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

9:03 a.m.

CHAIRMAN OVERTURF: Good morning, I would like to call the meeting to order and first of all, I'll turn it over to Christine Walsh, the Executive Secretary, for some administrative issues.

MS. WALSH: Good morning. I'm Christine Walsh, the Executive Secretary for today's meeting of the Vaccines and Related Biological Products Advisory Committee. I would like to welcome all of you to this meeting of the Advisory Committee.

Today's session will consist of presentations that are open to the public. I would like to request that everyone, please, check your cell phones and pagers to make sure they are off or in the silent mode. Due to a family emergency, Dr. Pamela McInnes will be unable to attend the meeting with us today.

I would now like to read into the public record the Conflict of Interest statement for today's meeting. "This brief announcement is in addition to the Conflict of Interest statement read at the beginning of the meeting on December 14^th and will be part of the public record for the Vaccines and Related Biological Products Advisory Committee meeting on December 15, 2005.

This announcement addresses Conflicts of Interest for the discussions of Topic 2 on the Safety and Efficacy of ZOSTAVAX manufactured by Merck and Company. Dr. Steven Self has recused himself from the discussion of Topic 2, Safety and Efficacy of ZOSTAVAX. In accordance with 18 USC Section 208(b)(3), waivers have been granted to Drs. Ruth Karron, Thomas Fleming and Daniel Scharfstein.

Dr. Ruth Karron for unrelated consulting with the competitor for which she receives less than $10,000 per year. Dr. Thomas Fleming for unrelated consulting with a competitor for which he receives less than $10,001 per year. Dr. Daniel Scharfstein for unrelated consulting with a competitor for which he receives less than $10,001 per year and ownership of stock in the sponsor currently valued at less than $10,001.

A copy of the written waiver statement may be obtained by submitting a written request to the Agency's Freedom of Information office, Room 12A-30 of the Parklawn Building. Dr. Seth Hetherington is serving as the industry representative acting on behalf of all related industry and is employed by Inhibitex Incorporated. Industry representatives are not special Government employees and do not vote.

In addition, there are regulated industry speakers making presentations. These speakers may have financial interest associated with their employer and with other regulated firms. The FDA asks in the interest of fairness that they address any current or previous financial involvement with any firm whose product they may wish to comment upon.

These individuals were not screened by the FDA for conflicts of interest. This Conflict of Interest statement will be available for review at the registration table. We would like to remind Members and consultants that if the discussions involved any other products or firms not already on the agenda for which an FDA participant has a personal or imputed financial interest, the participants need to exclude themselves from such involvement and their exclusion will be noted for the record.

FDA encourages all other participants to advise the Committee of any financial relationships that you may have with the sponsor, its product and, if known, its direct competitors." That ends the reading of the Conflict of Interest statement. Dr. Overturf, I turn the meeting back over to you.

CHAIRMAN OVERTURF: Again, I would like to welcome you to this meeting of VRBPAC for December 15^th and I would like to go around the Committee Members and ask them to introduce themselves again and, please, tell us where you are from. Dr. Karron, we'll start with you.

MEMBER KARRON: Ruth Karron, Johns Hopkins University.

DR. FLEMING: Thomas Fleming, University of Washington.

MEMBER WORD: Bonnie Word, Baylor College of Medicine.

DR. SCHARFSTEIN: Daniel Scharfstein, Johns Hopkins University.

DR. ROWBOTHAM: Mike Rowbotham, University of California San Francisco.

DR. GELLIN: Bruce Gellin, National Vaccine Program Office, HHS.

DR. WHARTON: Melinda Wharton, National Immunization Program, Centers for Disease Control and Prevention.

MEMBER ROYAL: Walter Royal, University of Maryland School of Medicine.

DR. HETHERINGTON: Seth Hetherington, Inhibitex in Alpharetta, Georgia.

MEMBER FARLEY: Monica Farley, Emory University School of Medicine.

MEMBER MARKOVITZ: David Markovitz at University of Michigan.

CHAIRMAN OVERTURF: And I'm Dr. Gary Overturf from the University of New Mexico. So we will begin the meeting today which is to evaluate the safety and efficacy of ZOSTAVAX and I'll ask Patricia Rohan to come forward and provide the introduction from the FDA.

DR. ROHAN: Dr. Overturf, good morning, personnel from Merck, invitees, Committee, I would like to welcome you all. I'm the medical officer and I'll be speaking later for this presentation, but first we would like to as usual go over the Committee's questions that will be considered later this afternoon.

Question No. 1: "Are the available data adequate to support the efficacy of ZOSTAVAX when administered to individuals 50 years of age and older in preventing herpes zoster, in preventing postherpetic neuralgia, preventing postherpetic neuralgia beyond the effect on the prevention of herpes zoster and decreasing the burden of illness and decreasing the burden of illness beyond the effect on the prevention of herpes zoster and, if not, what additional information should be provided?"

Question No. 2: "Are the available data adequate to support the safety of ZOSTAVAX when administered to persons 50 years of age and older, if not, what additional information should be provided?"

Question No. 3: "Please, identify other issues that should be addressed, including post-licensure studies. In particular, please, address the use of the vaccine in persons with co-morbid conditions. For example, those who might typically reside in assisted living residences and nursing homes. The use of the vaccine among persons taking chronic immunosuppressive agents, such as corticosteroids, the use of the vaccine in certain subjects of the sponsor's proposed age indication. For example, those 70 years of age and older, those 80 years of age and older. The duration of immunity and a sponsor's proposed pharmacovigilance plan." Thank you.

CHAIRMAN OVERTURF: Thank you, Dr. Rohan. We will begin now with the sponsor's presentation.

DR. GUTSCH: Good morning, Mr. Chairman, Members of the Advisory Committee, the FDA, ladies and gentlemen. My name is David Gutsch and I'm a Director in the Department of Regulatory Affairs at Merck Research Laboratories. Today I'm going to start by introducing you to ZOSTAVAX, the Merck vaccine, for the prevention of herpes zoster and its complications including postherpetic neuralgia or PHN.

As you will hear today, there is a medical need for a vaccine to prevent herpes zoster and its complications. Herpes zoster is common in those 50 years of age and older. There is no medical intervention to prevent herpes zoster. The acute and chronic pain associated with herpes zoster is often severe and debilitating. And with available therapies, management of the acute and long-lasting pain complicating herpes zoster can be frustrating.

The hypothesis for the ZOSTAVAX Program is that vaccination with the live attenuated Oka/Merck VZV vaccine will meet an important unmet medical need by reducing the incidence of herpes zoster, otherwise known as shingles, and by reducing the frequency and/or severity of herpes zoster of the complications of herpes zoster, including postherpetic neuralgia, the pain that can last for months to years after a rash heals.

As you will see in the following presentation, there are many definitions of postherpetic neuralgia in the literature, including pain persisting beyond rash healing through pain persisting beyond six months. Based on the literature and consultation with experts for the purposes of the ZOSTAVAX clinical studies, PHN was defined as the presence of clinically significant pain present 90 days or more after herpes zoster rash onset.

ZOSTAVAX is a live attenuated varicella-zoster vaccine, varicella-zoster virus vaccine, that uses the same Oka/Merck strain that is present in VARIVAX, the licensed vaccine for chicken pox, and the recently licensed ProQuad for measles, mumps, rubella and varicella. And while VARIVAX and ProQuad contain the same active ingredient, there are notable differences in these products.

VARIVAX is used for the primary prevention of VZV and, therefore, is administered to younger VZV naive population. The proposed use of ZOSTAVAX is for prevention of reactivation of VZV and the subsequent complications of that reactivation. So ZOSTAVAX would be targeted to an older population. ZOSTAVAX is a preservative-free lyophilized product that is administered as a single subcutaneous dose.

ZOSTAVAX is manufactured using the same process as VARIVAX and both vaccines contain the same excipients. When reconstituted and administered as instructed, ZOSTAVAX contains 19,400 plaque-forming units per dose, which is about 14 times the dose present in VARIVAX in order to list at the desired immune response.

The proposed indications for ZOSTAVAX are as follows: ZOSTAVAX is indicated for the prevention of herpes zoster or shingles, prevention of postherpetic neuralgia, reduction of acute and chronic zoster-associated pain. ZOSTAVAX is indicated for immunization of individuals 50 years of age and older. As you will hear in more detail, these three clinically meaningful indications are directly supported by having met the success criteria for key Shingles Prevention Study efficacy endpoints that were pre-specified and mutually agreed upon by the sponsor and the FDA.

The three endpoints that support the indications regarded the decrease incidence of herpes zoster, decreased incidence of postherpetic neuralgia and reduction of the pain burden of illness over a six month follow-up period after herpes zoster rash onset. Although the Shingles Prevention Study enrolled subjects 60 years of age and older, there is a strong case for vaccination with ZOSTAVAX starting at age 50. The next speaker will take you through the epidemiologic and clinical evidence supporting the proposed target age range.

In the ZOSTAVAX vaccine license application there are eight clinical trials in which ZOSTAVAX has been administered, including the Shingles Prevention Study of Veterans Affairs, Cooperative Studies Program, Multicenter Placebo-Controlled Study in which nearly 40,000 subjects were enrolled. And as you will see, these studies demonstrated that ZOSTAVAX is efficacious in preventing herpes zoster and PHN in reducing the overall burden of zoster-associated pain, including severe pain, and in reducing the interference with activities of daily living due to herpes zoster.

Furthermore, you will see that ZOSTAVAX is immunogenic in the VZV experience vaccinees and that ZOSTAVAX has an excellent safety profile. Collectively, the efficacy, immunogenicity and safety that results support a favorable risk/benefit assessment as an intervention to prevent herpes zoster and its complications, including PHN, ZOSTAVAX represents a major medical advance.

There are several collaborators present who are associated with the Shingles Prevention Study, the large pivotal study in support of ZOSTAVAX. Here today are Dr. Michael Oxman, Study Chairman for the Shingles Prevention Study; Gary Johnson, a Shingles Prevention Study Biostatistician; and Dr. Myron Levin, a key principle investigator. Also present as clinical consultants are Dr. Ann Arvin, Dr. David Cornblath, Dr. Robert Johnson and Dr. David Weber. And our statistical consultants are Dr. James Neaton and Dr. Janet Wittes.

A detailed briefing document was previously provided to the Advisory Committee Members. Dr. Jeffrey Silber from the Department of Clinical Research at Merck Research Laboratories will now present the highlights of the information provided in the briefing document. Following this, I will provide some concluding remarks.

DR. SILBER: Thank you, David, and good morning. This morning I have the privilege of sharing with you information on a number of topics. The epidemiology of herpes zoster and postherpetic neuralgia, an overview of the Clinical Development Program for ZOSTAVAX followed by a more detailed description of the study design and key results from the Shingles Prevention Study. I will also review available immunogenicity and safety data for the product before providing an overall summary of the clinical trial results.

As background, it is important to note that herpes zoster, commonly known as shingles, is a clinical manifestation of the reactivation of latent infection with varicella-zoster virus or VZV. Primary infection with VZV typically in childhood causes chicken pox. Thereafter, the virus establishes a latent infection in the dorsal root ganglion of the spinal cord where it remains quiescent for many years.

In the United States nearly all adults have evidence of prior VZV infection and therefore are at risk for shingles. During prolonged latency, VZV-specific cellular immunity keeps the virus in check. And for reasons that are not entirely understood, but are clinically associated with advancing age or immunosuppression, the virus reactivates.

The virus travels down the nerve root, reaches the skin and develops into the characteristic eruption of painful, erythematous, maculopapular lesions that evolve into clustered fluid-filled vesicles that are shown on the right hand side in a pathognomonic dermatomal distribution.

Herpes zoster is a relatively common disease. It is estimated that, approximately, 1 million cases of herpes zoster occur each year in the United States, of which nearly two-thirds occur in persons over the age of 50. And this number is expected to rise due to the aging of the population. An estimated 50,000 to 60,000 hospitalizations each year in the United States include a diagnosis of herpes zoster. And among these are an estimated 12,000 to 19,000 for which herpes zoster is the primary diagnosis.

Of note, 85 to 90 percent of all herpes zoster cases and 70 to 80 percent of hospitalizations occur in immunocompetent individuals. And it has been noted in recent studies that the lifetime risk of developing herpes zoster may be as high as about 30 percent and for those who attain the age of 85, up to 50 percent will have suffered one or more episodes of zoster in their lifetime.

Although herpes zoster has been noted to occur after stressful life events or the site of prior physical trauma, the only clearly established risk factors for herpes zoster are increasing age and immunosuppression. This figure is from a classic paper by Hope-Simpson showing the age-related contributions in herpes zoster and postherpetic neuralgia or PHN. I'll be speaking much more about PHN in subsequent slides.

In this figure, the X axis shows age in years and on the Y axis is the rate of disease. And you will note that there is a substantial increase in the incidence of postherpetic neuralgia beginning at age 60, whereas the incidence of herpes zoster begins to rise fairly dramatically at age 50. Similar findings have been borne out from more recently conducted population-based studies.

This slide shows the number and the proportion of all herpes zoster cases in the United States across the different decades of life, based on the most recent census data and the age-specific rates from the Hope-Simpson Study. And the results are generally similar when results of other population-based studies are applied. Note the preponderance of herpes zoster cases among the older adults with the number of cases among people in their 50s at least as high as among people in their 60s, a phenomena that is expected to continue.

The next two slides show typical herpes zoster eruptions. The first shows a herpes zoster case in a mid-thoracic dermatome. The lesions of herpes zoster are typically unilateral, but can cross the midline slightly and can also cross into adjoining dermatomes. The skin lesions usually evolve over about 7 to 10 days and then heal over the subsequent two to three weeks.

The following slide shows an episode later in its course, an ophthalmic zoster in the ophthalmic distribution of the fifth cranial nerve. After the thorax, the fifth cranial nerve is the most common location for herpes zoster to occur. Herpes zoster ophthalmic represents 10 to 15 percent of all herpes zoster cases and about 50 percent of those have ocular involvement. Sight threatening complications can ensue and so prompt attention to these cases is essential.

Numerous complications can result from an episode of herpes zoster. The most common neurologic manifestation is acute neuritic pain, which affects over 90 percent of all episodes of herpes zoster, and can be quite severe even in younger individuals. Postherpetic neuralgia, which is generally defined as pain present following resolution of the rash, is a relatively frequent complication that increases with age, and more on this later.

Other neurologic complications include loader motor neuron palsies, which can affect up to 5 percent of episodes, sensory deficits, autonomic dysfunction and more rarely meningitis, myelitis or encephalitis. A number of ocular complications can occur as a result of ophthalmic zoster as shown on the previous slide. Among the cutaneous complications of zoster are scarring and bacterial superinfection most commonly with staph and strep.

In immunocompromised individuals, visceral complications can occur, including disseminated disease, which carries a mortality rate of up to 40 percent. Although the rash is the most characteristic feature of acute herpes zoster, the most troubling symptom is pain. A majority of patients with herpes zoster first experience prodromal pain of varying duration and the symptoms can also include tingling, itching or burning.

As shown here, the pain during both the acute herpes zoster episode and the postherpetic phase can be quite severe. Patients frequently compare it to the pain associated with child-birth or passing a kidney stone. Early in its course, herpes zoster can be mistaken for a number of other clinical diseases that are common in older adults, including myocardial infarction, cholecystitis, kidney stone, migraine or other CNS condition or severe musculoskeletal pain.

Almost half of all patients with herpes zoster experience pain on a daily basis during the episode and a similar percentage described that pain as horrible or excruciating.

Postherpetic neuralgia is residual pain that is present after resolution of the acute cutaneous eruption of herpes zoster. The pain of PHN can be constant or intermittent, dull and achy, burning, sharp and stabbing or shock-like. And most patients with PHN describe more than one pattern of pain. A particularly common and distressing symptom, which affects a majority of PHN patients, is allodynia.

The exaggerated pain experienced in response to otherwise benign stimulus like the breeze, a bedsheet or just the touch of clothing, often leads to sleep disturbance, social isolation and depression. Overall 10 to 20 percent of herpes zoster patients develop PHN, but the incidence increases dramatically with age. The impact of PHN can be profound leading to physical, psychological, social and functional deficits as well as increased use of health care resources.

Particularly in older adults, PHN can last for months or even years. It is estimated that the prevalence of PHN in the United States is as high as 500,000 or more, which is nearly as high as the prevalence of diabetic neuropathy as a cause of neuropathic pain. As mentioned by Dr. Gutsch, the Pivotal Efficacy Study for ZOSTAVAX implied a specific and rigorous definition of PHN.

