR. LEVY: There are studies conducted in which urinary sodium excretion is measured over the first hours of dosing and the first 24 hours of dosing, a time period in which the effect of either a thiazide or a loop diuretic would be unmistakable. And we don't see anything at all.

DR. NISSEN: Okay.

I wonder if you could bring up slide 11. There are some things I didn't understand here, and I really want to explore it.

Let's look at the add-on group, or group 3. Now, the entry criteria for group 3 was what entry criteria? What did you have to have to be in group 3?

DR. LEVY: These patients had to be uncontrolled on antihypertensive therapy with blood pressures at JNC VI stage II, or a systolic blood pressure of 160 to 179 or a diastolic blood pressure of 100 to 109. These patients at randomization continued their existing antihypertensive therapies and added omapatrilat or enalapril.

DR. NISSEN: I thought that's what I heard, and then I was confused because the baseline blood pressures in this group are actually lower than the minimum requirement to get in that arm of the trial. When I read this last night, I just couldn't understand how that could possibly happen.

DR. LEVY: Well, that's a very good question. Remember, patients could enter the trial by satisfying either the systolic blood pressure criteria of 160 to 179 or the diastolic blood pressure criteria of 100 to 109.

Now, you made a very important observation. The systolic pressure is 166 which is within the target range, while the diastolic pressure is below. That reflects the difficulty in achieving systolic blood pressure control in populations. Failure to control systolic blood pressure is the primary reason for difficult-to-control hypertension. And Dr. Black is going to address this issue in more detail at the end of this talk. So, it's not a defect the study design. It really reflects the extraordinary difficulty that physicians have in bringing systolic blood pressure under control with existing medications.

DR. NISSEN: Well, let me tell you what I'm concerned about. Again, we're trying to tease out the group that might benefit here. So, this group 3 was going to be people who were just refractory. They couldn't be controlled on existing medications. When I see a group that's 166 over 97, it seems a lot less refractory to me than the entry criteria would look like. My guess is sometimes when you do a trial of 25,000 patients in less than a year, you've got to get patients in the trial, and so investigators tend to be a little more aggressive and maybe initial blood pressures were a little bit lower than you wanted them to be. So, I'm not sure how refractory that group 3 is. It does color my interpretation of the data when I see that the average blood pressures are pretty low and really below the targets in that group. Do you follow me?

DR. LEVY: Perhaps I'll ask Dr. Black to comment at this point.

DR. BLACK: Yes, thanks very much, Elliott.

Steve, I just want to say that I don't think those are really low at all. That's the world through a diastolic window, not through a systolic window. The problem we have, as I'll show you a little bit later, is not that we can't control diastolic pressure. It's that we can't control systolic pressure. In fact, arteries get stiffer in diastolic falls if you leave people untreated. So, pulse pressure widens and I think that's the group you're looking at.

DR. NISSEN: Well, the reason this is germane is you've said several times that you want to target this drug at those people that are very, very difficult to control with conventional regimens. What I see is in OCTAVE, a 25,000-patient trial, in each of the three arms the blood pressures are not extraordinarily elevated. So, I know you have some people in OCTAVE that were very, very high, but I'm interested in understanding whether there is, in fact, a target population identified here that would be optimally benefitted. It's harder when the average blood pressures in the trial are not as high as one might have expected.

DR. LEVY: Let me return to that point. Again, in these study group 3 patients, the systolic pressure at baseline is at least 27 millimeters of mercury above target on treatment. In those with diabetes, renal failure, heart failure where the treatment target is 130 millimeters of mercury, the blood pressure is 37 millimeters above target. If a patient walked into your office untreated with those pressures, you might be able to bring them down with one, two, or three medications. If a patient was already on two or three medications, the opportunity to reach target is very, very limited.

Again, we had patients in this group -- and I've shown you the results -- patients who were on two drugs at randomization, patients who were on three drugs at randomization. On three drugs at randomization with blood pressures in this range, the addition of very high-dose enalapril, making them on a four-drug regimen, plus other drugs, you still only get 28 percent of them to target.

DR. NISSEN: I agree that group is certainly a target group.

But I did want to look at the group that's really much more severe. I know you did a study, 137-049. I'm sure you have those slides. I'd like to see that study because I think it helps us here.

DR. LEVY: Sure. Perhaps I could just begin by going back to the slide from our core deck showing the results in patients with severe hypertension in OCTAVE.

DR. NISSEN: Sure.

DR. LEVY: If I could have the table displaying the results by comorbidity. That's slide 16. Again, for patients who entered the trial off therapy, we could assess their underlying blood pressure. That's study group 1. And 1,000 of those patients had severe hypertension. If you were to include, as we did by design, those who entered the study on treatment on at least two antihypertensive medications, then the number with severe hypertension goes up to about 7,000. So, it's a very large experience. And the confidence intervals around the estimate of treatment effect are very narrow.

Now, in 1998 and 1999, we conducted an exploratory study in patients with severe hypertension. That study included about 160 patients, about two-thirds of whom were on omapatrilat and a third on enalapril. That study was designed to determine whether the drug effectively reduced blood pressure in patients with severe hypertension, and it did. It did not demonstrate a statistically significant difference between the groups, and it was not intended or powered to do so.

DR. NISSEN: Do you have a slide with the data?

DR. LEVY: Sure. We can look for that and come to it a little bit later.

DR. NISSEN: I'd just like to see it because as I recall, the entry criteria for that -- there it is. That's the study.

DR. LEVY: This was a little bit different study population. The focus in registrational trials in hypertension has been on diastolic blood pressure, even though that's not the critical variable of the population. So, this study looked at patients with diastolic pressures of 115 to 130 off treatment. It's actually a very narrow segment of the severe hypertensive population.

DR. NISSEN: Okay, but nonetheless, these are pretty severe. So, it obviously does send us a signal that we'd like to see. Show us what happened with this group.

DR. LEVY: Can we see the primary efficacy results in this study? These are results at week 10. These are regimen comparisons. Virtually every patient was on multiple drugs by this time, many on three drugs, and blood pressure was reduced with both drugs. It's reduced about a millimeter of mercury more with omapatrilat than with enalapril in systolic blood pressure and about 2 millimeters of mercury more diastolic with omapatrilat than with enalapril.

DR. NISSEN: Did that result surprise you?

DR. LEVY: No. It's a small study and it was designed to compare regimen versus regimen. It wasn't designed to determine if omapatrilat reduced blood pressure more than enalapril. In fact, there was no planned statistical comparison in this study and it wasn't powered to make one.

DR. NISSEN: All right, fair enough.

DR. TEMPLE: Before you leave that, that's a pretty large antihypertensive study. It's not a small study.

DR. LEVY: I don't think that 60 patients in the enalapril arm is very large. In any case, it's a lot less than the 1,000 we have in OCTAVE.

DR. NISSEN: Actually there is a little discrepancy, Bob. In the FDA briefing package, the endpoints are shown, but they're actually opposite to that. They show actually that there was a little bit greater efficacy with enalapril than omapatrilat. I'm not sure which is right.

DR. THROCKMORTON: In this study?

DR. NISSEN: Yes, I think so. I'll pull it.

DR. THROCKMORTON: I don't think I included this particular study review from the original efficacy. I'm looking at my original package, and I don't remember doing that because that study, as Elliott said, when we looked at it originally, was very small and had no statistical plan even associated with it. So, we did relatively less with it. But we can double check that.

DR. NISSEN: Fair enough.

I'm exploring with you because obviously one of the things that we're trying to weigh here is risk versus benefit.

DR. BORER: Excuse me just one second. There is a mention of numbers and they do appear to be in the opposite direction.

DR. NISSEN: I thought so.

DR. BORER: Severe hypertension in CV137-049. These pages don't have numbers on them, so I can't tell you where in the review it is. But it does say that the change from baseline seated diastolic blood pressure was similar for the two groups, minus 26 for omapatrilat and minus 29 for enalapril.

DR. NISSEN: So, they're reversed from what's in there. I understand the limits of the statistical comparison here. Your point is well taken.

Let me tell you what I'm trying to explore with you. We're trying to weigh here risk and benefit, and obviously showing that in the very severe hypertensive, you can get them a much better chance to get them to goal has a real impact on our thinking about the relative risk and benefit of a drug. So, I was interested in 137-049, and I wanted to look at it with you because we just had one paragraph about it in our briefing book. And I wanted to understand what was done there, and I understand it wasn't a huge study. It doesn't compare to the 25,000 patients in OCTAVE, but I wanted to at least understand what it was all about.

Now, the next issue I wanted to go into -- and, Jeff, I won't take much longer because I think we want to move on -- is you compared to once-a-day enalapril. We had a rather extensive discussion yesterday on the issue of once-a-day versus twice-a-day drug dosage.

Now, the differences were about 3 over 2 millimeters, something like that, between once-a-day omapatrilat and once-a-day enalapril. One of the questions that I needed to have answered was, what might we have expected if the enalapril had been given as 20 milligrams b.i.d.? Remember now, we're going to try to calculate a benefit versus a risk. So, the differences between those two regimens is very, very important. What would the difference have been if we had given enalapril 20 b.i.d. rather than, say, 40 milligrams once a day? Any information about that?

DR. LEVY: That's a very good question. I really can't speculate about that. We didn't do a study versus b.i.d. enalapril. We chose an enalapril dose regimen that reflects the way physicians give chronic therapy to patients in practice, which is once a day.

Now, we do have a variety of studies against other agents, studies in which the optimum effect of omapatrilat was compared to the optimum effect of those agents, and that includes comparisons with not only lisinopril and losartan, but also with amlodipine which is an extremely long-lived, once-a-day drug. There again there is superior efficacy.

DR. NISSEN: Well, let me tell you what triggered me to ask the question. Since you're going to present OVERTURE and I don't want to presage that, it's interesting that in OVERTURE you gave the enalapril b.i.d. and in the hypertensive patients, there was exactly the same blood pressure reduction between omapatrilat and enalapril given b.i.d., 12.6 and 12.7 millimeters. So, I was left saying, gee, what if OCTAVE had done that? Could that have completely erased the blood pressure differences between the two regimens?

DR. LEVY: Again, it's hard to imagine it would do that in patients whose blood pressure remains uncontrolled. It's difficult to control with regimens like twice-a-day enalapril or twice-a-day enalapril plus a thiazide diuretic patients who need more therapy.

DR. NISSEN: Michael, you look like you have some thoughts about that.

DR. WEBER: I was going to suggest, Steve, that we take a look at the ABPM data because, in fact, that does show pretty good 24-hour efficacy for the ACE inhibitors as well. Do you have the ABPM data with the lisinopril study? Slightly different than enalapril.

DR. NISSEN: Wasn't lisinopril a bit longer-acting?

DR. WEBER: Yes. Do we have ABPM data for enalapril in the resistant --

DR. TEMPLE: Lisinopril I think is labeled for once-a-day only because it's got a very long half-life.

DR. NISSEN: So, I guess the lisinopril ambulatory blood pressure data I wouldn't consider relevant.

You know, it's really an important question, and I know I'm kind of being a stickler here. But if I'm going to calculate the potential benefit versus the potential risk, I've got to know how much the difference between enalapril and omapatrilat is. If enalapril is given in an optimal way, that might be b.i.d.

DR. PACKER: Steve, I think you're asking a very important point. I was in the audience yesterday and I know the committee was discussing what constitutes a fair comparison. Would it be appropriate to compare a once-a-day drug which is being proposed for once-a-day use against a drug which is most commonly used and includes a labeling for once-a-day use.

Having said that, there is an extensive experience with the comparison of once-a-day omapatrilat to twice-a-day enalapril in OVERTURE. I'll be reviewing OVERTURE, but I just wanted to address the question about blood pressure.

OVERTURE was a heart failure trial, not a hypertension trial. I think it would be fair to say that hypertension specialists tend to pay more attention to how they measure blood pressure than heart failure specialists who tend to think of blood pressure as being a general phenomenon and generally estimated. That creates a lot of noise in clinical trials.

Second is that the blood pressure measurements were made at trough in OVERTURE before the next dose of the drug, and there are considerable data from another heart failure trial called the IMPRESS study comparing omapatrilat once a day with lisinopril once a day, which is also approved once a day for heart failure, showing that, yes, the blood pressures with omapatrilat and lisinopril come together at trough, but there's a huge difference during the day. Therefore, if you look at the cumulative effect over 24 hours, there's still a major difference between omapatrilat and the comparator ACE inhibitor. We couldn't document that in OVERTURE because we only have trough blood pressures.

DR. NISSEN: Would it be safe to say, Michael, whoever -- let me ask you this. Would it be safe to say that a regimen of 20 milligrams b.i.d. of enalapril might reduce blood pressure over the 24-hour period more effectively than 40 milligrams once a day? Would it likely narrow that gap of 3 over 2 millimeters or would it not?

DR. WEBER: It probably could, but I can't be certain of that, Steve, because certainly there have been plenty of other trials with enalapril given once a day where, in fact, I thought it did rather well throughout the 24-hour period. In fact, our experience with ABPM would suggest that enalapril may be fractionally better twice a day, just as you could say the same with losartan. In fact, we know that would be true. But still, we're talking about a very, very minimal advantage.

DR. NISSEN: 1 or 2 millimeters?

DR. WEBER: 0 to 1, .5 to 1.

DR. NISSEN: I guess the answer is we really don't know. Is that a fair answer?

DR. WEBER: Yes.

The other thing too is omapatrilat is a long-acting drug. It gives you 24-hour efficacy, but you might have noticed from the ABPM data that towards the end of the dosing interval its advantage compared with the ACE inhibitor is getting less. You could argue that omapatrilat twice a day would be significantly better than omapatrilat once a day as well. So, I'm not sure how far we would want to take this particular argument.

DR. NISSEN: Well, I guess I wouldn't buy that necessarily, Michael, and the reason I wouldn't is that I'm a clinician and I've got a choice. I can give an agent with a more adverse safety profile once a day and take a risk of angioedema, or I can give a drug that's got a better safety profile twice a day. That's a very relevant consideration regarding approvability because if I could get the same blood pressure reduction by giving a safer agent twice a day, it would be hard to argue in favor of the less safe agent once a day I think.

DR. WEBER: Yes, but let me remind you of the patients who are resistant to ACE inhibitor, the study that Elliott showed before. The difference was really quite considerable between omapatrilat and enalapril in that setting, and I don't think giving enalapril twice a day there would have really compensated for those kinds of millimeters of mercury.

DR. NISSEN: I have two other brief questions, Jeffrey, if you don't mind. A couple of interesting things.

I was very struck by your slide number 6, if you want to show that. There's an interesting question that it raises. So, in 037 you were studying African Americans, and in 030 the comparison was amlodipine. So, given the fact that lisinopril didn't work as well in African Americans -- and neither did omapatrilat -- I'd be interested in whether you have any comparative data comparing omapatrilat to amlodipine in African Americans. Did you do any of those comparisons?

DR. LEVY: Well, there were small subset comparisons within each of these trials that are done, and about 10 percent of the subjects in each of the trials in unselected populations tend to be African Americans. In general, all those subgroup cuts are very consistent with the overall study results. There's a superior efficacy for omapatrilat.

DR. NISSEN: I seem to remember somewhere in Dr. Throckmorton's review some studies where that comparison was made where, in fact, in that subgroup omapatrilat actually produced less effect than amlodipine in the African Americans. I'm not surprised by that, but it's an interesting issue about choice of drugs in patients. There the risk-benefit really does shift quite a bit.

Doug, didn't you review that somewhere? Do you have that, Jeff?

DR. BORER: I think the statement is correct that in general the results look qualitatively similar by race. There may be a little bit more effect in non-black than black, but the results are qualitatively similar.

DR. LEVY: If I could comment, though, it's not our intention that omapatrilat should be used in patients who can readily be controlled with a safer agent. Particularly in black patients, we surely are not suggesting this drug should be used in place of a dihydropyridine calcium channel blocker in a patient who could be controlled on those drugs.

DR. NISSEN: I have one more brief question. The other questions I have on efficacy really relate to the issue of target organ protection, but I think I'm going to wait on those, Jeff, until after we hear from Henry and so on.

So, the one final question I had was on your slide number 3. I want to make sure I understand the entry criteria. So, this is the group you're proposing the drug is most likely to benefit. Was this criteria of presence of cardiovascular disease an entry criteria for OCTAVE?

DR. LEVY: It was not an exclusion criteria.

DR. NISSEN: But it wasn't necessarily an explicit one.

DR. LEVY: Well, I've shown you about 2,300 subjects had a history of MI or stroke or overt atherosclerotic disease at baseline. Heart failure was a small number, but there's of course a much larger number in the OVERTURE study.

DR. NISSEN: I want to come back to this later, but I do want to know subsequently. Since this is the population you're suggesting we should target with this drug, I will want to know more about studies done in such subgroups because, obviously, if you want to use a group in a subgroup, you've got to know a lot about it. So, I'll be interested later to hear about those people with known target organ damage, those people with post-MI, those people with three or more cardiovascular risk factors because, again, looking at risk-benefit, we need information about those groups if those are going to be the target groups that we're going to want to treat.

DR. LEVY: Right.

DR. BORER: Two final questions that I have for you. Again, you may not have specific information about this, and if so, you don't. But why were patients who were not adequately controlled stopped at two drugs? You had a number of patients who were given one additional drug, two additional drugs, or three additional drugs. And if they were not adequately controlled with two drugs, still a fair number continued on two drugs. Why was that or am I misunderstanding?

DR. LEVY: I'm not sure I understand the question. If you could point to a specific slide.

DR. BORER: Why if somebody's blood pressure isn't controlled would you not add additional drugs to try to control them? Was there something in the protocol that would have precluded that? Was there some suggestion in the selection algorithm that would have influenced that? I mean, if somebody's blood pressure isn't controlled, in general you'd want to continue to push the dose or push the number of drugs until you get it controlled. But I inferred from your slide -- and I'm sorry I didn't write down the slide number -- that a number of patients were given one additional drug or two additional drugs and still weren't controlled but continued on that regimen rather than being given an additional drug.

DR. LEVY: You don't know the slide?

DR. BORER: No.

DR. LEVY: I think there may be a misunderstanding, but I'll try to clarify that.

DR. BORER: I can probably find it easily enough here.

DR. LEVY: What I'd like to see is the slide from the core deck -- not this slide. Bear with me for a moment.

DR. PICKERING: I think it may be the protocol design. There were only two visits after week 8 -- is that right -- at which they could add additional drugs.

DR. LEVY: Let me first go to slide 20 in the core deck. I don't know if there's a misconception here. The number of meds. Those are the medications which the patient was receiving at study entry. Now, of course, there was no restriction on the number of medications that a patient could receive during the study.

And your point is a good one, though. If patients remain uncontrolled, physicians will continue to add drugs, and that's a very important point. They would do that. Obviously, most of these patients are not reaching target at the end of the study regardless of therapy. So, physicians would add drugs to both omapatrilat and enalapril.

DR. BORER: But did they? What I'm asking you is were there patients whose blood pressure didn't meet the target who were not on three drugs or more?

DR. LEVY: Yes.

DR. BORER: And why was that?

DR. LEVY: This is a 24-week trial. There are a discrete number of opportunities to add adjunctive therapy. Not every patient was brought to a three-drug regimen. Not every patient could be.

DR. PICKERING: Again, I think it was only weeks 8 and 16 that they had the opportunity to do that, so there was a limit to how many additional drugs you'd be able to add or dose-titrations you'd be able to do.

DR. LEVY: I think the larger question, though, is what can be accomplished with addition of a fourth, a fifth, or a sixth drug in patients who are multi-drug resistant. Maybe Dr. Black can speak to this question.

DR. BLACK: If I may, Jeff. This a practice-based study. You can, when you're doing a protocol, just encourage. You can't force necessarily a lot of physicians -- and there were lots of physicians in this -- to continue to add drugs. I'll show you some data late from our CONVINCE trial about what people used and where we ended up, another practice-based study with a fairly strict protocol, but we could not, in fact, insist that people went on. I think it's much like the question of why 7 milligrams of amlodipine and 20 of hydrochlorothiazide. I think people in practice dealing with individuals won't necessarily go to the top dose.

DR. BORER: No, I understand, and that's a very reasonable response.

My point only is that the fact that people don't -- and this is not a value judgment here, but you're proposing a very extensive and intensive education effort -- it's laudable; it's wonderful -- to try to make sure that pharmacists, patients, and doctors all know about the risks and minimize their impact, and I think that's wonderful. I'm just wondering if that same kind of intensive effort were used with regard to managing high blood pressure in the first place, we wouldn't have so many people on 7 milligrams or 20 milligrams of adjunctive drug and might have better blood pressure control.

And that's not your responsibility or anything like that, but I don't think we should judge the results here based on the fact that, well, this is a practice-based study and doctors don't always do what would be done in an academic medical center. That may not be the appropriate conclusion from all this.

But I'm sorry. Go ahead.

DR. BLACK: Yes. I think you reflect the frustration we had when we wrote the Joint National Committee report in 1997, looked at data on how poor control was in spite of a 25-year history of a very effective program. We did increase things. We've stopped and we need a much more aggressive physician and patient and pharmacist education program to improve control in general. We're not at all happy with 27 percent. NHANES IV looks as if we've improved things extremely little in spite of our awareness from JNC VI that this wasn't getting anywhere. We made some adjustments in JNC VI to try to make that more obvious, concentrate less on what drugs people use, but getting to a goal.

DR. BORER: Steve.

DR. NISSEN: Jeff, just to answer, I did find the comparisons that I was looking for with African Americans. If you want to see it, it's FDA table 7.12G.3. And I can't give you a page number, because there aren't any page numbers on there. But John Lawrence did the analysis.

What it shows is is that in the study 137-030, which was the amlodipine comparison, in black females omapatrilat was 7.9 millimeters worse than amlodipine with a p value of .01, and in black males it was 1 millimeter worse with no significant p value. So, there does appear, in fact, to be a racial difference, at least in the amlodipine comparisons, with omapatrilat being nominally worse in African Americans, but better in white males and females. So, it's a consideration here that I think probably needs to be out and discussed because obviously it's exactly that population where the risks of angioedema are the greatest.

DR. BORER: Okay, if there are no more questions, thank you that was very informative.

DR. FLEMING: On slide A-5, you're defining this configured target population. Can you show us -- because this, in essence, now is going to create a focal data set, I assume, from your perspective -- the population that meets these criteria in OCTAVE, baseline characteristics for the two arms and what the actual results were in terms of blood pressure control, as well as what the differences are in overall clinical endpoints in this group of patients in OCTAVE?

DR. LEVY: We've not prepared a pooled analysis in which all these patients are put together. I've shown you data regarding efficacy in patients with severe hypertension and data in patients with diabetes whose blood pressure is difficult to control with existing agents. We have data on efficacy in some of these other populations, which I'd be happy to show you as well.

DR. FLEMING: This is your target group that you're going to request be viewed as a group in which we will, hopefully, have a favorable benefit-to-risk. Correct? So that basically is it correct to say you would like to label the drug with this as the target indication?

DR. LEVY: The label is something that will be developed through discussions with the FDA. This is the intended target population.

DR. FLEMING: So, after the break, could you provide us, for this subpopulation of the trial, what the primary analysis would show for blood pressure control, differences in clinical events, and comparability at baseline?

DR. LEVY: Just to be clear, we've not done a pooled analysis in which we select all patients.

DR. FLEMING: I'm asking could you do so.

DR. LEVY: I don't know if we can do that between now and the break.

DR. FLEMING: Not between now and the break. Could you sometime after the break prepare that?

DR. LEVY: We'll certainly do our best.

DR. FLEMING: Have you not done this at all?

DR. LEVY: We have not prepared a pooled analysis of all these patients. I'll consult with the team. We'll do the best we can.

DR. BORER: What about each group individually? You've got four groups. Do you have data on each of the four groups?

DR. LEVY: Yes. I've shown you patients with severe hypertension who are represented here. I've shown you data for those with diabetes. There's data for other patient populations as well. I'd be happy to walk through all of that in detail. We have an enormous database. But as I say, we haven't put them all together.

DR. BORER: Would you accept that, Tom, looking at each subgroup individually?

DR. FLEMING: It's perplexing to me that we've done a major trial here. We're recognizing that risk was in excess of what we had anticipated. We make the logical conclusion that it might be that there is an important subgroup for which benefit could be particularly substantial. So, we define that subgroup, and we propose that this group be what we focus on as a retrospectively defined subgroup. And yet, we're not able to show what the overall benefit is and what the risk is in that subgroup. I'm assuming we can define whether or not the 25,000 patients individually would fit into this subgroup, so we ought to have been able to, in a fairly straightforward fashion, define what would be the primary efficacy outcomes and the safety outcomes in the subgroup.

DR. BORER: Doug?

DR. THROCKMORTON: Jeff, a minor thing. I looked back at the study 049, which was the relatively smaller study on resistant populations, and in fact, those two numbers that are in your briefing document are reversed. Again, I wouldn't make terribly large amounts out of them, but for what it was worth, the directionality was not different. That is, omapatrilat had the directionality towards a greater reduction than enalapril in that trial, which is the opposite of what's in the briefing document.

DR. BORER: Why don't we go ahead with the safety data and we'll come back to some of these efficacy issues later in the presentation.

DR. LEVY: In a moment, I'm going to ask Dr. Kaplan to come up to provide you with an overview of angioedema, but before I do, I'd just like to briefly provide a summary of the safety database.

The safety of the drug was characterized, as you know, in an extensive clinical development program, including about 35,000 hypertensive patients, 19,000 of whom were treated with omapatrilat. This, as you know, represents about 5 to 10 times the experience typically described in a hypertension new drug application. Large numbers of subjects were exposed to each of the proposed target doses. 13,000 were exposed for more than 3 months and about 1,500 for more than a year.

This extensive experience has provided an unusually clear profile of the safety of the drug. The overall incidence adverse events, serious adverse events, and discontinuation due to adverse events has been shown to be comparable for omapatrilat and enalapril. The risk of angioedema has also been clearly characterized and shown to be three times higher than with enalapril.

Because of the importance of angioedema in the assessment of omapatrilat, I'm going to ask Dr. Kaplan to come to the podium now. Dr. Kaplan is an angioedema expert who will provide a brief presentation on the pathogenesis and clinical spectrum of this entity before I return to complete the safety presentation. Dr. Kaplan.

DR. KAPLAN: Thank you very much, Dr. Borer, members of the advisory panel, and guests. It's a pleasure to be here today. What I'm going to try to do is give you a little overview about what angioedema is and what are some of the agents and circumstances in which it occurs.

I'm Professor of Medicine at the Medical University of South Carolina. I'm a clinical allergist, so I see angioedema all the time. And my research for 30 years involves the mechanisms of formation and destruction of bradykinin, which is directly germane to the drug that we are discussing today.

Now, angioedema is due to dilatation of small venules in the deep dermis of the skin. It's caused by a variety of vasoactive substances, but the vessels dilate, leak fluid, and cause swelling. And that's the common denominator of angioedema.

It has a predilection for various sites in the body, the most common of which are typically the face, particularly where tissues have low turgor. The most common site is the lip, but it often involves the eyelids, with periorbital edema, the cheek with an asymmetric swelling of the face. It can affect the tongue and it can affect the pharynx. When people have pharyngeal swelling, they will feel as if they are choking, even though their airway is not compromised. They will have difficulty swallowing and difficulty eating. On occasion angioedema will affect lower down and hit the larynx, and particularly we're concerned about vocal chord edema because then you're at risk of asphyxiating. It's uncommon but, nevertheless, there's a finite percentage who will have it. Other sites of angioedema are hands, feet, and genitals.

Among the common etiologies that we see of angioedema solo, without hives and without other manifestations, are a hereditary disease known as hereditary angioedema because the patients are deficient in a blood protein known as C1 inhibitor. In the absence of that C1 inhibitor, they overproduce bradykinin and that has now been proven to be the cause of the swelling and the hereditary disorder.

Similarly, the most common cause of angioedema that is exogenous -- that is, drug induced -- currently are ACE inhibitors. When you inhibit the angiotensin-converting enzyme, you not only prevent the conversion of angiotensin I to angiotensin II, but you're inhibiting one of three enzymes that are involved in the degradation of bradykinin. Therefore, by inhibiting degradation, bradykinin levels will tend to rise.

I should add that of those three enzymes are ACE, a plasma carboxypeptidase that is called carboxypeptidase N, and neutral endopeptidase. This drug inhibits two out the three, and that does distinguish it from ACE inhibitors because, given that information, the likelihood of bradykinin levels rising even more than you would see with an ACE inhibitor is at least theoretically possible and could account both for efficacy, as well as side effect.

Anaphylaxis and angioedema are different, and the reason I'll make a few particular comments about that is because they're often confused, and when patients present to the emergency room with angioedema, they often are treated for the other entity.

Angioedema, when it is due, let's say, to bradykinin in particular -- and that is in the hereditary deficiency, in the drug-induced -- typically evolves over several hours. An average might be 2 to 4 hours. But particularly severe cases may be more rapid and progress within an hour or two. We are not, however, talking a few minutes as is the case with anaphylaxis.

If you have facial swelling, in particular, I want to point out that patients typically are keenly aware of this even if they've never experienced it before in their life. A little lip swelling, a little eye swelling, just a little tongue swelling has people complaining early on. It's important because if we're going to talk about education, then it's important to have patient awareness early on to know that something is going wrong and be prepared for that eventuality.

How do you treat angioedema if it is due to bradykinin? Well, there are very few things that work and none of them are specific. People are often given antihistamines. That's, of course, worthless. They're given steroids, almost equally worthless, and it takes five hours for them to work. Epinephrine will work because it's nonspecific. It will constrict the vessels that are leaking and it will retard the angioedema from continuing. It will not take it away. It is just gradually then reabsorbed. So, the goal is to stop progression.

It is important also to note that the one that we're really worried about is laryngeal edema because it's the only one that causes airway compromise. I don't think I've ever seen a case of laryngeal edema that occurred solo without some other angioedema manifestation occurring with it. Usually lip swelling starts it. You may get tongue swelling, some pharyngeal swelling, and then the person complains of respiratory distress, first usually hoarseness, and then if it progresses, stridor.