This slide from the era before the availability of antivirals looks at postherpetic pain by age and makes several interesting points. The findings are not terribly different today for older adults with herpes zoster. First, note that postherpetic pain of at least a month's duration is rather common, even in middle-aged adults, but that it is very common in the oldest patients. Second, prolonged pain of a year or more becomes more common among the oldest individuals.

Antiviral medications have been shown to reduce the severity of acute herpes zoster and in some patients the medications can shorten the duration of the acute episode by a few days. However, the drugs need to be started within the first 72 hours of onset to have maximum effect. Also, antivirals have only a limited effect on the incidence or the severity of PHN once an episode of herpes zoster has begun.

Corticosteroids have often been used in acute herpes zoster, either alone or in conjunction with antivirals, and the corticosteroids may ameliorate the acute episode, but they have not been shown to affect either the incidence or the severity of PHN. Once PHN develops, finding effective treatment can be challenging, in part because of the wide variation in the type and intensity of the individual's symptoms.

Among the available therapies for PHN are a variety of topical and systemic analgesics, including opiates, tricyclic antidepressants, drugs with anticonvulsant properties and a number of invasive procedures. In general, these interventions have been shown to have limited benefit for patients with PHN and some patients are completely refractory to multiple interventions.

In addition, these agents often have narrow therapeutic indices. They are often associated with limiting side effects, particularly in elderly patients, that make their continued use problematic. The published literature shows that the risk of herpes zoster and PHN goes up substantially after age 50 and there are, approximately, 87 million people in the United States in this age group, and this is a number that will only rise with time.

As just pointed out, the handful of currently available therapies have only moderate benefits and sometimes significant limitations. No intervention can reliably prevent shingles or PHN. Because herpes zoster is more frequent and more severe as age increases and because VZV-specific immunity is known to decline with age, then if VZV-specific immunity could be boosted with vaccination, herpes zoster could be prevented or ameliorated.

For these reasons, ZOSTAVAX has been developed and is expected to have a dramatic public health impact in the United States. I would like to turn now to the ZOSTAVAX Clinical Development Program. The hypothesis that vaccination could prevent herpes zoster, ameliorate its severity and potentially prevent PHN comes from two proof of concept studies that were conducted by Ann Arvin and her colleagues at Stanford University using a heat inactivated formulation of the Oka/Merck VZV vaccine to vaccinate immunocompromised patients in a multi-dose regimen.

The first study published in 1997 show that the vaccine had good biological activity. Although the incidents of herpes zoster was not reduced among those who were vaccinated, the vaccine did reduce the incidence of PHN and significantly ameliorated the severity of herpes zoster. Based on the results of this study and other pilot studies, the efficacy trial was designed originally with pain-related primary endpoints and so-focused on the age group 60 and above in whom zoster-associated pain and PHN are most severe.

In a follow-up study, the results of which became available when enrollment in the pivotal study was complete and follow-up was continuing, Arvin's group found for the first time that vaccination could significantly reduce the incidence of herpes zoster outright.

The licensed application for ZOSTAVAX includes a number of studies that are outlined on this slide. The first two were dose selection studies that established the safety of the vaccine over a 35-fold range of potencies and also explored immune responses using a number of potential markers. Efficacy was evaluated in the pivotal Shingles Prevention Study, which will form the bulk of the remainder of my talk.

Other studies in the program included evaluation of a two-dose regimen and a booster study in individuals who received vaccine years earlier. Additional safety evaluations included vaccination of a small number of VZV-seronegative adults and a study that compared the vaccine at maximum potency with a potency similar to that studied in other clinical trials. In all, about 21,000 subjects received one or more doses of ZOSTAVAX and nearly as many placebo recipients were enrolled in well-controlled clinical trials.

I would like to spend a moment discussing the potency range that was studied in the program, and in particular, the prospect of assessment that led to the potencies that were evaluated in the Shingles Prevention Study.

Across the program, vaccine was administered across, approximately, a 100-fold range of potencies. In addition to demonstrating an adequate safety profile, the early studies suggested that potencies of, approximately, 17,000 plaque-forming units or higher resulted in a boost in VZV-specific immunity and thus formed the basis for selecting a target minimum potency of 19,000 plaque-forming units for the efficacy trial.

ZOSTAVAX has been studied in a large number of older adults reflecting the target population for the vaccine. The vaccine has been administered to individuals as young as 30 and as old as 99 with a wide array of underlying medical conditions. About 58 percent of the subjects enrolled were male. Over 95 percent of the study population was caucasian, but the database also includes over 400 African Americans, about 300 Hispanic subjects and a number of subjects from other racial and ethnic minorities. Except for some age related findings that we presented in subsequent slides, no differences in the efficacy, immunogenicity or safety of the vaccine were seen across demographic groups.

I would now like to turn to an in-depth description of the Shingles Prevention Study. The Shingles Prevention Study, the results of which were published in the New England Journal of Medicine earlier this year was a double-blind placebo- controlled multicenter trial conducted by the Department of Veteran Affairs Cooperative Studies Program in collaboration with the National Institute of Allergy infectious diseases of NIH and Merck.

The study enrolled 38,546 individuals 60 years of age and older. Enrollment was stratified by age to ensure that at least one-third of the subjects enrolled would fall in the 70 plus age group. Nearly 90 percent of the enrolled subjects had one or more underlying medical conditions, but those with known immunocompromised were excluded. Nearly half of the patients or subjects noted some limitations in their daily function from their medical illnesses with about 10 percent moderately or severely limited.

Enrolled subjects were randomized 1:1 to receive ZOSTAVAX or a placebo injection that was made up of the vaccine's stabilizer and uninfected cells. Most of the doses in the study were administered near the proposed expiry potency and after enrollment, follow-up to identify suspected cases of herpes zoster, monitor safety and ensure subject retention was undertaken through use of monthly telephone contacts and a final closeout interview.

As shown on the slide, although a majority of the subjects enrolled in the study were at VA medical centers, the overall enrollment was reasonably well-balanced by gender. The mean age in both vaccination groups was 69.4 years with 46 percent of the subjects at least 70 years of age and about 7 percent 80 years of age and older. And, as noted previously, the study population was largely caucasian.

The next two slides show the most common underlying medical conditions that were reported by subjects in the Adverse Event Monitoring Substudy, the Shingles Prevention Study. Of note, in a study this size, even a 1 percent incidence rate reflects enrollment of a fairly substantial number of patients with a given illness and these slides provide one measure of the heterogeneity of the population enrolled, this slide showing those conditions with an incidence rate of 5 percent or more and the following slide with the conditions that were seen in at least 1 percent of the subjects.

Embedded within the overall Shingles Prevention Study were a number of substudies. Among these were the Adverse Event Monitoring Substudy, a Cell-Mediated Immunity or CMI Substudy and a Persistence of Efficacy Substudy. The Adverse Event Monitoring Substudy, which was conducted at all 22 study sites, enrolled over 6,600 subjects who underwent a detailed assessment of local and systemic safety following vaccination.

The CMI Substudy, which was conducted at only the Denver and San Diego sites, enrolled almost 1,400 individuals who had blood specimens obtained at baseline, at six weeks postvaccination and at subsequent time points.

The Persistence Substudy, which is still ongoing at 12 of the original 22 study sites, is following, approximately, 7,500 subjects who had been randomized to the vaccine group. The substudy is expected to provide information on the performance of the vaccine through, approximately, 10 years postvaccination and the findings of this substudy will be reported at a later date.

This is a pictorial representation of the study and substudy enrollment. Randomization was quite successful with nearly equal numbers randomized to vaccine and placebo in each of the age cohorts and, importantly, the study enrolled nearly as many subjects in the 70 plus age category as in the 60 to 69 age category.

The average duration of follow-up in the study was 3.1 years with a range of up to 4.9 years. Remarkably, only 0.6 percent of the subjects in each vaccination group withdrew from the study or were lost to follow-up. This incredibly high degree of subject retention is a tribute to the effectiveness of the protocol-specified surveillance and the tremendous tenacity of the investigators and other study personnel at the 22 sites.

Over 95 percent of the subjects in each vaccination group remained in follow-up and conducted a closeout interview at the end of the study after accrual of all suspected herpes zoster cases was completed.

Shown here is an overview of the 42 day safety follow-up that was undertaken for all subjects enrolled in the study. More than 70 percent of the subjects either completed a vaccination report card if they were in the Adverse Event Monitoring Substudy or contacted the automated telephone response system, which was available to them for safety follow-up through day 51 postvaccination.

A variety of other types of subject contacts with the study sites were undertaken, including phone calls directly to and from sites, most of them shortly after day 42. In all, about 97 percent of the subjects provided safety information following vaccination, including 93 percent establishing contact by day 60 postvaccination.

The active surveillance for suspected herpes zoster cases cast a very wide net. Through monthly contact with the automated telephone response system and/or the study sites, subjects with findings at all suggestive of herpes zoster were asked to report to the study site within 24 hours and if the investigator could not confirm an alternative diagnosis, the subject was entered into six months of protocol-specified follow-up.

The study sites performed an initial clinical evaluation and were reminded to use a low threshold for calling a rash illness a suspected case of herpes zoster. Lesion and blood samples were taken for laboratory analysis. The digital photographs were obtained. A number of shingles-specific questionnaires were administered in order to define the impact of the illness on the subject, and treatment with famciclovir and analgesics was initiated.

And it's important to note that all of the enrolled subjects with suspected herpes zoster were urged to seek medical care immediately. These highly motivated subjects, well-educated about herpes zoster, received state of the art care by experts in the field with aggressive pain management and frequent, attentive follow-up by study personnel.

Thus, the vaccine's efficacy was not evaluated in the setting of a placebo group that received no treatment, but rather one that received optimal care for their episodes of herpes zoster and PHN. The study had three key efficacy endpoints, the incidence of herpes zoster, the incidence of PHN and the herpes zoster burden of illness or BOI. Subsequent slides will describe each of these endpoints further.

Pain throughout the period of follow-up was scored on a 0-to-10 scale using a validated instrument. The primary efficacy analyses were based on a modified intention-to-treat approach that excluded only those patients, subjects, who dropped out of the study or developed a case of herpes zoster within the first 30 days postvaccination.

This modified approach was employed to ensure that the primary efficacy analyses did not include those vaccine-associated rashes, a primary safety concern in the early days postvaccination, nor those cases of herpes zoster that may have already been in the prodromal phase at the time of vaccination and before the immune response could be elicited.

Although I will be presenting the MITT analyses this morning, the analyses were also performed using a full intention-to-treat approach, as were a variety of sensitivity analyses with virtually identical results.

As mentioned earlier in the presentation, the state of scientific knowledge when the Shingles Prevention Study began indicated that vaccination could prevent, might prevent PHN and lessen zoster-associated pain, but there was no evidence that vaccination could prevent herpes zoster altogether. Thus, the study was designed with two co-primary endpoints related to this important issue of pain, the herpes zoster pain burden of illness and the incidence of PHN.

Incidence of herpes zoster was considered a tertiary endpoint. Because of the age-associated increase in the incidence of PHN, the study was designed to enroll subjects beginning at age 60. Following publication of the second Ann Arvin Proof of Concept Study and prior to study unblinding, Merck and CBER agreed to the elevation of herpes zoster incidence to a key secondary endpoint with a prospectively designed and defined criterion for success.

Turning now to the endpoint definitions. Suspected herpes zoster was defined as any subject with a suggestive cutaneous eruption. These subjects were evaluated by the study physicians and underwent the six months of protocol-specified follow-up that I just mentioned to monitor the presence and the amount of pain and discomfort, development of PHN and the development of any other possible complications.

Although the study cast a very wide net to accrue the suspected cases of herpes zoster, in the end the protocol utilized very strict definitions of herpes zoster and PHN.

All suspected herpes zoster cases were clinically adjudicated by a Clinical Evaluation Committee made up of five independent, that is non-Merck, members of the study's Executive Committee. The Clinical Evaluation Committee adjudicated the cases in a blinded fashion according to a detailed Standard Operating Procedure with all laboratory data redacted from the clinical summaries.

Final confirmation of the herpes zoster cases was determined by a hierarchical algorithm that considered the results of PCR of skin lesions, viral culture and the decision of the Clinical Evaluation Committee in that order and, in the end, a large majority had final determinations based on PCR results.

For the purposes of the primary analysis, PHN was defined as zoster-associated pain with a score of 3 or higher on a 0-to-10 scale that was present for at least 90 days following herpes zoster rash onset. During earlier validation of the pain questionnaire, it was found that a pain score of 3 or higher was correlated with functional limitation on activities of daily living.

The co-primary endpoint, the herpes zoster burden of illness or BOI, was a composite endpoint that was designed to capture the entire burden of pain due to herpes zoster, a population measure that reflected the incidence, the severity and the duration of zoster-associated pain and discomfort over six months following onset.

This slide includes a graphic that shows a curve representing the pain scores over time for a hypothetical subject who developed herpes zoster. With time noted on the X axis and the 0-to-10 scale on the Y axis, an individual severity-by-duration score is thus generated and the BOI represents the scores of all subjects in a particular group.

For the BOI and the other efficacy endpoints, the primary analysis was performed on the entire MITT population. So for each subject who developed an episode of herpes zoster, a severity-by- duration score was calculated and an area under the curve constructed.

Those subjects who did not develop herpes zoster during the study were included in the analysis. They were assumed, however, to have had no zoster-associated pain and, thus, were given a severity-by- duration score of zero.

Incidence and severity-by-duration are both important to describing the overall burden of herpes zoster on patients and the outcome measure needed to reflect both of these components. And to help the Committee get a better grasp of the concept of BOI, which is a bit abstract, the following three slides give hypothetical examples of the BOI in action and I would like to thank Dr. Oxman for providing these slides to me.

In the first example, the putative vaccine reduces the incidence of herpes zoster, but the severity of those cases that do occur in the vaccine group are no lower than the severity in the placebo group, and the reduction in the BOI for the vaccine group is reflected at the bottom of the slide to show the impact on incidence but not severity.

Conversely, here we show the vaccine reducing the severity of the individual herpes zoster episodes, but no impact at all on the incidence of the disease. And the reduction in BOI once again shows benefit from the vaccine. In this third example, the vaccine reduces both incidence and severity-by- duration. And as you will soon see, this third example most closely reflects the outcome of the Shingles Prevention Study.

Shown here is a flow diagram of the 1,308 suspected herpes zoster cases that were followed during the course of the study. Of these, 481 were in the ZOSTAVAX group and 827 in the placebo group. Of note, across the two vaccination groups, similar numbers of subjects were determined not to have herpes zoster, 156 in the vaccine group, 161 in the placebo group, about 0.8 percent of the total population in each group, a finding that reflects the comprehensive and unbiased nature of case accrual.

Of the 322 and 662 herpes zoster cases respectively in the full intention-to-treat analysis population, nearly all were included in the primary modified intention-to-treat population and of these about 93 percent in each vaccination group were diagnosed by PCR, about 5 percent in each group by the Clinical Evaluation Committee and the remainder by viral culture.

This figure provides an overview of the key efficacy results from the study for the three main endpoints, herpes zoster incidence, PHN incidence and the herpes zoster burden of illness. The vertical line at 25 percent efficacy reflects the pre-specified minimum criterion for success that had been established for each of these endpoints in discussions between Merck and the FDA.

The blue bars reflect the vaccine efficacy that was observed in the study along with the 95 percent confidence intervals for each endpoint. The slide shows that for each of three endpoints, the efficacies exceeded substantially the minimum criterion established for the study's success.

The two key messages from these results are, one, that the vaccine was able to significantly reduce the incidence of herpes zoster among the vaccine recipients and, two, there was a significant impact on the severity of the herpes zoster episodes. And I will go on to discuss each of these endpoints in turn.

First, for herpes zoster incidence, this slide shows a Kaplan-Meier plot for the cumulative incidence of herpes zoster over time by vaccination group. The X axis indicates time of follow-up and on the Y axis, the proportion of subjects developing herpes zoster. 315 herpes zoster cases occurred in the ZOSTAVAX group compared with 642 cases in the placebo group.

The curve demonstrates a vaccine effect soon after vaccination. The two curves continue to diverge throughout the entire follow-up period. But note that follow-up beyond four years is rather limited because only a small fraction of the overall study population was followed for four years or longer.

One can see here that the most common complications of acute herpes zoster occurred at a lower rate among ZOSTAVAX recipients than among placebo recipients. The neurologic complications, as shown on this slide, exclude the acute neuritic pain.

Complications of sacral dermatome involvement include such findings as urinary retention or incontinence, constipation or rectal incontinence and across these categories, the vaccine reduced the frequency of complications by, approximately, 65 to 75 percent and the reduction in these complications reflects the vaccine's effect on severe cases of herpes zoster.

Shown here is another Kaplan-Meier plot for PHN incidence using the protocol definition of pain greater than or equal to 3 present 90 days or longer after herpes onset. There were 27 cases of PHN in the vaccine group and 80 in the placebo group. Supportive analyses using alternative time points to define PHN, 30, 60, 120, 182 days, showed generally similar results.

As was the case with the herpes zoster endpoint, the vaccine effect was demonstrated early and then throughout follow-up. And note again the relatively small proportion of subjects with follow-up extending out to four years or longer.

I would like to turn now to the herpes zoster pain burden of illness. As noted earlier, the BOI includes both the incidence of herpes zoster and the severity-by-duration of zoster-associated pain.

The overall efficacy for this endpoint was 61 percent with the 95 percent confidence intervals as shown. The BOI efficacy that was demonstrated in the study reflects a combined effect of both of these components. The 51 percent reduction in the incidence of herpes zoster was already described. With respect to severity-by-duration scores among those subjects who developed herpes zoster there was a statistically significant 22 percent reduction in the scores among those in the vaccine group.

Now, to put a human face on this reduction, because again these scores are a bit abstract, the reduction from a mean of, approximately, 180 to 140 creates a 40 point difference, which reflects nearly a two week reduction in the duration of clinically significant pain at a level of 3 or a four day reduction in pain at the maximum level of 10, the worst imaginable pain. This indicates a huge impact in preventing suffering over and above the vaccine's impact and reducing the incidence of herpes zoster.

This slide gives another perspective on the impact of ZOSTAVAX beyond its ability to prevent cases of herpes zoster. The slide shows a histogram with the subjects who had the highest severity-by- duration scores. The increasing scores are shown on the X axis, the number of the cases on the Y axis. And for the purposes of illustration in this exploratory analysis, the scores of 600 or higher are depicted.

To obtain a score of 600 or higher, the subject would have to have the maximum pain score of 10 for at least two months or a score of 3 to 4, that is clinically significant pain, every day throughout the entire six month follow-up period. So we're talking about very severe cases of PHN.

The slide overall shows that there were only 22 vaccine recipients of 600 or higher compared with 40 recipients and that's a 73 percent reduction. And if one looks at the slide beginning from the right and working toward the left, you can see the very dramatic effect of the vaccine at the farthest end of the pain spectrum.

Among those vaccine recipients who went on to develop PHN, the vaccine effect was equally clear. In an exploratory analysis, including those subjects who developed PHN that was in the license application, it was found that through the end of the follow-up, there was a 57 percent reduction in the severity-by- duration scores among those who received ZOSTAVAX compared with those who received placebo. This statistically significant benefit again shows evidence of the substantial role that the vaccine can play even in subjects who ultimately go on to develop PHN.

As noted earlier, subject enrollment in the study was stratified by age and this slide displays the vaccine effect for the three main study endpoints stratified by age. For the herpes zoster endpoint, there was 64 percent efficacy in the younger cohort and 38 percent efficacy in the older cohort. Despite this difference across the two age strata, vaccine efficacy for herpes zoster incidence remained substantial even for the older age group.

Importantly, since PHN frequency and pain severity increase with age, the vaccine efficacy for PHN was comparable across the two age strata as shown. In fact, among subjects with herpes zoster the vaccine reduced the risk of developing PHN by a statistically significant 38.5 percent, including 47 percent reduced incidence in the 70 and older age group.

Because the vaccine retains substantial efficacy for those subjects with more severe pain associated with a zoster episode, the overall effect on burden of illness was relatively well-preserved among the older age group. Although the point estimate for the burden of illness was a bit higher in the younger group, reflecting the effect on the incidence, there is wide overlap in the confidence intervals between the two age strata because of the benefits on severity-by-duration in the older group.

ZOSTAVAX also had an effect on the incidence of zoster-associated interference with activities of daily living. These analyses were based on the mean of the responses to seven ADL-related questions on a validated questionnaire using a 0-to-10 scale. The vaccine reduced the overall interference with activities of daily living by 66 percent in the overall population.

This combined score for the overall population is sensitive to the incidence, severity and duration of interference, and so it was analogous in many respects to the burden of illness for the overall population.

The vaccine also led to a 55 percent reduction in moderate-to-severe interference with daily living. Now, this reduction was of course influenced by the reduction in the incidence of herpes zoster. So in a pre-specified analysis to determine the vaccine effect on ADL interference above and beyond the vaccine's effect on the reduction of herpes zoster incidence, a reduction of 8 percent was seen, which was not statistically significant.

The duration of the vaccine efficacy was alluded to briefly in the prior Kaplan-Meier curves for efficacy, and this slide presents the efficacy for herpes zoster and PHN over 48 months of follow-up. After an initial decline in efficacy during the first year, the point estimates for efficacy remain relatively stable through 48 months postvaccination.

Now, the confidence intervals do get wider over time reflecting fewer subjects with long follow-up and few with clinical endpoints. So the interpretation at the later time points becomes limited. However, the follow-up for the longer term persistence of efficacy is currently being evaluated, as previously noted, at 12 of the 22 sites and so additional information will be accruing over time.

Thus, the Shingles Prevention Study has shown conclusively that vaccination can reduce the incidence of herpes zoster with better efficacy among the younger age cohort, reduce the incidence of PHN and reduce the burden of illness associated with herpes zoster pain.

For the pain-related endpoints, the vaccine efficacy was very well-maintained in the older cohort compared with the younger cohort. The vaccine also reduced the duration of pain and the risk of substantial interference with activities of daily living and, thus far, the efficacy has extended out to four years.

I would like to turn now to the immunogenicity results. Declining VZV-specific immunity most frequently associated with age is thought to be a precursor for the development of herpes zoster and, as such, the immune response to vaccination is thought to be reflected in efficacy.

The ZOSTAVAX clinical studies evaluated immune responses using two key validated assays of VZV-specific interferon-gamma enzyme-linked immunospot assay and a glycoprotein enzyme-linked immunosorbent assay that has been used to measure antibody responses in the varicella vaccine programs for many years. The VZV-specific antibody measured through the gpELISA is known to be T-cell dependent and is, therefore, felt to reflect the cellular immune response to vaccination.

In the pivotal efficacy study, the primary endpoints for immunogenicity by these assays and also for the responder cell frequency assay were assessed at six weeks postvaccination. And on the next few slides, these endpoints, the endpoints that will be shown, are the ratio of the geometric mean titers or counts in the vaccine and placebo groups, as well as the geometric mean fold increases from baseline.

At six weeks postvaccination in the Shingles Prevention Study, immune responses were seen for both of the key validated assays. Of note, of course, these are previously VZV experienced individuals and so even at baseline, rather high levels of preexisting VZV immunity were seen. And, as you can see, relative to the day zero levels, the VZV antibody measured by gpELISA increased 1.7-fold and the ELISPOT counts increased 2.0-fold, both of which were statistically significant increases.

In a regression model that looked at each of these immune markers as possible correlates for prevention of herpes zoster, both the gpELISA and the VZV interferon-gamma ELISPOT assay correlated with protection. However, the gpELISA correlated best with efficacy, as shown in the slide, with each log unit increase associated with a larger risk reduction.

Note importantly that this is a correlation for values across the population. No specific value in either assay can reliably predict whether an individual subject is protected from herpes zoster, and so the study was unfortunately unable to define a true surrogate.

Looking at the different age cohorts, one sees immune responses that are generally similar with slightly higher geometric mean fold rises from baseline and postvaccination geometric mean titers for the younger group. Immunogenicity in adults has been evaluated in the VARIVAX and ZOSTAVAX Program, including both seronegative and seropositive individuals, and the vaccine has been shown to be immunogenic in adults.

To lend further support to the utility of the vaccine, this slide provides a preliminary summary from a subset of subjects who were enrolled in a recently completed study that was not included in the original license application. And with all studies that have been initiated since 2003, this protocol, Protocol 010, enrolled subjects beginning at age 50.

The slide shows preliminary results for 113 subjects, 45 of whom are 50 to 59 years of age, 68 of whom are 60 to 69 years of age or, I'm sorry, 60 years of age or older. Note that in this study the postvaccination blood sample was obtained at four weeks postvaccination. Not surprisingly, the immune responses in the 50 to 59 group were as good as those in the 60 and older group.

So in summary, in the face of often high levels of preexisting immunity, ZOSTAVAX elicits an immune response by both gpELISA and ELISPOT. The VZV antibody response measured by gpELISA, a T-dependent phenomenon that reflects cellular immunity, correlates best among the assays evaluated with protection against herpes zoster.

I would like to move on now to the safety profile of the vaccine. It's important to remember that ZOSTAVAX is a high potency Oka/Merck VZV vaccine that builds on an extensive VARIVAX safety database. More than 56 million doses of VARIVAX have been distributed mostly in VZV naive individuals since the initial licensure of the product in 1995. VARIVAX has demonstrated an excellent safety profile in the ensuing ten years.

Within the ZOSTAVAX Program, the clinical evaluation includes over 20,000 subjects who receive vaccine and importantly over 19,000 placebo controls. So this assessment of safety was performed in a rigorous comparative setting that permitted a reliable enumeration of both common and uncommon adverse experiences. As shown here, the studies that were conducted had 97.5 power to detect an event with a rate of 1.8 per 10,000 and 80 percent power to detect an event with a rate of 0.8 per 10,000. The studies have demonstrated that ZOSTAVAX was generally well-tolerated in these older adults.

With regard to the safety evaluation in the Shingles Prevention Study, the following safety evaluation was undertaken for all enrolled subjects. Adverse experiences occurring day 0 to 42 were to be reported and assessed. Vaccine-related serious adverse experiences occurring at any time during the study were also to be reported, as were deaths at any time following vaccination.

As previously noted, all subjects enrolled in the study were to have contact shortly after day 42 postvaccination to ensure complete ascertainment of serious adverse experiences in the full cohort and 93 percent of them did by day 60 and 97 percent overall. The Adverse Event Monitoring Substudy again conducted at all of the sites and including over 6,600 subjects added two additional measures over and above the safety evaluation that was done for the overall population in the Routine Safety Cohort.

In addition to the standard safety evaluation that was on the prior slide, the subjects completed a diary, a vaccination report card through day 42 postvaccination. And in addition for this cohort, hospitalizations for any cause were to be reported through the end of the study. For the overall study population, the incidence of serious adverse experiences in each vaccination group was identical with a rate of under 1.4 percent.

In the Adverse Event Monitoring Substudy shown in the hash marks here, more serious adverse experiences were reported in the ZOSTAVAX group than in the placebo group, which was offset by a corresponding increase in the serious adverse experiences reported among placebo recipients in the Routine Safety Cohort.

Now, this table reflects the distribution of serious adverse experiences in the two safety cohorts. A review of the serious adverse experiences in the substudy found that no body system, no clinical syndrome, no diagnosis was responsible for this difference and there was no temporal clustering of these serious adverse experiences relative to vaccination.

Given the follow-up for and distribution of these serious adverse experiences in the overall population, the conclusion of the detailed review was that the imbalance and serious adverse experiences in the substudy was chance event. In further support of the safety profile of the vaccine, in the entire cohort of nearly 40,000 vaccinated subjects, there was a total of only five possibly vaccine-related serious adverse experiences reported. Two in the vaccine group and three in the placebo group. The number of deaths during both the first 42 days postvaccination as well as during the entire study were the same in the two vaccination groups and there were no vaccine-related discontinuations at all in the study.

In the Adverse Event Monitoring Substudy, the data recorded in the vaccination report cards demonstrated, not unexpectedly, that injection-site adverse experiences were more frequent in the ZOSTAVAX group than in the placebo group. The recording of intensity demonstrated that most of these, approximately, 85 percent were scored as mild by the subjects.

In this double-blind experience, the overall proportion of subjects with systemic clinical adverse experiences was the same in both groups, just under a quarter. An increase in vaccine-related systemic adverse experiences was observed in the vaccine group, but the rates in both the vaccine and placebo groups were low, about 6 percent in the ZOSTAVAX group and about 5 percent in the placebo group.

Among the vaccine-related adverse experiences only headache was seen more frequently in the vaccine group than in the placebo group. Hospitalization rates at any time during the study for any reason were comparable at 107 per 1,000 person- years.

I would like to turn now to Protocol 009, the safety study that was conducted at the estimated maximum vaccine potency. This double-blind controlled multicenter trial, which enrolled subjects 50 years of age and older, evaluated two lots of the vaccine that were administered at 58,000 and 207,000 plaque-forming units per dose. About 700 subjects were enrolled with nearly 200 of them 50 to 59 years of age. The vaccine was generally well-tolerated at both potencies as shown in the next slide.

Here we have the overall safety findings of the study by potency group and age cohort. The proportion of subjects with local adverse experiences was higher at the higher potency and the younger age cohort reported these local reactions more often than the older cohort, but these events were viewed almost exclusively as mild or moderate in intensity and of relatively brief duration of just a few days.

Importantly, the frequency of systemic adverse experiences was similar in the higher and lower potency groups and overall the vaccine was generally well-tolerated at both of the potencies administered. A relatively small number of VZV- seronegative adults has been identified and enrolled through our clinical trials. No seronegatives at all were seen among the 1,400 subjects in the CMI Substudy of the Shingles Prevention Study.

A different study, Protocol 003, was conducted in tropical countries specifically to enhance the potential for identifying VZV-seronegative adults, because published literature suggests that VZV-seroprevalence is lower and seropositivity obtained at a later age than in temperate climates. Despite screening over 1,100 individuals, few VZV- seronegative adults were found and enrolled.

In Protocol 049, from the VARIVAX Program, varicella history negative adolescents and adults were enrolled. Among these, 17 VZV-seronegative subjects 30 years of age and older were identified. In this small subset from the two studies, it appears that local and systemic adverse experiences as well as elevated temperatures occurred frequencies that are similar to those seen in VZV experienced individuals.

Importantly though, despite concerted efforts to identify such individuals, VZV- seronegativity is very rare among persons over the age of 30. Based on these findings, the criteria for enrollment in the ZOSTAVAX studies which did not screen for VZV-serostatus, there is no need to screen or otherwise assess pre-vaccination immune status in individuals who are otherwise candidates for ZOSTAVAX.

An adverse experience of particular interest in the ZOSTAVAX Program, as it has been in all of the varicella vaccine programs, was the development of rash within 42 days after vaccination. In that time frame, within ZOSTAVAX clinical trials, approximately, 0.3 percent of subjects reported a VZV- like rash across the database. A rate that is roughly 10-fold lower than that seen following administration of VARIVAX.

Those who developed VZV-like rash were requested to have sample lesions obtained for PCR analysis. Across the entire clinical database, two subjects with VZV-like rash were found to have the Oka vaccine strain in their lesions. Among the subjects in the Shingles Prevention Study, the Oka/Merck strain was not identified in any suspected herpes zoster case or in any postvaccination rash at any time point early or late in the postvaccination period.

So in summary, compared with placebo those who received ZOSTAVAX had a higher incidence of injection-site reactions, but a similar overall incidence of systemic clinical adverse experiences. The incidence of vaccine-related and systemic experiences was slightly higher among ZOSTAVAX recipients than among placebo recipients. Following a dose of ZOSTAVAX vaccine-associated rashes were uncommon and so we conclude that overall the vaccine had a very acceptable safety profile in those 50 years of age and older.

So to summarize, ZOSTAVAX is proposed for vaccination of individuals beginning at 50 years of age. Although the pivotal efficacy study enrolled subjects beginning at 60 years of age, the epidemiological data and the limitations of currently available therapies argue that there is a strong medical need to prevent herpes zoster and its complications starting at age 50.

Over 100,000 additional cases of herpes zoster and an additional 8,000 to 15,000 additional cases of PHN could be potentially prevented each year in a group of individuals who suffer as much acute zoster-associated pain as those 60 to 69 years of age. With the additional societal burden of being an age group in which a majority of the population is still employed, the data indicates that substantial benefit could accrue from vaccination beginning at age 50.

Efficacy for ZOSTAVAX had been demonstrated directly for those 60 years of age and older with a very high degree of efficacy against herpes zoster, 64 percent, among those 60 to 69 years of age. Efficacy in this age group should predict well the efficacy in persons 50 to 59 years of age. The vaccine has been shown to be immunogenic with generally comparable responses in the 60 to 69 and 70 plus cohorts in the Shingles Prevention Study.

Similar age-related findings were observed in other ZOSTAVAX studies. Most recently in the form of the supportive data that have recently become available from Protocol 010 for those 50 to 59 years of age. The vaccine has been administered in clinical studies to individuals 50 years of age and older and has been shown to be well-tolerated with only a moderate increase in transient injection-site reactions of mild to moderate intensity.