In this last slide, I'm contrasting anaphylaxis with a drug-induced or hereditary angioedema, meaning the bradykinin-induced process. Anaphylaxis can occur in minutes. Infuse somebody with penicillin who's allergic to it, be stung by a bee while you're gardening and you're allergic to bee venom, and within a minute or two symptoms can begin, are often with generalized pruritus, followed by urticaria, angioedema, and then other manifestations. The patient will also often complain of like something really bad is about to happen, and we call it an impending sense of doom, if you will. But angioedema of the sort we're talking about doesn't evolve in quite that way.

In addition to the cutaneous manifestations, the key to anaphylaxis is that you now have cardiovascular manifestations and the hypotension and shock. That does not occur in the hereditary angioedema, nor does it occur in the drug-induced swelling.

Anaphylaxis can cause two syndromes, if you will, with regard to respiratory embarrassment: classical asthma where the person starts to wheeze and really has difficulty expiring; and laryngeal edema. Laryngeal edema is theoretically common to both. Bradykinin can do that solo. You don't get asthma in patients with this, but as you know, you get cough with ACE inhibitors.

To our knowledge, bradykinin is the only mediator of the angioedema that we are talking about whereas in anaphylaxis you release histamine, leukotrienes, platelet-activating factor, an array of cytokines, and just multiple vasoactive factors.

The treatment for anaphylaxis is, of course, epinephrine. Anaphylaxis tends to rebound. You can have somebody that has anaplylacted, but they're in the emergency room and they're making it. You've given them treatment, they start to feel better. You can be seduced to think that they're okay and stop treatment, and then five hours later, the syndrome may come back, not quite as bad, but it's there. Steroids stop that which is why it should be given, but it's not the first thing that you do. They also need to receive IV fluids and, of course, IV antihistamine such as Benadryl which does counteract the histamine.

In the drug-induced, if they receive all of these things, the only one that does anything is the epinephrine. Therefore, as you'll probably see, many patients that recover, even in an emergency room setting because they have gone there, who do not receive epinephrine but received all those other drugs, have spontaneously resolved without any treatment.

Thank you.

DR. BORER: Does anyone have any questions? Yes.

DR. NISSEN: Given what you said, there's an obvious strategy here for risk limitation that I had wanted to explore with you. If you had a drug that you knew had the potential to produce this, would it be prudent to give these patients an Epi-Pen? I know many of my patients who have had reactions to bee stings and so on carry that around. Could the sponsor here mitigate against this by giving every patient who is given omapatrilat an Epi-Pen so that they could self-inject with epinephrine if they get stridor?

DR. KAPLAN: Number one, of course, it would be a possibility which would theoretically be helpful and, if you had a reaction, would certainly tend to stop it.

There you have to balance. Now, the patient population that you're dealing with, if we're going to talk about the use of this drug in the most severe hypertensive who may have heart disease, arrhythmias, and who knows what else, now having them self-administer epinephrine has some risk associated with it. You might not want to willy-nilly give it to everybody, and if you're going to do it at all, it's either all or none. Therefore, you'd have to somehow rationalize how many people would have the side effect of the epinephrine that was worse than what was happening to them. Perhaps it would be -- I'm just giving you the counter-argument -- better to select out those who have the most severe swelling, get them to the emergency room promptly and let some physician make a decision as to whether it's appropriate to give epinephrine or not. But I think it is a point well taken, and it is at least one of the things that could be considered.

DR. NISSEN: Suppose a patient is -- let's say, African Americans who had, I think, about a 1 in 18 or 1 in 19 chance of developing angioedema in OCTAVE. Would that be a high enough risk group that you might think about it?

DR. KAPLAN: Yes.

DR. NISSEN: And smokers again, it was about 1 out of every 27 smokers got angioedema. That might also be a good target population.

DR. KAPLAN: Yes. And I wouldn't argue the point with you. My only concern would be that I'm sure among the smokers and the black hypertensives are people with some of the most complicated other things that are cardiac that you would have to deal with.

DR. BORER: Let's go Tom and then Susanna and then Paul. Tom.

DR. PICKERING: Thank you. I wondered if anything is known about C1 inhibitor deficiency in African Americans as compared with whites.

DR. KAPLAN: A C1 inhibitor deficiency is slightly less statistically of African Americans than in caucasians. Of course, it's rare to start with. That's in the hereditary disorder. There's a second form that is acquired and there the incidence is equal. It relates mainly to lymphoma. There are some people with lymphoma who express tumor antigens to which you make antibody. So, you have an immune complex and you fix-complement, and you can do so in massive fashion. You can fix so much of the first component of complement that the C1 inhibitor, which is the inactivator now binds to the activated first component and gets consumed. If the level of C1 inhibitor drops below 25 percent of normal, you're now at risk for having angioedema. So, the acquired form in lymphoma is a second type -- a third, if you will -- of bradykinin-induced angioedema, and there the incidence would be proportional to the incidence of the lymphoma in the population.

DR. CUNNINGHAM: I was wondering what you know about why smokers and African Americans are at greater risk for angioedema.

DR. KAPLAN: Knowing what I do about bradykinin, I certainly have thought about it and I could not answer the question. I don't know. Particularly the smokers. There are some data comparing blacks and whites with regard to end organ responsiveness to bradykinin with some interesting data that might explain that, at least in part, but there's nothing on smoking.

DR. BORER: Paul and then Doug.

DR. ARMSTRONG: Dr. Kaplan, first of all, thank you for contributing to my continuing medical education.

I'm interested in your thoughts about the epidemiology of angioedema in the general population, especially in the aging general population. I'm interested in your comments about the frequency of new onset allergy in the aging population such as, for example, fish or medicines or pollens, and the implications of those phenomenon in a patient taking a medicine that would inhibit bradykinin.

DR. KAPLAN: The incidence of a food allergy goes down in an aging population and therefore allergic urticaria and angioedema due to a food allergy is actually lower.

The most common disorder that we see that is not related to a specific allergen -- somebody walks into your office and says I've had hives and swelling for five months. I have no idea what's going on. I saw my internist, and they find nothing wrong with me. That turns out to be an autoimmune disorder due to, in part, at least in half the people, of a circulating antibody to the IgE receptor. So, the antibody cross links the IgE receptor just as if you had an allergen and they have waves of urticaria and angioedema that can last months to years. That is common throughout the population in all age groups, but I think in terms of allergy per se, even though it's going up in incidence in our population, it's almost all hayfever and asthma. It's not allergic urticaria or angioedema, and foods, in particular, goes down as we age.

DR. ARMSTRONG: So, a patient who develops late allergy for whatever reason who's taking an agent that inhibits bradykinin is no more likely to develop angioedema?

DR. KAPLAN: That's a tough question, but it has to be focused now only on an allergen for which angioedema is one of the manifestations. In other words, if you have hay fever and asthma, it's no more or less likely to be affected by an ACE inhibitor, nor will the allergen cause angioedema per se just because you're on the drug. On the other hand, if you give me a circumstance in which angioedema might otherwise occur anyway and you are on an ACE inhibitor, you'd be more likely to get it, even though the pathogenesis then would be multifactorial.

DR. ARMSTRONG: Do you think that the exclusion criteria in OCTAVE -- and there was some exclusion criteria associated with a history -- I was looking for exactly the criteria. I can't find them at hand, but what I'm trying to get at is how effective the exclusion criteria in OCTAVE precluded a higher incidence of angioedema in a hypertensive treated population then would otherwise have been the case, if you follow my drift.

DR. KAPLAN: I know there's no way of predicting, which is ideally what you'd like to do, as to who will have angioedema to any of these drugs. I'm sure there's an explanation. It could be some subtle, genetic polymorphism in ACE or other things that are involved with bradykinin, but we just don't know. So, I'm not sure whether I can be more specific in answering your question. Others involved with the study might be able to chime in because I'm not that close to it.

DR. ARMSTRONG: There's a statement about any drug-induced rash of any kind would have been an exclusion criteria in OCTAVE, for example.

DR. LEVY: I'd like to clarify that because that's not correct. Patients with a history of multiple drug sensitivities with a history of drug rash to two or more drug classes were excluded with the study, not patients with a history of a rash to any medication.

DR. ARMSTRONG: Do we know how many patients were excluded for that reason, Dr. Levy?

DR. LEVY: There's no way to know.

DR. KAPLAN: I could comment on that. There is a syndrome not well understood, a multiple drug hypersensitivity syndrome. A patient comes in and gives you a list of 10 medications. They get rashes to all of them. They go from one antibiotic to another, phenobarbital, an antihypertensive, and it cuts across classes of compounds and so on. It's reasonable in a study to eliminate them because they always come in and react to something, and you're just going to get into trouble.

DR. BORER: Doug.

DR. THROCKMORTON: Just one quick question. The statistical reviewer from the FDA appropriately pointed out that the number of cases of angioedema in this data set offers an unparalleled opportunity to look at angioedema and did some modeling as far as risks and things like that. I wonder if you could comment -- and you may be talking about this later, in which case it can come up later. Is there anything about the angioedema that you saw in this data set that suggests that it's of a sort that's different than the kinds of angioedema that you've been talking about up to now?

DR. KAPLAN: It's a very important question, and the details you'll hear in a few minutes by Dr. Levy. The differences are quantitative but not qualitative. A severe patient on enalapril looks like a severe patient on omapatrilat. A mild patient looks like a -- I could not qualitatively -- and as a member of the review group, we tried to determine who has angioedema, is it drug-related, blah, blah, blah. I see angioedema due to ACE inhibitors all the time. That I could not distinguish. So, it's not qualitatively different, but it may be quantitatively different.

DR. BORER: Can I just follow that up? Because I was struck by the model also in reviewing this and I was going to ask the question later, but I think you're the right guy to ask.

When I looked at that model, my inference was if only we knew how, we could identify the people at risk. It was a three-group fit that best fit the curve. I'm inferring from what you said earlier, that we have no basis --

DR. KAPLAN: No marker.

DR. BORER: -- to identify risk. I don't know if any work is going on within the company to try to do that. I assume there is, but right now there is no basis. Is that correct?

DR. LEVY: Let me just mention that we've had some ongoing work in that area, and perhaps once you've seen the safety data, we can share some of that work with you.

DR. BORER: Tom.

DR. PICKERING: As a follow-up to Dr. Throckmorton's question, is there any suggestion that the rate of progression of symptoms might be different in the omapatrilat than enalapril patients?

DR. KAPLAN: I don't think so. I think that when you see the data, the number that were considered "severe" was greater, but in terms of rate of progression, they looked exactly like what I'm used to seeing with any ACE inhibitor.

DR. BORER: Why don't we go ahead then. Thank you very much, Dr. Kaplan.

DR. LEVY: I'd like to thank Dr. Kaplan for that very interesting presentation and go on and describe for you in more detail the safety and particularly the problem of angioedema with omapatrilat.

Because the procedures used to assess angioedema in studies prior to OCTAVE and in OCTAVE were different, I'll describe the findings separately.

In studies prior to OCTAVE, angioedema was reported using standard procedures for reporting adverse events. The investigator typically provided a brief text description of the event which was then assigned a diagnostic code for the purpose of tabulation. The diagnostic codes were assigned using a dictionary based on the International Classification of Disease, or ICD-9. These procedures for reporting and classifying angioedema, which were identical to those used for the classification of all other adverse events, introduced certain limitations.

The ICD-9 based coding system assigned potential angioedema events to several different coding terms, depending on the actual verbatim text provided by the investigator. The most commonly used terms are "angioedema" and "head and neck edema." And while the term "angioedema" appeared to be quite specific for the event angioedema, the term "head and neck edema" was not specific, and the adverse event reports themselves didn't provide sufficient additional detail to further assess these potential cases.

The findings of studies conducted prior to OCTAVE are summarized here. A total of 44 cases of angioedema were reported. An additional 40 cases of head and neck edema were reported, which may have been angioedema. These are shown on the right. 4 subjects experienced angioedema with airway compromise which required mechanical airway protection, and it was these findings reported in the prior new drug application which prompted the FDA to ask if the incidence and severity of angioedema were greater with omapatrilat than that historically reported with ACE inhibitors.

In reviewing these data, we observed that the rate of angioedema appeared to be lower in subjects who began treatment with a dose of omapatrilat of less than 20 milligrams compared to those who began treatment with a dose of 20 milligrams or greater, in this case .45 versus 1.35 percent. Moreover, all four cases in which angioedema resulted in airway compromise occurred in subjects who began treatment with a 20 milligram starting dose, shown here. This analysis suggested that the incidence and severity of angioedema, particularly angioedema with airway compromise, might be reduced if patients were to begin therapy with a lower starting dose of omapatrilat.

The four cases of angioedema with airway compromise observed prior to OCTAVE are summarized here. All occurred in patients who had begun therapy with a 20 milligram starting dose. Two occurred on the first day of treatment, one on day 6 and one on day 11. All occurred while patients were receiving treatment with omapatrilat 20 milligrams prior to any dose titration. None of these cases presented in a fulminant manner; however, all required mechanical airway protection prior to resolution and all patients recovered without residual sequelae.

In the presentation that follows, I'm going to identify black race and current smoking as the two major risk factors for angioedema associated with omapatrilat, and one or both of these risk factors was present in 3 of the 4 subjects who experienced angioedema with airway compromise prior to OCTAVE.

Based on the observation that the incidence of angioedema appeared to be lower in patients who had begun therapy with doses of omapatrilat less than 20 milligrams, OCTAVE was designed in part to determine whether the incidence and severity of angioedema with omapatrilat could be reduced to a level comparable to that seen with ACE inhibitors if the starting dose of omapatrilat were reduced to 10 milligrams. Enalapril was chosen as a representative ACE inhibitor. And of note, the study wasn't designed to directly compare the incidence of angioedema with omapatrilat at starting doses of 10 and 20 milligrams.

Because of the difficulty encountered in previous studies in the accurate classification and counting of potential angioedema events, a special evaluation process was created for OCTAVE. Investigators were actively solicited to report all potential angioedema events using a special case report form page, and then detailed follow-up information on each potential case was collected on a structured questionnaire to ensure a consistent and complete database. Potential angioedema cases were adjudicated by an expert committee without knowledge of treatment assignment. The analyses that follow are based on cases confirmed as angioedema by that expert committee.

As you know, angioedema occurred in 274 omapatrilat treated subjects, or 2.17 percent, as compared to 86 enalapril treated subjects, or .68 percent. And the relative risk of angioedema with omapatrilat versus enalapril was 3.17.

Corresponding to the scientific hypothesis that reduction in the omapatrilat starting dose would result in a rate of angioedema comparable to that of enalapril, a statistical hypothesis was prespecified in which a significant increase in the incidence of angioedema with omapatrilat relative to enalapril would be excluded if the upper bound of the 95 percent confidence interval for relative risk was less than 2. And clearly, this hypothesis was not confirmed, but nevertheless a fairly precise estimate of the relative risk of angioedema with omapatrilat relative to enalapril was provided with reasonably narrow confidence limits.

An important secondary objective of OCTAVE was the assessment of the severity, as well as the incidence, of angioedema. Because no established classification systems for angioedema severity existed, a classification system was created for OCTAVE. Since it's not possible to obtain direct assessment of severity as these events occur, this system utilized treatment rendered as a proxy for severity.

The assignment of subjects to severity classes was performed by the event adjudication committee as part of their blinded review of angioedema cases. In this system, subjects receiving no treatment were assigned to severity class I, as were subjects treated only with antihistamines. Subjects treated with corticosteroids or epinephrine but not hospitalized were assigned to severity class II. Those who were hospitalized but did not require mechanical airway protection were assigned to severity class III, while subjects who required mechanical airway protection or subjects with fatal airway compromise were assigned to class IV.

It became apparent early in the trial that hospitalized patients were not consistently more ill than nonhospitalized patients treated with steroids or epinephrine and that at times patients were hospitalized for observation or other reasons. As a result, we asked the adjudication committee to identify patients hospitalized with serious angioedema by determining if airway compromise was present and assigning patients to class IIIa or class IIIb accordingly.

As you know, in OCTAVE angioedema ranged in severity from mild and self-limited to life-threatening. No deaths occurred from angioedema in OCTAVE. The majority of patients, about 60 percent, who experienced angioedema with omapatrilat received no treatment or antihistamines only and were assigned to severity class I. One subject treated with omapatrilat experienced angioedema with airway compromise requiring mechanical airway protection and was assigned to severity class IV. A second omapatrilat treated subject experienced anaphylaxis with associated angioedema and transient airway compromise which resolved without mechanical airway protection, and this subject was assigned to severity class IIIb. No enalapril treated subjects angioedema with airway compromise. 17 omapatrilat treated patients and 2 enalapril treated patients were hospitalized for angioedema without airway compromise.

Analysis of the relationship between severity class and treatment group showed that patients who developed angioedema on omapatrilat had higher severity classes indicative of a more intensive treatment pattern than those on enalapril. And in our review of the clinical manifestations of angioedema, we found that an appreciable difference between omapatrilat and enalapril events was the somewhat more frequent occurrence of tongue swelling and associated symptoms of difficulty speaking or swallowing with omapatrilat. And the more frequent occurrence of this highly symptomatic presentation may have led to this more intensive pattern of treatment.

Of greatest concern, of course, were the cases in which angioedema resulted in airway compromise. The rates of angioedema with airway compromise in OCTAVE and in all omapatrilat studies including OCTAVE are summarized in this slide. In OCTAVE, 2 patients, or 1.6 per 10,000 treated, experienced angioedema with airway compromise. If one places 95 percent confidence intervals around this rate, an upper confidence limit of 5.7 is seen, suggesting a rate of 6 per 10,000 as a worst case estimate. If one were to include all cases of airway compromise observed with omapatrilat, regardless of starting dose, a point estimate for the rate of angioedema with airway compromise would be 3.2 per 10,000 and the upper bound of the 95 percent confidence limit 7.0.

Now, it should be noted that the rate of angioedema with airway compromise observed in OCTAVE with the 10 milligram starting dose was distinctly different from the rate of angioedema observed in prior studies with the 20 milligram starting dose. In OCTAVE, angioedema with airway compromise occurred in about 1 per 6,000 treated. In prior studies with the 20 milligram starting dose, angioedema occurred in about 1 in 600 treated.

While not definitive and not a direct comparison, these data do suggest that the rate of life-threatening angioedema is lower with the 10 milligram starting dose and that the estimate of angioedema risk obtained from OCTAVE is perhaps the most relevant estimate for considerations of benefit and risk based on the recommended dosing. But whether one uses the OCTAVE estimate or the estimate from the entire clinical development program, the worst case estimate that runs 6 to 7 per 10,000 is not meaningfully different.

The two cases of angioedema with airway compromise that occurred in OCTAVE are summarized here. The first occurred in a white female who developed edema of the eyelids, lip, and neck, difficulty speaking and swallowing, hoarseness, hypotension, and cyanosis within 15 minutes of the first dose of omapatrilat. This presentation with systemic manifestations, including cardiovascular collapse, as well as angioedema, within minutes of exposure to the drug is characteristic of anaphylaxis and was diagnosed as anaphylaxis by the treating physicians. No other cases of anaphylaxis have been reported in the omapatrilat clinical development program. This subject was treated with epinephrine and recovered promptly. She was admitted to the hospital for observation and discharged the following day with no complaints.

A second case occurred in a black female during the 10th week of treatment with omapatrilat. She had been treated with omapatrilat 80 milligrams for about 4 weeks prior to the event without difficulty or dose interruption. Over a period of several hours, she developed diffuse and massive swelling of the face and oropharynx, as well as difficulty speaking and swallowing. She presented to the hospital emergency room within 2 hours of onset of symptoms and, about 3 hours after symptom onset, underwent tracheostomy for mechanical airway protection, and subsequently recovered completely.

Of note, both cases of angioedema with airway compromise in OCTAVE occurred in subjects with major risk factors for angioedema.

Now, 17 other omapatrilat treated patients and 2 enalapril treated patients were hospitalized for treatment of angioedema. Upon review by the adjudication committee, none of these patients were felt to have airway compromise. As discussed previously, many of these patients had highly symptomatic and visible presentations of angioedema, including tongue and lip swelling, and in many cases angioedema was not the sole consideration in the decision to admit to hospital. None of these subjects had progression of their symptoms while in the hospital. 14 were discharged after 1 day, and 3 after 2 days in hospital. Thus, while the number of omapatrilat treated patients hospitalized for angioedema substantially exceeds the number of enalapril treated patients hospitalized for angioedema, the level of severity of these cases appears to be low.

Now, the rate of progression of angioedema, once it begins, is an important question. One case of anaphylaxis with associated angioedema was observed in OCTAVE, and this case progressed within a matter of minutes to a life-threatening condition. In general, as you've heard, angioedema that occurs outside of the syndrome of anaphylaxis progresses over hours rather than minutes.

To determine whether the rate of progression of potentially serious angioedema in OCTAVE with omapatrilat was consistent with that described for angioedema in other settings, we examined those cases that were considered serious enough to receive treatment with epinephrine or corticosteroids. We then characterized the length of time between the onset of symptoms and the receipt of treatment, and since no treatment was received during that period, any progression would reflect the natural course of the episode. Other than the two cases with airway compromise discussed before, no patient had progression of angioedema to airway compromise.

And while about 20 percent of angioedema events treated with epinephrine or corticosteroids occurred in the doctor's office and therefore received immediate or near immediate medical attention, about 80 percent occurred outside of the physician's office. Of these, about two-thirds were associated with an elapsed time of at least an hour between the onset of symptoms and the patient's arrival at medical facilities, while in a substantial proportion of patients, more than 6 hours elapsed between the onset of symptoms and the patient arriving at medical facilities. The lack of rapid progression to airway compromise during the period from onset of symptoms to presentation at a medical facility is consistent with a rate of progression of the underlying disease measured in hours and not minutes.

A related question is whether angioedema with omapatrilat is sufficiently symptomatic and characteristic to be recognizable by the patient and prompt them to seek medical attention. In general, angioedema that might result in airway compromise is a highly symptomatic event with visible and diffuse swelling. In OCTAVE, both patients who presented with angioedema and airway compromise were highly symptomatic with diffuse visible swelling and a constellation of other symptoms.

We examined the clinical presentation in all other cases to determine if there were any patients who presented with angioedema and potential airway compromise in an occult rather than clinically overt fashion. Perhaps the most worrisome presentation would be the patient who presented with nonspecific throat discomfort and no other signs or symptoms. In OCTAVE there were no patients who presented in this fashion. Every patient with angioedema had a clinically overt presentation with visible swelling. Many had accompanying functional complaints, such as difficulty swallowing or difficulty handling oral secretions attributable to the swelling, and no patients with angioedema had nonspecific lower airway complaints such as stridor, dyspnea, or hoarseness alone.

The time course of angioedema with omapatrilat is illustrated here. The risk is greatest during the initiation of therapy. 88 cases, about one-third of all cases, of angioedema with omapatrilat occurred on the first day of treatment, as opposed to only 3 cases on the first day of treatment with enalapril. Many of these occurred within 2 hours of administration of the first dose. Nevertheless, angioedema continued to occur although at much lower rates through the trial. In the last weeks of the trial, the rate of angioedema was low with both drugs, though still about twice with omapatrilat compared to enalapril. Based on the observed incidence of angioedema in the last weeks of OCTAVE, one might predict that the rate of angioedema of any degree of severity during chronic treatment to be about 1 or 1.2 percent per year.

Now, data from studies prior to OCTAVE identified two potential risk factors for developing angioedema with omapatrilat: black race, which has also been described as a risk factor for ACE inhibitor-associated angioedema, and smoking.

An exploratory analysis of the OCTAVE data was performed to determine the effect of demographic characteristics, comorbidities, treatment history, and personal habits on the risk of angioedema with omapatrilat. The results of these analyses are summarized in this figure. On the left, is the multivariate relative risk of angioedema with omapatrilat in subjects with the stated characteristic compared to those without those characteristics. For example, the relative risk for angioedema in omapatrilat patients who currently smoke is 2.58 times that seen in patients who never smoke. On the right side is the observed incidence of angioedema in patients with these stated characteristics.

These analyses confirmed the importance of black race and smoking as risk factors for developing angioedema with omapatrilat. These two characteristics were associated with at least a doubling in risk of angioedema shown here, and the observed incidence of angioedema in these patients was 5.5 percent in black patients and 3.9 percent in current smokers.

Several other characteristics shown here, not identified as potential risk factors in the prior database, were found to be associated with either modest increases or modest decreases in the risk of angioedema. Of note, while it was expected that a history of treatment with and tolerance of ACE inhibitors might be associated with decreased risk of angioedema, this was not observed.

In sum, through an extensive clinical development program, the safety of omapatrilat has been very well characterized. This program has identified an incremental risk of angioedema relative to ACE inhibitor treatment which must be weighed against the potential benefit of greater blood pressure reduction.

With omapatrilat, as in other clinical settings, angioedema has a wide spectrum of severity. Current smokers and black patients have been shown to have a substantially higher risk.

In OCTAVE, the rate of life-threatening angioedema was 1.6 per 10,000 patients treated, and the upper bound of the 95 percent confidence interval for this estimate was about 6 per 10,000.

With omapatrilat, as in other clinical settings, angioedema was a symptomatic event with a characteristic presentation. In those with severe symptoms, the rate of progression was rapid but not fulminant, and all patients who developed angioedema with omapatrilat were successfully treated.

Bristol-Myers Squibb has proposed a risk management plan for omapatrilat that would minimize the risk of life-threatening angioedema through a comprehensive system of education. As I've noted, angioedema is a condition with clinical features which facilitate its management through education. It has a symptomatic and recognizable clinical presentation, rapid but not fulminant progression, and effective therapy can help to prevent poor outcomes.

The objective of the plan is to ensure a favorable benefit-risk ratio for patients taking omapatrilat. The cornerstone of the plan is a multifaceted and comprehensive program of education for prescribers, pharmacists, and patients. The approved labeling and other educational modalities will be used to educate physicians about the appropriate use of omapatrilat. Unit-of-use packaging will reinforce key risk messages, and the plan includes a novel and mandatory risk counseling program for patients.

A post-marketing surveillance plan would include a prospective observational cohort study and a plan for ongoing assessment of program effectiveness, including the use of an expert panel. We're also committed to providing extensive pre- and post-marketing testing of risk message comprehension and are confident that the proposed plan would be effective in minimizing the risk of life-threatening angioedema.

At this point, I'd like to make a few comments on benefit-risk. In general, the target population proposed for this drug would include those identified by the WHO IHS classification system as being at very high risk for cardiovascular disease or at high risk for cardiovascular disease, with an absolute risk of major cardiovascular events of at least 2 to 3 percent per annum and perhaps higher.

In these patients, a greater reduction in blood pressure by 3 over 2 millimeters of mercury, such as that observed with omapatrilat relative to enalapril in OCTAVE, would be projected to be associated with a 10 percent relative risk reduction, which would correlate to the reduction of 20 to 30 major cardiovascular events per 10,000 treated per year. A greater reduction in blood pressure by 5 over 3 millimeters of mercury, such as that observed with omapatrilat over other agents in other studies would be associated with at least a 15 percent relative risk reduction, which correlates to a reduction in 30 to 45 major cardiovascular events per 10,000 patients treated per year. As I've described, the observed incidence of angioedema with airway compromise over 24 weeks in OCTAVE was 1.6 per 10,000, with a 95 percent confidence interval of 0.2 to 5.7.

Now, these observations suggest that at those at high or very high cardiovascular risk, the projected number of life-threatening cardiovascular events prevented would substantially exceed the number of life-threatening angioedema events caused by at least an order of magnitude and perhaps more. If one takes the worst case estimate, the upper bound of the 95 percent confidence interval, as the basis for comparison, the benefit-risk relationship is still favorable.

Special consideration needs to be given to black patients and to current smokers as the overall risk of angioedema is higher in these patients. While BMS recognizes the increased risk in these patients and recommends that omapatrilat be used with special caution, we believe that carefully selected black patients and current smokers may benefit from omapatrilat treatment.

To conclude, in patients at high risk for cardiovascular events, the number of major cardiovascular events prevented would be projected to exceed the number of life-threatening angioedema events caused by at least an order of magnitude and possibly much more.

Now, when projecting cardiovascular benefit based on blood pressure reduction, there may be a concern about any unintended cardiovascular consequences of the therapy which could undermine or diminish the benefit. In this regard, I'd like to introduce Dr. Packer to review available CV event data with omapatrilat.

DR. BORER: Let's hold that just for a second, if we can. First of all, we're going to want to ask you some questions before we hear from Milton about heart failure. But it is 10:13 and 32 seconds, not by the satellite, and we'll take a break until 10:25 right now and then come back, ask you some questions, and then we'll go on with the presentation.

DR. LEVY: Thank you.

(Recess.)

DR. BORER: Before we begin the questions, there are two issues we need to deal with. There were no requests for public comment, but I want to determine that there is no one here who wants to comment about the issues that we're discussing today.

(No response.)

DR. BORER: If not, one other matter. The statement about Dr. Beverly Lorell's involvement and the reason for her exclusion wasn't really precisely stated. She is one of the principal investigators in the OVERTURE trial. She has no direct financial interest. For reasons of public disclosure, I think it's useful to know that.

Let's go on then with questions about safety. I'd like to begin with a request for clarification on two slides, and then we can get into more substantive safety issues.

Slide number 43. The issue here is that the female patient is listed as having become hypotensive. Can you give us a little bit of detail here? How hypotensive? Was this a clinically evident problem or did somebody measure a low blood pressure and record the patient as being hypotensive? What happened there?

DR. LEVY: She lost consciousness and didn't have a measurable blood pressure, and with an initial epinephrine injection, she regained consciousness. She received a second subcutaneous epinephrine injection, and over about 3 to 5 minutes, she regained a blood pressure of 110.

DR. BORER: This was not presumably -- or was it -- one of the patients who would have been your high-risk groups that's being targeted for the drug in the proposed labeling, or in your proposal I mean.

DR. LEVY: No, she wouldn't have been.

DR. BORER: Just for my information, were there other patients who developed hypotension or lost consciousness with the drug?

DR. LEVY: The rate of hypotension was extremely low with the drug. It was on the order of a 10th of a percent of all patients.

DR. BORER: And how did that compare with the comparator?

DR. LEVY: They were very similar.

DR. BORER: And then a comment more than a question. The risk factors that you defined in your last slide included seasonal allergies. Now, that's not overwhelmingly surprising, but a little surprising in view of the information that Dr. Kaplan gave us. I don't expect that you could possibly have an explanation for it, but it suggests that the biology we're dealing with here is more complex than perhaps we fully understand at this point. Is there any comment you want to make about --

DR. LEVY: Can I have that slide again?