To conclude, in a very large clinical database ZOSTAVAX has been shown to reduce herpes zoster by one-half, reduce PHN by two-thirds and to reduce herpes zoster pain burden of illness by over 60 percent in older adults. The vaccine elicits a VZV-specific immune response, demonstrates efficacy that persists for four years postvaccination and has an excellent safety profile.

At this point, I would like to turn the podium back to Dr. Gutsch for a few concluding remarks.

DR. GUTSCH: In addition to the large and comprehensive database that went into the application for licensure, there are ongoing and future plans for further study of ZOSTAVAX that will shed light onto the vaccine performance. To answer the question what is the durability of ZOSTAVAX efficacy? There is continuation of the Shingles Prevention Study at 12 of the 22 original sites involving 7,500 subjects.

In addition, up to 18,000 of the placebo recipients in the Shingles Prevention Study and Protocol 007 are in the process of receiving vaccination with ZOSTAVAX and this will then provide further safety follow-up.

A clinical study is being conducted to assess a new formulation of ZOSTAVAX that allows refrigerator storage to increase the settings in which the vaccine will be available. Another study is being conducted to show that ZOSTAVAX can be administered concomitantly with inactivated influenza vaccine. Pharmacovigilance planning is important for a vaccine as it enters the postmarketing period.

A pharmacovigilance plan was developed that builds on the extensive VARIVAX experience with over 56 million doses distributed to the market and ZOSTAVAX for which a robust database has been provided in a licensed application. Proposed plans include extension of the postmarketing surveillance activities that are well-established at Merck for vaccines to monitor adverse events after licensure.

In addition, the VZV Identification Program determines by a preliminary chain reaction if wild type or vaccine strain, varicella zoster virus, is present in clinical specimens, such as vesicle fluid or cerebral spinal fluid from individuals with adverse experiences.

Finally, the Pregnancy Registry that was initiated with the VARIVAX Program in 1995 will also be applied to ZOSTAVAX in the postmarketing period. Collectively, results from our program indicate that the benefit/risk ratio for ZOSTAVAX is favorable. Herpes zoster and PHN are often debilitating diseases in need of better management. ZOSTAVAX would be the first intervention when licensed to prevent herpes zoster and its complications, including postherpetic neuralgia.

Beyond the benefit from preventing these two diseases, ZOSTAVAX also reduces the severe pain associated with herpes zoster and PHN. ZOSTAVAX has been studied in subjects 50 years of age and older and has demonstrated an excellent safety profile with no clinically important safety risks identified from a very large database of placebo-controlled clinical trials. So overall, the benefit/risk ratio is favorable and ZOSTAVAX, when licensed, will meet an important unmet medical need.

In closing, the proposed indications for ZOSTAVAX supported by the clinical data just presented are: ZOSTAVAX is indicated for prevention of herpes zoster, prevention of PHN, reduction of acute and chronic zoster-associated pain. ZOSTAVAX is indicated for immunization of individuals 50 years of age and older. Thank you very much. We can now entertain your questions.

CHAIRMAN OVERTURF: Are there questions from the Committee for the sponsors at this time? Dr. Wharton?

DR. WHARTON: I have a couple of questions about safety monitoring. I'm a little confused about the group who were in the safety substudy. Did they participate in the 42 day automated telephone call? From the first slide it seemed as if they didn't and then later it was stated that the supplemental safety monitoring was on top of other safety monitoring being done.

DR. SILBER: The question relates to the type of safety follow-up for the subjects in the Adverse Event Monitoring Substudy. And as initially designed, all subjects were to have the day 42 contact by phone or through other contact with the sites. And the vaccination report card was used for that subset of 6,600 individuals. It became apparent through frequent phone calls to the sites that the subjects found it a bit of an annoyance to have to go through all of this redundancy having already completed a 42 day diary card, so that the protocol through an operations memorandum permitted either of those contacts to be a suitable completion of the 42 day contact.

So the majority of those subjects who were enrolled in the Safety Monitoring Substudy had the vaccination report card in lieu of the phone call. But for those who didn't turn in the vaccination report card or for some who did both, there may have been, and actually in other respects, more than one form of contact. In the pie chart that was shown any given subject was only counted once with the vaccination report card and the ATRS being prioritized.

DR. WHARTON: Okay. And I want to follow-up to that. For the subjects who didn't have contact with the investigators within 60 days of vaccination, which I think were about 7 percent on your pie chart, when was information on those subjects attained and how was it obtained?

DR. SILBER: Yes, that was variable. First, I should say for the persons who were involved in the Adverse Event Monitoring Substudy, about 97 percent returned those vaccination report cards. For the remaining individuals, the small subset who had follow-up beyond the day 60, it was highly variable. Many of those were in contact by day 90, but some went on longer.

I recall that after this initial follow-up period for safety, the monthly contacts for efficacy were continuing and so there were reminders to the subjects on a monthly basis. Despite all of these efforts and despite the very careful attention to follow-up for efficacy and safety by the investigators, we still had 3 percent who ultimately had no follow-up for safety.

CHAIRMAN OVERTURF: Dr. Hetherington?

DR. HETHERINGTON: How are the patients in the Safety Substudy selected and recruited and how do they compare to the general population in the study?

DR. SILBER: Yes, thank you. That's a question about the selection of subjects in the AE Substudy. What happened was, as you can imagine, this was a very huge endeavor to undertake to have a study of this size at 22 sites. And the way that the Adverse Event Monitoring Substudy was conducted was that basically after the first several months that allowed the sites to sort of settle in with their procedures and do the routine activities, each of the sites was then asked to consecutively recruit the next 300 individuals into the Adverse Event Monitoring Substudy.

And so there was no cherry picking or preselection that happened. And then at the time when that cohort was filled, the routine cohort continued. And so that through the randomization and through the way that the timing intervals occurred, there were no differences, overall differences demographically or in other ways between the two cohorts.

DR. HETHERINGTON: Did the patients have a -- were they able to elect whether to participate or not? Could they decline the long-term or the Safety Substudy participation? And if so, what was the rate?

DR. SILBER: I believe I'm going to have to turn to Dr. Levin for confirmation. But my understanding is that subjects could opt out of the Adverse Event Monitoring Substudy and remain in the routine cohort, but the communications that I have heard from Dr. Oxman and Dr. Levin and others over the years is that it was a very small number who did so.

DR. HETHERINGTON: Could we get that exact number some time this morning?

DR. SILBER: We will certainly try.

DR. HETHERINGTON: And did you ever compare the demographics and the age distribution in that subgroup versus the general population?

DR. SILBER: They were similar.

CHAIRMAN OVERTURF: Dr. Farley?

MEMBER FARLEY: I have some questions about what is known about the more detail of the epidemiology in the 50 to 59 year-old group that has sort of been added into this expanded request. And do you know whether those who develop herpes zoster in that decade are more likely to be immunocompromised, more likely to be HIV with reconstitution syndromes or people on steroids or with malignancies and therefore might not be in the targeted group, at least initially for this vaccine?

And also, do you know anything about the epidemiology of zoster in the last decade of general use of the varicella vaccine? Has that had an impact not having natural chicken pox out there as much by far as previously and in terms of if there is a boosting effect of exposure as adults?

And finally, do you have any information about the durability of the immune response that you showed us the small numbers of 50 to 59 year-olds, because, of course, the issue will be will they not have the benefit when they are at maximum risk later?

DR. SILBER: Okay. So there were at least three questions there, so I'll try to take each of them. I'll start with the VARIVAX question and influence of varicella vaccination on the incidence of herpes zoster. In fact, the few population-based databases that have been available long-term suggested the incidence of herpes zoster has been increasing for at least the last 50 years. Both in terms of absolute numbers and in terms of age adjustment.

The data that are available thus far, realize are only 10 years out from onset of varicella vaccination and only about seven years out from widespread use of the vaccine. And so it may be too early to see anything, but the fact is that the studies that have been completed to date, some of them at CDC, at Group Health Cooperative out in Seattle have indicated that thus far there does not appear to be any differences in the incidence of herpes zoster with the use of varicella vaccine.

Now, there are mathematical models that have predicted that this will happen over time and that with less boosting from exogenous exposure, assuming that there is not endogenous boosting to make up for that, that the incidence of herpes zoster may increase. And, in fact, that the age of zoster may shift to an earlier age. But, at this point, it remains speculative and the available data do not indicate that this is happening as yet.

With respect to the demographics and characteristics of persons 50 to 59 years-old who develop herpes zoster, the fact is that the population has a higher percentage of immunocompromised with age. And so in terms of the overall population, there were fewer immunocompromised individuals in their 50s than in their 60s, 70s and later. There are few data looking specifically at immune status in the 50 to 59s, but such that is available suggests that the large majority of the cases of herpes zoster among people in their 50s are in the immune-competent population.

Your third question had to do with the durability of the immune response and the durability of protection that would ensue. And what we have seen with vaccination in terms of the CMI Substudy that went out to three years and also following an episode of herpes zoster actually, that there is an early and large increase, but that within six months the markers of immunity tend to decline and head back toward baseline, as one might expect even with silent exogenous or endogenous boosting.

And it appears that following an episode of zoster and following vaccination, people sort of settle out at a level and this is what one might expect when we are dealing with a memory response in the face of prior immunity. And so in terms of the actual immunologic markers, they do head back toward, but still remain above the baseline values. However, in terms of the vaccine efficacy, as one looks at year two, year three, year four, there was no decline seen in the point estimates for efficacy.

So at least from the clinical protection standpoint, even for those 60 and older, a decline in durability has not yet been observed. And one would expect that vaccination of an even younger adult cohort 50 to 59 durability would be at least as good as is seen with other vaccines.

CHAIRMAN OVERTURF: Dr. Markovitz?

MEMBER MARKOVITZ: Yes, I have two questions. The first one is in terms of safety in the 50 to 59 year-old group, unless I missed it, the only slide we saw was with these higher doses. And of course, there is a lot of reactogenicity, you know, systemic reactions around 40 percent in all age groups. But yet, it is stated that the vaccine was "well-tolerated." Can we have some elaboration on that? And then I have a second question.

DR. SILBER: Sure. If we can pull up the safety table?

MEMBER MARKOVITZ: 74.

DR. SILBER: Across the ZOSTAVAX Clinical Development Program, most of the studies were placebo- controlled. And as you might recall from the Adverse Event Monitoring Substudy in the Shingles Prevention Study, roughly a quarter of those individuals receiving placebo had one or more systemic adverse experiences. And so just to sort of put a frame of reference around it, that's what is seen after placebo injection.

And across ZOSTAVAX studies, we tend to see systemic adverse experience rates in this range. And as you see, we can look at it both ways across the potencies horizontally across the age groups vertically. There was really not a large difference across the potencies. In fact, in the younger age cohort, the rate of reporting of systemic adverse experiences was actually lower in the higher potency than it was in the lower potency.

And in this study the reporting rate was within a couple of percentage points of those for the 60 plus. And again, other than a slightly higher incidence of headache in the younger group, there were no differences seen by body system, by clinical syndrome, by any other diagnostic criteria and a very small percentage of any of these adverse experiences were rated as severe in intensity by the subjects.

CHAIRMAN OVERTURF: Just a point of clarification before Dr. Markovitz asks the second question. Both the higher and the lower potency dose in the 50 to 59 year-old age group are considerably higher than the overall dose that you are asking for approval. Is that correct or am I confused on that issue?

DR. SILBER: I'm sorry, the potencies administered?

CHAIRMAN OVERTURF: Yes, the potencies.

DR. SILBER: Is that what you are asking? Yes, the question is about the potencies administered here. The lower potency within Protocol 009 was 58,000 plaque-forming units and doses of around 50,000 plaque-forming units were the highest potencies administered within the Shingles Prevention Study, which had 12 different lots, and 50,000 was also the potency that was administered in several of the other clinical studies.

And so we selected a lower potency within this study really to help us benchmark to the other studies, because we didn't have a placebo control here. And prior to Protocol 009, the highest potency we administered in any prior clinical trial was 67,000 plaque-forming units, and so the higher potency group here was about a 3-fold higher potency that had been administered previously. And we don't expect that very many people would ever receive a potency this high out in clinical practice.

CHAIRMAN OVERTURF: In fact, I was wondering --

PARTICIPANT: I don't think that answered the Chair's question. The Chair asked a very important question here and I just --

CHAIRMAN OVERTURF: Yes. To me what you're asking for is approval of licensure for the 50 to 59 year-old age group, but you're asking for approval for a dose that is considerably lower than either one of these. Is that correct?

DR. SILBER: Well, the specification for this, as for other live virus products, is built around a minimum expiry, a minimum potency that would be observed at expiry and so the clinical experience from the time the vaccine is manufactured and released would be at a variety of potencies higher than that, and so there would be a spectrum of potencies administered.

And so much as the Shingles Prevention Study evaluated efficacy at the lowest, largely at the lowest potency, so we are looking at safety at the highest potency to provide a buffer, if you will, for what might be seen in terms of the efficacy experience on the one hand and the safety experience on the other hand when a vaccine might be administered in practice. Did that answer it not?

CHAIRMAN OVERTURF: We may come back to this issue, but let Dr. Markovitz ask.

MEMBER MARKOVITZ: Yes. Actually, that's what I was trying to get to also because, essentially, most of the data you showed us for the older people, meaning 60 and over, were based on I believe a 19,000, was it, plaque-forming unit dose and here it's much higher.

So I guess what I'm really asking or suggesting is you have no data that deals with the actual vaccine going into 50 year-olds in terms of safety, is that right, or you haven't presented it?

DR. SILBER: No, this -- I'll take a step back on this. The Shingles Prevention Study included 12 lots of vaccine that ranged in potency from roughly 20,000 plaque-forming units up to roughly 60,000 plaque-forming units. The dossier included a number of other studies, including but not limited to Protocol 009, that also used vaccine at a range of potencies up to 67,000 plaque-forming units.

Across that dose range, let's take it separate from Protocol 009 for a moment, across the dose range seen and actually reflecting what has been seen for many years with VARIVAX across a wide range of potencies for seronegative individuals is that other than a potency-related increase in injection- site reactions, no difference was seen in the safety profile.

So that is what was seen for all of these other studies and, in fact, the dose selection studies prior to the Shingles Prevention Study were looking at whether there was any dose-related effect.

Jump now to Protocol 009 and, again, to create the highest hurdle, if you will, for safety because it is at the maximum potency that someone might expect, this is what was seen. And, again, with systemic safety consistent with what was seen in other trials with local reactions at a somewhat higher reporting rate.

MEMBER MARKOVITZ: But these other protocols actually dealt with 50 to 59 year-olds before, too?

DR. SILBER: No. The shingles?

MEMBER MARKOVITZ: When you talk about that you tested a wide range of doses of the vaccine, what I'm trying to understand is what percentage of those people who got vaccine, what's in the range of what is going to actually go into people in a clinical setting, were in the 50 to 59 year-old range?

DR. SILBER: Right. The data that were included in the original dossier included, approximately, 200 to 300 subjects, most of them from Protocol 009, some of them from Protocol 049 that was alluded to, which was actually a VARIVAX protocol but included what for VARIVAX is a high potency lot that was actually used in the Shingles Prevention Study at, approximately, 50,000 plaque-forming units also.

And so in terms of the database within the dossier, again because enrollment starting at age 50 only began in the studies in 2003, that is what is in the dossier. The studies that have been conducted since, including the Refrigerated Vaccine Bridging Study and the Influenza Concomitant Use Study, both of which are at or near completion, also included the vaccination of individuals beginning at age 50 and has gone through a range of safety doses.

And, again, younger individuals in general seem to report adverse experiences more often than the elderly. And so with this population and with the highest maximum potency, we consider that the data here provide comfort that vaccine administered at lower potency than that maximum would be with an acceptable safety profile for the age group.

MEMBER MARKOVITZ: What will be the actual dose of the vaccine going into people in a clinical setting if license is granted?

DR. SILBER: Do you want to answer that one? I think I will turn to Dr. Gutsch for a moment to answer that.

DR. GUTSCH: Just for clarification, this is a live virus vaccine so there is a shelf life during which the vaccine has a decaying potency that occurs over time just due to storage conditions. And so there is not one dose that anyone is going to get at any given time.

What we want to assure is that at the expiry potency, the potency at the very end, that it never goes below that so that we have an efficacious vaccine. But we have to put sufficient virus in there of this live virus to ensure that at the end of this shelf life, there is sufficient left over. And in order to do that we need a little bit of a range there and, therefore, we test the lower extreme for efficacy and the upper extreme for safety.

So we can't really say that you're going to get one specific dose, but the label indicates that you will get greater than 19,400 plaque-forming units. Does that clarify things?

MEMBER MARKOVITZ: Yes, thanks, that does.

CHAIRMAN OVERTURF: Dr. Karron?

MEMBER KARRON: So I have three sets of questions, one related to dose and potency, one related to safety and then finally to the gpELISA. So, first, I just wanted to know in terms of understanding about dose, actually really two sets of questions, one to follow-up on Dr. Markovitz'.

So in the VA study, were there multiple lots of vaccine used and were they of different potencies? And in terms of looking at efficacy, did you stratify by potencies to see if there was any difference in efficacy according to potency of vaccine?

DR. SILBER: Shall I answer that one first?

MEMBER KARRON: Yes, yes.

DR. SILBER: Okay. The question had to do with the lots that were used in the Shingles Prevention Study. In all there were 12 lots that were used, as I mentioned previously, at release potencies ranging from 20 some odd thousand up to about 60,000 with doses administered going from about 18,700 up to close to 60,000.

Included within those 12 lots was a three lot consistency series. What we found, what was found across the lots and across the potencies, and the trial was not powered based on the efficacy endpoints to demonstrate formally consistency, but when one looks at the lots by potency across the 12 lots and, in particular, when one does a comparison pair-wise within the three consistency lots, in both cases there were no differences observed with respect to efficacy for any of the three endpoints and with the potencies.

MEMBER KARRON: Okay. And then to follow-up on that issue, when the original studies, the dose ranging studies, were done to choose a dose to go forward with, had you done that looking over an age range?

Did you look at the young elderly and the very elderly in terms of making that decision?

DR. SILBER: Yes. The early dose selection studies that included the immunogenicity assessment were done in individuals 60 years of age and older. There were a couple of earlier pilot studies that enrolled people beginning at 55, but Protocols 001 and 002, which were the ones that were alluded to in the Clinical Development Program overview slide that I gave, included individuals 60 years of age and older.

At that point, the interferon-gamma ELISPOT assay was not operative. There were a number of cytokine ELISAs, the responder cell frequency. A number of candidate markers were used with slightly different patterns in terms of when immunity developed, but across the assays that were being used at that time it was in the range of 17,000, 19,000 for two of the potencies that we did administer that we started to reliably see across the different markers that there was an immune response.

MEMBER KARRON: Okay. And I just want to understand better the rationale for the high dose study in the younger individuals, so somewhere between 5 and 10 times the dose, the minimum dose, the minimum 19,000 PFU dose.

I mean, is that because there are potentially future plans to try to have a higher potency vaccine? What was the rationale for that higher dose?

DR. SILBER: The question surrounding the rationale for Protocol 009 was really just to frame at the very highest end of what might be manufactured, because the manufacturing will be targeted, again, to ensure that all doses administered within human certainty will be above a certain minimum potency, that there will be a target and variable potencies, as Dr. Gutsch had alluded.

And, again, this is really just to frame what could be acceptable at the very high range. If you're asking if we're specifically targeting 200,000 plaque-forming units as a dose for this or future studies, the answer is no.

MEMBER KARRON: Okay. Okay. So those were my potency questions. Safety questions. I was actually wondering if we could look at that slide 36 again, which is the pie diagram? I don't know if it's possible to pull up.

DR. SILBER: Sure. Can we get 36? Thank you.

MEMBER KARRON: Okay. So I think if I understood the briefing documents correctly -- so, first of all, I want to clarify. The blue are the individuals in the detailed Safety Substudy?

DR. SILBER: That is correct.

MEMBER KARRON: Is that right? Okay. If I read the pie, the briefing documents, correctly, I think that there are about 25 percent of the people in this large study for whom I would say that safety data collection was not absolutely optimal. It was after day 42. In some cases is was well after that time. It had to be sort of sought by study personnel and such.

And my question is for those people not in the blue and the green --

DR. SILBER: Sure.

MEMBER KARRON: And I would sort of like detailed information about this. Are they comparable in terms of age, in terms of underlying conditions, in terms of whatever demographics we can measure to the people in the blue and green?

DR. SILBER: Sure.

MEMBER KARRON: Whom I do think we have good safety data on.

DR. SILBER: Okay. So the question has to do with the characteristics of those people who had other than an ATRS contact by day 50, other than a vaccination report card.

Before answering that, I would just like to turn attention to particularly this magenta or purple and gray. These two bars or two pieces of the pie represent over 20 percent of the individuals in each group, and these represent either the staff calling ATRS, basically on behalf of the subject, due to some contact or the staff following up on an ATRS fax, because again at day 50 or 51 it is shut down.

And just as with the vaccination report cards, not everybody comes in exactly on day 42. It might be sometime later. Out of these 21, 22 percent of individuals in the gray bars, the timing of when those contacts took place is known and roughly 80 to 90 percent of those additional subjects also had that contact prior to day 60.

So it could have been for any number of reasons. Without going to the ATRS, a subject may have called the site directly to say, hey, I had this really bad sore arm or, hey, I remember that I was supposed to check in six weeks later and I'm calling you or I have this rash, maybe it's shingles, can I come in? There are going to be all different ways in which this sort of contact might have occurred, and so been done in lieu of the green or the blue.

So having said that, I don't have at my fingertips any of the demographics or the characteristics of the persons in the bars, in the pieces other than green or blue, but we could check on that.

MEMBER KARRON: I think that would be very helpful. My last question actually just has to do with the gpELISA and the comment that that's the best correlate protection. And my question is really so when we look at responses in the sort of younger, the under 70s and the older 70s, in fact, over 70s have higher titers.

So what my question is is so does it correlate best, say, with protection against postherpetic neuralgia or burden of illness as opposed to incidence because, in fact, the incidence efficacy is much less in the over 70 group?

DR. SILBER: Sure, yes. The question has to do with the surrogacy of or, I'm sorry, the correlation, there is no surrogacy, the correlation of the gpELISA.

Now, recall that these assays were conducted just on the CMI Substudy representing 1,400 of the 38,000 individuals enrolled in the study. And so there were relatively few clinical endpoints among the primary efficacy endpoints that occurred among the individuals within the substudy. Unfortunately, there was no blood collected from the other 37,000.

And so based on that, and I can't recall, I think there were only one or two or just a handful of PHNs at all across either of the treatment groups within the substudy. So the analysis looking at the correlation with protection was built on the protection against the incidence of herpes zoster.

CHAIRMAN OVERTURF: Dr. Fleming?

DR. FLEMING: I'm trying to pare down questions here. There are three areas of questions I would like to pursue, the first relating to generalizability, the second relating to safety, the third relating to the BOI for efficacy.

Starting on generalizability, your label is very broad. You're asking in slide 5 for immunization of individuals over age 50 and then note in the EpiData that the only known risk factors are age and immunosuppression. And, yet, you completely excluded patients, for example, that were on regular use of inhaled corticosteroids.

You have also excluded other high risk patients, patients that are homebound or non-ambulatory or have cognitive impairment and your representation of those patients over the age of 80 is only 2,500, a very small fraction, a fraction that in fact shows very little efficacy when you look at the age relationship to efficacy, and a group that shows on slide 71 to have a particularly higher excess rate of SAEs.

So the first of the generalizability questions is why such an under-representation of those very groups that your risk factor analysis says are the people in greatest need?

DR. SILBER: Okay. That was a dense question. Can I answer that one before any other questions? Okay. The clinical studies for ZOSTAVAX, as for the other live virus vaccines, have routinely systematically included or excluded, excuse me, those with known immunosuppression.

And so with respect to generalizability, the studies were conducted in this way and the proposed package circular that has been submitted would contraindicate the vaccine in those with known immunodeficiency, as is consistent with a label as it had been for VARIVAX. And particularly because of the high potencies administered, there were potential safety concerns with using a high potency vaccine in individuals with known immunodeficiency.

Having said that, there were a small number, a handful, of subjects who did enter the clinical trials with cancer or on steroids and one cannot infer anything definitive, obviously, from just a handful of subjects, but no safety concerns accrued.

The second, there were a number of individuals in the Shingles Prevention Study and in the other studies who developed immunocompromising conditions or required corticosteroids or other immunosuppressors shortly after vaccination, and no adverse experiences were noted there over and above what was seen in the general population.

Third, I would like to go back again to VARIVAX where the vaccine has been used in a very large experience over many years. Recall that although not indicated as such in the United States, the vaccine was initially developed for use in leukemic children and through studies that have been conducted by ACTG and others, the vaccine has demonstrated a very acceptable safety profile in immunocompromised populations.

And so in terms of that particular aspect of generalizability, what we are proposing for our initial package circular is consistent with what we have studied. In fact, we will, beginning in 2006 now that we have analyzed the full safety database for immunocompetent individuals, be looking judiciously to expand the populations for whom the vaccine would be developed, but those are for future studies. That is not for now.

In terms of the other exclusions from the Shingles Prevention Study, those were in some cases due to the immunodeficiency exclusion criteria, but several of the other criteria were more practical considerations for that study only given the fact that people needed to have frequent contact with sites, needed to have long-term follow-up, needed to get back and forth.

And so the enrollment for that study was built around largely ambulatory subjects. There were two sites that did some recruitment and enrollment at nursing homes. We went back and tried to verify exactly how many and who these were, could not get exact numbers.

But separate from the ambulatory issue, per say, I want to go back to the functional status that was evaluated and, based on a functional status measure taken at baseline, there were, approximately, half of the enrolled subjects who were mildly, moderately or severely limited at baseline, about 10 percent moderately or severely limited at baseline, and the profiles that were seen were largely the same across those groups.

DR. FLEMING: Let me ask for maybe more concise answers, because I have got several questions and I know time is limited. But the bottom line is it's unfortunate not to have more data on these particularly important high risk groups.

I am confused about the exclusion of the 50 year-olds. When you were giving slide S-13 and S-39 you were justifying their exclusion logically when the study was designed in part because, as your data do show, PHN risks are very low until you're age 60. But your closing slide, S-79, then says EpiData strongly established the clinical need above age 50. It seems inconsistent.

DR. SILBER: Yes. The question is about age 50 and the decision to not enroll or not target age 50 initially, but to target age 50 subsequently, is actually consistent with the scientific data knowledge regarding --

DR. FLEMING: Well, basically, you only thought you were going to affect pain at the beginning and then you actually more affected zoster risk and, hence, now you believe that 50 years-olds are going to potentially benefit? Is that the concise answer?

DR. SILBER: Let's separate out the medical need on the one hand and then the ability of the vaccine to meet that need on the other hand.

Several population-based epidemiologic studies suggest that on the order of 200,000 episodes of herpes zoster occur each year in this country among people 50 to 59 years of age. The acute episode for people in their 50s is as severe, requires as much medical therapy, requires as many doctor visits and results in on average five work days lost. So based on the magnitude of that, the need is there.

DR. FLEMING: Could you put up S-13? I need a much more quick answer, S-13. As you were describing when you designed the trial and you were focusing on PHN and severity of zoster cases --

DR. SILBER: Correct.

DR. FLEMING: -- you noted that risk is small below age 50, below age 60. That is still, in fact, what you would view to be the truth, correct?

DR. SILBER: That is correct, that the risk of PHN begins to rise substantially at 60.

DR. FLEMING: Next question. Why only 2 percent blacks?

DR. SILBER: 2 percent blacks? Is that what --

DR. FLEMING: Right. Why are there only 2 percent in the study population, blacks?

DR. SILBER: Yes. The Shingles Prevention Study was open to the general population and was advertised in the general community, and this is something that we acknowledge and have been making efforts to increase enrollment of minority populations in studies. It was a bit of a surprise to us as well, I must admit, and the recently completed studies have included substantially more minority individuals.

DR. FLEMING: Two more very quick questions. One safety question and to make it short, I will just ask for some data we can be presented later on. When we look at the Safety Substudy, there is a 60 percent relative increase in SAEs and it's an 80 percent relative increase in SAEs when you're above age 70.

Could we get a summary of what those SAEs are? And, secondly, could we get a summary of overall hospitalization by arm for the entire cohort of 38,000 and, specifically, zoster-related hospitalization by arm and serious morbidities by arm? If we could get those data, that would be helpful.

My last question up front here relates to the burden of illness score. My understanding is you're looking at, in essence, the average. You look at that burden of illness score and the duration of time that you have that score and the product then, in essence, gives you that total burden.

So if somebody had a score of 4 and somebody had a score of 3 and they were the same duration over 182 days, then that would be a ratio of 4:3. If somebody has a score of 3 and somebody else has a score of 2, it should be 3:2.

But if, in fact, someone has a score of 3 over the 182 days and someone has a score of 2 over the 182 days, does that come out as 3:2 or is the 3 counted for all 182 days and the 2 only counted for 30 days, in which case you're getting a radical misrepresentation of that 3:2 ratio?

DR. SILBER: The question relates to the use of the BOI and the BOI was designed to look specifically at pain scores of 3 or higher because of the validation that suggested that pain scores of 1 or 2 were not clinically meaningful in terms of daily activity. So, yes, scores below 3 --

DR. FLEMING: So is what I'm saying correct?

DR. SILBER: -- were not included in the scores, the overall scores for the vaccination or placebo groups.

DR. FLEMING: So you impute a score of zero after day 30 for people whose scores aren't above 3. Is that correct?

DR. SILBER: That is correct.

DR. FLEMING: And could you show us on S-52, you come up with a P of 008 for the contribution, in essence, of this burden of illness beyond the actual reduction in incidence of herpes zoster cases.

What is your statistical test here that you're using? Is it one that, in fact, exaggerates emphasis of the right hand tail and how do you justify that you have integrity of randomization since it appears this is an initial analysis?

DR. SILBER: Sure. Okay. I would like to call Dr. Chan, our biostatistician, to talk about the statistical methodology here.

DR. CHAN: So the question is relating to the supplementary analysis result, the severe duration among zoster cases as shown in slide 52. First of all, all the pre-specified analysis would support the indication based on the intention-to-treat analysis. So we did this pre-specified analysis just as a way to quantify the -- a second component about the duration of pain. And this, obviously, was done based on post-randomization comparative --

DR. FLEMING: What's your test statistic?

DR. CHAN: The test is based on a normal approximation stratified by age based on comparing the mean between the two groups. And in that comparison, we also adjust for the age between the two groups.

DR. FLEMING: And how do you address the fact that this is not based on all people? This is based on a conditional sub-cohort of people that, in fact, had diagnosis of HZ. Integrity of randomization doesn't hold here. What's the validity of your P-value?

DR. CHAN: That's true. So what we have done is also done a couple of the sensitivity analyses that, one, are based on the bootstrapping techniques and that sort of doesn't take into account the distribution of some things.

DR. FLEMING: Well, that doesn't address the issue of integrity of randomization. Dan, did you want to comment?

DR. SCHARFSTEIN: Do you have data showing the --

CHAIRMAN OVERTURF: Please, identify yourself and use the mike.

DR. SCHARFSTEIN: Sorry. Do you have data presenting the demographic characteristics or health status characteristics of these two groups?

DR. CHAN: Yes, we --

DR. SCHARFSTEIN: So you can present?

DR. CHAN: Yes. The question is whether we have demographic data about the zoster cases and we do.

CHAIRMAN OVERTURF: I'm going to ask that we suspend discussion after he answers this question. If you want to wait until this afternoon when we have time to address it, that would be fine.

DR. SCHARFSTEIN: Yes.

DR. CHAN: Can we -- can I have slide 1015? This slide shows the demographic characteristics by trimming group among the zoster cases that developed. And as you can see, there is slight imbalance in the age distribution between the two vaccine groups. Apart from there, all the other characteristics are very similar across the trimming group. And one thing to know, obviously, the age imbalance, actually what you can see is there are about 61 percent of older individuals in the vaccine group that have zoster compared to 48 percent.

So that in effect, there is actually in the comparison that was performed that is reflected in the other slide that comparison takes into account the age differential in there.

DR. SCHARFSTEIN: Are there other clinically relevant factors that you are not including on this slide that could explain differences between populations?

DR. CHAN: The questions are there in the clinical characteristics.

DR. SCHARFSTEIN: But relevant characteristics you're not including on this slide that could explain differences between these two cohorts.

DR. CHAN: We haven't identified any other characteristics that are different between the two groups.

DR. SCHARFSTEIN: Can I just ask one more important question about this? I'm sorry. You considered cases as being evaluable or non-evaluable. The rate at which you considered cases as being non-evaluable was much higher in the ZOSTAVAX group as opposed to the placebo arm. Can you explain?

DR. CHAN: Yes. For that question, I would probably turn to Dr. Silber.

DR. SILBER: Now, this question relates to the rate of non-evaluable or cases that were not found to be evaluable cases of herpes zoster in the Shingles Prevention Study. And I think that it would be useful to turn quickly to slide 45, please. What one sees is that the fraction of cases that turned out not to be herpes zoster is the same in the two groups if one uses as the denominator the entire population who would come in with anything.