DR. BORER: And perhaps Dr. Kaplan wants to comment on it.

DR. LEVY: As I mentioned, there was a prior hypothesis, a relatively strong one, regarding black race based on both our experience and --

DR. BORER: I'm specifically talking about seasonal allergies.

DR. LEVY: Right, I understand. My point is that there was no prior hypothesis for seasonal allergies, nor was there a reason to believe that this would be identified as a risk factor. We looked at a wide variety of characteristics and showed these modest changes, increases or decreases, in risk in some. The information that's reported here makes no attempt to correct for multiplicity of analyses. The confidence intervals are nominal, 95 percent confidence intervals. And in the absence of some pathophysiologic rationale or prior hypothesis, this should really be regarded as hypothesis-generating.

I'll ask Dr. Kaplan to comment on it, if he would.

DR. KAPLAN: I think in terms of the issues there in terms of where the relative risk was higher or lower, I don't think I could have predicted any one of them in particular. The incidence, if you look at the seasonal allergies, was a little bit higher, but I don't know I could have related the two or necessarily predicted that the risk would be somewhat higher.

Female gender. The incidence of angioedema, irrespective of cause, is higher in women. That might be consistent with it.

Nor could I tell you why somebody with diabetes or atherosclerosis would have lesser risk. So, I can't help much with the way those data came out in terms of what we know or what we could have predicted.

The only possible one would be in terms of blacks perhaps having more risk for angioedema. The only data related to that is responsiveness to intradermal bradykinin seems to be heightened blacks. Therefore, they may have end organ responsiveness that's a little bit higher than caucasians, and that would predispose to more angioedema.

DR. BORER: Thank you.

Tom, do you have any specific safety issues before we go on to Susanna and to Steve?

DR. PICKERING: Well, there are some questions I'd like to ask relating to the generalizability of the findings and the OCTAVE design.

I wonder if you could tell us a bit more about how the centers were selected. I believe there was something about being close to a major medical center.

Also, were any of the episodes occurring during the first 2 hours while the patient was still in the hospital setting? And finally, what information was given to the patients about the risks and symptoms that they might expect?

DR. LEVY: Well, those are very good questions. The first question regarding selection of study centers, this is an enormous trial with 3,300 centers in 12 countries. They represent both experienced clinical trialists as well as physicians skilled in the treatment of hypertension, but without prior experience in clinical trials. The issue you've just cited, in prequalifying we did require that they be within 1 hour of a medical facility with resuscitation equipment.

Your last question, if I could ask you to restate it.

DR. PICKERING: What information were the patients given about expected symptoms or side effects?

DR. LEVY: Patients were provided with a detailed informed consent, and that informed consent described the phenomenon of angioedema, swelling of the anatomic sites, provided rather detailed information about the quantitative risk of angioedema, as it was known at the time so they could evaluate the risk of study participation, and concluded with a sentence instructing them to seek medical attention should it occur. It's very consistent with what's done in trials. We propose in our risk management program a level of patient education that goes very far beyond that.

DR. PICKERING: The third part of the question was relating to episodes during the first 2 hours after the dose.

DR. LEVY: Yes. As I showed you, a total of 88 episodes occurred on the first day of treatment. 56 of those occurred within 2 hours of administration of the first dose.

DR. BORER: Susanna.

DR. CUNNINGHAM: You've defined the target population that you anticipate using this drug in. I want to know what percentage of that target population is African American, what percent are current smokers, have renal disease, seasonal allergies, et cetera. So, what's the risk profile going to look like in your defined high risk population?

DR. LEVY: I can certainly refer you to the trial data. Those are excellent questions.

For instance, overall, 10 percent of subjects in the study were black, 13 percent of those with diabetes were black. So, there is some association, but overall, the vast majority of patients with diabetes who would be candidates for the drug are not black.

Prevalence of smoking overall in the study was about 18 percent, and it was fairly consistent across all study subgroups, including those that we've identified as potential target populations for the drug.

So, I guess the short answer is that blacks would probably be represented somewhere between 10 to 13 percent in the potential target population, perhaps a little bit greater than their overall prevalence in the population, and smoking probably around 18 percent.

DR. CUNNINGHAM: And how about those other potential new risk factors that we don't know for sure about, the other ones, the seasonal allergies, the former smokers?

DR. LEVY: The population is 51 percent women; former smokers, maybe another 20 percent. Again, those are characteristics that are, at this point, hypothesis-generating associated with small differences in risk.

DR. BORER: Blase.

DR. CARABELLO: You indicated the proportion of patients that developed angioedema at which dose and that it was much higher at 20 milligrams than at 10. But of the patients that developed angioedema, how many did not have it at lower doses and then subsequently developed it as the dose was up-titrated?

DR. LEVY: Well, it's a great question, general question about the dose relationship of angioedema. This study, of course, was not designed to really characterize the relationship of incidence to dose. You'd need a true parallel group study to do that, in which patients started off at each dose and were titrated upwards, so you didn't filter people.

What we saw is that over time the incidence of angioedema decreased despite the up-titration of patients to higher levels of drug. But there were a significant proportions of patients who did develop angioedema on 80 milligrams, having tolerated 10, 20, 40 milligrams.

DR. BORER: Steve.

DR. NISSEN: I want to explore one of the principal hypotheses of the trial which was that by starting at a low dose and then gradually working our way up, that we could avoid the more catastrophic problems. It's difficult to answer that question obviously because the way that the angioedema was adjudicated is different in the two trials, but help me a little bit, if you will.

The raw rate of angioedema in percent in the pre-OCTAVE trials I have at about 1.96 percent. Do you agree with that? It looks like angioedema 1.03 and then head and neck edema, another .93. So, your slide number 36 would suggest that the rate of was around 1.96 percent pre-OCTAVE. Is that right?

DR. LEVY: Well, we didn't know exactly what it was pre-OCTAVE because not all those head and neck edema cases were angioedema, and conversely, there might have been other events that were called allergic reactions that were angioedema. But to the best of our knowledge, that's a reasonable, very rough estimate.

DR. NISSEN: Do you think that's a high estimate?

DR. LEVY: If we knew exactly what the incidence was before OCTAVE, we wouldn't have done OCTAVE. I think it's a reasonable rough estimate.

DR. NISSEN: All right.

And then in OCTAVE, the rate was 2.17 percent. So, again, obviously there's an issue here, but it looks to me like the actual incidence, about 1 in 50 patients pre-OCTAVE had angioedema and about 1 in 50 patients in OCTAVE had angioedema. So, it looks like the strategy of starting low and working up may not have been effective. Is that a reasonable assumption?

DR. LEVY: I think that's quite possible. Just bear in mind that the study wasn't designed to compare 10 and 20 milligram doses. There's an enormous difference in the way in which physicians were solicited to provide angioedema reports in OCTAVE. We know from other trials, like the SOLVD trial, what when you ask physicians to report this event, the reporting rate goes up dramatically.

DR. NISSEN: Yes. The reason I think it's relevant is that there was a difference in the number of very severe cases pre-OCTAVE and in OCTAVE. But because those numbers are so small, the confidence intervals are quite wide. So, I wanted to go back and look just at the raw rates of any angioedema to get a sense for whether the strategy of starting low would be protective or not. To me, there doesn't look like there's any evidence that that strategy is going to work in protecting patients, at least not from what we can see in the data.

Now, just before we broke, you said that you thought that this drug would be acceptable in smokers and blacks. The word you used is you said in "selected" smokers and blacks. What I guess I would like to know is how are we to select those people. The incidence was about 1 in 18 or 1 in 19 in blacks. So, what criteria should I use to select those African American patients that can successfully be given omapatrilat?

DR. LEVY: They would be patients with very high cardiovascular risk and hypertension that can't be controlled with existing medications.

DR. NISSEN: But that's the same criteria you told us for the rest of the population. So, you'd apply the same criteria to the African Americans that you would to the non-African Americans.

DR. LEVY: The same principles, but one might set the bar higher.

DR. NISSEN: All right, fair enough.

Now, I guess I had a question for Mike Weber because you obviously spend your life treating this, and I know you deal with this. The issue relates to compliance in a clinical trial versus compliance in practice. In the great State or country of Brooklyn --

(Laughter.)

DR. NISSEN: -- what are compliance rates like?

DR. BORER: Tread lightly there.

(Laughter.)

DR. NISSEN: As we know from those who live in the great State of Manhattan.

(Laughter.)

DR. NISSEN: But what are compliance rates like among populations with severe hypertension in your setting?

DR. WEBER: Well, as you know, Steve, the largest population group by far in our setting happens to be African American or Caribbean American. They do actually extraordinarily well with hypertension treatment because the African American community is, in fact, highly educated about hypertension and takes it very seriously. In fact, even if you look at NHANES data, there is a suggestion that blacks overall have very comparable adherence to treatment as compared with non-blacks.

So, I would say compliance is good. Now, what do I mean by good? I would say that about 50 percent of patients who start on a medication are still taking it about 6 months later or taking some sort of appropriate treatment 6 months later.

DR. NISSEN: How frequently, in your experience, do patients miss a few doses, skip a weekend, go off somewhere, and stop the medication, and then restart it again?

DR. WEBER: I would say about 70-75 percent of hypertensive patients make those sorts of errors or omissions.

DR. NISSEN: The reason I get to that is because I'm worried about a risk here, and the risk is you've titrated somebody up to 80 milligrams of omapatrilat. They take a long weekend with their spouse somewhere and they forget to take their medicines with them. They've been off the drug for three or four days. They come back home and they restart it. I'm trying to assess what the risk is going to be in clinical practice compared to the risk in a clinical trial. So, I need your thoughts about that.

DR. WEBER: Well, it is going to happen and it does happen all the time in clinical trials as well as in regular clinical practice. So, we do know that starting almost de novo on an extraordinarily high dose of a treatment, omapatrilat or anything else, is happening all the time, presumably with relatively little side effects or adverse effects that we are aware of.

We had quite a few people in the early omapatrilat experience who, in fact, did start directly on higher doses or were accelerated quite quickly to higher doses in the parallel group studies, and to the best of my knowledge, with the exception of some people who had some hypotension -- and there were not many of those -- in fact, it was pretty well tolerated.

DR. NISSEN: But you made the case that the incidence of those severe cases was worse in the pre-OCTAVE experience, and the suggestion is here that we can prevent those. I guess I'm worried here that in the general population where people start and stop drug, that the risk of somebody being off the drug for a few days and then going back to an 80 milligram dose might be pretty significant over a period of years. See, the question is whether the risk of angioedema is going to tail off with time and kind of get vanishingly small or whether we're going to see year after year an ongoing risk of this. And that relates to whether intermittent therapy is likely I think.

DR. WEBER: I don't think there's an answer to that question, Steve.

DR. LEVY: I might just provide a few facts from the trial that you might find helpful. We did ask patients at each visit if they had been compliant with medications. Compliance was defined essentially as taking at least two-thirds of their prescribed medication from the previous visit, and at each visit about 3 percent of patients admitted that they hadn't been compliant, which is a small number, but it's still 300 to 400 patients at each visit on omapatrilat who admitted they had missed at least a third of their medication from the previous visit.

We very carefully characterized dose interruptions in subjects who developed angioedema, and we found 3 subjects who developed what was essentially mild angioedema following a period of dose interruption. So, there certainly is no signal that there's an increased risk in patients who take their drug intermittently.

DR. NISSEN: The reason I ask is one of your really bad cases was a patient that missed a dose. One of your severe cases of angioedema in the database that I reviewed, the patient took a dose about 8 hours late and immediately got into trouble. I can refer you to that.

DR. LEVY: No. That is the subject in OCTAVE. She typically took her dose at 8:00 in the morning. She reported she took it at 4:00 that afternoon instead. I think given the half-life of the drug, 14 to 19 hours, it would be difficult to link those two.

DR. NISSEN: A second question I guess relates to how to assess the risk in general use. I'm sure many other members of the panel have the same concern, that when you administer a drug in a clinical trial, there's a certain kind of a protected environment that's involved. You know, you strictly mandate that the patients stay for 2 hours after every dose titration. The physicians know they have to look for this side effect. They've been educated at an investigator meeting. There's a lot of stuff that goes on.

What I worry about is what happens in Sioux Falls, South Dakota when a patient kind of goes in a rural office where it's a much less protected environment. Because once you let a drug out of a clinical trial environment, you're less protected.

Given the fact that this is a pretty serious side effect, my worry is that patients won't make it in time or won't be recognized in time because they're not going to be as protected as they would be in a clinical trial. I'd appreciate any insight about what kind of a risk that represents here.

DR. LEVY: Maybe I can just make a comment on it. It's an excellent question. One of the reasons why we did a 25,000-patient trial at 3,300 sites in 12 countries was to provide as much information as possible about how the drug would be used and what the results would be in real-life practice. Of course, there were many clinics in places like Sioux Falls, South Dakota, very remote locations in Russia, all over the world.

It's also worth pointing out that by and large, when patients experienced angioedema, they sought medical attention at a facility other than the investigator's office. So, the question is whether those facilities in a small town can provide epinephrine and, if necessary, in the rare cases mechanical airway protection.

DR. NISSEN: I guess the final comment -- and perhaps it's a rhetorical one -- is on page 114 of your document, you say that treatment of life-threatening angioedema does not require specialized training. Angioedema associated with omapatrilat is managed in the same fashion as angioedema due to any other cause. Treatment of serious allergic reactions is a core skill for physicians and nurses, and airway protection is a routine procedure for emergency personnel, et cetera.

Well, one of the things that was most troubling, in reading the case narratives, is that 3 of the 6 patients required cricothyroidotomy. Other than my friend here, Blase Carabello, who does everything well, I would doubt if any of us on this panel with a Bic pen -- I mean, I'm glad to hear that Dr. Temple is skilled in this.

(Laughter.)

DR. NISSEN: But I'm standing in the shoes of being somewhere and giving the drug and having a patient get laryngeal edema. And those 3 patients could not be intubated. So, somebody that was skilled enough to take a scalpel and pierce the cricothyroid membrane was necessary to save the life of the patient. So, to say that this is a core skill I think is to trivialize the problem. I guess I would like your comment, but 3 of the 6 serious cases had to have a cricothyroidotomy in order to protect their airway.

DR. LEVY: Perhaps one of the clinicians on the panel would care to speak to that.

DR. NISSEN: Mike Weber, do you do these cricothyroidotomies?

(Laughter.)

DR. WEBER: Well, I have done them in rabbits.

(Laughter.)

DR. WEBER: I'm working my way up to humans.

But no, this clearly is an issue. I think the most important thing I can say about it is what Dr. Kaplan and Elliott have also pointed out, that fortunately these cases do not suddenly announce themselves as sudden respiratory embarrassment. There is a fairly long prodrome. So, as long as the patient knows that they ought to be going to an emergency room, hopefully that will allow us to deal with those patients. But if you can ask for some sort of a guarantee that there would be a 100 percent system to get absolutely everyone taking an ACE inhibitor who's going to have angioedema, I guess we can't guarantee that. But luckily, we do seem to have those several hours for the patient, as long as they know that they ought be doing it, to get to the emergency room.

DR. NISSEN: I'm going to, Jeff, hold further questions. I did have some further questions on the risk management program, but I thought it would be better not to do those now because, obviously, there's an issue about how do you manage the risk here.

DR. PACKER: Jeff, I just wanted to comment one thing about the need for a cricothyroidotomy. I actually have done a couple, having trained in a city hospital, but it has been a while.

I just wanted to emphasize that the core message, I think, which needs to be conveyed to physicians is the importance of epinephrine because epinephrine is the most effective treatment to prevent progression of this disease. Antihistamines don't work and steroids work but they work too late to have an impact on progression. And what is striking is the fact that in so many cases the use of epinephrine was delayed. In all the cases you're talking about, epi wasn't even given or epinephrine was delayed. I think part of the educational program is to remind physicians as to what really is the appropriate treatment for a serious and potentially life-threatening angioedema.

DR. BORER: I think that's a very important point. The only problem is, as Dr. Kaplan pointed out, if the drug is actually given to the people that you're targeting, there's going to have to be more known than that you give epinephrine. There's going to have to be something known about how you deal with the problems that may develop when you give epinephrine to that target population. So, it's a somewhat more complicated problem.

DR. PACKER: But we're talking about what might be called a risk-benefit relationship. You're not giving epi to everybody. You're only giving epi to people in whom the risk-benefit relationship is favorable. Someone who is going to die from angioedema -- the risk-to-benefit relationship is extremely favorable.

DR. BORER: Right. I'm not suggesting you wouldn't give epinephrine. I'm suggesting that you have to know how to do more than give epinephrine. You have to be able to deal with the consequences of it.

DR. NISSEN: It's a little more complicated also. Let me just tell you that you have an educational program. You educate people like me that treat hypertension on the importance of epinephrine. But the patient goes to an emergency department somewhere where there's not been any omapatrilat education given, and that doctor there has to know that the first thing you've got to do is give epi to the patient, not steroids or something else. I question. Because so many of these patients were treated elsewhere for their angioedema, the ability to educate people about this is challenging.

DR. PACKER: See, the patient and the patient's family play such an important role here because they can have a card that says I'm at risk of angioedema or whatever. This is the appropriate treatment.

DR. BORER: Paul.

DR. ARMSTRONG: I have a couple of questions for Dr. Levy and perhaps for Dr. Kaplan.

Dr. Levy, I may have missed it, but if you look at the 95 percent confidence limits on the estimates of angioedema in the 10 milligram versus the higher dose, do they overlap? You showed that there was a difference in the frequency, but I didn't see the confidence estimates around those.

DR. LEVY: Again, we've not directly compared the incidence. We didn't intend to. We provided two estimates of risk.

DR. ARMSTRONG: The second question is that you reminded us that this was a trial of international scope and very large. As someone who's had the experience of doing some of these trials, one of the things that one finds amongst events that are of fairly low frequency is that there's sometimes a difference in the surveillance detection when it's left to physicians who are participating. We've been finding, for example, that things like bleeding detected in Russia are less frequent with the same exposure and have hypothesized that that might lead to a better understanding of how different countries survey these phenomenon. So, with that background, what is the difference in the frequency of angioedema across the countries which participated in this 25,000-patient trial?

DR. LEVY: We looked at the incidence of angioedema by region comparing North America with Europe, which is where almost all the other patients were treated. And the incidence of angioedema was a little bit lower in Europe than in North America, as you'd expect, since there are essentially no patients of African descent in Europe.

DR. ARMSTRONG: And in Russia?

DR. LEVY: We didn't look at it by country.

DR. ARMSTRONG: Perhaps you or Dr. Kaplan can help me then. You've identified that Afro-Americans have a higher frequency. Do other ethnic groups also have a higher frequency of angioedema if one looks at Southeast Asians or Chinese or Japanese? What do we know from the ACE inhibitor data and other data vis-a-vis ethnicity and angioedema?

DR. LEVY: We're not aware of any other described ethnic associations, and there aren't sufficient data in OCTAVE to look at that question.

DR. ARMSTRONG: Dr. Kaplan, when you inject bradykinin subcutaneously to other ethnic groups, what do you find? You commented on that being a detection --

DR. KAPLAN: Yes, but it hasn't been done. That study was strictly Afro-Americans versus whites. I think the answer to your question, like people in Southeast Asia or Japanese or so on, has not really been looked at. There are just no data on that in terms of the incidence of angioedema. I know of nothing to suggest that it's accentuated in some way, but there's basically no data on it.

I'd like to make a comment with regard to when angioedema occurs because we mentioned that there was no way we could predict. There's no test. There's no way you could tell who was at risk. I'm going to make a statement that's really just theoretical, but just think about it a little bit because part of it has a certain randomness to it. It would be logical that if you take more, that you might see more angioedema, but that doesn't necessarily hold uniformly.

I have seen patients on ACE inhibitors who had a few multiple episodes, say, of facial angioedema and it was not recognized that it was due to their ACE inhibitor. And they come to me, now taking it for 3 months more, and they haven't had a swelling. When they come in, once I see that there's no other available cause, I immediately stop the drug. So, there's something that we truly don't understand about when the angioedema occurs.

I'll tell you what I think is going on, but it's right out of my head, and that is obviously it's got to relate in some way to bradykinin levels, which has to do with the rate of formation versus the rate of degradation. If you're taking a drug and you've reached a reasonable steady state, there's no question on an ACE inhibitor that you get some elevation of the kinin level. But if you're measure blood levels, they're a little bit up but you're not struck that it's tremendously high. I'm suspicious that when the angioedema occurs, something that is not yet identified is occurring to the person that produces bradykinin. They have a cold. They have an infection. They fell. They bumped their hip against the corner of their table, something of that sort. Then it doesn't take much to have levels soar sky-high.

And let me emphasize the lability of it. If I measure a blood bradykinin, just put a tourniquet on, stick the needle in, versus do the same procedure, get the needle in, remove the tourniquet, take 10 mls of blood and throw it in the garbage, and measure the bradykinin in the 11th ml, the difference between those two is 50-fold in bradykinin level, just from the needle stick and a little pressure. So, it's exquisitely labile.

So, I have a hunch that there are unknowns here that relate to when the actual attack of angioedema occurs, and that's why it has such a random feel.

DR. ARMSTRONG: If I may, Mr. Chairman, with Dr. Kaplan, I'm sure one of the easiest places to develop consensus today will be what's not known. But as we pursue this, can you just again help me with the epidemiology of angioedema that's not drug-related, that's spontaneous as it relates to age? With several hundreds of thousands of patients treated with ACE inhibitors, is the distribution by age any different with patients on ACE inhibitors than it is spontaneously, sir?

And when you explore co-factors or factors that you believe produce bradykinin and then engender an episode of angioedema, do you reckon that those co-factors are any different in patients on ACE inhibitors as opposed to other agents and, by inference, with the drug that we're discussing today?

DR. KAPLAN: To my knowledge, the angioedema that one would see with an ACE inhibitor is not going to vary particularly, let's say, between the ages of 20 and 80. I don't think anyone has looked at it in terms of age groups but I don't think it would be dramatically different. The most common form of angioedema that we see, regardless of etiology, is that autoimmune one that I mentioned to you. It persists for a long time. It's recurrent. It's like there all the time, and it's often associated with hives. First of all, it's two-thirds women and one-third men. So, it's skewed by sex. And the peak is between 20 and 40, and it's at both tails. As you get older and in youngsters, it's quite a bit less. I'm positive that although I don't know the details, that the ACE inhibitor situation would not parallel that. My best estimate is that it would be fairly level among age groups.

DR. BORER: We have a question from Mike and then a comment from Doug.

DR. ARTMAN: This may be more theoretical and perhaps Dr. Kaplan might be the one best to address it, but I'm just wondering if these risk factors for angioedema are additive. In other words, if you're a black female, smoker with renal disease and seasonal allergies, is your relative risk up to 10 to something?

DR. LEVY: No. The answer is no. The two major risk factors identified were black race and current smoking. You put them together and the incidence of angioedema is identical to that you see in blacks. It's 5.6 percent.

DR. THROCKMORTON: And yet, Elliott, the timing of those angioedema events for those two particular populations was quite different, as I recall. Could you show those two curves? The time to angioedema events for blacks and for smokers.

DR. LEVY: Yes.

DR. THROCKMORTON: Because those seemed very different. Again, going to the question of are all risk factors equal and are we talking about a single angioedema thing or are there different kinds of angioedema.

DR. LEVY: They're certainly not additive. There is a difference in the time to onset of angioedema amongst blacks and current smokers. In current smokers, the risk is greatest at the initiation of therapy. There were 45 cases in smokers on the first day of therapy, and then the rate declined fairly dramatically to a level that was near that seen in other patients.

In blacks, on the other hand, the risk was not dramatically greater on the first day of therapy than it was in whites, but it remained at a higher level for a longer time and the risk decayed more slowly.

So, there is a difference in the time course of angioedema in patients with each of those risk factors.

DR. CUNNINGHAM: Yes. You got me to wondering. What are the risk factors for angioedema in the group that was on enalapril. Are they the same?

DR. LEVY: The risk factors were quite similar with the exception of current smoking which did not appear to be a risk factor for enalapril associated angioedema.

DR. CUNNINGHAM: Because one of our questions is whether or not the two are the same, and if they have different risk factors, that makes you wonder.

DR. LEVY: It's a little hard to look at the enalapril group because of the relatively small number of events. The profile is quite similar with that one exception.

DR. BORER: Tom and then Steve.

DR. PICKERING: Yes. I'd like to pursue the question of the definition of black a little further. In this country it usually refers to African American, but I practice in northern Manhattan and a lot of patients look black to me but define themselves as Latino or Hispanic. And the distribution of risk factors is not necessarily the same as in African Americans. Can you tell us what the definition was and also how many of the blacks were U.S. African Americans as opposed to some other dark-skinned group?

DR. LEVY: Well, they're almost all U.S. The investigators were provided with one of four categories and simply asked the subjects to identify which of the four they belonged to. They were white, black, Asian Pacific, and other. So, it's not possible to tell you where the black subjects came from, whether they were Afro-Caribbean or of Spanish descent.

DR. BORER: Steve.

DR. NISSEN: None of us has asked you about cough, and obviously cough is an ACE inhibitor side effect that I think we believe is bradykinin related. Was there a difference in incidence of cough across all your trials in ACE inhibitors and omapatrilat?

DR. LEVY: No. They're pretty much spot-on, identical.

DR. NISSEN: Can anybody give me an explanation for that? It seems surprising.

DR. KAPLAN: The data on cough are not as good as the data on angioedema in terms of relating a kinin level to the actual event. Most people think that it is related to bradykinin, however.

DR. PACKER: I think that from the understanding that I have, there may be multiple mediators of cough. Bradykinin may be one. Substance P, a whole host of other factors have been implicated. So, I think it's probably much more multifactorial, which is why we're not seeing a signal here.

DR. THROCKMORTON: Steve, if you're interested, the incidence of cough was looked at by Dr. Pelayo in the original safety review, and that's on page 23 of his tab, which I guess is tab 4. As they said, the numbers are fairly small, but there does seem to be an ordering where the majority of the events were in the omapatrilat group and not placebo.

DR. NISSEN: Statistically speaking, there's no difference?

DR. THROCKMORTON: It's 2.1 percent versus 0.3 percent. It was a safety analysis. So, we wouldn't have normally don't statistical.

DR. NISSEN: I see, okay.

DR. LEVY: I'm sorry. I couldn't hear that data. Could you repeat that?

DR. THROCKMORTON: It's page 23 of Dr. Pelayo's. This is comparing against placebo. Is that what you were interested in, Steve?

DR. NISSEN: No.

DR. THROCKMORTON: You were interested in enalapril.

DR. NISSEN: Yes. Again, tolerability compared to enalapril. It sounds like it's a wash.

DR. LEVY: Well, you can see there's --

DR. NISSEN: No difference.

DR. LEVY: -- no difference.

DR. NISSEN: Very good. That's helpful.

Let me just ask one more final question for me, and then I'll pass this along. Part of your risk management program is to try to keep patients in physicians' offices for a couple of hours after they get that first dose. I assume that that's going to be a recommendation. Am I correct?

DR. LEVY: It's a consideration. The program is under development now.

DR. NISSEN: But I guess one of the things that I know about physicians and their levels of patience is -- I'll ask you a question and see if the clinicians agree with this. When you do something like that and you have an event that's relatively rare, like angioedema, physicians may start out keeping patients for a couple of hours. They won't see an event and they will start to get a little complacent, and they'll start letting people go sooner. And I guess I'm worried that in a big program that goes on for a while, because the events themselves are rare, any individual physician is not likely to see one. And there's going to be a tendency to get increasingly complacent until something catastrophic happens. It's a just a question of behavior and it's something that worries me. Any thoughts that any of the clinicians have about whether this is a real concern or not a real concern I'd be interested in.

DR. BLACK: I haven't even done trachs in rabbits, so I'm not sure I'm really qualified to talk about it. But I've had angioedema that, in fact, had to do with something else in Charlotte Hungerford Hospital in Torrington, Connecticut, which is near Russia actually.

(Laughter.)

DR. BLACK: My own feeling just in general is every emergency room, in fact, can do this procedure. The care that I got was exactly what you heard. It was shotgun. I got the right stuff and it got better. But I think, in fact, this program will really improve the care and awareness of angioedema whatever the cost. And we know there are cases from ACE inhibitors also. So, I think it's going to really help out. The people who are going to do most of the care are going to be people in ERs. It's not going to be in the first few hours in the doctor's office. So, I'm not as concerned.

I think a program that asks you to stay there for a while is probably going to be, as you say, not in fact -- and it probably wouldn't make too much difference. The anaphylactic case was the only one, and those are clearly by chance.

DR. LEVY: I think it might be useful for the committee just to know a little bit more about the risk management program at this point, if you'd be interested, because the topic has come up a few times.

DR. BORER: If we can hold that just a little bit because that ultimately will be part of our discussion in terms of risk-benefit and we will want to hear a little bit about it. You know, we got a lot in our handouts and materials about what you submitted.

Why don't we just go through this OVERTURE data quickly and then we can come back and clean up.

DR. FLEMING: Jeff.

DR. BORER: Oh, I'm sorry. Tom.

DR. FLEMING: I had two or three questions on safety. I'd like to pursue a little bit more what Steve and I think Paul were getting at earlier about what is the evidence that there is, in fact, a relationship here and a safety risk with starting dose.

Can you put up slide 36? As Steve was alluding to, in slide 36 there appears to be evidence that there may be a two- or three-fold lower risk of angioedema when you're starting below a 20 milligram dose. In fact, in this experience, there were no cases of airway obstruction, airway compromise, in the less than 20 dose. So, in a certain sense, the OCTAVE study is a disappointment when you look at the fact that the 10 milligram starting dose gave a higher overall occurrence rate of 2.17 percent.

Yet, as you point out, that readily could be under-detection in this setting here. One piece of evidence of that is when you look at the rate of airway compromise, it turns out that in OCTAVE it's 1.6 per 10,000. Here, if you look at the greater than 20 milligram group, it's almost 10-fold larger. It's 15 per 10,000. So, there really is evidence when you look at airway compromise that there really is a relationship with dose.

To try to get a better sense about this, beyond just relying on the airway compromise rates, we know that in OCTAVE there were these two cases, but there were overall 19 cases that were hospitalized. Can you give us for these two columns here, the below 20 and the greater than or equal to 20, how these cases break out relative to hospitalization? Because that may give us further reinforcement to the airway compromise data that there really is a dose-response relationship.