The ones who develop suspected herpes zoster, in fact, were more likely to be in the placebo group than in the vaccine group. So the fact that the rate of non-evaluability is higher in the vaccine group reflects the smaller denominator, because the vaccine was, in fact, efficacious.

CHAIRMAN OVERTURF: I'm going to ask that we go ahead and take a break, because we're a little bit limited on time, I ask everybody to be back at 15 minutes after the hour for the FDA presentation. Thank you.

(Whereupon, at 11:09 a.m. a recess until 11:24 a.m.)

CHAIRMAN OVERTURF: I would like to ask the Committee Members to, please, take their seats, members of the audience, sponsors, please. Dr. Rohan will provide the FDA presentation.

DR. ROHAN: Good morning again. I would like to just give you a brief overview of my presentation. We will discuss the proposed indication, a little bit about the introduction and background behind this disease, the ZOSTAVAX Clinical Development Program and particularly we will discuss Protocol 004, the pivotal study and Protocol 009, finish with a summary and then presentation of the questions, which will be discussed later today.

I'm going to kind of skip over some of the slides, Dr. Gutsch, Dr. Silber did a great job in presenting a lot of the background, the indication as well, and just point out that, you know, the very serious problem with postherpetic neuralgia, the pain that can be debilitating and it can last for months or a year or longer, particularly in the oldest age cohorts and that pain control may often be inadequate in those with the most severe cases and there may be complications or side effects from the treatments as well.

VARIVAX was licensed in 1995 and I might be able to answer some of the questions here, but I know that we have some CDC colleagues as well present in the audience. By 2003, the United States had achieved an 85 percent vaccination rate nationwide in the population for whom this vaccine is recommended. At the same time, CDC had been monitoring varicella zoster virus disease and had seen a decrease of primary varicella infections over that same period by, approximately, 85 percent.

And I would like to point out that the epidemiology of this disease may be changing and that future adult populations, these young people that are coming up that have had vaccination and we don't have circulating varicella zoster vaccine, chicken pox, out there, they may be relying on vaccination for protection from primary VZV infection.

In addition, the CDC has been interested in rates of herpes zoster. Again, we discussed earlier that there is a concern that if circulating varicella zoster virus is not out there in the community, we may see some impact on the HZ incidence as well as its manifestations. We saw that from 1999 to 2003, age standardized rates for overall herpes zoster have increased in the adult population and that upward trends in both the crude and adjusted rates were both statistically significant with P-values of less than .001 in specifically the 25 to 44 year age group and the 65 year and older age groups.

And this is from Dr. Yih and Dr. Seward, who I believe is here today, too. This is just a schematic you've seen before that older age groups have higher rates. This is a bar graph of some of the data that Merck presented earlier. I'm showing the large number of elderly subjects with pain with duration of at least one year or 6 to 12 months, 1 to 6 months and at least one month. And I would say that, you know, this data is derived from a study that was published, it was noted earlier, in 1957. And I don't think you will see this high rate of pain with long duration from the studies that we will be discussing later.

This is an overview of the ZOSTAVAX clinical development. As I mentioned, VARIVAX was licensed in 1995. Lydick published an article in 1995 comparing subjective responses to area under the curve based on the brief pain inventory measure. The ZOSTAVAX IND was actually submitted in the fall of 1996. The last vaccination in the pivotal study Protocol 004 was administered in September of 2003.

The last case of herpes zoster was accrued in that pivotal study in the fall of 2003. We would note that the PHN definition was changed from at least 30 days to at least 90 days in December of 2003. Protocol 004 was completed in April of 2004. The incidence of herpes zoster, the duration of herpes zoster pain and the interference, significant interference with activities of daily living were all endpoints that were elevated from tertiary to secondary endpoints and success criteria were provided in June of '04.

Protocol 009 was completed in June of '04. The publication validating the use of the HZ BOI instrument was published in August of 2004. And in April of this year, the ZOSTAVAX BLA was submitted. This is a comparison of the dose ranges for VARIVAX, which is licensed for primary infection versus ZOSTAVAX. A summary of the clinical trials that are included in the BLA and again the focus will be on the pivotal study, 004. That's the study that has efficacy data.

I can probably go through this quickly, but the Committee can stop me if I'm going too quickly to try to not be too redundant with the previous speakers. The primary objectives were to reduce the incidence and severity of herpes zoster in those at least 60 years of age as measured by the BOI and to reduce the incidence of PHN.

Secondary objectives included reduction in the incidence of herpes zoster, reduction in the duration of HZ pain and reduction in interference with activities of daily living in subjects with HZ.

The ZBPI was published, as I mentioned, in 2004. It was based on 121 subjects who were enrolled within 14 days of the onset of their rash and showed that the ZBPI severity duration associated with severity duration as measured by ADLI and worsening of quality of life measurements, also that a score of at least 3 on a 10 point scale occurring at 90 days or more after the HZ rash had high agreement with pain worse than mild using a modification of the present pain intensity scale taken from the McGill Pain Questionnaire.

As you can see, there were nearly equal randomization in the older and younger age cohorts. Although, the majority in the older age cohorts were in the younger range of that, 70 and above. There were 12 clinical lots. Nine were accelerated-aged to mimic end-expiry potency.

Pertinent exclusion criteria included any subject that had more than intermittent use of topical or inhaled corticosteroids, life-expectancy less than five years, bed-ridden or homebound, cognitive impairments or severe hearing loss and those two latter criteria were not specifically defined and weren't specifically measured or tested.

This is to point out that this study employed the 12 clinical lots. When you entered the study, you weren't necessarily randomized to one of the 12 lots. They were ruled out in sort of a dose de-escalation fashion, if you will. However, I would point out that the group 2, 3 and 4, which each include three of the clinical lots, these were the accelerated-aged.

We don't really know how the parental lot and the aged lot compare. So by saying that the nominal potency was 50,000 to 62,000 PFUs in group 1 versus 21,000 to 26,000 in group 4, we don't know whether the effect of what the parental lot was measured at may have an impact, because these are measuring live virus, but there is also a proportion of no longer living virus in the lots.

You can also see that the follow-up time is different. That the number of subjects enrolled in each of these different sort of dose ranges are different. The CMI was combined to the last group 3 and group 4. And there were different proportions of subjects in each of these groups across the dose ranges enrolled into the Adverse Event Monitoring Substudy as well.

We talked about the vaccine report card, the automated telephone response system. We also talked a little bit about the HZ rash onset, which then triggered additional follow-up, immunogenicity and other test instruments, particularly the IZIQ and ZBPI, which measured the pain severity, which was used to calculate the BOI.

We talked about the populations used and analyses as well. I would point out that one of the very nice aspects of the study is that all potential HZ cases were evaluated by the Clinical Evaluation Committee and then the analyses were compared using these different populations and that the results were very similar. So that subjects that were evaluated and determined to have HZ by clinical laboratory methods, PCR or culture, were also evaluated in a blinded fashion by the CDC and the results were really quite similar.

We can see that the groups were balanced in terms of gender, race and age, although it says previously mentioned, there are very few non-white subjects in this group, in this study. There was a high proportion of follow-up. Only 0.3 percent and 0.2 percent respectively of the subjects were lost to follow-up at the end of an average of three years of follow-up.

The PCR detected not only wild type VZV, but it also could detect the Oka/Merck attenuated strain in the vaccine and HSV. And as previously was mentioned, no Oka/Merck attenuated strain was isolated from any of the lesions in this study. You can see that the majority of the cases that were determined to be evaluable HZ were determined by PCR and very few by viral culture and the remainder by the Adjudication Committee.

The co-primary endpoint, co-primary is used in the study not meaning as might be thought that you would have to win on both endpoints. It's really an alternative endpoint so that you could win on this endpoint or you could win on the other co-primary endpoint and the study would be declared a success.

There was a decrease in the HZ burden of illness of 61.1 percent in the ZOSTAVAX group. There was a decrease in incidence of PHN of 66.5 percent in the ZOSTAVAX group. And as I pointed out, the definition of PHN was changed after the last HZ case. The secondary endpoints included a decrease in incidence of HZ of 51.3 percent. And this endpoint was elevated from a tertiary to a secondary endpoint after the last HZ case was approved, but prior to formal unblinding.

The duration of clinically significant pain was found to be 20 days in the vaccine group. 22 days in the placebo. It was using clinical scores of at least 3 on a 0-to-10 point pain scale. The P-value was less than 0.001 in the MITT group overall. The P-value was 0.041 in evaluable cases only. And again, this was an endpoint that was a tertiary endpoint, but was elevated to a secondary endpoint after the last case of HZ, but prior to formal unblinding.

And this is a statement from the Clinical Study Report and I think this to me sums up the issues, the impact, all the endpoints. This is related to the secondary endpoint substantial interference with activities of daily living or SADLI. "Because Substantial ADLI can only occur among HZ cases, the benefit of vaccination in reducing the incidence of Substantial ADLI was confounded by the benefit of vaccination in reducing HZ incidence."

As you can see, the rates were 36.2 percent in the ZOSTAVAX group versus 39.4 percent in the placebo group with an 8.2 percent reduction beyond the reduction in HZ incidence with a non-significant P-value. And as I mentioned in the last slide, this is the one endpoint that does not include the impact of HZ incidence on the vaccine effect of this endpoint. It was also elevated from tertiary to secondary endpoint after the last HZ case was accrued.

The sponsor did a number of sensitivity analyses and modeling very nice to develop more information about something we don't know enough about yet. And as you can see, the thing that stands out is that age is the big factor in the severity-by-duration scores. Obviously, analgesic use, but that's sort of not considered causal and that the vaccine versus the placebo group had a significant P-value as well.

There was a question about immunosuppression in the two groups. And as you can see, they were fairly well-balanced as far as difference, causes of potential immunosuppression. I think if you look at the rates at the top, this is sort of the flip side, the issue with looking at rates when you already have a big difference in the incidence of HZ. You have a higher rate of subjects that are immunosuppressed getting zoster, herpes zoster in the ZOSTAVAX group. It doesn't mean that the vaccine is causing that, it's just that you have fewer cases, so the rate is higher in this case, in this group.

This is a table showing some analyses by the sponsor looking at the efficacy for the major endpoint of HZ BOI by year. And, obviously, up to the first year, the subjects aren't randomized, but you can see that there is a decrease in the differential between the placebo and the ZOSTAVAX groups over time. You can also see that after year three, year four, half of the subjects are no longer in the population and the follow-up is maybe about half a year or a little over half a year. And then in the fifth year, there are even relatively fewer subjects and even relatively less follow-up.

You can see with PHN again a decrease over time in the comparison between the two groups for this endpoint. And finally, you can see HZ efficacy, which was the secondary endpoint, you see a decrease after the first year, but then it seems to not be dropping, just from a non-statistician's perspective, as much as the others.

This is a somewhat truncated summary of the mean worst pain in both of the groups over time. And you can see that the numbers are fairly similar. On day one, the rash onset there are not too many subjects in that group relative to the other time points. Most people were seen relatively rapidly after the onset of their rash. But you can see that day 2, 3, 4, 5, 9 to 11, week 4, 6, 8, 12, 16 and 26, there are not huge differences on that 10 point pain scale.

This is looking at the effect of age on efficacy. We can see that there is some decrease in the 70 and above age group in terms of the BOI endpoint. The incidence of PHN is similar, but the incidence of HZ appears much lower in the group that's 70 years of age and older.

FDA did a number of exploratory analyses. We wanted to try to see if we could understand further this difference in the incidence of HZ efficacy endpoint. And this is exploratory and the numbers as you get into the oldest ranges are very small, but you see a consistent decrease, a trend towards decreasing efficacy as you increase age.

We also looked at the major endpoints looking at the impact of the vaccine beyond its affect on the incidence of herpes zoster. The median HZ BOI score among the HZ cases were fairly similar and didn't appear to be statistically significant. The percent of HZ cases with PHN slightly higher in the placebo group, but did not appear significant. And the duration of clinically significant pain did not appear significant, either.

This is a graph looking at the rates of the BOI among the HZ cases, not the rates, excuse me. And you can see that among cases those in each -- instead of looking at numbers, you are looking at proportions of subjects with HZ among each of the treatment arms and they look relatively similar.

Comparison of the BOI between a vaccine and placebo group, I would like to point out that the median HZ BOI among the HZ cases is very similar and not significant. And using a variety of approaches, the comparison of the BOI between the placebo and ZOSTAVAX group among the HZ cases, except for the Log-Rank, the age-adjusted P-value, the other P-values don't appear significant.

Comparison of the PHN incidence between the vaccine and placebo groups, the percent of PHN among HZ cases 8.57, 12.5, this was presented in an earlier slide. And this is looking at the distribution of the BOI scores between the placebo group and the ZOSTAVAX group. To get an idea, very few people had high scores. Very few people had scores, you know, as time went on. Most of the cases resolved. So, you know, when you are looking at 90 or 120 days, you are looking at a relatively small proportion of subjects.

Comparison of the BOI among the PHN cases, you can see that the median HZ BOI among the PHN cases again is not that different and it's not statistically significant. Comparison of the BOI between the placebo and ZOSTAVAX groups among the PHN cases using a variety of approaches, the P-values, even age-adjusted, don't appear significant.

Now, I would like to switch gears and look at the immunogenicity. There was a lot of very interesting data. The sponsor used three different assays, but it didn't seem the Responder Cell Frequency or the gamma interferon ELISPOT provided any additional data, at least at this point, compared to the gpELISA. So I'm going to focus and limit my comments to the gpELISA data.

Here we can see the various -- these are the clinical lots that were -- then the data were pooled between two of each of these lots to represent the parental lot from which these age lots were derived. Then the efficacy data from the paired lots were then combined and then compared pair-wise representing the three parental lots for the lot consistency.

No differences were seen. There were based on clinical efficacy endpoints. As well, this was not part of the endpoint analysis, but you can see that the geometric mean fold rises were similar between these accelerated-aged lots.

Looking at the gpELISA by HZ status, you can see that in people that developed HZ, whether you are in the ZOSTAVAX group or the placebo, the GMTs at 6 weeks were much lower than the people who did not develop HZ. Even among the placebos there is a difference in the gpELISA levels. Looking at the geometric mean fold rise from day 0 to week 6, you can see that there is a 1.7 GMFR in the ZOSTAVAX group in those subjects who didn't develop HZ and that that's much different than the GMFR seen in people that went on to develop HZ.

We then wanted to look at the GMTs that were observed and look at the risk of HZ by 6 week gpELISA titer. And as you can see, once you reach about 400, there are very few cases. There are three. I did go back and look at those three cases and at the end of the study they were not considered to have been immunocompromised. During the study there were no adverse events reported with these subjects. So there wasn't any particular explanation, but it looks like the majority of the cases are occurring about that level, on the chart below as far as in terms of the quantitation of the gpELISA titers.

There were no clear differences in the rates of the various reported complication among the HZ cases in the treatment groups. Again, when you account for the difference in the incidence of HZ. Also, there didn't appear to be any HZ association with immunosuppression differentially seen between the two treatment groups. The absolute numbers were similar. And as was previously mentioned, there were three subjects, two placebo and one ZOSTAVAX, who developed two cases each of herpes zoster, evaluable cases, but data from the first case, data were used in the analysis.

Comparing the immune response in terms of gpELISA in subjects who were vaccinated to people who developed HZ on study, whether they received ZOSTAVAX or placebo, if you look, the GMT 6 weeks after vaccination in the ZOSTAVAX recipients compared to the GMTs seen 6 weeks after their onset of herpes zoster rash in the two groups are quite different, as well as the GMFR.

The safety follow-up was quite a huge undertaking in such a large group. Most of the subjects were randomized and that were randomized to the two treatments were followed in the Routine Monitoring Cohort. However, a subset were in the Adverse Event Monitoring Substudy. As well, the CMI Substudy, this is really the immunogenicity subgroup, they were a subgroup of the routine monitoring. So you couldn't be in both. They were separate.

The Adverse Event Substudy, I have to watch where I'm leaning, involved 6,600 subjects who used vaccine report cards, which specifically queried for solicited adverse events on day 0 to 4, specifically queried for temperature day 0 to 21 and allowed for subjects to report other complaints on that vaccine report card.

The automated telephone response system, which was to be conducted on or around day 42, specifically queried for the occurrence of rash, any unusual reactions, hospitalizations, disability, life-threatening events, new diagnoses of cancer, overdose of any medication. ATRS follow-up was conducted monthly for surveillance of suspected HZ and in the AE Monitoring Substudy for hospitalization.

In addition, investigators could review available medical records on or around day 42 to look for other information related to adverse events or possible herpes zoster.