DR. LEVY: Let me just see if I understand. You want to know from these data in this program what proportion of patients required mechanical airway protection, what proportion were hospitalized.

DR. FLEMING: Yes. We know it's 0 and 4 for airway compromise. So, in these two columns, of the 18 cases in the less than 20 milligram setting, how many of them were hospitalized, and of those 66 in the greater than 20 milligram, how many were hospitalized?

In essence, what I'm getting at is if there's under-detection, as I'm almost certain there is here, it's less likely to be under-detected in the most serious cases. Airway compromise I'm assuming you're going to see. Hospitalization I would think you would be more likely to see. So, we'll get a better clue, along with the airway compromise, that there really is a dose response.

DR. LEVY: In those who were started at 20 milligrams or more, there were 4 patients who were hospitalized for angioedema without requiring airway compromise. I'll ask my team to verify it for me. My recollection is that in those less than 20 milligrams, it was 1 patient hospitalized, but I'll ask them to check for me.

DR. FLEMING: Okay, and they can give that to us later after they check.

Let me go on to a second question.

DR. KAPLAN: Could I make a comment on the 19 hospitalized patients? In looking those over, I read all of them to see what was the criteria for hospitalization. If you look at it carefully, you will see that about 8 or 9 out of the 19, upon arrival to the emergency room, were almost asymptomatic, had either a little bit of lip swelling that was left or had nothing, but gave a history of having had tongue swelling or pharyngeal swelling or drooling or something that had happened hours before and they were then hospitalized for observation. That's a safe thing to do and it's exactly what you might consider doing if it were anaphylaxis.

But the fact is, if you read them individually, of course, they were all hospitalized overnight. Nothing happens. They're discharged the next morning. And the fact that about 8 or 9 of them, by our criteria I think and by my judgment as an allergist, ought not to have been hospitalized because if you understand what happens with an ACE inhibitor, you get the swelling, it crescendos. That time may vary depending upon the person and severity, and then it finally abates, and it does not recur. So, it doesn't rebound, which is the reason why steroids are of no value actually in treating them in contrast to anaphylaxis.

So, I think that those who have respiratory embarrassment on arrival are the obvious. But I think hospitalization may not be the best criteria as we look at this study for the actual incidence of the "severity" because a substantial proportion of those patients resolved spontaneously and really didn't need hospitalization.

DR. FLEMING: Let me go on to the second question and that's slide 39. Having seen in the prior experience before OCTAVE no cases of the airway obstruction and evidence of lower rates, the intention here was to see if we could show that the rate was below 2. So, the null hypothesis was a rate of 2. The alternative was something discernibly less than 2. Ultimately what we see here in the bottom confidence interval is that we cannot only not rule out that the rate is less than 2. We can't even rule out the rate is less than 4, and the data suggests that the rate is actually 3.2.

I see Jim Neaton here. I don't know if it's because he was on a DSMB for this study. I'm just guessing.

How was the DSMB monitoring this phenomenon as the study was ongoing? Because it appeared your null hypothesis was 2 and the alternative, I'm assuming, was 1 or 1.5 or something like that. And you're entirely way inconsistent with that with these data. How was this being factored in during the monitoring of the trial?

DR. LEVY: Let me comment on that. Jim was actually not on the DSMB.

But the DSMB was provided with these data, as well as safety data. In their view, it was very important to weigh both potential harm and potential benefit in assessing whether this study was to continue or not, and they didn't apply a simple stopping rule based on whether or not the prespecified hypothesis for angioedema was reached. In their view, there was clear evidence not only of increased risk of angioedema, but also of greater blood pressure reductions.

DR. FLEMING: So, the protocol simply said the null hypothesis is 2, alternative is less, and there was no stopping guideline specified in the protocol.

DR. LEVY: There was no prespecified stopping rule.

DR. FLEMING: The last question. When we look at angioedema by severity, you've given us that data in the aggregate. The add-on group with 4,751 patients is an important subgroup here. In this subgroup, do you have the breakdown of the cases of angioedema by grade?

DR. LEVY: Yes, we do, but let me just make a point and that's that in that group and all other groups, there's a remarkable consistency across this database. What you'll see is that the incidence of angioedema in group 3 is similar to that seen overall, and that 60 percent of the cases received no treatment or antihistamines only as they did overall. So, we'll be happy to show you those data, but they're quite consistent with the overall data.

DR. FLEMING: Okay, and please do so, though. At some point bring those back to us.

DR. BORER: If there are no other questions of fact here, maybe we can go on here about OVERTURE, and then we'll come back to some of the other safety issues.

DR. PACKER: Before I begin, I just want to note that in light of my status as an SGE but also in light of my role as principal investigator of the OVERTURE study, the Advisors and Consultants Staff of the FDA has consented to my participation and presentation in today's meeting.

I also wanted to correct Steve's comment, and I think this is particularly sensitive to both Jeff's and Tom's views. I think those who live in Manhattan neither characterize it as a State or a country. I think they characterize it as a universe.

(Laughter.)

DR. PACKER: With that in mind, at yesterday's meeting on candesartan, the advisory committee indicated it was comfortable believing that an incremental decrease in blood pressure would be translated into a reduction in cardiovascular events if it could be reassured that the experimental drug did not exert an adverse effect independent of its antihypertensive action that could increase the risk of a cardiovascular event. Therefore, the committee implied it would feel comfortable, assuming that a decrease in blood pressure would produce a predictable reduction in cardiovascular risk, if the drugs being compared were in the same class, but they might not feel such comfort if the drugs were in different classes. And I think Steve in particular made this point.

DR. FLEMING: Some of us, though, might not have been as comfortable with such a broad generalization as you have stated.

DR. PACKER: Even in the same class. Right.

So, I'd like to consider the present situation which is that both omapatrilat and enalapril are both ACE inhibitors and that's in part reassuring, but omapatrilat differs from enalapril in also being a NEP inhibitor. So, the question is, how comfortable can the committee be that NEP inhibition does not produce adverse cardiovascular effects that could negate the cardiovascular benefits expected from its incremental ability to lower blood pressure?

This table shows the cardiovascular events that were observed during the 6 months' treatment with omapatrilat and enalapril in the OCTAVE study. Now, although this was a prespecified analysis, the study was not designed to compare the two drugs on the risk of cardiovascular events. So, I think these data need to be interpreted very cautiously. Having said that, there were 105 cardiovascular events in the omapatrilat group and 121 in the enalapril group.

This slide shows the Kaplan-Meier plots for these events. The hazard ratio of omapatrilat to enalapril is 0.87, with an upper bound of the 95 percent confidence interval of 1.13, I think in and of itself suggesting that NEP inhibition is unlikely to exert a meaningful adverse effect that might detract from the expected clinical benefits of the drug.

Now, although these data might be considered to be reassuring, my own view is that these data need to be interpreted very carefully since the duration of follow-up in the study is only 6 months.

I also think that it is likely that Tom might ask for an analysis of these data according to the characteristics that the sponsor is proposing. It might form the basis of use of the drug. And I just want to let you know we are working on that as we speak, including trying to address the issue of the blood pressure lowering effects in all of those individual subgroups at high risk.

Well, in light of the limitations of these data, I think it's important to consider the results of OVERTURE. Preliminary results of this trial were presented at the ACC in March. Final results will appear in Circulation online in about a week from now, and before reviewing the results, I want to emphasize that although these data have been presented to the FDA, they have not been reviewed by the FDA. Therefore, they are being presented with the proviso that if they have any influence on your judgments, they will need to be confirmed by the agency.

The OVERTURE trial evaluated 5,770 patients with class II, III, or IV heart failure. All patients had an ejection fraction less than or equal to 30 percent. All were hospitalized for the treatment of heart failure within the past year. All patients were receiving excellent background therapy for heart failure, including beta blockers in 50 to 60 percent of patients and spironolactone in over 40 percent.

Importantly, about 1,300 patients, or about 20-25 percent of the population, were hypertensive. I just want to mention that hypertension is a particularly important problem in patients with heart failure since it is so critical to lower blood pressure in these individuals. Yet, there is a sizeable risk for frequency of hypertension in people with heart failure. It's 20-25 percent in moderate to severe heart failure. It's over 40 percent in milder degrees of heart failure. And these patients are already receiving diuretics, ACE inhibitors, beta blockers, and they can't take calcium channel blockers. So, I think that an analysis of that subgroup would, in part, address Tom's request for additional data, including outcomes data, in high-risk individuals.

Now, eligible patients for this trial had any prior with an ACE inhibitor discontinued and were randomized in a 1-to-1 fashion to either omapatrilat or enalapril. The target dose of omapatrilat was 40 milligrams once daily, which had shown promising results in earlier heart failure trials, and the target dose of enalapril was 10 milligrams b.i.d., which was the target dose used in the SOLVD Treatment trial. I think this remains the most definitive study showing a favorable effect of ACE inhibitors on morbidity and mortality.

What I'd like to do is to make two points about these doses. First, the target doses of both drugs was half the target dosage used in the OCTAVE trial, and second, because this was a heart failure trial, enalapril was given twice a day, whereas the drug is conventionally given only once a day in the treatment of hypertension and, as Steve has mentioned, the use of a b.i.d. regimen arguably provided a tougher test for omapatrilat.

Now, the primary endpoint in this study was the combined risk of all-cause mortality or hospitalization for heart failure. This endpoint was used prospectively in the original protocol to test two hypotheses, a non-inferiority hypothesis and a superiority hypothesis. According to the original protocol, omapatrilat would be considered non-inferior to enalapril if the upper bound of the 97.5 percent one-sided confidence interval was less than 1.09, and if this were achieved, we would have been able to conclude that omapatrilat would have retained at least 80 percent of the effect of enalapril seen in the SOLVD Treatment trial, which was the protocol-specified reference standard, greater than 80 percent. Of course, if the upper bound of the one-sided 97.5 percent one-sided confidence interval was less than 1, then we would have concluded that omapatrilat was superior to enalapril.

Now, here are the results on the primary endpoint. There were 973 patients who died or were hospitalized for heart failure in the enalapril, 914 such patients in the omapatrilat group. It translates into a 6 percent lower risk of the primary endpoint in the omapatrilat group. The upper bound is 1.03, which is greater than 1 but less than 1.09. Therefore, we could not conclude omapatrilat was superior to enalapril, but we could conclude that omapatrilat was not inferior to enalapril.

Now, this slide shows the effect of omapatrilat and enalapril on the combined risk of cardiovascular death or cardiovascular hospitalization. This was a prespecified secondary endpoint in the study, and it represented the most comprehensive cardiovascular endpoint specified in the original protocol. For this endpoint, omapatrilat had a 9 percent lower risk of a cardiovascular event which was nominally significant.

Now, as I said at the beginning, over 1,300 patients in OVERTURE were hypertensive in that they had a systolic blood pressure that was greater than 140.

Now, this slide shows the influence of baseline systolic blood pressure on the magnitude of the difference between omapatrilat and enalapril on the primary endpoint of death or hospitalization for heart failure, and on the secondary endpoint of cardiovascular death and cardiovascular hospitalization. And as can be seen, the higher the systolic blood pressure, the greater difference in favor of omapatrilat, and this was true for both endpoints. The difference in favor of omapatrilat in patients with a systolic blood pressure greater than 140 was a 16 percent lower risk of death or hospitalization, and a 21 percent lower risk of cardiovascular death or cardiovascular hospitalization.

I guess, Tom, these are probably the best estimates we now have with respect to outcomes data in hypertensive patients, albeit it in hypertensive patients with heart failure.

I would like to close with a brief note about safety. This slide lists selected adverse events that were seen in the OVERTURE trial. As can be seen, omapatrilat had more reports of hypotension and dizziness, but fewer reports of heart failure and fewer reports of impaired renal function. Angioedema was seen in 14 enalapril patients, 24 omapatrilat patients, and of these, 3 patients were hospitalized, 2 in the enalapril group and 1 in the omapatrilat group, and none had airways compromised.

Now, in summary, I think the results of OVERTURE are at least suggestive and certainly I think consistent with the hypothesis that in patients with hypertension and heart failure, omapatrilat might reduce cardiovascular events when compared with enalapril even when enalapril is given twice daily.

But I want to emphasize a much more important point, and that is, I think these data provide considerable reassurance that NEP inhibition does not detract from the cardiovascular benefits one can expect from the incremental antihypertensive effects of omapatrilat.

With that, I'd be delighted to answer any questions the committee might have.

DR. BORER: How were heart failure events defined in the protocol, Milton?

DR. PACKER: Heart failure was defined by the investigator, which in most heart failure protocols, heart failure is defined by the clinician. The qualifications for heart failure are based relatively on the severity of the disease. So, they had to have class II, III, and IV symptoms limited by dyspnea and/or fatigue.

DR. BORER: No, no. I'm sorry. That's not what I'm asking.

DR. PACKER: Oh, I'm sorry.

DR. BORER: These are adverse events. Everybody in the trial had heart failure.

DR. PACKER: Oh, I understand. I think, as you may appreciate, in a trial where the -- and we see this all the time in heart failure trials. Investigators are asked to report all AEs. There is no guidance given to investigators as to how they should report AEs or not. In general, heart failure as an AE is by far the most frequent AE reported in heart failure trials. In general, in drugs that work in heart failure, the reports of AEs in heart failure tend to be lower in the active treatment than in the placebo. But there is no quality control here. There is no guidance as to how heart failure as an AE should be defined. It's really up to the judgment of the investigator.

DR. BORER: And similarly I assume for hypotension.

DR. PACKER: Similar for hypotension. All the AEs are reported at the discretion of the investigator in a spontaneous manner without any specific instructions as to what they should or should not report or how to define specific terms.

DR. BORER: Can you tell us what doses of the two drugs actually were achieved? I see the design, but what was actually achieved?

DR. PACKER: I know the estimates, and Jeff, we can give you the actual numbers, but it's in the range of about 80 to 82 percent in both treatment groups received target dose. We will check on whether that's -- that's correct? It's 82.7 percent and -- we'll get you the data, but that's the range.

DR. BORER: Steve, do you have any questions?

DR. NISSEN: I just wanted to come back to the blood pressure issue since what's on the table here is the application for approval of this drug for hypertension. I want to hear again your thoughts, Milton, on why there was no blood pressure difference between omapatrilat and enalapril in the hypertensive heart failure patients because, again, this does shed some light on whether b.i.d. enalapril might be as good as omapatrilat.

DR. PACKER: I just want to, again, emphasize the points, but let me supplement them as well since you're asking me to do that.

First of all, again this wasn't a hypertension study. This was a heart failure trial, and heart failure investigators in general view blood pressure as a range as opposed to a number. I don't know another way of saying that. There's a complete difference in the quality of the blood pressure data in the context of a hypertension trial than in the context of a trial done for another indication.

Having said that, I think that the most important point is the trough blood pressures were similar, but there is evidence from other trials in heart failure, not from OVERTURE, that during most of the day the blood pressure is considerably lower in the omapatrilat group than in the ACE inhibitor group. And the difference, by the way, in previous heart failure trials has been in the realm of about 7 to 8 millimeters of mercury greater in omapatrilat than, for example, in the previous trial with lisinopril.

In that trial, Steve -- and the trial I'm referring to IMPRESS. Lisinopril is a once-a-day drug. The blood pressures came down and were very similar at trough in that trial, but during the day the blood pressures were dramatically different in the two treatment groups. I think that reinforces the point that the committee made yesterday, which is it isn't just trough blood pressure that affects cardiovascular events, it's the delta blood pressure throughout the day.

DR. NISSEN: I'm not sure I get the argument. What you're sort of saying is blood pressure isn't measured as well by heart failure docs as it is hypertension docs. But that variability would occur in both arms of the trial. By most blood pressure standards, it's a pretty big trial. The number of patients with hypertension. OVERTURE is 5,700 patients and of that, what, 1,500 of them are hypertensive. That's a pretty big sample. So, when you see spot-on same trough effects -- I recognize there might have been differences in peak effects, but the most important metric that's used in hypertension evaluation is that trough blood pressure. When given b.i.d., these two drugs had an indistinguishable effect on trough blood pressure. So, it's troubling me.

DR. PACKER: Obviously, there are other hypotheses, but the other hypothesis, at least suggested by the data, is that NEP inhibition has cardiovascular benefits independent of blood pressure lowering. Obviously, we can't say that from the data. Both of those hypotheses are possible.

I actually feel more comfortable with the delta blood pressure during the day than I am suggesting to you that NEP inhibition has an incremental effect on the biology of this disease that is independent of blood pressure.

DR. BORER: From the AEs, at some point during the day, 8 percent more on omapatrilat are having a lower blood pressure. They were hypotensive.

DR. PACKER: Steve, the blood pressures had to be lower at peak because hypotension and dizziness was much more frequent in the omapatrilat group than in the enalapril group. I know we didn't measure it, but it had to be that way.

DR. NISSEN: I agree although, again, conceivably there is a very early effect. It doesn't last very long. The patients get kind of dizzy and syncopal for an hour or two, but then the levels track together. Without having ambulatory blood pressure data, we really don't know. But again, at least at trough, which is what you measured, there really wasn't much difference.

DR. BORER: Are there any other questions? Tom.

DR. FLEMING: Milt, could you put your last slide 13 up again?

You seem to be saying that we're looking at two mechanisms that omapatrilat would have. One is through NEP inhibition and the other is through whatever mechanisms that lead to the incremental antihypertensive effects, and that somehow this study is telling us that the favorable benefits on cardiovascular endpoints mediated through that second mechanism aren't in some way offset or compromised by NEP inhibition. And where does that come from --

DR. PACKER: Oh, no, no, no.

DR. FLEMING: That's what the technical wording seems to say.

DR. PACKER: This addresses specifically the concern that you raised yesterday, which is if you compare an ACE inhibitor and ACE inhibitor -- and let's assume for a moment that one reflected the committee's view that they would feel comfortable doing that. That may not precisely reflect your view, but ACE inhibitor and ACE inhibitor -- then if the one ACE inhibitor or an angiotensin II antagonist lowered blood pressure and another one lowered blood pressure more, that the delta that one observed in blood pressure would be translated into a cardiovascular benefit is because there was no other mechanisms that these drugs had that had been identified that might detract or modify the relationship between delta blood pressure and delta events. That's a hypothesis, but that's the concept that I think was promulgated yesterday.

If you go across classes, you're less certain. What I wanted to emphasize here is that there is an overlap between the mechanism of omapatrilat and an ACE inhibitor. Everyone is comfortable with what an ACE inhibitor might do. So, I want to put forward the OVERTURE data as reassurance that the incremental action of omapatrilat -- there is no evidence that that would have an unfavorable effect on cardiovascular events especially if you think that blood pressures were the same. Therefore, whatever you see in hypertension, that you could translate the delta in blood pressure to the delta in events without being concerned that there's some other action of the drug that might be adversely affecting cardiovascular events.

DR. FLEMING: Milt, it would seem, to follow through on this argument, you would have to be saying you know somehow that if you take away NEP inhibition, that the remaining mechanisms that omapatrilat would have would yield overall better antihypertensive effects than an ACE inhibitor alone.

DR. PACKER: No. I'm actually suggesting that if this drug were not a NEP inhibitor, it would look like an ACE inhibitor.

DR. FLEMING: The argument that we were saying yesterday is if you're comparing two agents that yield different antihypertensive effects and we want to infer from that difference a difference in cardiovascular benefits, that is a perfectly acceptable inference so long as there aren't any other mechanisms out there that would offset that.

So, therefore, for the logic to carry over to here, what you're having to conclude here is that omapatrilat has mechanisms relative to enalapril that yield a better antihypertensive effect and NEP inhibition is not in any way compromising the corresponding beneficial effects you would expect to see on the endpoints.

Let's move on, though, to maybe an even more fundamental question. This is sort of a negative in a certain sense. Basically when I'm looking at omapatrilat against enalapril, another way of interpreting this is to say, well, at least with omapatrilat we didn't make things worse, or we're not less effective than enalapril. And there's a little bit of that even in your hypothesis of non-inferiority. Yes, we're trying to maintain at least 80 percent of the benefit.

I'm always troubled in a non-inferiority argument, though, when the experimental arm is not anticipated to be more favorable in some way. I believe strongly in non-inferiority when I have an experimental intervention that has a safety profile or a convenience or a cost profile that would make it more favorable in that domain such that if efficacy is the same, then I come out ahead. And as a result, because of that, I'm willing to potentially give up a little bit of efficacy.

So, bottom line here is for this trial to be interpreted as positive, it's positive only in the sense that we can say we're ruling out that omapatrilat is meaningfully worse, and hence that's a win as long as in the safety domain we're all convinced omapatrilat is better than enalapril. But I think what this whole discussion is about today is that that's not where we are. So, shouldn't you have expected to be required to show at least superiority here for it to be win?

DR. PACKER: Could I have my backup slides, please, on the SOLVD Treatment definition and the slide that follows that?

DR. BORER: As you go through this, I think it's important to remember you did a heart failure trial, and we're not evaluating this drug for its efficacy for heart failure. We're trying to evaluate it for its efficacy as a treatment for people with high blood pressure. So, I think that's really Tom's point.

DR. PACKER: I think what Tom is saying -- and we have certainly learned this lesson many, many times in heart failure trials -- is that in spite of the prior hypothesis of non-inferiority, one would be a lot more comfortable if this trial had met its primary endpoint. In light of the fact that it didn't meet its primary endpoint, one has to be particularly cautious of subgroup analyses on either primary or secondary endpoints.

In light of that, I just want to mention one aspect of OVERTURE which is new. This was not presented at the ACC, but it does appear in our publication in Circulation.

Let me emphasize that the primary endpoint was death or hospitalization for heart failure. This was the definition of hospitalization used in the OVERTURE trial. It included all hospitalizations attributable to heart failure as adjudicated by the endpoint committee which required IV treatment and had a duration of more than 24 hours. This was exactly what was said in the protocol.

The reference standard for this trial was SOLVD Treatment. This was the reference standard for non-inferiority. We recognized only after the trial was over that the definition for hospitalization for heart failure in SOLVD Treatment was different than for OVERTURE. In SOLVD Treatment, the hospitalization for heart failure was all hospitalizations attributed to heart failure by the investigator regardless of treatment or duration. And there was no adjudication process in the SOLVD Treatment trial.

So, I just want to, for purposes of curiosity, show you what the data would look like if one had used the reference standard definition.

DR. FLEMING: If this is in interest of answering my question, just because time is short, I don't know that we have to go into this because I don't think this is getting at that separate issue that I was asking.

DR. PACKER: Jeff, I'll be done in one second.

This is the results you've already seen, a 6 percent lower risk with a p value. This is the primary endpoint using the SOLVD definition, 11 percent lower. This is obviously a post hoc analysis. But I offer it only to suggest the fact that had we been wise enough or whatever, if we had used the same definition used in our reference standard, maybe things would have worked out better. I don't want to put too much emphasis in it. I only provide it for whatever reassurance it would give you.

DR. BORER: Bob, did you have a comment?

DR. TEMPLE: Only that, I guess while Milton is suggesting there might be something really good going on here, the main purpose I think was to make the case that at least nothing bad happened other than the angioedema, so you don't have to worry. And that point would be fairly strong I think.

DR. BORER: Yes. I think that the issue that we're trying to focus in on here is that we're considering this drug as an antihypertensive. We want to be sure that the safety is acceptable for the intended use. It certainly is nice to know that it might turn out to be a real good drug for people with heart failure where the benefit-risk issues are very much different. But in the hypertensive population, what are we going to see?

And what we saw was that, for whatever reason, the measure that was used showed no difference in the efficacy of the drug for the hypertensive population here and perhaps no additional cardiovascular risk. So, we're still talking about the angioedema as being our primary concern. And that's reassuring to know. I mean, that's useful.

DR. FLEMING: Just in a single sentence, Bob, in view of the angioedema, all I'm saying is it's not enough to convince me that nothing bad is happening. I want to see something good happening.

DR. TEMPLE: Right, and I don't think it's being alleged, although perhaps it's being suggested, that there was any finding like that. All it does is give you some assurance that it doesn't do anything bad.

DR. BORER: That there's no new problem.

DR. TEMPLE: Given the choice of primary endpoint, you really can't say much more than that probably.

DR. PACKER: I'd like to introduce Dr. Black for the next presentation, if that's all right.

DR. BORER: Henry, just tell me approximately how long do you think you'll be taking?

DR. BLACK: Well, if I use my Manhattan speed, it will be 5 minutes. I do want to bring us back to blood pressure and I think this is a good way to do it.

DR. BORER: Okay. Why don't you go ahead.

DR. BLACK: Thanks. I do appreciate it. I realize how late it is and how tired everybody is, but I do think it would be useful to talk a little bit about where we are on high blood pressure now and to answer one question in particular, which is whether omapatrilat's greater efficacy does add to the value of current agents. I'm not going to talk about safety at this point.

In order to do this, I want to review what we did in the Joint National Committee to try to improve hypertension care. You heard yesterday from Dr. Kannel that overall we were controlling 27 percent of hypertensives in America. This is actually considerably better than any of the rest of the world, and this is only people from 18 to 74. The data for older people are considerably worse.

In order to educate physicians and also patients about how we would do this, we borrowed somewhat from ATP II and we talked about goals rather than control, understanding this was dichotomous and you could be a millimeter above or beyond or not and be at goal. But that's what we thought was easier for people, in fact, to operate with.

The goal for most hypertensives was less than 140 and less than 90. For high-risk individuals like diabetics or people with heart failure or chronic renal failure, we set that goal lower, even though at that point in time, with the possible exception of SHEP, there was no trial that confirmed that more aggressive therapy was beneficial in diabetics in particular. Syst-Eur, UKPDS, HOT, and other studies as well, LIFE, have really suggested this was a good call even though it wasn't at that time evidence-based. And for those with proteinuria, it was even lower still. This goal was not dependent on age, gender, or other forms of comorbidity.

What I want to do is show you, with that in mind, three clinical trials and my own clinical experience as to whether we can achieve that goal and why we can't.

I'll begin with LIFE, which was completed this year. This was a comparison of two regimens, one beginning with an ARB losartan, one beginning with a beta blocker atenolol, and only about 11 or 12 percent of individuals took only those drugs. It was a large trial. It was a long trial. And the goals here are shown, as you see it.

Overall, those who reached diastolic goal of less than 90 for both arms was quite impressive, almost 90 percent. However, for those who reached the systolic goal -- and as you heard yesterday, again it's systolic pressure, especially in older people, that's a better predictor of outcomes -- it was under 50 percent. And those who reached both goals, it was also about 45 to 48 percent.

In the diabetics, the highest risk group, you had quite similar data or you didn't do quite as well, 85 and 82 percent for losartan and atenolol, respectively, but under 40 percent for both arms to get systolic pressure under 140. That's not the 130 goal that we're talking about.

Two other studies, one of which is published and one of which is not yet published, I also want to show you. This is the ALLHAT trial, which was just completed. It's 42,000 high-risk hypertensives. Everybody enrolled was over 55 and had another risk factor. There were 15,000 diabetics in ALLHAT. There were about 15,000 African Americans in ALLHAT. And everybody had to have something else.

What I want to call your attention to is not the outcomes, because those aren't available yet, but how we did with respect to blood pressure. In this study, 90 percent of people were on treatment when they started and only 27 percent of that 90 percent overall were at the JNC VI goals, not even again using the diabetic goals. What happened here was you got switched to one of the treatment regimens which was a diuretic or lisinopril or amlodipine or doxazosin. And there was very, very careful nagging of our clinicians to titrate to a goal, and the goal was less than 140 over 90.

We accomplished a lot. In one year, we got 86 percent to diastolic goal, 58 to systolic goal, and this was maintained throughout. Now, this suffers from patients we can no longer follow and not having blood pressures, from people with events not being followed, people who died not being followed, but it's a good look at what happens.

However, for systolic blood pressure, which began at 31 percent under 140, we got only up to 70 percent. So, there's still a large number of high risk older people whose systolic blood pressure we could not get to below 140 in spite of these efforts, and overall, 69 percent at 6 years reached both. These numbers approximate 30,000 hypertensives.

In the CONVINCE trial, which we just presented in May, we see very similar data. Here we were comparing a non-dihydropyridine verapamil to diuretics or beta blockers as the comparators. 16,000 individuals, 13 countries. Began with 20 percent at the JNC VI goals of less than 140 over 90. Once again, no problem getting diastolic under control in this older high-risk group, but a lot of difficulty getting equally good results for systolic pressure. Started with 20, got to 67 percent. That's quite good. That's as good really as any study so far, but there's a large group of people untreated.

Now, what we did -- and you don't have this slide in your book. We added it after some of the earlier discussion -- is to show how we did it. At the end of titration, almost by definition, people were on one drug. Step 1 is monotherapy. But with time, the number of people who could reach and maintain that goal has slipped. So, by 30 months, which is the last data we have, only about 24 percent were on single agents, 44 percent were on one or two agents, many were on third agents or open label. Our physicians could use just about anything they wanted. We nagged them unmercifully to get there, and this was the best we could achieve.

Well, that's fine. Let's look at how we do in a specialist clinic. These are clinical trial patients at one end of a spectrum. They're watched closely. What about people who are refractory? And that's what we see mostly in our clinic.

We used HEDIS criteria here to see if they were reasonable, and then we had to follow everybody for at least a year. We just looked at that visit to see how well we were doing. This is 437 consecutive patients and we saw how often we achieved the goal of less than 140 and less than 90. This is where we started.

These are people sent to us because their doctors couldn't control them, and I would want to parenthetically say that main reason in two studies we've done of refractory hypertension why that doesn't happen is that people do not in practice use the right drugs in the right doses. That's simply a reality. Those are two studies separated by 10 years with exactly the same findings.

So, we started with 35 percent at systolic goal, 51 percent at diastolic goal, and that's same interesting 28 percent at both. When we got done -- and we think we're pretty good at doing this -- we came very close to the clinical trial results. 86 percent were under 90, 63 percent were under 140, and 60 percent were at both. But that's still as good as we can do.

If you look at the diabetics, it's a little more interesting. HEDIS goals at that point are less than 140 over 90, just the way the trials were. This is how we did. 87 of those 437 had diabetes. 52 percent at both goals, but if you look at JNC VI now, which was less than 130 over 85, we were only controlling 22 percent. And the biggest gap was, of course, in systolic pressure. Diastolic, we weren't doing too badly.

If you look at ADA or NKF, it's considerably worse. Now we can only get 15 percent of this high-risk subset at the goals set by expert committees.

And how did we do this? We weren't afraid to use drugs. Most of our patients were on three or four or two. Occasionally we could use non-drug therapy, but very rarely. So, of the diabetics, 50 percent were on three or more and 30 percent were on two drugs at least. And we used everything. We didn't have the restrictions you have in a trial of not having availability of a class. We used diuretics. We used calcium antagonists. We used ACE inhibitors in about 60 percent, ARBs in about 20 percent. So, we're practicing according to guidelines. That was nice to see. We looked at the few people who weren't on one of those and there was a good reason in almost every one. And we used minoxidil, central acting agents, beta blockers, alpha blockers without any particular bias.

So, I think right now we can conclude -- and I don't know if I've quite made my 5 minutes -- that we've failed to reach systolic goals in a substantial number of patients despite what we currently have to do and despite expertise and despite what happens in a trial. So, I think regimens that include omapatrilat will greatly improve our ability to achieve goals, especially systolics.

Thanks.

DR. BORER: Thank you, Henry. I didn't mean to suggest that you had to hurry. It was just that we have to take a break at some point so people can check out of the hotel and then come back.

DR. BLACK: I do understand.

DR. BORER: Why don't we take just a few minutes to ask you questions. Then we'll break for lunch.

DR. BLACK: Sure.

DR. BORER: Steve.

DR. NISSEN: Actually, Jeff, I had questions that are probably more complex than we can do in a few minutes. I think I would prefer everybody take a break. I think what Henry raises is the issue of benefit to risk. You've now given us the benefit side and I want to explore that, but I don't think I can do that quickly.

DR. BORER: Why don't we then break now and we'll come back here at 12:50.

(Whereupon, at 12:00 p.m., the committee was recessed, to reconvene at 12:50 p.m., this same day.)

AFTERNOON SESSION

(12:55 p.m.)

DR. BORER: We'll begin again.

It occurs to me, Dr. Waclawski, perhaps you want to make your concluding statements and then we'll get into the issues that we want to get into in terms of the questioning about safety and risk-benefit, to the extent that we have questions about these things.

DR. WACLAWSKI: If you would just let me clarify one thing. Did you intend to have questions specifically with respect to Dr. Black's presentation?

DR. BORER: Yes.

DR. WACLAWSKI: And you'd do that after the concluding statements.

DR. BORER: Yes. We'll wait for you to finish.

DR. WACLAWSKI: Very good.

Good afternoon again. I'm Anthony Waclawski. We'd first like to thank the committee and FDA for their kind attention and the chance to present these data to you.

As you continue your discussion today and you consider the target population, we would welcome the committee to consider our proposal for a target population. But we fully realize that there may be other subsets of this population where the committee considers the benefit-to-risk ratio for omapatrilat to be favorable. We're looking forward to your continued deliberations on these points.

That concludes our formal presentations for the day. Thank you again.

DR. BORER: Thank you.

DR. WACLAWSKI: Elliott Levy can now return to the podium.

DR. BORER: Let me ask you, because I think we want to be absolutely fair in hearing everything you think is important. You presented us or we were presented with extensive documentation of the risk management plan. Is there something that's changed since the document that was submitted to us? Because if not, I don't think we need an extensive presentation here. If it has changed in some substantial way and you think that's important, then you should be able to tell us about it.

DR. WACLAWSKI: I think we'll agree that what you've seen we've sufficiently clarified through our presentations and made the points that we felt we needed to make, mainly about the objectives of the plan and why we think that an education-based program is one that could have some success, and that we're confident with working with the agency going forward.

DR. BORER: With that having been said, let's get into the questions of the committee, the remaining questions regarding safety and the risk-benefit issues. We'll start with the committee reviewer. Steve.

DR. NISSEN: Thanks, Jeff.

Again, we left with Henry up there and we didn't get a chance to interact with you, Henry, and I really would like to. I'm going to make a couple of statements and then ask you some questions.

I assume you would agree with me that the long-term effects of vasopeptid ACE inhibitors on morbidity and mortality are not known.

DR. BLACK: Yes, I would agree.

DR. NISSEN: And that we've seen demonstration of very perhaps superior blood pressure reduction with omapatrilat in these studies. So, what I'm grappling with is how or whether we can translate the blood pressure differences for this new class of drugs into estimates for event reduction. And I want to ask you a hypothetical question because in my mind what we're all balancing here is benefit versus risk, which you were obviously addressing in your presentation. So, here's my question.

If we had a diuretic that reduced 24-hour ambulatory blood pressure by 12 millimeters of mercury and an ACE inhibitor that decreased ambulatory blood pressure by 10 millimeters of mercury, would we be confident that the diuretic arm would result in reduced events? So, you have a diuretic that decreases by 12 and you've got an ACE inhibitor that reduces by 10, using the most elegant 24-hour measures available.

DR. BLACK: This is if compared to each other?

DR. NISSEN: Yes, compared to each other.

DR. BLACK: I think in general I can't answer that specific question without having some real data to back it up.

I do think -- and I think this point was talked about a lot yesterday -- that incremental drops in blood pressure, even small ones, seem to result in considerable reduction in outcome events.

DR. NISSEN: I guess what I'm trying to get at is whether, in fact, one can predict that a drug that has a modestly greater blood pressure reduction will result in a greater reduction in events. Because isn't that what we're being asked to assume here in terms of the benefit of this agent?

DR. BLACK: Yes, I understand. There have been some attempts to do this. Some have used epidemiological estimates. I think you saw some of that yesterday from Dr. Kannel. Dr. Stamler has used similar things to predict reductions in mortality, of small reductions in systolic pressure leading to fairly large reductions in mortality. And there's been a large meta-regression done by Jahn Staessen suggesting that small differences in systolic pressure could result in 20 to 25 percent reductions in cardiovascular mortality. Now, those studies are always up for some interpretation, but there's a consistency about them based on 31 clinical trials that have been done so far.

DR. NISSEN: Mike Weber, do you want to offer us some advice here? Because there are some calculations that appear in here about this relationship between how many events are prevented versus the risks.

DR. WEBER: Yes, but underlying that, Steve, is exactly the conversation you had yesterday morning and that the difference in blood pressure has its greatest meaning when you're comparing the same kinds of pharmacology. So, if you had one ACE inhibitor that was minus 10 and the other was minus 12, then I'm going to favor the one that's minus 12. But you gave us a diuretic at minus 12 and an ACE inhibitor at minus 10, and that's a very difficult situation because they clearly have very different profiles, and I suspect there are some patients who are going to do a lot better with one than with the other. But I think what we're talking about here, of course, is within the ACE inhibitor family.

DR. NISSEN: Well, see, it was a deliberately difficult question because for me to make a judgment here, I have to believe that vasopeptide ACE inhibitors fundamentally will act on events in the same way that the ACE inhibitors act on events. Yet, we're talking about the first drug in a new class that does different things. So, I'm asking you guys -- I respect both of you. You've done tremendous work over the years in hypertension -- whether you can justify that sort of assumption, and if so, how you can justify that assumption.

DR. WEBER: Well, I think we would both depend quite heavily actually on what Milton showed us just before lunch because a concern you would always have is that what it is that is different about omapatrilat didn't just mean more blood pressure reduction but everything else stays the same, but whether there is something about adding in this NEP inhibition that's going to cause something that's unexpected or adverse.

I think the two things that Milt showed us, first of all, the major cardiovascular endpoints in OCTAVE were certainly moving in the right direction, for whatever that's worth, but certainly not moving in an adverse direction, and secondly, in the OVERTURE study, the heart failure trial, where again you could have a pretty strong level of confidence, particularly in the hypertensive patients, that if anything, things were favoring omapatrilat, for whatever that's worth. But I think we can at least lay to rest the concerns that Tom Fleming was expressing yesterday about comparing different classes.

DR. NISSEN: Well, I guess the problem we see with using the OVERTURE data isn't the problem that that's not the population that this drug is being proposed to treat. It's not being proposed for heart failure. It's being proposed for hypertension. So, the reassurance is obviously going to be limited, is it not, by the fact that it was studied in a different population than is being proposed to be used here?

DR. WEBER: That's right, albeit a high risk population with hypertension, but I acknowledge that.

Dr. Hennekens, I wondered, would you have a comment?

DR. HENNEKENS: Well, on this point, Steve, I'm a recent addition to this advisory group because of work I had done at Harvard on hypertension, some of which included collaboration with the Oxford Group. We looked at 14 randomized trials, including over 30,000 subjects that were treated for 2 to 3 years with blood pressure lowering agents. We predicted going in that the 3 to 5 millimeter reductions would be associated with about a 40 percent lowered risk of stroke and a 30 percent lowered risk of CHD and about a 20 percent lowered risk of cardiovascular mortality.

What we found is that 2 to 3 years of blood pressure lowering led to the predicted 42 percent lower risk of stroke, about a 16 percent reduction in heart disease, and about a 21 percent reduction in vascular mortality, so that the stroke and the vascular mortality reductions, over 2 to 3 years with this amount of blood pressure lowering, were very similar to the epidemiology. Where there was the shortfall was in CHD.

We speculated that chance in the trials might explain it. We speculated that there might be a more immediate and direct effect on the brain, a more delayed and indirect effect on the heart via atherogenesis. We also speculated that the first-line drugs, the diuretics and beta blockers, which have a 5 percent adverse effect on LDL, might be increasing the risk of coronary heart disease events.

The issue here becomes complicated in terms of the application of those data to the risk-benefit ratio on this drug, and I think that's why the sponsors have correctly tried to define a target population, all of whom have a 10-year risk of about 20 percent or greater, and if we add uncontrolled hypertension to that risk, it's probably closer to a 40 percent 10-year risk of adverse cardiovascular outcomes. And it's in these patients where I think the claim is that the benefits will outweigh the risk, but that does presume a 2- to 3-year sustained difference in blood pressure of that amount.

DR. NISSEN: I've looked at these data very carefully as the primary reviewer here, and as I think we all at this table know, there is no long-term exposure data to omapatrilat available. So, we don't know what happens down the road.

DR. HENNEKENS: No, but what we do know, though -- and I will yield to Dr. Levy in just a second -- is that you see a sustained advantage over 24 weeks in 25,000 subjects, and that's not the same as a 2- to 3-year reduction, but it's at least heading down the right path.

DR. BORER: Bob, did you have a comment about this?

DR. TEMPLE: Slightly different.

DR. LEVY: May I just respond? A point of clarification. We do actually have quite a bit of long-term experience with omapatrilat. We have patients treated for up to 5 years. Patients have been treated in controlled trials for up to a year, and the antihypertensive effects are sustained and they're superior to comparator. Over 5 years, there's no indication that the antihypertensive effect is lost. In fact, we did a withdrawal study in which patients who were maintained on the drug for over a year and had stable blood pressures were withdrawn from therapy to demonstrate that it retained its antihypertensive effect.

DR. NISSEN: But let me just make sure I understand what you know. The differential effect against, say, enalapril is sustained in those longer-term trials?

DR. LEVY: We have comparative data versus losartan not enalapril in a trial that lasted a year, and the difference between the two drug regimens is sustained.

Our longer-term experience is in primarily open-label, uncontrolled trials.

So, of course, we can't speculate what would happen if patients were to be followed for 5 or 10 years. On the other hand, in the patient population we've identified who don't seem to be able to get to target with existing meds, it seems highly likely that there would be some lasting benefit if they can stay on this one.

DR. NISSEN: But there's no hard data on differential effects beyond 12 months.

DR. LEVY: Yes, that's right.

DR. NISSEN: Now, the second question --

DR. BORER: But just before you go on to that, Steve, Bob.

DR. TEMPLE: I have an observation about this. Steve is obviously asking the fundamental surrogate question. You always, when you rely on blood pressure, are making some assumptions, and they're not always right. I'm absolutely positive you get better blood pressure control with 100 milligrams of chlorthalidone than you do with 25 milligrams of hydrochlorothiazide, and there are even well-controlled studies that show improved survival or improved stroke anyway in those people. But it turns out you pay a price for the better control in the form of arrhythmias and other things. So, it turned out there was an additional effect in addition to the one that you were relying on that was a worry.

Nonetheless, for what's it worth, we do act -- and the whole community acts -- as if lowering blood pressure to goal is a desirable thing however you do it, with whatever drugs you do it, even though they can't prove that. ALLHAT is supposed to get you some further insight on that question, is lowering blood pressure equivalent, to the same extent the same, no matter how you do it?

DR. NISSEN: I guess the spirit of my question relates to trials like LIFE where a similar blood pressure reduction has different effects on events. So, what I said yesterday I'm kind of repeating today, which is when you cross classes, there may be class-specific effects that are unknown, uncertain, and it's particularly germane when it's the first drug in the class were we don't really know, over a period of time, what the effect of morbidity and mortality are, let alone know what it is relative to some other agent.

DR. TEMPLE: Yes. But of course, every time you approve a new drug, especially of a new class, you don't know that. Some people would tell us we should make people know that and do an ALLHAT each time, but we have not adopted that policy.

DR. NISSEN: Well, Bob, again, the difference here is that there's risk. If the drug had a similar risk profile, we wouldn't have this conversation.

DR. TEMPLE: That's fair.

The other observation I guess -- I've written this, so I want to say it -- is it doesn't seem out of the question about you can learn about some unexpected bad news from studies in different populations. So, I've always felt some of the concerns about calcium channel blockers in hypertension were, to some extent, resolved by the post-infarction studies, a fragile group, and it didn't seem to do anything bad in those, other than cause heart failure. And I believe that's the argument they're making about OVERTURE, that it should reassure you that nothing unexpectedly awful is happening even though it's a different population.

DR. NISSEN: So, for me the most powerful evidence would be obviously direct evidence on morbidity and mortality, but there is less powerful evidence that might be germane here, and I wanted to give you an opportunity to give us your perspective on it, and that is the issue of target organ protection. So, I wanted to know if there is any evidence here of superior target organ protection for omapatrilat in comparison to other available agents.

So, this would obviously be another kind of surrogate, but it would be one that we probably ought to weigh in our deliberations. So, I'd be interested in anything you can provide that might help us there.

DR. LEVY: Some of this information was cited briefly in the briefing documents. Let me just review that.

DR. NISSEN: But we haven't discussed them. That's what I want to discuss now.

DR. LEVY: Yes. There's a presentation on this issue.

We conducted a number of studies designed to examine the effect of omapatrilat on target organ damage. Three of them were cited in your briefing document.

DR. NISSEN: Now, keep in mind now we're talking about in comparison to other agent, not just against placebo.

DR. LEVY: Yes. I'm going to show you the findings from three studies and briefly cite a fourth.

In your briefing document, the first is a trial that was conducted in patients with chronic stable angina. We had noted in preclinical studies that omapatrilat had an anti-anginal effect that wasn't shared by ACE inhibitors. And we conducted a trial -- it was a placebo-controlled trial -- in which omapatrilat was shown to improve exercise tolerance. That was this study.

And if I could go to the next slide. We demonstrated significant improvements in various measures of ischemia, including maximal exercise duration, time to onset of angina, and time to ST segment depression. So, not an active-controlled trial, but a novel finding that hasn't been described with ACE inhibitors.

DR. NISSEN: Unfortunately, it wasn't germane to what I was asking because I'm looking for evidence that there is some target organ protection here not afforded by an active control agent. In other words, is there anything that says that omapatrilat improves angina in comparison to enalapril or lisinopril or amlodipine or anything like that?

DR. LEVY: Right. Well, again, this hasn't been described with the ACE inhibitors, so we thought it was worthwhile.

DR. NISSEN: Also, certainly with amlodipine it's been described.

DR. LEVY: If I could have the next slide. We conducted a study to examine the effects of the drug on proteinuria. This was actually a study conducted early in the clinical development program. We used amlodipine as the comparator because of its potent effects on blood pressure and its apparently neutral effects on proteinuria. We found in this trial, if I could have the next slide, that omapatrilat produced significant reductions in urine albumin excretion rate that in magnitude were about comparable to that seen with the ACE inhibitors.

DR. NISSEN: Were there any direct comparisons made between omapatrilat and, say, ARBs, which I guess are about to be labeled for this indication, or ACE inhibitors?

DR. LEVY: Not for this purpose.

Now, the next study we have a comparison with losartan in patients with left ventricular hypertrophy. The primary endpoint here was change in echocardiographically determined LV mass after 24 weeks of therapy with omapatrilat or losartan, and then patients remained on therapy for up to a year. At the primary time point, at week 24, both drugs reduced left ventricular mass to a significant degree with a trend towards greater reduction with omapatrilat.

DR. BORER: What were the blood pressure responses to the two drugs?

DR. LEVY: Can I see the tracing, summarized blood pressure changes over the full duration of the study?

These are the blood pressure changes over the full 52 weeks of the trial. In the first 24 weeks, patients remained primarily on monotherapy, and then they went on to add adjunctive therapy. About 34 percent of those treated with omapatrilat received another agent; about 60 percent of those treated with losartan. And there's a difference in systolic blood pressure of about 4 millimeters of mercury that's pretty well sustained from week 24 in the trial on.

You had asked me before about what evidence we had that there's a long-term superiority. This is an interesting trial, much smaller than OCTAVE, but one in which, despite a much greater discrepancy in the rate of adjunctive therapy use, you still see a preserved difference of about 4 millimeters of mercury in systolic blood pressure.

DR. NISSEN: But I guess I was looking more for evidence. I guess what I'm trying to understand is it would help me if there were evidence that in comparison to ACE inhibitors or calcium channel blockers or diuretics, that some organ system was protected in some way.

DR. LEVY: Let me show you one more study, if I could just have the primary finding from the CHOIR study. These are data that were not included in the NDA, and so the FDA hasn't reviewed them. If there of interest to you, I'd certainly like them to review the study.

But this was a study that we conducted in patients with systolic hypertension, randomized to treatment with omapatrilat or enalapril, in which we assessed the effect of the drug essentially on conduit vessel stiffness, which is a major finding in older patients with primarily systolic hypertension and is thought to have a pathogenic role. Now, in animal studies the natriuretic peptides were shown to have a favorable effect on the large arteries.

In this study -- can I just have the primary results?

DR. NISSEN: I'm not sure I would call that an end organ, though.

DR. LEVY: Well, there's a degenerative change in these vessels over time that seems to be associated with poor outcomes.

Anyway, the drug produced a reduction in central pulse pressure that's not seen with enalapril, and it indicates that there's a distinct effect on the pathologic change in these conduit vessels.

So, we certainly have a variety of information about target organ damage. At the very least, the drug appears to share the beneficial effects of existing drugs. It may be superior in some areas.

DR. NISSEN: Yes, I would agree with that conclusion, from what I've seen in the documents, that it does appear to share those properties. But again, looking for superiority as a way to justify the increased risk, that was what I was really probing for.

That's all I have.

DR. WEBER: I just wanted to remind Steve that, in fact, that in previous trial with losartan, the differential effects on left ventricular hypertrophy were really quite clear. And that's interesting because in the LIFE study, if you remember, losartan was clearly superior to the beta blocker in regressing LVH. So, this is, if you like, one good example of a target organ difference to the favor of omapatrilat against a standard comparator.

DR. NISSEN: Well, Mike, the differences were highly significant compared to placebo, but there was not a significant difference compared to losartan. The p value was nonsignificant. So, again, it was a demonstration of equivalence, not necessarily of superiority.

DR. WEBER: I think it was 7 versus 4.

DR. NISSEN: Well, but the p value was greater than .1.

DR. LEVY: You're correct. There was a trend towards greater reduction with omapatrilat.

I have data on some of the subgroup analyses that were requested before the break.

DR. BORER: Why don't you go ahead and then we'll get on to some other questions.

DR. LEVY: Dr. Fleming had asked about efficacy and safety in the proposed target population. If I can have the first slide there. This is the proposed target population. These patient populations are identified based on review of the clinical guidelines to determine patient populations that would increase CV risk and therefore might stand to gain the most from incremental reductions in blood pressure. Of course, the second criteria, hypertension difficult to control with existing agents, patients who can't benefit elsewhere.

We presented these data by subgroup because they're post hoc analyses, and it's very important to be able to examine each of the subgroups for consistency.

Can I have the next slide? I showed you these results earlier. I call your attention to the right-hand panel. There's a very consistent reduction in blood pressure in all these high-risk groups, ranging from 3 to 5 millimeters of mercury more with omapatrilat than with enalapril.

But for the sake of clarity, we've prepared a pooled analysis in which we put together these populations. This is what a population looks like. Again, the two largest risk groups that were represented in OCTAVE were diabetes and atherosclerotic disease with 3,300 and 2,300 patients respectively, and then smaller numbers with renal disease and heart failure. So, there are about 6,000 patients represented in this analysis. They tend to be a little bit older than the overall study population, but otherwise they're not remarkable in terms of demographic characteristics.

DR. FLEMING: And do all of these patients also satisfy the criterion of having had a difficulty to control hypertension?

DR. LEVY: This is all subjects. I wanted to show you the largest group possible. We've also done these analyses for those who entered the study uncontrolled on medication and the results are very similar.

DR. FLEMING: If you have it, because time is short, it would be adequate just to drill down to that target group rather than including this bigger group that includes a number of people who wouldn't be in your target, if you have it.

DR. LEVY: I'll call it up, but there really are only two slides to show and one is that, as you'd expect, when you see groups that are consistent, you see a consistent difference in efficacy of about 4 millimeters of mercury in the target population at week 24. As I showed you earlier, the rate of angioedema in the study was lower in those with diabetes or atherosclerotic disease than in others. So, the risk of angioedema in the target population is also lower. The two events that are subject to that airway compromise were not in the target population and the number of patients hospitalized was also quite small.

DR. FLEMING: So, it's just not been possible at this point still to produce the actual target population subgroup? I'm presuming that the target population subgroup would only be half that size or two-thirds.

DR. LEVY: I'm sorry. If we were to focus on those patients who entered the trial uncontrolled on therapy with the same comorbid characteristics, there are about 2,000 subjects in the analysis. Again, the reduction in blood pressure is 3.6 millimeters of mercury more with omapatrilat than with enalapril.

DR. FLEMING: And do you happen to know what the distribution is for the clinical events and also for the safety events?

DR. LEVY: Yes. As you know, there were 226 clinical events in the trial. In this group there were 102, 58 in subjects randomized to enalapril and 54 in subjects randomized to omapatrilat. So, the hazard ratio is .91. It's very consistent with what we saw overall.

DR. FLEMING: And the angioedema? Did you have that data?

DR. LEVY: Well, this is the angioedema.

DR. FLEMING: That's still a bigger group. Right? That doesn't focus or drill down on only those people that were difficult to control hypertension at baseline.

DR. LEVY: In those who had these comorbid characteristics and who entered the study on medication uncontrolled, there were 18 angioedema events out of 1,140 subjects on omapatrilat, an incidence of 1.58 percent, and 8 events out of 1,053 subjects on enalapril, .76 percent. In both cases, most of the events were severity class I.

DR. FLEMING: Do you have how many were at III-IV?

DR. LEVY: I'm sorry?

DR. FLEMING: Severity class III-IV.

DR. LEVY: Well, as you can see here, in the larger group, there were 2 patients who were hospitalized, neither with airway compromise, and there were no patients who required mechanical airway protection. In the smaller group, there were also 2 subjects hospitalized without airway compromise.

Does that answer your question?

DR. BORER: Not quite. Just so it's on the record here, people who entered the trial uncontrolled are not actually the group that Tom is focusing on. I'm sure you don't have these data, and nobody expects you to put them together in 2 minutes. But it's the people who couldn't be controlled, not the people who weren't controlled.

Henry Black showed us that there are people in his own clinic -- and he's an expert -- on maximal therapy who aren't controlled. So, they exist but those aren't the people who came in uncontrolled into this clinical practice population for a study.

The question we would really have to define the risk-benefit ratio we want most precisely would be the people who, on maximal medical therapy under optimal care, could not be controlled without omapatrilat and now could be controlled with omapatrilat. What's their risk?

I don't think you have that group, but it's different from your group 3 in the OCTAVE trial. Again, you may have those data. I don't know.

DR. LEVY: We've showed you data from a variety of groups that are relatively difficult to control. I think what you can conclude is that the efficacy advantage is preserved, no matter how difficult the patient is to control, and we got some very difficult-to-control patients represents in OCTAVE. In patient populations where there's a higher risk of diabetes or atherosclerotic disease, there's less angioedema.

DR. BORER: All the data you've shown us are consistent with what you're suggesting.

One of the reasons that I'm sort of not fully satisfied -- and it may be impossible without some additional trial to provide that satisfaction -- is that the argument that routine clinical practice does it this way and they don't make it just isn't really a very good argument to me, as I suggested earlier.

One of the reasons that I'm concerned about this and the underuse of appropriate medications and whatever are the data that Ray Lipicky presented I think the first time two years ago, although maybe he put them together earlier than that, about the dose-response curve of antihypertensive drugs showing that, by and large, probably everybody underdoses most antihypertensive drugs, and if you just push the dose a little bit more, you'd get the blood pressure down with conventional agents that have already been approved.

So, to assuage my concern about that, the most convincing thing I've heard and seen was the slide that Henry showed earlier from his own clinic where there are people who are really expert at this who have --

DR. BLACK: I can perhaps try to give you some idea. I don't think it's possible with what you saw here to make that guess. But in our diabetic group, which is not large, where we are very aggressive using the real guidelines for diabetics -- and diabetics are one of the groups that this is being recommended to use -- we could only achieve goal in about 20 percent of people with what we currently have. And we use large doses. We use four or five or six drugs if necessary. So, there's a big gap in that group alone.

DR. LEVY: If I could show one slide just to clarify a point. If I could have the slide from the 73 study.

We're proposing that the drug be used in the patients that Henry Black is talking about, the patients who can't be brought to control despite very honest attempts to get them there. In the right-hand panel here, you've got patients who were on very high dose ACE inhibitor therapy plus one or two or three other medications. They remain far from target with systolic pressures at baseline in the 150s. The substitution of omapatrilat for their prior ACE inhibitor therapy produced further reduction in blood pressure of around 10 millimeters of mercury. There are few alternatives for these patients.

The question was raised before about whether one could achieve the same results with b.i.d. enalapril or with the addition of a thiazide. In patients who can reach those with those manipulations, this is not the role of the drug. But there is a very substantial incremental blood pressure reduction which may be of value in patients who are very difficult to control with existing drugs.

DR. BORER: You don't really believe that if the drug were approved and marketed, that it would only be used in the group that Henry Black couldn't control with six other drugs.

DR. LEVY: I think Henry has shown data that there are 30 to 40 percent of patients who can't be controlled.

DR. BORER: I understand, but I'm asking you do you really believe that that's the way the drug would be used if it were marketed. It's a rhetorical question.

(Laughter.)

DR. CARABELLO: But it's a question I'd like to go into a little bit further. I see this drug perhaps as somewhat akin to amiodarone where you have the nettlesome problem of atrial fibrillation, very few drugs to control it, and we have a very toxic, not particularly safe drug, but its risk is mitigated by the fact that it's only prescribed by a few people, those people who have a special knowledge of the drug and of arrhythmias. My question would be is, could we limit the drug in terms of who prescribes for what, where, and when? And would that not be yet a strategy we haven't talked about for mitigating risk?

DR. WACLAWSKI: Excuse me, Dr. Borer. Could I just add to your rhetorical question perhaps? It's certainly something we have been discussing on the risk management side within the company for some time, and it certainly is one of our concerns as well, which is that if and when the drug were to be approved with a target population, it would be necessary to show that we could limit the use to those patients where the benefit-to-risk is clearly favorable.

And that's important to us not only for the good of the patients, but also because even if the benefit-risk was to be expanded beyond that later, it's important to focus on a group that has the highest benefit-to-risk initially for the initial marketing of the product. We recognize that as a risk and we've worked internally try to work through that. And there may be some tools, some ways to build the risk management plan around that, and that's something certainly we would welcome input on. But your concern is well taken.

DR. BORER: 15 years ago -- and I guess it's 17 years ago now -- this same discussion revolved around amiodarone, and the consultants who were speaking for the drug insisted that it should be approved only for use by experts who were the six of them sitting in the front row --

(Laughter.)

DR. BORER: -- and shouldn't be used by anyone else. I don't think we achieved that, but perhaps that's okay.

Bob.

DR. TEMPLE: I have a slightly different question. Let's say we were willing to assume that it really was good for outcome to be able to lower blood pressure 3 millimeters of mercury more than you otherwise could. That's sort of what Henry is saying in some ways.

There are two sets of data. One is moderately convincing evidence I think that this works a little better than other ACE inhibitors. At least with lisinopril, how many times a day you give it probably doesn't matter, and they seem to have some data there. So, that's one thing.

What I hear Henry saying is, look, if this is as well as you can do with available therapy without omapatrilat, you're going to do 3 millimeters of mercury better when you substitute this for your other ACE inhibitor. So, that's one line of argument.

The other line of argument is that OCTAVE actually showed that you could get 3 millimeters of mercury better with this than without it, but I have some questions about that. These are all points raised by Dr. Stockbridge in his review.

It's quite striking that even though people were allowed to increase the dose of enalapril to gain control, only about 40 percent of people got on the maximum dose, and just to save Steve from having to say it, they didn't get an opportunity to have it twice a day. So, you don't really know what would have happened if they had gone to the right dose. Maybe that 3 over 2 would be 1 over .5.

In addition, the fairly simple expedient of adding another drug was only used in a very small fraction of patients. So, I recognize the idea that is implicit in what Henry said, which is, well, it works better, so you've got to end up better. And I guess I raise the question, don't you have to know in practice how different these resistant patients will be when you actually do it as opposed to sort of the theoretical advantage which is, well, how can it not, which is I think what Henry's argument is.

So, I'd be interested in some response to why, given the opportunity to use the proper dose or the maximum dose, if you like, of enalapril and given the opportunity to add therapies which they could have, nobody really did it. So, do we really know how much better this is than conventional therapy in an actual "I can't control this patient" setting?

I take your point. A lot of people can't be controlled, but if this were available, do we actually know in a hands-on way and a demonstrated way how much difference it would make? That's really what Norm was asking in this review.

DR. BLACK: Bob, if I could, I'd try to give you two impressions. We've done two assessments of our clinic when I was in New Haven and again in Chicago about 10 years later to look at our patients who were resistant and see what the reasons were and what we could about it. The most common reason both times was that the patients that we got were not properly dosed, did not get the right drugs in the right order, didn't have them long enough, exactly the practice gaps we see.