The Routine Monitoring Cohort, which is the remainder of the study, basically relied on the ATRS monitoring for 42 day safety follow-up, the same questions asked as I have mentioned before, available medical records could be reviewed for adverse events or herpes zoster, and otherwise the safety monitoring in this cohort was basically passive and, as I mentioned before, the monthly ATRS monitor for HZ in this group and monitor for HZ and for hospitalizations in the AE Monitoring Subgroup.

Looking at the day 42 ATRS data set that was submitted to us, we note that there were 38,546 subjects enrolled. However, there are only 25,613 subjects accounted for in the day 42 ATRS subsets. 66 percent of the subjects. We saw that there were 55 percent of the total cohort who called in or either called or called the ATRS as was planned by the protocol. We also see that an additional 11 percent of the total study population had data answered by the staff for the subjects into the ATRS system.

We also noted that only 9 percent of the subjects from the AE Monitoring Cohort had data included in this database. And in addition, although subjects were to have a report on or around day 42, there are subjects, there are 1,240 additional reports for subjects that already have data in this data set and the data were entered sometimes one, two, three years after the initial report.

In looking at the source and time course of reporting, you can see that most subjects called in on or around day 42 and that their reporting rates dropped over time. After this point, we have been told that subjects -- the system was disconnected or unplugged and so subjects couldn't call in. You can see also that staff were calling for the subjects over time and that after this point, over 4,600 subjects are having data entered by the staff. This can go out two and three years afterwards for the day 42 safety reporting. In addition, you can see that some of these, not very many, are entered well before day 42.

This is a table showing the AE rates based on the vaccine report card monitoring. You can see that temperatures didn't seem to be wildly different between the two groups as far as high temperatures or even feeling like your temperature was abnormal, even if the documented temperature was less.

The statistically significant differences were seen in erythema, pain and tenderness and swelling. And I would point out those were three things that were specifically queried for on the vaccine report card. All had P-value for the difference of less than 0.001. Regarding unsolicited adverse events, I would note that there was a higher rate of pruritus in the ZOSTAVAX group and some increase in reporting of warmth.

This is looking at systemic adverse events reported in the AE Monitoring Substudy between 0 and 42 days. And you can see that in the different body systems, they are fairly similar. Now, looking at serious adverse events, in the Routine Monitoring Cohort, the group that was more passively monitored, the large portion of the study subjects, there is a slightly higher rate of SAEs reported in the placebo versus the ZOSTAVAX group.

If you look in the Adverse Event Monitoring Substudy, there is a higher rate of serious adverse events reported in ZOSTAVAX versus placebo. And although this difference is not as marked in the younger group, it is even more notable in the older age group. The rate of deaths reported in the first 42 days were similar between the two groups.

These are the report causes of death. Obviously, things could be coded in multiple ways, so something that was coded, reported to us as a myocardial infarction sometimes it would also say heart arrest or cardiovascular disease. So these are sort of in a hierarchial, if you will, exploratory, obviously for safety. There were similar rates of serious adverse events resulting in death in the first 42 days. The cardiovascular causes appear fairly similar.

So we didn't really see a difference as far as cause. I know that was a question that had come up earlier. But I would note that of the deaths that are reported, 26 of these are coming from the Routine Monitoring Cohort. They were more intensively monitored in that first 42 days, but not afterwards. We also had a question about hospitalization and you can see that the overall rate of hospitalization was similar and didn't seem to be a huge difference in the rate for HZ-related causes as well.

Deaths overall for the entire study period appeared similar in both age cohorts and overall. There is no information that has been submitted to date on the proportion of subjects with ATRS contact at each month overall by group and by site. And this is important in terms of safety follow-up, because the AE Monitoring Cohort were being queried for hospitalizations. Also, "Due to the passive and inconsistent nature of the safety data collection in the Routine Monitoring Cohort from day 43 through the study end, caution should be exercised when interpreting these particular data."

Looking at the AEs that occurred at rates of at least 1 percent in either group from day 43 to study end, again we don't see huge differences in all the various body systems that were being monitored.

And now, I'm going to have a few comments on Protocol 009. The objectives, comparison of the safety and tolerability of a higher potency ZOSTAVAX vaccine with that of a lower potency dose. And also, that among adults 50 years of age and older, the higher potency ZOSTAVAX will be generally well-tolerated as compared to the lower potency of ZOSTAVAX. And I think most of this was already reviewed.

The endpoints were, the primary endpoints included the difference between the higher and the lower potency vaccine groups in the risk of investigator-determined serious adverse experiences occurring up through day 42, postvaccination, and the other primary endpoint is that the upper bound of the 95 percent confidence interval for the incidence rate of moderate or severe injection-site pain, tenderness, soreness or swelling, composite endpoint of those events occurring on days 1 through 5 postvaccination would be higher -- in the higher potency vaccine group would be less than 21.5 percent. And this is based on the historical rate reported for PNEUMOVAX²³.

Secondary endpoints included monitoring varicella, varicella-like rashes, HZ and HZ-like rashes and fevers as well. The primary endpoints, there were no investigator-determined vaccine-related serious adverse events. The rate of the composite local adverse events in high potency group was 17.2 percent. The upper bound being 21.0, which met the pre-specified criteria that it be less than 21.5 percent.

The secondary endpoints, there were no occurrences of varicella or varicella-like rashes. The zoster or zosteriform rashes were similar in the two groups and temperatures were similar as well. These are a listing of the serious adverse events that were reported day 0 to 42. You can see that there was one in the low potency group. There were four in the high potency group. And these two were in the lower age cohort. These two were in the 60 and above age cohort.

There were no deaths reported in comparing the injection-site reactions based on the composite endpoint. You can see that there was a higher rate in the high potency group, not surprisingly. And also that higher rates of injection-site reactions were seen in the younger cohort compared to the older cohort, but this was particularly notable in the higher potency vaccine comparing the two age ranges.

So in summary, the ZOSTAVAX issues and summary of the data. Reduction in HZ incidence is 51 percent in relatively healthy adults age 60 years and older, postvaccination. 64 percent in those 60 to 69 and 38 percent in those 70 years and older. There is a reduction in the PHN incidence of 67 percent at 90 days following HZ rash onset. There is a reduction in the HZ BOI score of 61 percent over the six month period following HZ rash. And the effect on PHN incidence and BOI appear relatively small after accounting for the affect of the vaccine on the incidence of HZ.

In persons with HZ, there is no clear correlations seen between the reduction of BOI scores and measures of clinical benefit. For example, things like mortality, serious morbidity, hospitalization, use of pain and medication of or interference with activities of daily living. In the completeness of the safety, the ATRS and study termination follow-up is unclear at this point.

And age appears to be the strongest factor in determining vaccine effect and in an exploratory analysis, efficacy appears minimal starting in around the 75 years of age and older, which I would suspect would be the age group with potentially the largest burden of disease as far as prolonged and severe pain.

The relative increase in the rate of serious adverse events seen between day 0 and 42 in the AE Monitoring Substudy was most notable in subjects age 70 and above. However, there was no specific pattern of serious adverse events seen. Exclusion criteria, those not expected to live at least five more years, not ambulatory, chronic use of corticosteroid use, cognitive impairment, make it difficult to draw conclusions as to the generalizability of Protocol 004 efficacy and safety analyses to a typical population aged 60 years and older.

Protocol 009, this includes a younger cohort of subjects 50 to 59 years of age, but there is no comparison of the older age strata to previous age groups based on ZOSTAVAX studies, particularly the pivotal study. The vaccine dose is four times higher than any previously studied, but again there is no comparison or bridging to the previous ZOSTAVAX studies.

The clinical relevance of the study endpoints that were chosen, the primary endpoints are not clear. And I wanted to also acknowledge there are too many people to thank that have made this project what it is, but I would like to specifically thank Dr. Ahnn, the statistician, Dr. Pratt, Dr. Finn, who I wouldn't be here without their help, Dr. Goldenthal and Captain Matrakas.

So I would like to go back to the questions for the Committee to reiterate from earlier today. Are the available data adequate to support the efficacy of ZOSTAVAX when administered to individuals 50 years of age and older in preventing herpes zoster; in preventing postherpetic neuralgia; preventing postherpetic neuralgia beyond the effect of the prevention of herpes zoster; decreasing the burden of illness; decreasing the burden of illness beyond the effect on the prevention of herpes zoster? And, if not, what additional information should be provided?

Are the available data adequate to support the safety of ZOSTAVAX when administered to individuals 50 years of age and older? If not, what additional information should be provided?

And finally, please, identify any issues which you feel should be addressed, including post-licensure studies and, in particular, please, address the use of the vaccine in persons with co-morbid conditions. For example, those who might typically reside in assisted residences, excuse me, assisted living residences and nursing homes; use of the vaccine among persons taking chronic immunosuppressive therapies including corticosteroids; use of the vaccine in certain subsets of the sponsor's proposed age indication, for example, those 70 years of age and older, those 80 years of age and older; the duration of immunity and the sponsor's proposed pharmacovigilance plan, which we really didn't discuss in our presentations, but we could expand on during the discussion this afternoon.

So that concludes my presentation. Thank you.

CHAIRMAN OVERTURF: Are there questions from the Committee Members for Dr. Rohan at this time? Yes?

DR. SCHARFSTEIN: The follow-up of the pain data says that it's measured day zero through 182 after the occurrence of the rash. How was that data collected and can you comment on the completeness of those data?

DR. ROHAN: Well, the sponsors might want to answer that.

DR. SCHARFSTEIN: Yes.

DR. ROHAN: Because they probably would be able to do a better job, but subjects, once they had suspected HZ, were seen. And, as I pointed out, Dr. Oxman and his colleagues did a phenomenal job in the surveillance, evaluation, treatment, determination of the HZ cases. So within usually a couple days within the onset of the rash, subjects were seen and that's like two or three days a lot of the subjects were seen.

Then they were followed. There were a couple, maybe about every several days, there were some windows for the time points that they were asked again and again how is your pain? You know, there were a number of instruments that were involved. I think in my briefing document I have a list of those.

But then they went out after the first week or two to weekly evaluations for several months, and then they went to monthly evaluations. If that's -- you know, if people had pain, they continued to have weekly follow-up.

After day 30 if their scores fell below 3 at two consecutive time points, then they weren't followed again, but they did then have the monthly follow-up so that should they have pain that increased or recurred or occurred, then they would be captured in the area under the curve with those monthly time points and again go back to the weekly follow-up to capture the full burden of the disease.

CHAIRMAN OVERTURF: Dr. Royal?

MEMBER ROYAL: I also have a question concerning the pain data. It was said earlier that significant postherpetic pain was that which was scored as 3 or worse on the pain scale. And going back to the BOI data, if you calculate those numbers using just those patients who had that severity pain or worse, what do your comparisons show?

Also, it was stated that the frequency of neurologic complications was decreased in the immunized group. And could you speak to whether that -- what the complications were that were seen and whether there was an even decrease across that list of complications or whether some were lower than others within the vaccinated group?

DR. ROHAN: I don't know if the sponsor wants to answer the question about the neurologic complications?

DR. SILBER: Yes, one point of clarification. The surest way to prevent a complication of herpes zoster is not to get herpes zoster. And so I think our presentation and Dr. Rohan's presentation basically said the same thing in two different ways in that there was, indeed, across the population a reduction in the neurologic and the other complications of herpes zoster.

What Dr. Rohan was presenting was that once the herpes zoster developed, the fraction of cases that went on to develop the neurologic complications was neither higher nor lower. And so the reduction in complications is from the reduction in the case outright.

DR. ROHAN: You had a second, actually your first question. Could you repeat that?

MEMBER ROYAL: It concerns the group of patients with significant postherpetic neuralgia with pain scores of 3 or better. When one does the severity duration calculations and compares the two groups, do you see the same, especially if you take away the most severe pain duration scores?

DR. ROHAN: When you're saying when you take away the most severe pain duration scores, can you clarify?

MEMBER ROYAL: Well, a bar chart was shown of those individuals who had the most severe with the highest pain duration scores. If you take that group away and look at those who are left who had significant postherpetic neuralgia, do you see a difference between the treated and placebo groups?

DR. ROHAN: I don't know if a subset analysis, if you will, was done looking at different ranges of pain.

DR. AHNN: I think he is mentioning page 47.

DR. ROHAN: Slide 47?

DR. AHNN: Yes, slide 47. Yes, that's just the comparison of the distribution of area under the curve only among the PHN cases. So it's -- the comparison is in a non-random subgroup, so it's hard to make any conclusion statistically, but it's just for the exploratory purpose that we just want to show the distribution of the area under the curve.

MEMBER ROYAL: Right, but --

DR. AHNN: Between the two groups.

MEMBER ROYAL: Theoretically, those values could represent individuals with low pain scores and long severity duration, so it represents a mixture.

DR. FLEMING: Can I comment on that, because I think this is an important point that I wanted to pursue as well, and I would like to walk through a few slides in progression to amplify this point.

If we start with slide 45, what we're looking at here is an alternative attempt from what the sponsor presents to get a sense about whether, given that you have an HZ case, is there a difference in the BOI? So is there a difference in severity? And while the sponsor had a P-value slightly below 01, these P-values predominantly are showing little association or a fairly comparable balance.

Whenever you see a Log-Rank P-value lower than the other P-value, it's suggesting if there's a difference it's in the right hand tail, these are importantly rank-based analyses where the sponsor had a parametric analysis that is going to be heavily sensitive to this difference in the right hand tail.

So if you go to slide 47, what we're seeing is the difference between the FDA analysis showing really no meaningful difference in distribution and the sponsor claiming that there is, I suspect is driven by these cases here in the right hand tail.

And I suspect that it's an artifact or it's a result somewhat of the definition, because the definition if someone is 3 versus 2, but the 3 is only counted for -- the 2 for 30 days and the 3 for 180 days, then it's really giving the 2 versus 3, not a 3 to 2 weighting, but a 10 to 1 weighting. So my read on this is that the BOI really becomes tantamount to is there a difference in the fraction of people who have sustained level 3 or higher and not something more general than that.

So if you go ahead to slide 40, the question is is there a difference in the message in BOI from the message in HZ, and the answer is at least in part. When you're below age 70 where more than half the patients were, there is no difference, but there is a suggestion that there may be, in fact, more benefit than just prevention of HZ when you're looking at those people above age 70.

Then if you could go to slide 41, what we're seeing here is there is strong evidence that there is an age effect of HZ. It dwindles as you get older. Now, to the extent that the BOI data are interpretable, and it's complicated by this oddity and the way BOI is calculated, but if you put any interpretation on it, it seems to me it's in there. Beyond preventing HZ, is there some added evidence that the most severe prolonged cases are occurring less frequently?

If you put that interpretation on from the analysis on page 40, if the answer is no, not at all in the age 59 to 69 categories, but above age 70 maybe, so could we go back to 41, could someone produce this slide for the BOI by age? If we go back a slide, you have pooled together all the people above age 70 to suggest there is some difference.

It would be interesting to see how this breaks out by half decades, next slide, as you do for the HZ. Does that slide exist?

DR. ROHAN: It doesn't exist. I also wonder whether the right hand tail is the older age group or not.

DR. FLEMING: That's exactly a rewording of my question. That's exactly what I'm trying to get at. I suspect if we go back to slide 40, that it must be predominantly at least people above age 50 -- above age 70, excuse me, because you don't see any difference between the HZ and the BOI for ages 60 to 69.

It's showing up for people above age 70. That is suggesting to me that if you go back to slide 47 that these people might be the older people. Where are they in that distribution of older people?

MEMBER ROYAL: I would still like to see the data for those, the BOI data for those with significant postherpetic neuralgia pain only, given someone with lesser pain with longer duration would be scored equally as someone who has more severe pain and shorter duration.

DR. ROHAN: So your comment is based on this 10 point scale, scores of 3 and above were included, but your question is what about people 8, 9 and 10. Is that what you're asking?

MEMBER ROYAL: Well, this chart shows individuals with pain scores of 0-to-10. I would like to see those with pain scores from 3-to-10.

DR. SCHARFSTEIN: Can we get a figure that demonstrates the completeness of the data for day zero, you know, the evaluation times for pain and how it differs between treatments groups because I --

DR. ROHAN: I think it's small and I don't think we have analyzed it, but I think there might be some small differences, but there are small numbers of subjects in the follow-up not -- in the proportion of subjects who had follow-up for the whole 182 days, the numbers are fairly similar, but there appears to be slightly less follow-up in the ZOSTAVAX arm in sort of the mid range versus the placebo group.

DR. SCHARFSTEIN: I think there is a lot of imputation going on here.

DR. ROHAN: Right.