We were able with our manipulations to get control in about 60 percent, very similar to what you saw here, both times. In 1990, it was using diuretics when people didn't know how to use them, and there are newer things now. That wasn't quite the problem. So, I think that's going to be an issue. I don't think you can address exactly what would happen if omapatrilat were around, but I think that's the pattern of practice.

In our trials, where we do lay out a protocol, we reward people for getting control and punish them when they don't and let them use whatever they want. We still can't do any better than, in fact, what we're seeing.

DR. TEMPLE: But I'd still like to hear a little bit about -- I mean, you had a difference of 3 over 2, or thereabouts, with people on inadequate doses of enalapril. You've got to imagine that if the dose had gone up or if it had been b.i.d., the difference would be less than that, and you certainly have to imagine if they'd added a drug, which in many cases they did not, the difference would have to be reduced. Now, Norm had an estimate based on what happened when you did add a drug, that it wouldn't be very hard to get control by adding another drug, and yet they didn't.

Obviously, the question is, okay, on the one hand, I can get away without this other drug. On the other hand, I have the angioedema. So, you sort of have to know how you do with another drug. Or maybe you don't think you do. So, what do you think about that?

DR. FLEMING: Henry, just for my understanding of your response to Bob's question just now, you've made the strong point that we have to do better in being able to get a larger fraction of patients to target and that it's hard. A lot of the existing agents in a lot of patients are unsuccessful. Yet, it seemed in this trial that there was a large fraction of patients in which there wasn't even the attempt made or the additional -- in the population here in table 5 in the FDA briefing document, it looks like there are approximately 20 percent that received adjunctive treatment in weeks 9 to 14 and large fractions in the first 8 weeks didn't get to maximal dosing.

So, that seems to be perplexing. It seems like in the majority of cases we're not taking advantage of what there already is available to us to achieve what you're saying we need to achieve. It's not a matter of they're trying it and it's not getting there.

DR. BLACK: I think that worked in both arms of the study. You didn't get to 90 or 100 percent control rates in the omapatrilat arm either.

DR. FLEMING: In both arms.

DR. BLACK: In either arm. And that's only 6 months. Ours is a year. The trials are 4 and 5 years. So, there are several different ways to look at that.

I don't think that's how people practice. We have an education message for that as well, but I'm not sure we can manage even with what we do in trials or what happened here to get people to practice as we would like them to do it. I think a couple of extra millimeters for each one of our steps is going to get us much further along the road. We've got a lot of education to do beyond what you see here to get people to use what we have right and to have a more powerful tool to use that right.

The NHANES IV data is kind of almost done. NHANES III is where the 27 percent comes from. There's been a big national campaign to improve control rates, and it's been, at best, only modestly successful. We've improved control rates in men to about 30 percent, which is about what women get, but nowhere near what we ought to. So, there's a lot to still be done. I think we need better agents. Even with the optimal circumstances of a trial or of a specialist clinic, we're still not where we want to go.

DR. TEMPLE: That's what I understood your argument to be. You're basically saying how can we not do better if we have a drug that works a little better than the thing it's going to substitute for. And maybe that's sufficient. I'm not trying to prejudge this, but that's not the same as saying in a population where we couldn't get control, here's what we know about substituting versus, say, adding another drug. It's a question of whether you need data on that point or have data on that point as opposed to making some not necessarily unreasonable assumptions.

DR. LEVY: Let me just make two points. We freely concede that in patients who can readily be brought to target with the addition of a thiazide or up-titration of an existing drug or a switch to b.i.d., that this drug is not needed. The drug is being proposed for use in patients in whom the option of adding another drug or adding another drug and reaching target isn't available.

Now, physicians in this trial, as you pointed out, didn't invariably add adjunctive therapy when it was available to them. They did so about 40 percent of the time, and they were much more likely to do so in patients whose blood pressures were well above target on monotherapy than in those who were closer to target, just as physicians in practice would. But there's a wide variety of practice represented in the trial.

What we did here was just to classify the investigative sites according to the aggressiveness of the physician, how frequently did they add adjunctive therapy at a visit where a patient's blood pressure was above target? These are 20 percent of sites, the quintile, where the rate of adjunct use was highest, and these physicians I think were very diligent adding adjunctive therapy at least two-thirds of individual visits where patients remained uncontrolled, up to 100 percent at some sites. And even at these sites, you see the same difference in blood pressure reduction between omapatrilat and enalapril that you see overall. You see the same difference in control rate.

This speaks to Henry Black's point. In the difficult to treat patient, if the physician decides to become more aggressive, add therapy, up-titrate, they're going to do that whether the patient is on omapatrilat or enalapril, and the results will be better with omapatrilat.

DR. FLEMING: But that's not in your targeted subgroup. That's all patients at the sites that had the highest adjunct use?

DR. LEVY: These are roughly 2,000 patients at the 20 percent of sites where the physicians were most aggressive.

DR. FLEMING: So, what were the results at those sites that had the highest adjunct use that were in your target population? What were the results?

DR. LEVY: Well, we haven't done that analysis. You've seen the results were extremely consistent across this database.

DR. BORER: Tom and then Steve.

DR. PACKER: Jeff, if I could, I just want to address directly Bob's question. You've seen the data that Dr. Levy just showed you, but this shouldn't be too surprising and I wouldn't be surprised if it were shown in the patient population that was being proposed. The concept I think is that if you have a patient who has easy-to-control blood pressure and you can get people to target, then I think it would be a very easy assumption to say that the difference between omapatrilat and an ACE inhibitor could be made up by adding a diuretic. No big deal.

But if patients are very far away from target, then a diligent physician, a really good physician, would then add incremental therapy or give b.i.d. or however you want to do it to both groups. The difference in blood pressure only disappears if there's a differential use of adjunctive therapy or intensive therapy. But that differential use would never occur if you're not to target. You would expect a patient who is resistant to therapy to -- the physician would add adjunctive therapy or optimize dose or would do whatever you would want them to do to both groups. As long as they're far from target, the difference in favor of omapatrilat would always persist.

DR. TEMPLE: So, that's an argument of logic, although it's not necessarily an argument of data, although that last thing was interesting.

DR. PACKER: Yes. But if you think about it as the fact that in order for the difference to disappear, there has to be a differential use of adjunctive therapy, which cannot -- and by the way, to a substantial degree, the calculations that Norm Stockbridge made was 50 percent difference in differential use. Regardless of what the number is, the number isn't important. As long as you're far away from target, that differential use won't happen, especially if you're requiring people already to be on multiple drugs at the beginning.

DR. TEMPLE: It's still a question of whether that makes sense, which I wouldn't say it doesn't, and whether there's actual data so you can see actual numbers about how much better you do.

DR. PACKER: I think what I'm addressing is not whether what I'm saying absolutely has to be true, but I think you were struggling with how it could be true. What I'm proposing is that's the explanation for why it would be true.

DR. BORER: Why don't we move on to another issue. Tom.

DR. FLEMING: Can I have one last comment?

DR. PICKERING: It is actually a related issue. Starting with the possibility that if OCTAVE had been continued longer, the difference might have diminished, my reading of the protocol was that there were only two clinic visits at which it was possible to add additional drugs, which in my opinion is not very long.

Henry, you mentioned the ALLHAT trial. I don't know if we could see your number 6 slide which shows that using conventional therapy in not perhaps quite as high a risk population but certainly in an enhanced risk population, you can in fact do extremely well, going from 27 percent to 69 percent control at 6 years. I guess it's a question of whether you see the glass is half full or half empty, but I would say that's a huge increment just with the use of conventional therapy.

DR. BLACK: I would agree. I think we did make a lot of progress, but I think there's still a lot more progress to be made especially in what we don't have the tools for. These people are seen every 4 months.

One thing too, when you begin to add third and fourth and fifth drugs, you get into less well-tolerated drugs, drugs with their own inherent problems too. It's not just adding drugs that are free of side effects when you get to that, and patients don't comply well with that. So, we begin to run out of well-tolerated options a little earlier than maybe has been implied.

But we're not doing badly here, nor did we in CONVINCE, but we're not doing nearly as well for systolic as I would like to do. That's where the problem is I think.

DR. BORER: Bob.

DR. TEMPLE: Well, Henry, if I remember ALLHAT, the additional therapy was sort of peculiar. For example, if you were in the diuretic group -- you tell me -- you couldn't use an ACE inhibitor because that was another group and that would have confounded. So, you were limited in the number of specific additional drugs you could use.

DR. BLACK: The trial protocol -- the artificiality of any active comparator where you can't use what you would ordinarily use --

DR. TEMPLE: No, I'm not blaming the trial. But it means that may not be as good as --

DR. BLACK: What you could do, though, if you felt you needed it, was add an agent of the blinded class at half the initial dose, drop down your drugs. So, you could do it. I don't have the drug use in ALLHAT yet as to how many people got on open-label, how many people crossed over, but it was in fact substantive. It wasn't just that you couldn't do it. People did use their own drugs.

DR. BORER: Steve.

DR. NISSEN: I feel compelled to point out for Bob and for others, which I know everybody here knows, but we don't know whether that extra 3 over 2 millimeters for this drug actually lowers events. So, this is all hypothetical, what if we did this and what if we did that. The reality is if there were a very large difference between the arms, I think one could be more confident, but these are small differences. Therefore, without knowledge of the relationship for this class of drugs on the amount of blood pressure control versus morbidity and mortality, we're really speculating about what the ultimate impact is going to be.

DR. BORER: Are there any other questions, issues the committee wants to raise? Because if not -- Tom.

DR. PICKERING: I'd like to return to the issue of adherence. I think we're all agreed that omapatrilat is a more effective antihypertensive than the other existing drugs. One of my concerns is what the adherence will be with this medication, given all the warnings and education about potential risks. It could be that the potential benefits in its potency might be offset by decreased compliance.

On that note, I think it's just worth mentioning that patients' self-report about compliance with antihypertensives is generally thought to be next to worthless. Pill counts are maybe a little better, but a lot of the pills end up in the parking lot of the hospital of people in studies. So, I think it is a real issue, and I would be concerned about this.

DR. BORER: We'll structure the remainder of our discussion around these questions. I want to frame them before we begin. I think, although the questions are reasonably straightforward, there are some key issues about which we need to make judgments.

First -- and Tom just mentioned this -- is are we convinced that omapatrilat is more effective within the labeled range than other available antihypertensive agents so that one might expect that it could add something. And a subset of that is does it add to other drugs that might be given at the same time.

Second, is there a population that could be controlled by omapatrilat that cannot be controlled with all the other conventional therapies and approved therapies that we have?

Third, if there is such a population, can we define it?

Fourth, if there is and we can and we do give the drug and it does lower blood pressure by some amount, is there a clinical benefit associated with that, something what we really have never required anybody to show and maybe we can take on faith. But the issue still remains as Steve has said several times.

Finally, assuming that lowering blood pressure is equivalent to a clinical benefit, what are the risks associated with gaining this benefit in the specific group that we have defined that couldn't be controlled in any other way than by adding omapatrilat? That's really the sequence of issues that we have to face and we're going to face them in different ways in these questions.

Bob.

DR. TEMPLE: If you got to that point, you would also really need to address the risk management program and whether you think that will do it. We have people here who reviewed the risk management program who could comment further than what they've already written, if that were helpful.

DR. BORER: Ultimately, presumably that mitigates to some extent or may mitigate to some extent the total risk, and if you believe there are FDA comments about that, then perhaps when we get to that question, we'll ask for the comments.

DR. TEMPLE: Well, you've seen some of them.

DR. BORER: Yes.

DR. TEMPLE: There's a fair amount of skepticism probably noted.

DR. THROCKMORTON: It's important for us to hear some comment around that I believe not perhaps specifics that this one thing you think works or whatever it is, but what you said, Jeff, was perhaps it reduces the risk or changes the risk somehow. We need to understand whether there's sort of a belief that that is possible or is not. Again, not necessarily that you know the exact right answer but that you are optimistic about those things as possible would be something we need some help on.

DR. BORER: You'll hear it.

Why don't we then begin. What I'd like to do, I'll read the preamble here, and then as we go through these questions, I'd like to hear Tom's response first, Tom Pickering because he is a nonvoting member, and then we'll go to Steve who's our committee reviewer and then to Tom Fleming and then the rest of us.

The committee is asked to provide an opinion on the approvability of omapatrilat for hypertension. Omapatrilat is an inhibitor of angiotensin-converting enzyme and neutral endopeptidase. Reviews of chemistry, pharmacology, toxicology, and biopharmaceutics present no apparent barriers to approval. Omapatrilat clearly lowers blood pressure.

During its initial development, an increased risk of life-threatening angioedema was noted for patients taking omapatrilat compared with other antihypertensives, including ACE inhibitors.

To characterize this safety finding and to gain additional information on the relative antihypertensive efficacy of omapatrilat, the sponsor conducted the OCTAVE trial.

OCTAVE was a randomized, double-blind study in which 25,302 subjects with hypertension were randomized to once-daily enalapril or omapatrilat and followed for 24 weeks. During the first 8 weeks, subjects were titrated to a maximum dose of 40 milligrams enalapril or 80 milligrams of omapatrilat as needed, after which subjects who did not achieve the blood pressure goal could have additional antihypertensive agents added through week 24. At 8 weeks, 41 percent of subjects in the enalapril group and 33 percent in the omapatrilat group were on the highest recommended doses. Between weeks 8 and 24, 19 to 36 percent of the enalapril subjects and 13 to 26 percent of the omapatrilat subjects added antihypertensive therapies. At 8 and 24 weeks, omapatrilat had a significantly greater effect to lower trough blood pressure compared with enalapril, but angioedema, including serious angioedema, was significantly more common in subjects taking omapatrilat. And we have a table outlining those data.

With these results and the data from the other trials of omapatrilat, the committee is being asked to characterize the risks of omapatrilat, to identify and characterize the benefit to which this risk needs to be compared, and to discuss whether omapatrilat's benefits outweigh its risks.

So, we'll begin. How should one best characterize the risk of angioedema with omapatrilat? 1.1. Are the clinical features of angioedema associated with omapatrilat similar to those associated with approved ACE inhibitors? Tom.

DR. PICKERING: Well, I think we've heard quite convincingly that the clinical features are generally similar, although with omapatrilat, the extent of the angioedema is more likely to be severe.

DR. BORER: Is it not true that the angioedema tends to occur earlier also with omapatrilat than with the ACE inhibitor?

DR. PICKERING: Yes, since most of the episodes did occur during the first day.

DR. BORER: Are there any other comments about the characterization? Steve.

DR. NISSEN: I want to emphasize that there's a pretty steep gradient here compared to enalapril in terms of the risk ratio for mild, moderate, and severe. So, it really looks like there's a significant shift from the more mild forms to the more severe forms. So, it's not just a quantitative measure. It's really also I think a qualitative measure which I think you were saying also, Tom, that there is disproportionately more severe angioedema with omapatrilat.

DR. BORER: Any other opinions?

(No response.)

DR. BORER: Then let's go to 1.2. In the original development program, about twice as many subjects were exposed to omapatrilat 20 milligrams than to 10 milligrams as an initial dose, and the rate of any angioedema was about three-fold higher in subjects initially receiving 20 milligrams. OCTAVE's primary safety hypothesis was that starting omapatrilat at a low dose and titrating up would reduce the risk of angioedema of any severity to no more than twice that of enalapril. Was this hypothesis supported by the study?

I don't think this needs much discussion. It wasn't. Is there any dissent from that?

(No response.)

DR. BORER: No.

1.3. In OCTAVE, there were two cases of life-threatening angioedema among 12,000 subjects treated for about 6 months. In the original development program, there four such cases in a population about one-third as large. Estimate the risk of life-threatening angioedema to expect post-marketing and estimate the upper confidence limit for that risk.

I think in fairness maybe we better begin with Tom Fleming for that one, and then go back to Steve and Tom.

DR. FLEMING: Well, let me respond to this and add a little bit of response that relates to question number 1.2.

The sponsor has provided us the estimate with these two cases that the upper limit of the confidence interval is 5.7 per 10,000.

Let me just add to this answer to question 1.2 that as the question indicated, there certainly was evidence before OCTAVE that those patients at 10 would have had a lower rate of angioedema than 20. Roughly, it appears 1.1 percent as opposed to the 2.4 percent. So, hence the design of the trial to rule out that the rate could be as high as 2 percent, hoping it's around that 1.1 percent, and obviously, disappointingly the rate was at 3.2 percent. So, as you said, Jeff, the answer to 1.2 in that sense is no, although I'd go on and say when one looks at these life-threatening cases, the rate appears to be by estimate 10-fold higher in those in the historical experience who had received 20 as a starting dose.

So, the bottom line is, as I see it, the rates that had been hopefully reduced to levels below the 2 certainly were not. The rate was 3, but it might have that there was systematic under-detection in the previous experience. My own read of this is that there probably is a dose response, and again the best measure of that would be looking at the fact that we see approximately an estimate of 2 with an upper limit of 5.7 per 100,000 life-threatening cases in the OCTAVE trial in contrast to a 10-fold higher rate than that in the previous experience at 20.

DR. BORER: How about the 1.3? I'm sorry.

DR. FLEMING: I started off by saying the estimate that I accept that was given by the sponsor that the upper limit of the 95 percent confidence interval is 5.7 or about 5 to 6 cases per 10,000.

DR. BORER: Steve.

DR. NISSEN: Well, I think that, of course, Tom is right statistically, but I think there are other factors that we have to think about here. Let me see if I can make this very clear.

I'm concerned that when the drug is administered in the community outside of a setting of a clinical trial, the rigor, the discipline of giving 10 and waiting 2 hours, and then waiting 2 weeks before up-titrating, that we may lose some of that discipline in administration. So, I would tend to raise that estimate somewhat because I'm not convinced that in general use you achieve the discipline that you do in a clinical trial. So, I think I've got to make some upward revision of that estimate based upon the fact that in clinical use recognition may be a little bit less because people may be further from tertiary care centers, and therefore the risk of a life-threatening event, which is what you're asking us, Bob and Doug, I think could be a lot higher because I don't believe that this plan can be as tight as it was in the clinical trial.

DR. BORER: Blase.

DR. CARABELLO: I see the potential actually for the opposite of that, that in post-marketing that we don't release this to general use. I agree with Steve. I think that's looking for trouble. The risk of an angioplasty is not predicated so much upon the equipment or the atherosclerotic lesion, but rather the judgment of the angioplaster. If you have an experienced person with a lot of judgment, his or her complications are usually less than someone who doesn't do it very well. I think if you limit the use of this to people who have, because of the kind of practice they're engaged in, extraordinary judgment, the complication rate actually could be less.

DR. BORER: Yes. The problem is making that limitation. But let me ask a question and you can comment on that and whatever else, Doug. Dofetalide. In order to use dofetalide, it's necessary to go through an educational program and then be approved for use. So, I guess the precedent exists. I assume we could do that.

DR. THROCKMORTON: Yes. That's probably not a precedent, unfortunately, that we've had a broad amount of success with. We understand there has been movement towards other pharmaceuticals perhaps in some areas because people have been reluctant to use that.

DR. TEMPLE: But that is because there's an alternative. You can get sotalol without it, unfortunately, for the same use.

DR. THROCKMORTON: Right.

DR. TEMPLE: But here that might not be true.

DR. CARABELLO: But going back to amiodarone or adriamycin, we have plenty of drugs that, even if the agency doesn't restrict them, people restrict their own use because they're scared as hell to use the drugs. A generalist is not going to prescribe adriamycin to the next person who walks into his clinic.

DR. THROCKMORTON: Right, and we of course don't restrict that at all.

And there is precedent for restricted distribution, which is I gather what you're sort of talking about. Normally it's drugs that have extraordinary toxicities that the Office of Drug Safety people are convinced can be managed using those kinds of restricted distributions. But it's very hard. And for an antihypertensive where we have several dozens of alternatives, it seems like you'd need to be able to make a clear case for doing that. That would be a hard thing, maybe not an impossible thing, but not straightforward.

DR. NISSEN: Just a comment, though. The analogy here of, say, the angioplasty population, Blase, is not in my view a good one, and I'll tell you why. Somebody doing an angioplasty looks at the lesion and all kinds of characteristics and can kind of profile the patient. The problem is other than skin color, I can't look at a patient and know who's going to have it and who's not. So, I could be the world's greatest expert in hypertension, but I'm not sure that I can pick the patient out who's going to have this side effect. So, it's not quite the same as deciding who you're going to put a stint into.

DR. CARABELLO: Yes, but you can, as an expert in hypertension, pick out the group of patients for whom the benefit is the most, at least alter the risk-benefit ratio that way and say, okay, yes, you may get angioedema, but in your case my judgment is you're the patient who's likely to benefit from this drug. People like Dr. Black and others might well want to have this drug in their armamentarium when the other stuff doesn't work.

DR. TEMPLE: We're developing numerous models for risk management. You're familiar with one of them, bosentan, Tracleer, where there's a central distribution system, and it is shipped directly to the patient from a single place, which allows you to assure that people get the information which could include a video or a lot of different things. That's relatively extreme, manageable with a relatively small population, somewhat intrusive, but in that case we thought it was worth it. And there is a wide range of others. Anne or Julie could probably tell you some of the details.

One that might be considered highly relevant would be what we did with Lotrinex, which has risks of somewhat the same order of magnitude and severity, but we thought it was appropriate for a well-described population of people who were made miserable by their irritable bowel syndrome. That's no so different from people who are put at very high risk because their blood pressure is uncontrollable. So, perhaps a description of that.

But in fact, we'd like to hear what you suggest. Obviously, shipping directly from a central pharmacy is burdensome in one sense but worth it if the risk is bad enough and if you really want to be sure all of the people involved get educated. As an example, if you wanted to be absolutely sure that patients knew what angioedema was like, well, the only way to make really sure is to tell them, enable them to ask questions, show some pictures of it, or something like that. So, I'm skeptical of whether your local pharmacist will be able to do this in a reliable way, but there might be other people who could do that.

So, there's a very wide range of possibilities. We're happy to hear any suggestions you have, and if that was considered sufficient to make this available, we would work with our own staff and them to figure out what that is. They've ranged from giving good advice, having a patient insert, putting a black boxed warning, all the way to specialized distribution systems.

DR. BORER: Well, that's exactly the point we're up to here, 1.4. I'm sorry. Tom.

DR. FLEMING: I was just going to briefly say, as we're leaving 1.3, I accept Steve's comments as amending my response. My response is based on what this trial shows, and I understand clearly your concern and endorse that concern that this, in fact, in clinical practice could be worse. Blase makes the relevant point that we may well have a differential benefit-to-risk, but other than blacks and smokers we don't know how to prevent this. So, I accept your point, Steve.

DR. BORER: 1.4 requests some opinion about the risk management plan. The sponsor has proposed a risk management plan focusing on patient education by pharmacists. To what extent can a risk management program based on patient education be expected to reduce the risk of death from angioedema?

Why don't we begin this time with Steve, and then we'll go to Tom.

DR. NISSEN: Well, I think it's a start and I think it certainly does help. The question is how large a magnitude of reduction in risk will we get. It's not going to prevent angioedema. What it might do is get people to seek treatment earlier, and I'm convinced by the arguments from Dr. Kaplan that that could make a difference.

I actually think to really reduce risk, a much more comprehensive program of risk management will be required. I was a little surprised there wasn't much more here for physician education because, in fact, I must tell you that I didn't understand the subtleties of differences in what drugs work and what drugs don't work and so on. So, in my view we would have to get information to most emergency department physicians. We'd have to make sure they understood about omapatrilat, they knew what to do, how to do it, and so on. I didn't see that in here. So, I think the effect here of this is likely to be modest rather than more than that.

DR. BORER: Tom.

DR. PICKERING: Yes, I generally agree. I think it's a great idea, but my impression is that brief pharmacy education programs are not in general terribly effective. Again, I guess I would be concerned about what happens in a sort of busy pharmacy in Harlem where there's a long line of patients waiting for their prescriptions, how much time there would actually be to do this. I think it's an untested possibility.

DR. BORER: Mike, do you have any thoughts about the education program, the risk management program?

DR. ARTMAN: Yes, I agree with what Tom just said. I think, in a practical sense, it's impossible to say that this is going to be useful or not. It's interesting talking about where the trials were done. I assume if this is going to be used overseas, we heard from Dr. Black how he had an experience in a remote emergency room in Connecticut and everything came out of fine, which he said was like being in Russia. But I wonder what it's really like in Russia. So, I have little confidence that this risk management plan is going to do anything.

Now, that said, the sponsor also sort of made the case that it really doesn't matter because these are obvious signs and symptoms. The patients recognize them. They come on slowly. So, it's six of one, half a dozen of the other. I don't know which side of the fence they're sitting on.

DR. WACLAWSKI: Dr. Borer?

DR. BORER: One second. Once we start this part of the session, we don't really want input unless we ask for it.

Susanna.

DR. CUNNINGHAM: Is there any evidence that actually a pharmacist-based program can do risk reduction? It's a very theoretical thing and I think, really, we've already heard how it probably is quite impractical. So, I would be interested if somebody has evidence that it would work.

DR. BORER: Is there any experience that we can refer to?

DR. TEMPLE: Julie or Anne may want to comment, but I'm sure they'll tell you the answer is probably not. But come on. Don't let me speak for you.

DR. TRONTELLE: I'm Anne Trontelle from the Division of Surveillance Research and Communication Support in the CDER Office of Drug Safety.

We do have some evidence. It's mostly known at this point to be published by Duke Center for Education and Research in Therapeutics on the dofetalide program which involves not only education of pharmacists, but also of practitioners and specially staged introduction of that drug. I think there's some suggestion of improved efficacy over comparable drugs, but that's again a highly specialized program and one that has probably resulted in substantial voluntary restriction on use of the product perhaps because that program has been perceived to be burdensome. We really are lacking data from any of the other programs at this point.

DR. TEMPLE: But a crucial distinction. You start dofetalide in the hospital or some ambulatory equivalent.

DR. TRONTELLE: That's correct.

DR. TEMPLE: So, it's not your local CVS pharmacist who's responsible for this. It's the hospital pharmacist. Well, no. That was no offense. It's not your local pharmacist. It's a highly specialized group of people and there's a lot of exchange of information. But do we believe that a complicated information system will be handled by local pharmacies? I'm not aware of any information, and as a profound user of many drugs, I think that's very unlikely.

(Laughter.)

DR. TRONTELLE: I think it's hoped for but we don't have data at this point.

DR. TEMPLE: Actually we know that patient package inserts are not well handed out by local pharmacies. That's a not a very hard thing to do. But we even know that. That seems unlikely.

DR. ARMSTRONG: Maybe I'll speak to that point and also give an opinion on this issue on the table.

There is data in western Canada on pharmacists and cholesterol lowering and adherence to medication which engages pharmacists and actually demonstrates enhanced adherence to the guidelines. So, there is some data.

On the point about the education program on the table, notwithstanding the great intentions and efforts of the sponsor, my worry would be that the infrequency of the event, even with the best efforts of educating the physicians and the patients would lead to desensitization, and I use that term advisedly in relationship to recognition of the event as time elapses.

DR. NISSEN: Would the pharmacist be reimbursed for this activity, or would this be sort of gratis?

DR. THROCKMORTON: Yes. It's important to note and we may want to ask the sponsor to comment on it. The program they've proposed, the outline of the program -- again, remember, the specifics would be worked out were you to give us some level of comfort that it was possible. The program that they've outlined involves a central pharmacist who would interact with the patients initially and I wouldn't want to guess how that person would make their living. The sponsor might want to comment a little. Is that enough, Elliott, that there's a central pharmacist and then a follow-up with the dispensing pharmacist?

DR. LEVY: Well, there's no intention to make the local pharmacist the focus for patient education. We would use pharmacists in a program that would require that every patient receive counseling before they went to the local pharmacy and got the drug. That, of course, is in addition to physician counseling and to a host of messages that are provided them through the packaging and educational materials.

DR. CARABELLO: I personally don't think we'll have too much success in mitigating risk, although patient education to let them know what angioedema is would clearly be valuable. Going back to my original plan, I think our best bet is trying to improve the risk-benefit ratio by focusing on the prescriber so that whatever the risk is it's only matched by the higher benefit by prescribing this to some specific patients.

Finally, I'd like to distance myself from any group that would compare both South Dakota and Connecticut to Russia. I think that's a very dangerous thing to do.

DR. BORER: Because you have asked for a lot of opinions about this, I want to say a couple of things here in addition to what's already been said.

My perception of the risk has been modified in an important way by Dr. Kaplan's comments and the fact that there are perhaps several hours off usually before major sequelae of angioedema might be expected to occur. On the other hand, that means you have to be within a couple of hours of help and you have to be aware that you need help.

I share the concern that has been raised by a couple of people on the committee that a patient education program just isn't going to cut it, or by itself it won't. It's an important component if one could do this at all, but the physician education is crucial. I don't know how one could achieve adequate physician education without limited distribution, which is a tall order, as we've heard. So, I have real concern that a practical program could be developed.

But I do not believe that a risk management plan focusing on the patients and the pharmacists would be sufficient to deal with the risk if we perceived the risk to be an important risk relative to the benefit and if -- on and on and on. Again, I think we still have to discuss what the risk is in view of the information we received today, but I don't think this kind of a plan would do it.

Was there another comment?

DR. THROCKMORTON: Susanna, do you have a comment at all?

DR. CUNNINGHAM: No.

DR. BORER: Tom, I'm sorry. I didn't ask you specifically for a comment about this.

The sponsor has shown the results of OVERTURE --

DR. TEMPLE: Jeffrey, before we leave that, could we hear other people's views of that last point? I mean, that's not unimportant, after all. You are skeptical that anything but a limited distribution system which allows you to interact clearly with the physician -- I assume that's the point --

DR. BORER: That's right.

DR. TEMPLE: -- is likely to do anything. Is that everybody's view? That's potentially very critical depending on your answer to the rest of it.

DR. BORER: Well, let's hear.

DR. NISSEN: I also concur.

DR. CARABELLO: I think I made my own point clear.

DR. CUNNINGHAM: I agree.

DR. ARTMAN: Well, I agree but yet even that -- there are two issues. One is controlling the use of the drug and restricting its use, and the other is managing this adverse event and that we can't really predict who's going to get that. There are some that are at little bit higher risk than others, but anybody who's on this drug for any given time at any dosage may develop significant angioedema. So, yes, I agree with that that you can maybe minimize it by minimizing the number of people you expose, but I think it's a whole different issue to say that that's going to reduce the risk. I don't think it will.

DR. TEMPLE: The theory I think -- Blase has said this several times -- is that you make the risk acceptable because the benefit is particularly good. That's certainly the theory of Lotrinex. If you only give it to people who are willing to accept and understand all the various risks, then the benefits might outweigh the risk for that portion of the population, whereas if you just gave it to everybody with IBS, you wouldn't feel that.

DR. ARTMAN: Well, let's extend that logic that Blase used to amiodarone and angioplasty. There are knuckleheads out there using amiodarone and angioplasty. And some of these things started out being sort of restricted. So, once the cat is out of the bag, all bets are off.

DR. NISSEN: There's one other part of the equation again I think that's very important to understand, and that is that the suggestion here is that we can optimize risk-to-benefit ratio by selecting patients that are most likely to benefit. Tom Fleming tried to drill down very hard to find that group that's most likely to benefit and to see whether or not there's very good evidence here that lets us identify such a group and then to estimate the magnitude of benefit.

The problem I have is I'm not quite sure how to drill down to that group that's most likely to benefit, and I'm certainly not sure how to estimate the magnitude of benefit of a 3 over 2 millimeter blood pressure difference in a new class of drugs. So, this is all predicated on the assumption that we know that and therefore, by picking these high risk individuals, we can somehow optimize overall benefit, and I just don't think the data gives us that information.

DR. BORER: Let's go on to question 2. The sponsor has shown the results of OVERTURE, a comparison of omapatrilat and enalapril in the treatment of chronic heart failure. If the results of this study are as presented, how relevant are these data to the approval of omapatrilat for hypertension?

Tom and then Steve.

DR. PICKERING: Well, I certainly think the results are relevant because I guess we were hoping there was some evidence of additional benefit on outcomes from omapatrilat above those of enalapril. But my own interpretation is that the study was negative. In terms of the blood pressure reduction, there doesn't seem to be any convincing evidence that there was a difference using the b.i.d. dose. On the other hand, I guess you could argue that the doses weren't necessarily the maximal that are used for blood pressure control, and it wasn't designed as a blood pressure study. So, the bottom line is I don't think it adds a whole lot of support for the case.

DR. BORER: Go on to 2.2. How reassuring are these data with respect to the use of omapatrilat in a hypertensive population? I think you've just answered that, but do you want to elaborate a little bit? I think you've answered it.

DR. PICKERING: Yes.

DR. BORER: Steve.

DR. NISSEN: I guess I think the OVERTURE data are minimally relevant. I'm glad we have them, and I agree with Milt Packer's suggestion that it does tend to make us believe that there's not being harm done by this drug with respect to that particular group of patients. But because it is a different group of patients, it's hard to interpret it.

I think I'm convinced that probably there was a blood pressure difference in OVERTURE, particularly given the greater incidence of hypotension. But I think the amount of weight I would put on OVERTURE in our decision is very, very small, and I wouldn't urge us to consider it very much in our overall decision making.

DR. TEMPLE: Let me ask a specific question. You've raised the issue several times now that you don't know what a new drug that hasn't got outcome data does. You don't know whether the expected benefit of a 3 millimeter of mercury systolic pressure difference -- okay.

Well, there are a number of possible reasons for that, but one of them is maybe it's like a high-dose diuretic and it kills you in some way.

One might have argued -- I'm not trying to put words, but we need to understand this because, without understanding, we might go off and say the wrong thing. One might think that OVERTURE is somewhat reassuring on that point. You have a fragile population susceptible to getting MI, sudden death, all those things, and that didn't happen to them. So, why doesn't that, to some extent, contribute to your answer on that?

DR. BORER: Let me respond to that. It does for me. I don't think that the OVERTURE data are the panacea to respond to our concerns, but it's hard not to begin to develop the belief that there isn't a smoking gun in terms of other cardiovascular toxicity out there. When you look at the OVERTURE data, as Tom said, it's not an antihypertensive study and the doses were half the maximal doses that were given in OCTAVE and half the maximal doses that would be used in clinical practice, though they were relevant for a heart failure study. But remember, of the people with heart failure, more than half had coronary disease as the basis, and we didn't see a jump in myocardial infarction, sudden death, and whatever.

So, I think these are useful data. They're not dispositive, but I think they're useful data in making us focus in on the safety concern that we have to balance against the benefits, specifically the angioedema. It wipes away some of the potential peripheral noise.

DR. NISSEN: Bob, I wanted to also respond directly. Let me see if I can be clear about this.

I agree with Jeffrey and I agree with Milton that it does, in fact, help us in the belief that there is no harm being done by the drug. But I was speaking to the question of what inference can we place on that 3 over 2 millimeter blood pressure decrease upon the likelihood that that will produce an incremental benefit on events, and I can't tease that out of OVERTURE. I still don't know yet whether the better blood pressure reduction that we get with omapatrilat will, in fact, ultimately translate into a reduction of events.

DR. TEMPLE: You obviously have no direct information on that. And Henry may be invited to talk to this too. The people who do meta-analyses and look at various drugs generally reached the conclusion that, if you like, it's the blood pressure, stupid, and that a fallen blood pressure, unless it's balanced by something toxic or some failure to treat a component, maybe like doxazocin, will have the expected effect. That's not proof, but I'm curious to know where your skepticism comes from. Do you doubt that or where does it come from?

DR. NISSEN: I do doubt it, and I doubt it because of some studies. I doubt it because LIFE is a study where similar blood pressure reductions produced different effects on events. Looking at all the different events, I'm pretty well convinced, for example, that the same amount of blood pressure lowering with a calcium channel blocker may lower stroke to a greater extent than, say, ACE inhibitors, whereas lowering with ACE inhibitors may do a better job of preventing heart failure. I think there are lots of clinical trial data that suggest it is not just the blood pressure, stupid.

I think that given that fact and given the fact that we've got an entirely new drug in a new class, I cannot estimate -- and OVERTURE doesn't allow me to estimate -- the magnitude of benefit of a 3 over 2 blood pressure increase on the long-term morbidity and mortality in this population. It was even a different population from OVERTURE.

DR. BORER: Are there any other comments from the committee about OVERTURE? Susanna. I'm sorry. Tom.

DR. FLEMING: I find OVERTURE very relevant to the setting in which it was conducted. In this chronic heart failure setting, we see a 6 percent reduction in these primary endpoints, and in my sense, a nonsignificant modest difference here is not an adequate difference when you're looking at angioedema risks as part of the overall spectrum of side effects.

To get to the question, though, how relevant is this to the antihypertensive setting? As I understand it, in essence the way it's supposed to be relevant is we -- and I'm going to oversimplify the world probably, but omapatrilat, let's say, is an inhibitor of ACE and NEP, and let's say that it's the inhibition of NEP that's the causal mechanism by which we're achieving a higher level of blood pressure control. So, what I really want to do is be able to be reassured that the only effect that that inhibition of NEP has is mediated through this blood pressure control. That's essentially, I assume, the reassurance I'm supposed to get out of this study. And yet, I don't see the magnitude of blood pressure difference in this setting that I have in the antihypertensive setting.

I struggle to see the logic behind how I'm going to be able to then use this as a way to conclude that in the antihypertensive setting, that these data provide me compelling evidence or reliable or even useful insights that I'm not going to have unintended effects on clinical endpoints which, in fact, wouldn't have to occur in the magnitude that they would occur in a chronic heart failure setting to be important on relative risk. Events are much more rare in an antihypertensive setting. So, if something doesn't show up in a setting where events are frequent, it doesn't mean that there isn't a signal there that's being lost in the background of a lot of other naturally occurring events that in the antihypertensive setting would be important. So, I don't see how this is informative for our setting.

DR. BORER: May I just ask you to elaborate a little further, Tom? I wonder whether the data from OVERTURE have to be understood in context of the data from OCTAVE about cardiovascular events. There was no significant reduction in cardiovascular events, but they went in the right direction and we didn't see a problem with excessive cardiovascular events in this very high risk population of OVERTURE. Taken together, do they or do they not give you some confidence that we don't have to be overly concerned?

DR. FLEMING: There's not nearly enough known to draw that conclusion. One could argue -- and maybe it's entirely wrong -- that what small trends that are there are, in fact, due to the intended mechanism that is the added blood pressure control.

What I want to be reassured about is the point Steve has articulated on several occasions, and that is if we now have a new agent that not only is an inhibitor of ACE but NEP and it may be through that mechanism that we get this additional 3 millimeter blood pressure control, that I can now conclude I'm going to achieve the full benefit in reduction of clinical events, and there won't be any other unintended effects on cardiovascular events. I can't glean that from the OVERTURE data.

DR. TEMPLE: But, Jeff, this is really critical. If the committee as a whole doesn't believe that lowering the blood pressure more, even if that were well documented, is of any value, then we can stop now.

DR. BORER: But I'm not sure we're at that point.

DR. FLEMING: We're not at that point.

DR. BORER: We'll answer that question.

Blase.

DR. CARABELLO: Just to make the point that in OVERTURE, I think it is somewhat reassuring that in what is surely a very sick group of patients, we didn't see increased cardiovascular events.

But the other point I was going to make is that in some respects this is comparing apples with freight trains. Remember, cardiologists only have to remember two things at once. In this case it's total peripheral resistance and cardiac output. And the two things supporting blood pressure and heart failure are so vastly different with every system known to man revved up and screaming at one another. Whatever the difference between that is and in essential hypertension I don't know, but I suspect that we're talking about two very different pathophysiologic settings.

DR. BORER: Paul.

DR. ARMSTRONG: Just to respond to Bob's point, I'm not sure that the prevention of the degradation of AMP, adrenomedullin, and bradykinin long term aren't harmful. So, I'm not prepared to accept blood pressure lowering with this agent as a likely mechanism for the prevention of long-term cardiovascular morbidity and mortality.

DR. BORER: Bob, has pointed out that we do have to deal with this issue as to whether we accept blood pressure lowering as a surrogate. I think that what's developing from this discussion may not be that everybody wants to tell the FDA to junk the surrogate -- or maybe they do and we'll ask specifically -- but that when the surrogate is achieved by the use of a new agent that acts by a different mechanism, are the risks associated with the use of that new agent sufficiently modest so that even though there may be mechanism-specific differences in the magnitude of benefit from a given degree of blood pressure lowering, we can assume that the blood pressure lowering causes a benefit sufficient to overcome those risks that we don't understand so well. That's sort of a complicated statement, but I think you get the idea.

DR. TEMPLE: But it is fundamentally untestable. You can't use this in a 10,000-patient study because it wouldn't be even ethical to even give those people the drug.

DR. BORER: Right.

DR. TEMPLE: Just remember.

DR. BORER: Just to answer your question, is there anyone here at the table who wants to tell the FDA today that we just cannot accept the blood pressure lowering as a surrogate anymore?

DR. THROCKMORTON: In this case, you must say -- since yesterday you said you could do that for drugs within the same class. Specifically you must now be saying comparing drugs not within the same class.

DR. FLEMING: Jeff, we don't really want to make a blanket statement. Right? I mean, I think as has been articulated by many people, it depends on the circumstance. There are an awful lot of surrogates that I wouldn't put much stock in at all. This one is one that stands out among the few that really has some considerable credibility, and yet you don't blanketly apply it. And there are certain settings, for example, within drugs within the same class where there are no detected concerns in safety where you're going to be more confident in relying upon it than in other settings where you have different drugs in different classes or, in particular, as is in this case -- I'm not saying you wouldn't give it some credence, but there is a higher bar that you have to hit when you have to overcome an important significant side effect.

DR. BORER: I think that that statement probably stands for the committee here. The committee isn't suggesting that the surrogate has to be scrapped. We don't have the database to be able to suggest such a thing. But with this particular agent, there is information that indicates a risk that's higher than we might have expected for some new antihypertensive drug, and now we have to explore whether the risk or other risks that we haven't quite fully fathomed yet outweigh the putative benefits of the blood pressure lowering. And that's what we're sort of grappling with here.

DR. TEMPLE: I think I'm hearing that you think under these circumstances, it's sufficiently uncertain as to whether there is a benefit, that there cannot be anything that outweighs the well-known risk. The risk is documented. That can't go away. You can manage it, but you can't make it go away. But in this case, you can't know with enough certainty that there's a benefit of a 3 millimeter of mercury difference, so that there is really no presumed benefit from that outcome.

DR. BORER: I don't think that's the consensus here yet.

DR. TEMPLE: I'm being provocative. I want to hear what you do think.

DR. NISSEN: I know you are and I really want to try to directly answer that a little bit.

The reason that I think Tom Fleming said that it depends is because it depends on the magnitude of the difference in blood pressure and the magnitude of the risk. So, if you give me a drug that has no defined risks above that of comparators and has a fairly robust and substantial blood pressure advantage over a 24-hour period of time, we're going to probably be just fine. And we did that yesterday. We took a few hundred patient trials, a couple of trials, and we said a drug in the same class with no special risks that has a couple of millimeters better blood pressure reduction is superior.

So, if this drug could produce an 8 or 10 millimeter increase, very, very robust differences, that would shift the equation a little bit. And if the risks here were a bit smaller, if it were only a twofold increase in the risk of angioedema, not a threefold, and there weren't these racial issues.

And so, the reason it's context that makes a difference here is we have to as clinicians and you have to as an agency balance the magnitude of the benefit with the magnitude of the risk. What I think we're saying is that for a 3 millimeter over 2 millimeter blood pressure difference, we know that there are interclass differences in effect on events, and those could potentially overwhelm that 3 over 2 benefit, particularly in the context where safety is a problem.

DR. TEMPLE: I just want to ask one other thing. To my best knowledge, no placebo-controlled trial of any class of drugs, which includes calcium channel blockers, reserpine, hydralazine, even high-dose diuretics which are lethal, has failed to show a favorable effect on stroke in other matters. That doesn't mean there can't be differences between the classes. There can.

But I would have said the general observation of lowering blood pressure, barring some bizarre thing like causing arrhythmias, is always good for you was fairly well established. You're absolutely right. That doesn't mean there can't be interclass differences. But you don't think that's necessarily good enough because you can't really quantify it.

DR. NISSEN: Well, it's true, true, and unrelated. I mean, the fact that drug X is better than placebo, because it lowers blood pressure, isn't the same as saying that drug X is better than drug Y because it lowers blood pressure by a little bit more. I think that's the problem. The problem is we have shown interclass differences. So, it's true that every drug that lowers blood pressure has been better than placebo. I think that's right.

DR. TEMPLE: So, across class, even something that was a little bit better at lowering blood pressure might not be better because of other factors.

DR. NISSEN: And that's what ALLHAT is testing in an enormous population. If ALLHAT is a wash, then okay. But I would make you a prediction that different endpoints in ALLHAT may go in different directions based upon which agent you use independent of blood pressure.

DR. BORER: Today, however, we have to make our judgment based on what we know. Of course, ALLHAT isn't available yet.

And I want to hear Tom's comment about this, but I think to put it in a slightly different context, I'm willing to accept that 3 over 2 is good. If omapatrilat is what it takes to get there, that's a good thing. The issue is do the benefits outweigh the putative risks. At the end of the day, we're going to have to come to a qualitative judgment of that because there is absolutely no way we can quantify these things. And we're going to get there, I promise. I don't know what that judgment will be, but that's what we're going to have to do and that's what we're sort of moving towards.

Tom.

DR. PICKERING: Yes. I'm one of those who says that a 3 over 2 reduction in blood pressure is extremely important clinically for the reduction of risk.

I'd like to sort of clarify what we're talking about in terms of risk here. Maybe angioedema is one that's clearly defined, but I think I'm sort of hearing insinuations that there may be other risks that we really don't know about, and if so, I think that's unfair to the sponsors. What we should be judging is the blood pressure and the known risks at this stage.

DR. BORER: Yes, I agree with that statement to the extent that we have looked at the adverse event profile for this 25,000-patient study plus the OVERTURE data for a different population. That's true. If anything, it would tend to reassure one that bad things aren't happening. I think Tom's point, which is absolutely right, is it also doesn't tell you you're clearly benefiting in terms of event reduction from the blood pressure lowering. On the other hand, that's not what the trial was designed to do, but it does make a pretty reasonable case that there's not a smoking gun out there that some horrible thing is going to happen besides the angioedema, the relative importance of which we're going to have to judge at the end of the day.

Why don't we try. Unless anybody has anything else to say about 2, we'll move on to 3, which is fairly specific. Consider the antihypertensive effects of omapatrilat relative to other drugs. 3.1. Is omapatrilat superior to enalapril? What results support this?

Tom, can you give us an opinion about that?

DR. PICKERING: Yes. I would say the answer is yes and I accept that these studies against twice-daily enalapril might have reduced the superiority a little, but I would expect it would still be there.

As I said earlier, I'm sort of disappointed that there aren't head-to-head studies between omapatrilat and enalapril or lisinopril plus a diuretic. I would very much like to see what those data would show. I know there's an additive effect when you add omapatrilat to a diuretic, but I think the head-to-head studies would have still been helpful.

DR. BORER: Steve, what do you think about omapatrilat versus enalapril?

DR. NISSEN: I'm convinced. If you show it in a 25,000-patient trial of this strength, the evidence is just overwhelming that it is superior at lowering blood pressure to enalapril. Not controversial.

DR. BORER: Tom, do you have any concerns about that? No. Anybody else?

(No response.)

DR. BORER: So, we're willing to accept the answer to 3.1 as being yes.

How about 3.2? Steve, why don't you start.

DR. NISSEN: Two adequately controlled trials against lisinopril in reviewing Dr. Throckmorton's material -- and I think they were well done, and there's also ambulatory blood pressure data. So, I think that in fact there is adequate evidence of superiority to lisinopril.

DR. BORER: Tom.

DR. PICKERING: I agree.

DR. FLEMING: Well, I just have a comment, additional thought that will apply to 3.2, 3.3, and 3.4, and that is I certainly agree the data are there to establish a superior antihypertensive effect. All of these studies, though, that will be relevant for 3.2, 3.3, and 3.4 had starting doses of 20. So, these superior antihypertensive effects were established in settings where, from a point estimate perspective, the angioedema rates were maybe an order of magnitude higher than what we see in OCTAVE, although I suppose it could reasonably be presumed that had these studies also been done with the lower starting dose, that they still would have yielded comparable improvements in antihypertensive control or effects.

DR. BORER: So, you've extended now to 3.3 and 3.4 and accepted omapatrilat as superior to the other drugs as well.

Tom Pickering, would you agree with that?

DR. PICKERING: Yes.

DR. BORER: And Steve?

DR. NISSEN: I do not. I reviewed this pretty carefully. Let's take 3.3 first. There were two trials against amlodipine, one of which was positive with a delta of minus 2.1, and one of which showed a delta of minus .3 and a p value of .6. So, if you say it takes two trials, the two trial rule was not achieved against amlodipine. Now, both trials were adequately done, but I don't think there are two positive trials. Doug, correct me if I'm wrong, but when I read your review here, the study CV137-032, your review said, failed to detect a significant difference between omapatrilat and amlodipine, and the delta was minus 0.3, p equals .617. So, I would say not proven.

And for 3.4, I don't think there's adequate data against losartan. I think the two trial rule requiring well-performed trials simply isn't there. But correct me if I'm wrong, Doug. You've reviewed this.

DR. THROCKMORTON: Review the losartan data, maybe.

DR. NISSEN: You'll have to point me to the --

DR. BORER: While you're looking for --

DR. NISSEN: I think the only trial I saw was the LVH study, but maybe there's something I don't know about. I don't think that was adequate.

DR. BORER: While you're looking for them, though, remember that the replicability of effect principle is really an approvability principle. We're talking here about whether we have data that would convince us rather that for approval purposes for moving forward with an opinion about whether this drug adds something that the drug actually was more effective than losartan or amlodipine, in which case it might be reasonable. I'm not suggesting you should do it, but it might be reasonable to add the data together and look at the average. Both trials went the same way, for example. I'm not suggesting that that should be the opinion, but one might look at it that way.

DR. NISSEN: Let me tell you why it's relevant.

DR. THROCKMORTON: Let's ask the sponsor to just briefly review the losartan.

But just to comment, the general notion here was, is there any superiority that you discerned for any comparative antihypertensives? Where you define none, then we're done.

DR. LEVY: Just two points. First of all, in addition to the 38 study, there were two adequate and well-controlled trials versus losartan, one of which is shown here and the second of which is shown on the next slide, which was an ambulatory blood pressure comparison shown on the right.

I'd just like to briefly comment on the -32 study. Dr. Throckmorton's comment on the office trough diastolic blood pressure results are correct. This was an ambulatory blood pressure study, powered for ambulatory blood pressure. Primary outcome measure, ambulatory mean blood pressure, which was positive, as were ambulatory systolic, diastolic, and office systolic pressures. The sole outcome measure in the entire program that I described to you that was not positive was the office diastolic blood pressure in this study which was not even a primary outcome variable.

DR. NISSEN: That's helpful I think. That might have been pruned, Doug, from your packet because in the material we got, the losartan studies were not in here. So, I didn't get a chance to review them.

DR. THROCKMORTON: No. That's correct. They were not part of my original review.

DR. NISSEN: So, when I said there wasn't adequate evidence, it was based upon what I was given to review. So, I stand corrected. It sounds like you've done the two adequate trials against losartan. So, I think the answer to that is yes.

DR. THROCKMORTON: Jeff, we've heard what we need, I believe, on this question. You've identified agents where superiority is adequately demonstrated. That allows you to go the next step I think.

DR. BORER: Which we will now do. Question number 4. And here we're going to need a vote. With what potential benefit should the risk of angioedema be balanced? We may need a little clarification here from the FDA, but let me read through the question. The sense of it is reasonably clear, but we may need some specific clarification so we answer you correctly.

With what potential benefit should the risk of angioedema be balanced? OCTAVE allowed the addition of no new antihypertensive drugs during the first 8 weeks, at which time the blood pressure was about 3 over 2 millimeters of mercury lower on omapatrilat. During the following 16 weeks, other drugs were to be added to meet blood pressure goals, but at the end of 24 weeks, the blood pressure difference was still 3 over 2 millimeters of mercury. What explains the persistence of the differential effect at 24 weeks?

4.1. Is a regimen including omapatrilat able to lower blood pressure to an extent that combinations of enalapril and other drugs cannot? Which is one of our key issues here. If so, is the risk-benefit comparison between the risk of angioedema and the expected reduction in cardiovascular events attributable to this blood pressure difference?

DR. TEMPLE: I think the question is should the risk-benefit comparison be based on that difference. That's sort of the question that Norm emphasized in his review and that I raised before. What if they had added another drug or gone up in dose, which didn't happen? So, that's the question.

DR. FLEMING: Which the second part of the question does more get at.

DR. THROCKMORTON: That is the second option. The first option, to phrase it another way, is that omapatrilat has a property that allows a regimen using it to lower blood pressure 3 millimeters of mercury more than any regimen containing enalapril by some means. Even if you add additional medications, you can't obtain those additional 3. If so and you concluded that 3 millimeters of mercury matter, then it seems that what you want to know is the potential benefits of those extra 3 millimeters of mercury compared versus the risks of angioedema.

If, on the other hand, you're not convinced or on the other side, that you believe that perhaps just adding one more drug in the OCTAVE trial would have sufficed to bring the enalapril group to the same blood pressure control as the omapatrilat group, then perhaps -- and that's what we need to have some discussion about -- the comparison is the risk of adding that additional approved medication compared with the risk of the angioedema.

Does that help to clarify things?

DR. TEMPLE: But also, you don't know what would have happened had they done that because it didn't happen.

DR. THROCKMORTON: They may be convinced that they do.

DR. TEMPLE: Yes, okay.

DR. BORER: I'll tell you what let's do here because you said you wanted a specific vote on the components of this and we'll restructure it so you get one. Let me start by asking Tom for his opinion because he can't vote, and then we'll move on from there.

DR. PICKERING: My problem with this question is that I don't think this was a question that OCTAVE was really designed to answer. It was designed to look at the relative incidence of angioedema. The word is "cannot." I'm not convinced that if the study had continued longer and additional drugs had been added to the enalapril group that the difference might have become smaller. I don't think I know. It may have persisted, but as I say, it wasn't really designed to get at this question.

DR. BORER: Steve.

DR. NISSEN: My sense here is that this was an artifact of the trial design, and let me see if I can be clear. There were only two opportunities for dose titration in the trial, in a relatively short-term trial. If you think about patients and physicians and how they care for them, there's a little bit of inertia here, and if you see a patient over time and your blood pressure is not falling, eventually you come around to adding another agent. Now, should we be quicker on the draw? Maybe. Are we a little lackadaisical? I think Henry and others have taught us that we are.

But my guess is that the reason there was such a small amount of additional drug use in the enalapril arm related to that artifact of only having two opportunities to do so, and if you had carried this trial out for a year and had five or six or seven attempts, or opportunities rather, to add additional drugs, that eventually you would have seen some upward creep in the additional drug use in the enalapril arm and that would have equalized. We don't know that. I'm just trying to help explain why that difference persisted.

DR. TEMPLE: Some trials, of course, insist that you titrate and insist that you add if a criterion isn't met. This didn't do that.

DR. NISSEN: That's right.

DR. TEMPLE: The question is how important it is.

DR. NISSEN: Yes, and I guess what I'm saying is I think this was a design issue, not an efficacy advantage because I'm convinced that if those physicians had been instructed to do so and given time to do so, they would have closed the gap between the two regimens. Or they might have.

DR. CARABELLO: But, Steve, why wouldn't it work with the other arm as well?

DR. NISSEN: I'm not sure I understand your question.

DR. CARABELLO: Well, why wouldn't the ability to titrate omapatrilat more aggressively continue to maintain the gap? Why would they only titer to one of the arms?

DR. NISSEN: Well, because there was a differential. So, the group that's in the differential with the higher blood pressures is going to naturally get more adjunctive therapy. I think that there's a tendency. If you're treating to target and you give the same target to both arms and one arm has omapatrilat, then the group that doesn't have omapatrilat is going to end up getting more adjunctive therapy and is going to tend to close that gap. So, I'm going to guess that a 1-year trial with five or six opportunities to titrate would have -- now, whether it would have closed it completely or not, nobody knows because it wasn't done. But I think that's the explanation for the difference.

DR. BORER: Yes, I think an important issue here is that, for better or for worse, we treat to goal and once you achieve the goal, perhaps inappropriately there really isn't an aggressive attempt to lower further. If you weren't able to achieve the goal in either arm, presumably you'd give more and more and more drug until you did, and then we would have known the true impact of omapatrilat.

Susanna, do you want to talk about 4.1 and 4.2?

DR. CUNNINGHAM: Well, I don't think I know for sure what would have happened, just as Steve has just outlined. So, I think it's unfortunate.

DR. BORER: Do you have any other comment?

DR. CARABELLO: No.

DR. BORER: Mike, Tom.

DR. FLEMING: Should we be answering both questions?

DR. BORER: Yes.

DR. FLEMING: Actually there are two parts, as I see it, to 4.1. The first is relating to the answer that Steve was just giving about whether omapatrilat is able to achieve better blood pressure lowering than other combinations would be able to do. I support Steve's answer and I would ask one other scenario that could justify why more aggressive dosing might have closed the gap is that if a lot of patients at baseline that we've had reported to us had been on ACE inhibitors and had not, in fact, achieved adequate response, if I'm going to randomize those people to something else to achieve a better result, omapatrilat is a very logical option as something that provides a more aggressive approach. If I'm going to randomize in control to enalapril, one of the ways that you could have gotten better response there would have been a b.i.d. or more aggressive dosing in that as a control arm.

What I don't know -- and I've already said I won't put too much stock in the OVERTURE trial, but maybe the OVERTURE didn't show as much difference in blood pressure dosing because of that reason. It's uncertain, and I would agree with Tom's original answer. I think the trial was not designed in a way to truly address this question. It may not be true and it may be true that more aggressive dosing with enalapril and then with other adjunctive therapies might, in fact, have closed the gap.

The second part, as I understand the rewording of question 4.1 in the second part, it's -- and I'm going to read it as I understand the rewording -- what is the risk-benefit comparison between the risk of angioedema and the expected reduction in cardiovascular events? At least I'm going to answer the question as I've just worded it.

The risk of angioedema at the most serious level, as has been approximated here, the upper limit of the confidence interval is around 5.7 per 10,000, although it could be considerably larger if one, in fact, starts with a dose of 20.

What is in fact the benefit? And there are two ways of getting at the benefit. One way is through the surrogate and extrapolating from a 3 millimeter reduction in blood pressure and essentially using estimates from HOPE and other sources that would say we would expect per 1,000 person-years 30 clinical events of cardiovascular death, MI, heart failure, stroke, and using the HOPE trial with the 3 millimeter reduction, maybe a 15 to 20 percent relative risk reduction, that would translate to something on the order of 40 to 80 events per 10,000 person-years.

We actually observed much less than that. Granted, the data are limited, but we still had 170 events, and these are from the very trial on which we're trying to make our assessment. The actual event rate was maybe half what was expected, and in turn, the actual relative reduction was half of what was expected. So, we ended up with maybe a quarter of the number of reduced events. The data of 89 versus 82 cardiovascular deaths, MI, heart failure, stroke at 6 months translates into roughly 10 to 15 events prevented per 10,000 person-years.

So, I stand back and basically make the assessment of what's prevented based on two sources of information, one, what the data actually said, and that's 10 to 15 percent, against what you might extrapolate if you truly believed in the surrogate. And it's probably twice that size. So, that's what we achieve in the context of what we are seeing as serious events of life-threatening events of angioedema which are roughly 5.7.

What it indicates to me is that there is a favorable benefit to risk in those analyses, although the serious events of angioedema are not trivial in the context of what we're trying to achieve, hence the concern that can we achieve what we're trying to achieve in ways without raising those events.

DR. BORER: Paul.

DR. ARMSTRONG: I would say that based on the doses of amlodipine and diuretic we heard were used in the adjunctive therapy, that there was additional opportunity for enhanced blood pressure control in the comparator arm.

I would say in relationship to 4.2 that the obvious blood pressure lowering superiority of the new agent may translate into a long-term benefit, and if the risk of angioedema was not a player, I would be comfortable in that proposition.

But given that there are three separate neurohumors that are affected by this agent and at least one and perhaps others that we don't know about is modulating the angioedema, and that the risk of angioedema is not likely to diminish over the lifetime of a hypertensive patient once the early first or 2 days is obviated, I have meaningful and real concerns.

DR. BORER: Bob and then Doug.

DR. TEMPLE: I just want to be sure we separate the two things out. The first part of the question is about do you believe this difference would persist if people had titrated or added other drugs, and what I heard from a number of people is the study wasn't designed to tell that, therefore you can't know in a hands-on way. You might suspect, but you can't know.

And the second was Tom's observation that even though there was a better control of blood pressure, you didn't the events. But I have a question for you on that. This was a relatively short-term study. Is that a question that could be answered in a study of this duration, or does it really take a little longer before you even have a shot at showing a benefit from that change? So, that's a question of how negative the failing to find that difference is.

DR. FLEMING: If what you're saying is we can estimate a relative risk reduction in these clinical events, but recognizing this is a small study, how wide is the confidence interval?

DR. TEMPLE: No. I'm saying it's short. It's true the benefits of antihypertensive therapy are observed relatively quickly, but I don't think they're usually seen in 6 months or a year very prominently. So, I'm not sure what the expected benefit would be even if it had the usual effect. So, I'm really asking how negative should one feel about the failure to see the reduction in actual risk in that study. Obviously, my implication is I'm not sure you would have expected it in a study of that duration.

DR. BORER: I'd like to provide a response. Everybody else has responded to this. I look at it just a little bit differently than some of the other commenters. And I'm going to divide it into parts here if I can.

Is a regimen including omapatrilat able to lower blood pressure to an extent that combinations of enalapril and other drugs cannot? I believe that it can, and the reason I do is not that there are direct data. There are not direct data of the kind that I would have liked to see to come to a firm conclusion. But in every comparison that we saw, regimens containing omapatrilat were better than the comparators. So, I believe that a regimen containing omapatrilat would be able to lower blood pressure to an extent that combinations of enalapril and other drugs cannot.

However, I'm not sure what group of patients that's referable to because that study wasn't done. The truly refractory patients weren't identified here. So, if you asked me to write a label, I would be hard-put to do it because I'm not entirely sure what group we're talking about, and I'd like a little bit more evidence that in such a group, this drug actually does provide a benefit of the magnitude we're talking about, though I believe it probably does.

But having said that, I believe that the judgment should be based on the risk-benefit comparison to angioedema because I haven't seen evidence that there is any other meaningful risk that we ought to be worried about. There don't seem to be other problems coming up with this drug.

Having said that, if I accept the 3 over 2 millimeter fall in blood pressure, additional reduction in blood pressure, if I accept that, in a truly refractory group that I didn't really look at here, but if I accept that for the moment as the benefit, or the surrogate for the benefit, and I compare the angioedema risk, I would come to the same conclusion that I think Tom did, that the benefit outweighs the risk.

Why do I say that? It's not just on a basis of event rate, but the fact that the risk of angioedema is not immediate closure of the airway and sudden death, but that in most cases the problem is not so severe as that, that there's some time to respond, and on and on and on. So, the really meaningful mortal risk is I think relatively low although we probably underestimated it here because of the fact that this was a study constructed as it was.

But my real sticking point here is figuring out who these people are that the drug would be used for, how you would define that. I'm saying it's people who were refractory, but I'm not sure exactly how I would define that.

Then the issue of the persistence of the blood pressure difference at 24 weeks. Is it a consequence of trial design goal, the blood pressure goal or the goal blood pressure, inadequate use of additional drugs? I think it's all of the above.

DR. THROCKMORTON: But you just said that you believed in your heart of hearts that despite difficulties in interpreting the 24-week data, that there was some population there for which omapatrilat alone had a greater blood pressure lowering effect than combinations of enalapril and other agents?

DR. BORER: No, no. That a regimen including omapatrilat would achieve better blood pressure control than a regimen of multiple drugs that didn't include omapatrilat.

DR. THROCKMORTON: I think we're saying the same thing, that you couldn't get to the place that you could get with a regimen containing omapatrilat with a regimen --

DR. BORER: Without it.

DR. THROCKMORTON: -- without. Then your trouble is you're not sure you can identify the population, but that the data from that trial are sufficient for you to believe that.

DR. BORER: Well, no. I said the data really are not sufficient. I'm making an inference. I'm making a leap of faith here.

DR. THROCKMORTON: I want to understand that leap.

DR. BORER: I'm looking at the data here and I'm saying in every comparison that was made, omapatrilat was superior to the comparator. Within the OCTAVE trial, there were, I believe, undoubtedly people who would be within the population for whom the sponsor is suggesting the drug should be used, people who were on multiple drugs probably at reasonable doses who just didn't respond and who were given omapatrilat instead of enalapril as part of the regimen and who did better.

It's just that I haven't seen precisely those data. The sponsor may be able to tease them out. I don't know if the documentation is sufficient to do that, but certainly one could look at the subpopulation that was on multiple drugs at at least such and such a dose of each of the components and omapatrilat rather than enalapril and could show that that group had a greater blood pressure reduction than the comparator. I mean, you could do that. You could ask them to do that.

DR. TEMPLE: Well, they actually did do that in some sense. What's missing I think is what would happen if they added guanfacine or something like that. That you don't see.

DR. BORER: Well, that's true. We don't know what range of drugs they gave, but we didn't see how much of each component they gave. So, we really don't know whether the maximum appropriate dose or the maximum labeled dose of all the components was given. We don't know that.

That's a tough row to hoe, and I'd like to see those data. I think it would be useful for the company to go back and tease them out because that would allow us to begin to answer one of the key questions that Tom raised. But if you asked me, do I believe we would find it? Yes, I do believe we would find it. And if we did, and if the blood pressure dropped 3 over 2 or greater, then I would say the risk-benefit relationship would favor the use of the drug for the reasons that I stated about risk. It's just that I'm having a hard time identifying the population.

DR. TEMPLE: I won't ask it now but I might later. Obviously, there's not complete agreement on that 3 over 2 because I don't think Steve would say the same thing, but that's what makes horse racing.

One possibility I guess if you did believe that a bona fide advantage of that amount was meaningful is that there could be another study in people who are refractory in some well-defined way in which people were randomized to two different approaches, including adding another drug, and you got to see if there was a persistent difference. So, maybe it's available in those data, but if it were not, I take it, that's another possibility.

DR. BORER: That would be an alternative solution to the problem.

DR. NISSEN: You kind of took the words right out of my mouth. I was going to suggest that -- I mean, this would be a very useful piece of information for us -- to take patients and to do everything you can to get them to goal using conventional agents, used aggressively with multiple opportunities for titration, and then randomize either to have them switched to omapatrilat or to continue on the ACE inhibitor that would be part of their regimen and see whether or not, in fact, you could do better.

DR. TEMPLE: But that only works if you believe that lowering blood pressure more with this drug is good. I just want to remind you.

DR. NISSEN: I understand. I'm not disagreeing with you, Bob, that lowering blood pressure is a very good thing. I'm not disagreeing at all. But I'm trying to say that in the context of a drug with very significant risks associated with it, we just can't accept that as being sufficient.

DR. FLEMING: If I could add to the comments you were just making. You had noted, as I had stated, that I believe there is a favorable benefit-to-risk in terms of cardiovascular events prevented against life-threatening cases of angioedema, which I do believe. I don't believe that the data are as strong as the sponsor said a couple of times when they said there's an order of magnitude difference in frequencies of those events. I would have put it more as a twofold larger number of cardiovascular events, and if you truly believe in the blood pressure lowering surrogate, maybe it's two- to five-fold. But then we have the uncertainties we've been discussing about the full validity of the surrogate in this setting and about the durability of maintaining that 3 millimeter difference.

All of this would be adequate from my perspective; i.e., I would consider those uncertainties of not sufficient magnitude to cause concern to me if it weren't for the life-threatening angioedema. And it's in the context of that life-threatening angioedema then that what I worry about is even though I do see a favorable benefit-to-risk here, it seems entirely plausible that you could readily alternatively achieve the benefit without the risk.

It's speculation -- but for what we've been talking about, and I think this is the essence of question 4.2 -- whether or not it was a design feature, so to speak, that led to these observed differences. I have serious concerns that we might have been able to have provided alternative management that would have had much lower differences in blood pressure without the corresponding risk of life-threatening angioedema.

DR. BORER: Okay. I think we've given you a great deal of opinion.

Depending upon the committee's answer in question 4, identify a target population and estimate the magnitude of expected benefit. I think we've discussed that. You don't really want us to define precisely what it means to be refractory, and we've all said refractory is what we're talking about.

DR. THROCKMORTON: Well, you've all at various times sort of said that you believed that -- and obviously, the sponsor has made proposals about target populations where the benefits were greater. I guess one useful thing would be to comment on how you would go about doing that. The sponsor has made one set of proposals, and you may find that credible. You may have alternative ways that you might use to identify a population that might most benefit from this drug.

DR. BORER: We'll split this into two parts then and get some opinions about the target population and then make separately a comment about how one would estimate the magnitude of the expected benefit.

Why don't we start out with the target population issue. Tom, do you have an opinion about that?

DR. PICKERING: Well, obviously, I don't think anybody is suggesting that it should be indicated as first line treatment for the general population. So, there has to be some selection of people who are at increased risk. I guess my problem with this is, again, OCTAVE was designed I guess on the assumption that this was going to be something that was approvable as a first line drug, and it wasn't intended specifically to focus on any high-risk target population. So, any information that's provided is a sort of retrospective analysis. Some of the other studies that were done comparing it with other agents in the high-risk population such as the one with people with very high diastolic pressures gave data that were less convincing than some of the other data that we've heard.

So, the other area where I have a real problem is the issue of the increased risk in blacks who obviously are going to be in this country a very big portion of any high-risk population. And I'm concerned that the angioedema rate in these patients is I think 1 in 19, and I don't know how one can separate out, to say that you shouldn't give this patient this medication to blacks. I think that opens up a whole nest egg of problems both political and how do you define who's black and who's not and also other things like that. So, I have a problem trying to define a specific high-risk population at this time.

DR. BORER: Steve.

DR. NISSEN: I don't think I can identify a target population based upon what we now know. Part of the reason I say that is that if we tried to do so by doing some kind of a subset or subgroup analysis from OCTAVE, then what we're talking about is we're trying to make a decision based upon lumping together some non-prespecified subgroups and saying, all right, well, if you had all this and this and this, it looks like you would benefit. I think that the level of evidence that you should look for here is a target population should be prospectively defined not retrospectively defined. So, I just don't like the idea of trying to carve up the data we have and use that carved-up data to try to define a population that would benefit. I don't think that's a proper in a drug with this kind of risk level, and therefore I cannot identify a target population.

DR. BORER: Susanna.

DR. CUNNINGHAM: All I can define is who probably shouldn't have it by the risks of the angioedema, but I don't think I can really define who should have it.

DR. BORER: Blase.

DR. CARABELLO: Yes, I agree. It seems silly to give this to blacks and smokers unless there's a compelling reason to do so. And the sponsor has already suggested that it be targeted at a population in whom other therapies have failed to control their hypertension. I think that's a start, but it has to be defined more carefully than that.

DR. BORER: Paul.

DR. ARMSTRONG: I think Steve and Blase have articulated my opinion.

DR. BORER: Mike.

DR. ARTMAN: Well, I think one population would be the population that Dr. Black showed us, his patients in his clinic that are on multiple drugs in a well-controlled setting that can't be controlled. I think that would be a place to start with this, and I think that would be a very sort of quick and easy study to do. So, I think that you could try to get your hands around that defining patients who are on adequate doses of at least three drugs and still are not controlled, that sort of thing.

DR. BORER: Tom.

DR. TEMPLE: Jeffrey, just on that last point, that's a different proposal from what the company said.

DR. BORER: I understand.

DR. TEMPLE: It says the blood pressure is what defines the population, not all these other things.

DR. BORER: I was going to comment on that just now myself.

I just made a suggestion a few minutes ago that maybe the data could be teased out of the 25,000-patient study, which is a pretty big study. But I must agree with my colleagues here who say that that may not be the proper way to go. Maybe another trial is necessary although I hate to say that, given the resources that would have to be lavished on such a trial.

But I don't think there's anything unreasonable about the target population that the sponsor has proposed, assuming that refractoriness of blood pressure is defined. They said difficult to control. I'm saying refractory which is qualitatively maybe a little bit worse. I too would say look at Henry's population that he can't control and see which one of them have the other problems and study them and show that omapatrilat adds something that wasn't added by the comparator.

I guess it would be very difficult to be reasonably certain that the benefit that I'm willing to impute to the drug, if it lowers blood pressure by 3 over 2, can be imputed unless we actually see it in the appropriate population. I think the best way to do that is to study it prospectively. It may be that the sponsor can come back with a subanalysis by searching its data that would be compelling. I don't want to rule out the possibility that that could happen, but I think the better way would be to do a trial.

DR. THROCKMORTON: Jeffrey, standard advice, when talking about sponsors about resistant populations, has been three classes of drugs, maximum doses, one ACE inhibitor, one CCB, and one diuretic, that sort of thing. Is that the sort of general thing that people -- I mean, does that define Dr. Black's clinic in some reasonable sense?

DR. BORER: That kind of construct would be reasonable to me. That specific algorithm might not be, three drugs, three different classes. But I would ask Tom for his opinion about that.

DR. PICKERING: I think that's a reasonable and commonly used criterion for resistant hypertension.

DR. THROCKMORTON: Just to go back one more time. What Bob said was important. That's a different way of thinking about a high risk or a population that might benefit from drug, the one that the sponsor has been putting forth which was looking at cardiovascular risks and then possible benefit.

DR. BORER: No, that's not different. If I understood their proposal --

DR. THROCKMORTON: No. There were two separate sort of things.

DR. BORER: I thought it was not exclusive but inclusive. You had to have the high blood pressure and you had to have the problem.

DR. TEMPLE: Yes, but that's the question. Do you have the problems or is a systolic pressure of 160 uncontrolled by three classes of drugs bad enough?

DR. BORER: Sufficient.

DR. TEMPLE: That's what I'm asking.

DR. BORER: Why don't we start with Tom and let's go around.

DR. PICKERING: Well, I guess it's a question of what risk you want to start with. Obviously if there's the additional risk factors as well, which I suspect there will be in most of these patients, the risk will be higher.

DR. BORER: Must we demand that those additional risk factors be present to give the opinion that the drug is effective and has safety acceptable for the intended use if all they have is refractory hypertension rather than refractory hypertension plus these clinical evidence of these other problems.

DR. TEMPLE: For those people who meet SHEP entry criteria, we know what blood pressure lowering does in those people, and they weren't selected because they had other risks particularly.

DR. PICKERING: I guess I would say other risk factors present as well.

DR. BORER: Steve.

DR. NISSEN: I actually agree with you, Bob. I think that you could define a population here. We know that people with refractory hypertension do very, very badly, and I think you could lower the bar here a little bit and I would still be very comfortable if you said show us that in a group of people we just can't control with the best drugs we've got, three of them, are still above some threshold and make that threshold significant, not above 130 over 80. I'm not sure where to set that. Because we know that people that can't be controlled despite everything we can throw at them do badly, and a drug that could get those people to goal would have enough likely advantages that it might well, in my view, if it were a robust study, outweigh the disadvantages of angioedema. I think you could define such a trial, but it would have to be very rigorously done.

DR. BORER: Susanna.

DR. CUNNINGHAM: So, let me clarify. We're not answering number 5 here. We're defining what a future study might look like.

DR. BORER: What's the population that we want to define as appropriate for getting this drug if it works in that population? Is it just a refractory hypertension population, or do these refractory patients also have to have cardiovascular problems or other end organ problems as was proposed by the sponsor? Do we want to be that rigorous or just blood pressure problems?

DR. CUNNINGHAM: I think just refractory hypertension would be enough.

DR. BORER: Blase.

DR. CARABELLO: I don't know how to answer the question. I'm not too concerned about the other risk factors.

I am concerned about how much demand we put on previous control of blood pressure. If we say that the population to be tested in one in which we've minutely titrated every last drug before we add this one and that's the only way in which we'll accept efficacy, I don't think that's a very good standard because we've got a whole bunch of folks out there who have had reasonable medical therapy and still are hypertensive. Now, could you get them a little lower if you beat the hell out of them and their providers? Maybe? But is that actually going to translate to benefit to the patient? I don't know. I think that this has to be taken in the context of current good but not impossible-to-reach standards.

DR. BORER: Just to ask for a further statement on that, the way it's been set forward -- and Tom indicates that this is routine practice in this area -- what we've now suggested is that to be declared refractory you should be on three classes of drugs, not necessarily five or six, but three, and that the drugs should have been titrated up to their maximally tolerable or maximally labeled doses. Would that be too much of a standard?

DR. CARABELLO: Well, unless of course going to three drugs generates so many additional side effects that now the patient won't take them.

DR. BORER: Then it's not tolerated, and I would think such patients could be included in such a trial.

Steve.

DR. NISSEN: I was just going to say, Blase, there are an awful lot of people out there that are on ACE inhibitor, diuretic, and amlodipine. There's a world of people like that and some of them aren't controlled on that. Some are still greater than 150 over 100 on 10 milligrams of amlodipine and 40 a day of enalapril and 50 of hydrochlorothiazide, let's say. I think if you could get those people down significantly with a drug, it might mitigate the risk involved in a drug like omapatrilat. So, I think such a trial would go a significant way toward making that an approvable drug.

DR. BORER: Paul.

DR. ARMSTRONG: The patient that Steve just described is often controlled with the addition of a beta blocker. So, the notion that three should be the standard for refractory hypertension, and the addition of a new class of drug is not necessarily for me adequate, but if the patient was truly refractory to best medical therapy in a supervised hypertension clinic environment and had evidence of target organ damage, I could see taking the additional and unknown risk of adding a drug such as this, especially if that patient was being supervised by experts who understood the side effect implications and were following the patient carefully.

DR. BORER: Mike.

DR. ARTMAN: Well, I sort of made the proposal and you guys refined it a little bit. So I stand as suggested.

DR. BORER: Tom.

DR. FLEMING: If the question is in what population is it most plausible that this agent could be established to have a role, I would accept the logic of the sponsor saying that you would want to target a population that simultaneously satisfied two criteria. First, it would be patients at high risk for major cardiovascular events, i.e., so the setting where the benefit is substantial; and where it's difficult to control hypertension, i.e., where it's less likely that alternative available therapies could yield that benefit. So, in the context of having a risk of life-threatening angioedema, I want to identify a population where there is substantial up-side benefit and simultaneously a population in which I can more readily achieve that up-side benefit with this agent, even though it carries the side effect.

DR. TEMPLE: Jeffrey, there really have been two slightly differing themes, and it's important. We know from SHEP that being 70 and having a systolic of 160 is a high-risk situation by definition. How much more than knowing the blood pressure do you all think that the entry population in this other study would have to have?

DR. BORER: You've heard a couple of opinions already. I think there's a slight preponderance in favor of blood pressure alone being sufficient. And I'm going to add my voice to that. While I'd love to see people with all the end organ problems in the population, I'm sure they will be, as Tom pointed out, just by the nature of the beast, but I would accept uncontrollable blood pressure alone as the population to study, uncontrollable by the definition that we've used.

DR. NISSEN: Jeff, if I could just add. It would be nice also to have that trial not be a 24-week trial because I think one of the things that really limited the current design was there really wasn't enough time to see potential differences emerge here. If we're going to do it, it ought to be pretty solid.

DR. TEMPLE: That's a very important question. If you're not looking for end organ damage and you're really just now looking at the substitution of one drug for something else in the regimen, you don't need a very long trial to do that. Why would it have to be more than 8 weeks?

DR. NISSEN: Well, because I would not want to preclude the addition of other drugs to either regimen. In other words, some people like the idea of adding a beta blocker, and I'd like both arms to have that open to them.

DR. TEMPLE: No. That should be done before they even get into it.

DR. NISSEN: That means four drugs, though.

DR. TEMPLE: Whether it's three or four is no matter. They're refractory and then you randomize to substituting this and keeping the same regimen.

DR. NISSEN: That would be certainly one way to approach it. I don't think we can design the trial at this table, but I think what you're getting from us I think is the sense that showing efficacy in a truly refractory population, well studied would be meaningful.

DR. TEMPLE: No. I've got that. It was the 26 weeks that threw me.

DR. NISSEN: Well, I guess part of it is that I'm always more comfortable when the exposure is a little longer and when you have a little more observational time. Part of it relates to the fact that I think there are some major risks associated with this drug. I guess I don't think we've characterized them very well.

DR. TEMPLE: Yes, but the study we're talking about is going to be small compared to OCTAVE, as is any study in the world. So, it's not going to get you much more safety information, at least as I'm hearing it. It's going to document unequivocally that you control people that were uncontrolled before. That's a very limited thing. It's not very hard to do if it's true and it shouldn't take 26 weeks if that's enough. But you need to be clear.

DR. NISSEN: Let me see if I can be more specific. Again, I hadn't really thought of this before I came in here. But also we'd like to know that the differences are stable, that they're not differences that are closing with time. Tom Pickering raised this question earlier. So, I personally would be a lot more comfortable if I knew that you could sustain for a year an advantage in blood pressure because it's going to be a smaller population, so it's easier to follow them for a little bit longer and see if you can get that kind of sustained benefit in this population. If they escape after 24 weeks, then you don't gain very much.

DR. BORER: I'd like to comment on that also. I'd love to see a very long trial. On the other hand, that's not a standard that we've ever set for an antihypertensive drug because there hasn't been any compelling evidence that problems develop late because we didn't run the trials long enough.

I would be perfectly happy with a 6-month trial, which is longer than the usual antihypertensive trial. I'm concerned, as Steve says, about exposure for safety, but if it's a small trial, we're not going to get much from that. I'd like to know, though, from the sponsor of the 25,000 patients involved in the trial, 12,000 were on omapatrilat. Have they been continued in an open-label experience or any subset of them?

DR. LEVY: No, they haven't.

DR. BORER: You did show us, though, some 1-year data in a withdrawal study that showed persistence of effect at 1 year, which I find sort of compelling.

DR. TEMPLE: I was going to ask you that. There's a lot of data that goes long. There isn't any suggestion that whatever effect there is goes away. It's really important. For one thing, I would be damned uncomfortable allowing a trial to continue that showed a difference and not doing something else to get those people under control. I think that would be dubious. We allow 8-week trials against placebo, but we don't allow 26-week trials against placebo.

DR. NISSEN: But, Bob, you've already said these are people that can't be controlled any other way.

DR. TEMPLE: No. There's a fifth and sixth and seventh and eighth drug too. Somebody would have to go do something.

But my fundamental question is, what's the question? I thought, from all the previous discussion, the question was can you take people who are bona fide refractory and who are refractory at entry -- they're on all these drugs at entry -- and get control when you couldn't before, a question that I would have thought could be answered in 6 to 12 weeks tops.

DR. ARTMAN: I think that's right. I think that is the question and I think it could be done in a relatively short trial because I think under that circumstance, we would be happy to see that incremental reduction in blood pressure if it were true in that population.

DR. FLEMING: In the absence of angioedema, I could be persuaded to accept that, but part of the fundamental issue at hand here is to be convinced that we're going to get a difference, sustained for adequate duration, that it will offset a very real and important side effect. So, coming back to what you were saying, Jeff, this isn't the standard situation. If there wasn't angioedema, one would be more permissive here and would be less concerned.

DR. THROCKMORTON: Absent angioedema, we wouldn't be here.

DR. TEMPLE: But you already have data on whether the effect of this drug is evanescent. So, why do you need to answer that question again?

DR. NISSEN: We have that data relative to placebo I think. We have it relative to an active control arm with three drugs. Do we know that?

DR. TEMPLE: Well, you could say OCTAVE answers that question to a degree, couldn't you? They didn't add on other drugs or push the dose, but you do have that difference persisting.

DR. BORER: I may have missed something, but I thought the withdrawal study you did was in a trial against enalapril, was it not?

DR. LEVY: No. Let me just clarify. I did show you data from a 1-year-long losartan controlled trial in which the blood pressure differential was maintained. I commented on the results of a withdrawal study. That was conducted in patients being followed in a long-term open-label extension, and they were randomized to continue treatment with omapatrilat or to withdrawal to placebo, and a difference in blood pressure emerged very quickly.

DR. BORER: That's fine. Thank you for the information.

To me those are pretty important pieces of information if you were able to maintain for a year a differential against an active comparator and you were able to show a difference on withdrawal to placebo after a year, I'm not sure that we have any information from any body of data that would suggest that people who are really severely ill would be less likely to maintain the effects of a drug. They may be less likely to remain in control, but then again that's what blood pressure cuffs are for, to find that out, and then you can alter the regimen.

But, Tom, is there any reason to expect that a severely hypertensive population put on a drug and responding to the drug will lose their responsiveness to that drug after a year as compared with somebody with mild to moderate hypertension?

DR. PICKERING: I don't think so necessarily, no.

DR. BORER: I would say, although I absolutely agree we'd like long-term experience so that we can get a better sense of safety and all those things, I don't think it's necessary to mandate a long-term trial to show the persistence of effect. I think they've already done that.

But you want to get the opinions of everybody around the table. You've just heard Tom's. Susanna.

DR. CUNNINGHAM: I think I'd take 6 months' worth.

DR. BORER: Steve, you already gave your opinion. Mike, Tom. Okay.

I think we've sort of dealt with 6, but I think we need to do it formally. Should omapatrilat be approved for the treatment of hypertension? Let's deal with that first, and then depending upon the answer, we may go to 6.1, 6.2, and 6.3. Steve.

DR. NISSEN: My answer is no, and I think the reasons have come out in the discussion. Mainly they relate to the fact that we have limited data on the effect of the drug on events which would have strengthened the argument and that the risk of angioedema is substantial enough to mitigate against the potential benefit of the lower blood pressure.

I would like to say that I see two potential routes to approvability from my perspective. One we've already discussed, which is a trial showing that people that are absolutely refractory can be controlled or have a better blood pressure result when omapatrilat is substituted for an ACE inhibitor adequately done.

But the other which I think should not be ruled out is to show that in fact in a broader population, there is a morbidity and mortality advantage that outweighs the angioedema. Tom and others and the company have tried to estimate the ratio of benefit to risk, but measuring it would be the most compelling evidence of all. So, if one could define a population -- if I were going to do such a study, I'd probably pick people at pretty high risk, and I'd try to show that there was a morbidity and mortality advantage in that population that really quite significantly outweighed the very real risk of angioedema.

The advantage of such a development program is that it would allow this drug to be used in a broader population, not just in the absolutely refractory patients, but in other patients because it may be that the company is right. There may be an order of magnitude benefit greater than risk in a relatively broad population. And if that were proven, then I think it would be a slam dunk that we'd accept the angioedema and we'd counsel about how to do everything possible to prevent it and all that risk management stuff, but we would be pretty comfortable that that overall would help the health of this population.

DR. TEMPLE: Just to be sure, would this be a trial in which people were allowed to be better controlled on omapatrilat than on the other drug, or would this be a case where they would have to show that using that is better than the other thing even though there's equal control?

DR. NISSEN: You know, I'd have to think about this. Again, I'm not sure I can design the trial at the table here, but I guess I would like to see in such a trial a design -- because it would be a longer-term trial -- then the opportunity for titration of the active control arm would continue as would the opportunity for titration of the omapatrilat arm. So, you'd really be looking for whether or not a regimen containing omapatrilat, when optimally delivered over a long period of time, would have a better efficacy at preventing morbidity and mortality than a conventional regimen titrated to optimal effect over time.

DR. TEMPLE: But their hope is that the way it would do that is by giving better control, and if you obliterated better control, then it would have to be through some magic thing.

(Laughter.)

DR. TEMPLE: They don't have a proposal like that.

DR. NISSEN: And if it obliterated better control, then there's no reason to ever use omapatrilat. So, I guess the point would be if you can't produce a persistent differential in blood pressure, then there's no reason for us to use this drug in a population we can achieve that control with conventional agents in.

So, I think you have to let both arms be titrated and you'd probably want to do it more rigorously than was done in OCTAVE, meaning mandate up-titrations. I have to think about it a little bit, but the idea is to try to show that there is a clinical advantage to regimens containing omapatrilat with respect to some harder endpoints other than just blood pressure.

Again, I'm not putting down the surrogate. I'm with you, Bob. I think the surrogate is a good surrogate, but it's only relevant when you have a drug that has comparable risks to what else is out there. So, when you don't have that, then I think you've got to go to those harder endpoints. I think it could work. I think you could find that over 3 years or 4 years that a regimen containing omapatrilat will end up with a blood pressure differential that's maintained and that leads to a difference in hard events, in which case this drug would be a good drug to make available for a broader population.

DR. BORER: Tom Fleming.

DR. FLEMING: I agree very much with the essence of what Steve has said, and in the interest of avoiding repeating that and other things that I've said before, I'll just state that for those reasons I vote no.

DR. BORER: Mike.

DR. ARTMAN: I would vote no at this time.

DR. BORER: Susanna.

DR. CUNNINGHAM: No.

DR. BORER: Blase.

DR. CARABELLO: I'll be the one dissenter and vote yes. I believe that if the drug were added to the community now, it would result in a substantial fall in blood pressure in our hypertensive patients. For me, it's not the question can it, under the rigorous controls of the trial, make a difference, but what difference would it make in the community in which apparently we as a medical group are doing a lousy job of controlling people's blood pressure. And I think it's that group of people that I'm most interested in. I think that the risk of truly life-threatening angioedema could be controlled.

DR. BORER: Reluctantly I'm going to vote no, and this is a very difficult vote and sort of a close call because my intuition is that this drug would offer a benefit that we don't get with other agents. But I'd like to see the data in a refractory population, defined as we've all discussed, that it does indeed improve blood pressure control because there's a countervailing risk which I think can be superseded by the benefit of the additional blood pressure lowering, but I'd like to see that. So, I don't think this is not an approvable drug, but I don't think it's an approvable drug today.

We don't have to go to 6.1, 6.2, and 6.3.

Bob.

DR. TEMPLE: I just want to thank the committee for struggling with what was a very difficult set of issues.

DR. BORER: Doug.

DR. THROCKMORTON: Yes, I'd just echo that. The materials and the issues you've been asked to look at over the last couple of days have been very challenging, and the agency really thanks you for your help. It's truly valued.

DR. BORER: And it's only 3:53 and 52 seconds.

(Whereupon, at 3:53 p.m., the committee was adjourned.)