DR. SCHARFSTEIN: And there is probably a fair amount of missing data along the way and then there is some impute, you know, trapping in these lines between time points and I would be interested to look at the distribution of the number of people who showed up at each visit or, you know, provided data at each of those post-rash evaluation points.

DR. ROHAN: Right. But I guess it's also sort of confounded by the design in which if you fell below a score of 3, you weren't going to be asked again the way people that had 3 or above will be asked every week until they went below for two times points. If you were below 3, you wouldn't then be asked until the next monthly, whenever that occurred, sort of protocol-mandated follow-up for everyone with PHN.

DR. SCHARFSTEIN: Right. So there is sort of structural missingness and then there is unstructural missingness. I'm really interested in the unstructural part of that.

CHAIRMAN OVERTURF: Dr. Markovitz?

MEMBER MARKOVITZ: Yes. I would like to ask some advice so we can ruminate properly over lunch. I don't know, Gary, if you want to tell us or FDA, but it seems like we're heading towards having to decide. There may be people who are in favor of 60 plus and not 50 to 60.

So one question is are we going to be able to separate those out on the vote? And then the second thing is what is the precedence for accepting an argument based on logic rather than data?

I know we have rejected things based that way, but they were going the other way, and my previous experience on the Committee was not wanting to extend things to older people, but this question of extending it to younger people is somewhat different.

And I'm wondering does the FDA or you, Gary, have anything to tell us to guide us as we think about this?

CHAIRMAN OVERTURF: I think if the discussion this afternoon generates concern and controversy about various age groups, then I think we may need to split it out and we'll probably split the vote on that. The other issue, I would make a strong recommendation based upon the consensus of the Committee for approval of the vaccine at a given group.

Now, the other issue is that I'm not sure that there is going to be sufficient agreement also on the way the question is currently read, is it's we're actually -- the question is whether efficacy is supported by at least three parameters and I'm not sure that there is support for all those parameters. But that may not affect the discussion as much as the issue of the age. So we could get clarification of that during the lunch period. Dr. Hetherington?

DR. HETHERINGTON: One other issue that might help when we try to deliberate on that point is the persistence of immunity by age group. I realize that it's rather short in duration, what you have now, but to look at whatever data there is available.

Between the 50 and 60 year group, 60 to 70, 70 to 80, etcetera, what is the data that we have today on persistence particularly with regard to titers above 400 or 500, which was a cutoff that I think was implied in the FDA presentation, and also by geometric mean fold rise. If we had those data after lunch, that might help us a little bit.

CHAIRMAN OVERTURF: I think I'm hearing also everybody is in favor of taking a break for lunch, so we'll break for lunch at this point and reconvene at --

DR. FLEMING: Should we ask the additional questions for the FDA after lunch then?

CHAIRMAN OVERTURF: Yes. There will be time for additional questions for both the FDA and the sponsor after lunch. So we'll reconvene at 1:30.

(Whereupon, the meeting was recessed at 12:21 p.m. to reconvene at 1:33 p.m. this same day.)

A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N

1:33 p.m.

CHAIRMAN OVERTURF: I would like to ask the Committee Members, members of the audience and the sponsors to, please, take their seats. I would like to call the afternoon session to order at this time. At this time, we have time slotted for an open public hearing, so I will turn the meeting over to Christine Walsh.

MS. WALSH: Good afternoon. As part of the FDA Advisory Committee Meeting procedure, we are required to hold an open public hearing for those members of the public who are not on the agenda and would like to make a statement concerning matters pending before the Committee.

I have not received any requests at this time. Is there anyone in the room who would like to address the Committee? I see no response. Dr. Overturf, I turn the meeting back over to you.

CHAIRMAN OVERTURF: Thank you. We'll proceed further with the FDA presentation and the questions after we give some allotted time to the sponsors to address some of the questions that were asked this morning. So I will ask the sponsors to come forward now and address those questions.

DR. SILBER: Thank you, Mr. Chairman. I would like to just spend a few minutes touching on several points that were recurring themes in the questions this morning in the hopes of bringing some clarity and closure to them.

I would like to preface the comments by reminding the Committee that at the primary analyses on the modified intention-to-treat population of subjects enrolled, in those primary analyses for the key endpoints for support of the labeled indications that the sponsor and CBER are in agreement on the primary endpoints and where the different analyses and cuts of the data are leading to sometimes different interpretations comes on the conditional supportive analyses specifically among subjects who developed herpes zoster.

And so it's important to remember also that the burden of illness is a composite that includes the incidence and severity-by-duration. Once incidence is removed, it's still a very clinically important issue to deal with, severity of illness or severity-by-duration or area under the curve as we call it, but we need to make sure that we don't call that burden of illness, because the way that the study was set up from the outset was with an understanding that both incidence and severity-by-duration were important clinical components of the disease.

And so with that in mind, I would like to touch on five points quickly that have come up and I would like to first address this issue of whether there was a benefit of the vaccine over and above the incidence of herpes zoster and particularly for the severe cases.

You will recall that in the presentation this morning, I showed a histogram starting with those individuals with scores greater than 600. I would like to show slide 1026 now, which uses different cut points. And so, as you see here, 600 is what we looked at this morning.

What we have now is using different cutoff scores from 400 out through 1,000, meaning with each cut we are dealing with successively smaller subsets, but those people with the more severe cases were out at the tail. And what we see here, if we go to the right hand column, the relative reduction in the likelihood of having this high score goes steadily up as the bar gets raised. The more severe the case, the greater the percentage reduction.

So I would like to go then to 1028 because, as Dr. Fleming pointed out this morning, in the younger age cohort a lot of what we are seeing is based on incidence. In the older cohort it's the issue of pain. And so if we look specifically in the older age group, what we are dealing with here is, again, increasing benefits with the increasingly severe cutoffs. I would like to turn next to slide 654.

DR. FLEMING: These are nested, so when you say increasing benefits, what you really have is--

DR. SILBER: It's a relative reduction.

DR. FLEMING: -- a signal at the highest level and then, of course, lesser, in fact, imbalances as you then increment next down scores.

DR. SILBER: That or, alternatively, the more severe the case, yes. Okay. And in terms of the reduction in the incidence of PHN among those subjects who developed herpes zoster, what we see here is a 38.5 percent reduction. So this is among those with zoster, 38.5 percent reduction. In those 70 and older in whom the incidence of PHN is greater, it is a 47 percent reduction in the incidence of PHN among those who have developed herpes, excuse me, herpes zoster. So that is point number one.

Point number two is an issue that Dr. Fleming had raised about the scores less than 3. And in the presentation this morning I had commented on the fact that a number of sensitivity analyses had been conducted and that those were virtually identical to the primary, and so I would like to put up slide 630 which gives the sensitivity analyses on the herpes zoster burden of illness and I would like to focus specifically on the third line here, the MITT using the full AUC scale over the six month follow-up.

And, again, you see a point estimate of 61.2 percent, the same as what we had overall. It does not address all of Dr. Fleming's points, because the frequency of follow-up was less when pain was less. Nobody was lost to follow-up, but the frequencies were different. But in terms of taking all of the scores that were obtained, there was no impact on burden of illness.

The third point that I think got lost was a question from Dr. Fleming on the 80 plus population. And we agree with Dr. Fleming that there were 2,500 subjects enrolled, which was actually a fairly sizeable population in this age group, and I would like to focus first on the efficacy in this population because there had been questions about efficacy in the older age group. So if we could start first with herpes zoster on slide 248.

We have the age stratification at 60 and 70. We have further split this out not in five year increments, but at least 60s, 70s and then 80 plus. And what we see here for the herpes zoster analysis is the 64 percent efficacy that we saw earlier today and among the individuals 70 to 79, roughly 40 percent efficacy with a lower bound of 27.6 percent.

And for those over 80, the efficacy for herpes zoster did fall to 18.3 percent, some reduction, but a confidence interval now below zero. We're dealing with a relatively small percent, again about 7 percent of the overall population, and the study was not powered to observe efficacy at this level.

But, again, as you get older it's the pain that becomes more severe and adds even more to the burden. And so if we could pull up slide 250, please, which is on the herpes zoster burden of illness, first we see that with the 61 percent overall, we have got 65.5 percent for the 60 to 69s, 59 percent, very well-preserved in the 70 to 79 group, and a point estimate of 38 percent. So still preservation on a very clinically meaningful and clinically important endpoint. Although, again with the small numbers, the confidence interval going below zero.

Lastly, on the PHN endpoint, slide 252, we have very good preservation of efficacy in the older group, 74 percent, and with increasing numbers, tight confidence intervals in the 70 to 79 group. Again, somewhat lesser efficacy, point estimate 39 percent, with the wide confidence intervals.

So the conclusion from all of this is that whereas the incidence of disease and the prevention of the herpes zoster is the critical parameter for the younger cohort, in the older group who suffers disproportionately with severe and long-lasting pain, the vaccine's effect is strong and persists for these endpoints.

With respect to safety, there is a question that also came up, and I would like to put -- we have a question about the serious adverse experiences. So if we could get to slide S-2, please? This is the overall cohort and the split in serious adverse experiences was 27 versus 21.

In this slide we have it broken out by body system and it's a fairly symmetrical mix here. I will get to the cardiovascular in a second. The skins were a couple of skin cancers. The metabolic/nutritional were a couple of dehydrations. But, in general, for many of these serious adverse experiences it was just one in one or the other groups.

Slide S-1 gives the specific breakdown among the cardiovascular events, so if I could get S-1. Oh, it's S-3 now. Okay. What we have here is, again, a mix of different events. There were three atrial fibrillations, three myocardial infarctions, but then there is a coronary occlusion in the other group and one MI. So it does not look as though there is any particular specific serious adverse event.

There was among the individuals 80 years of age and older one possibly vaccine-related serious adverse event. It was an 80 year-old male who developed some symptoms shortly after vaccination, was not diagnosed ultimately until about day 80 with polymyalgia rheumatica. This is a fairly common condition in older adults, often takes a long time until diagnosis, but that was the only possibly vaccine-related event in that group.

The next question is on persistence. The question on persistence came up shortly before lunch. As I had mentioned, after zoster and after vaccination, the immune markers tend to fall back toward baseline relatively quickly and I would like to refocus the discussion for a moment on persistence of efficacy, which in the end is what we really need to be thinking about.

So if we could start with slide 704. There is an analysis in the study report that was submitted in the dossier looking at efficacy by year. We detected a drop, as I pointed out in the presentation this morning, a drop in the first year followed by steady decline. And so I'm going to be showing you some tables now showing year-on-year, but also splitting out the first year into months 1 through 6 and months 7 through 12.

So here for incidence of herpes zoster, and we'll focus ourselves on the right hand column of these slides for the vaccine efficacy, what we find is that in the first six months postvaccination an observation that will recur in the next several slides, which is 75 percent efficacy in that early time period.

In the second half of year one we're at 51 percent, which by chance was exactly the estimate that we saw over the entire period of follow-up and, as I think you saw earlier in the day, that we have got point estimates of 47, 43, 51 throughout. So, really, from month seven on there is no indication from the trend that there was any decrease or any waning of the efficacy of the vaccine.

Now, I would like to split this out by the two age cohorts, so if we could go to 708. This is the year-by-year efficacy for herpes zoster in those 60 to 69 and you will note that there is actually no drop-off at all in this younger cohort. So this is, again, part of the 64 percent overall and persistence that we think will be predictive of vaccination in the adult population under 70.

And then for age 70 and up, slide 709, please, here we do, in fact, see a decline from year one, 60 percent efficacy, relatively flat point estimates. We're dealing with smaller cuts of the data. These confidence intervals do get wider, but at least the trend among the point estimates is for stable efficacy over time. So that is for herpes zoster.

Now, if we could turn to 714, please, and we'll get to the PHN endpoint. Again, you'll recall 66.5 percent overall, 93 percent efficacy in the first six months following vaccination. Remember, smaller numbers here, fewer cases of PHN, but again dropping somewhat in the second half of the first year. No clear pattern or suggestion that there is a waning of efficacy. Directionally, there is still a reduction in the vaccine group.

So if we could then split this out by age. 718, please. Among those less than 70 years of age, there were relatively few cases of PHN. You see very high point estimates the first two years, small numbers, 4411 thereafter, but again getting to the important point. In the older individuals, slide 719, where there is much more PHN, again very high efficacy in the first year, 83 percent, with very stable point estimates in excess of 50 percent from year two and thereafter.

Lastly, on the burden of illness, slide 724, 92 percent reduction in burden of illness over the first six months dropping to 70 percent. And, again, you will recall the 61 percent overall. From that point forward, one sees again no indication of clear waning over time and we can cut this by age also.

Slide 727, please. Burden of illness beginning at 83 percent in the first year for the younger cohort and then the percentages as shown, again not reflecting any clear waning. And, lastly, for 70 plus, a similar pattern after year one, stable estimates at 40.

So the last point that I would like to address refers to some of the safety follow-up. There were a lot of questions about that. I would like to call back slide 36, the pie chart, from the main presentation.

You will recall a number of 66 percent that was offered by CBER, which I think reflects the green and the magenta, but the fact is that if we combine the different means of follow-up, we come back to the fact that by one or another of the methods, 93 percent of the subjects did have follow-up.

There was a comment that these staff calls began or were clustered after day 50, and that is because the ATRS was open to the subjects until day 50. When that day came and went, the sites received faxes to inform them that the patients, the subjects, had not called ATRS. And within virtually a week or 10 days thereafter, a large majority of the calls were made to follow-up where the subjects had not.

Importantly, the information that was captured, the script that was used by the sites in their discussions with the patients, with the subjects, was exactly the script that was used from the ATRS. And so the follow-up was comprehensive and it was consistent across the population.

In terms of timing, I would like to turn to slide 407, and this is the distribution of staff calls for the routine cohorts, so these are the calls to the ATRS. You will see that it was about 13 percent of the subjects overall, but a very large majority of these happened before day 60.

And, again, these would be clustered from day 51 to day 60 because of the way the ATRS was structured, and it was a relatively small percentage of the overall enrollment of 31,000 subjects who had any follow-up that was after day 60.

For slide 408 we have got the follow-up to the faxes and, again, you see that a large majority of these were occurring before day 60 shortly after the ATRS was turned off for the subjects, only a little over 1 percent of the subjects beyond that.

There was a question about the demographics of these people who had the later follow-up.

DR. FLEMING: Could you go back a slide just before we lose the thought?

DR. SILBER: 407, is that --

DR. FLEMING: I think it was the previous slide to this.

DR. SILBER: I'm sorry. I can't see who is even asking. Oh, okay.

DR. FLEMING: Are these to be -- basically, are the bottom five rows mutually exclusive as they appear to be and should they add up to the staff called ATRS line?

DR. SILBER: These --

DR. FLEMING: I mean, I'm sorry, into the placebo? Should the bottom five lines add up? I think they do. The percentages don't add up.

DR. SILBER: Should be.

DR. FLEMING: Okay. All right.

DR. SILBER: Should be.

DR. FLEMING: Go ahead. The percentages don't add up, but go ahead.

DR. SILBER: Maybe some rounding. Okay. There was a question about the demographics and that is still being looked at, but in terms of gender and age, and we are looking now at functional status and other health markers, we have seen no differences at all yet among any of these parameters for those who followed-up for safety in the various different ways.

The last point briefly was about opting out of the Adverse Event Monitoring Substudy and we have Dr. Levin who could speak for the Shingles Prevention Study investigators on this point.

DR. LEVIN: Dr. Silber is correct in that there was a delay until the sites got started, but once they were set and ready to go, that the patients who were offered sequentially the opportunity to be in the substudy, they were not selected. And all I can report is in my experience and that of Dr. Oxman, who I was in close contact with, that roughly 95 percent of people accepted it at that time, and there was no bias in individuals not being in the study. Everybody who was offered it, essentially, was willing to be in it. Questions?

DR. FLEMING: You said everybody that was offered was willing to be in it?

DR. LEVIN: Well, 95 percent, and we had no reason to believe that a select group of people were choosing not to be in, but we don't --

DR. FLEMING: And that choice was made at time zero, at the very beginning?

DR. LEVIN: At the time that they were offered it.

DR. FLEMING: And remind me, that time was?

DR. LEVIN: At time zero. When they entered the larger study, they were asked if they would be willing to be in the special substudy. I'm sorry, I can't speak for --

DR. FLEMING: And people didn't drop out beyond that point based on willingness to participate. So at time zero, 95 percent agreed to be in?

DR. LEVIN: Now, that is my experience and Dr. Oxman's. I can't speak for the other sites and we do not have records on that, but that's our perception.

DR. FLEMING: And once somebody was in, you had or you retained them for long-term safety assessments in what fraction of cases?

DR. LEVIN: They were retained the way all the other subjects were retained for the long-term assessment. Actually, they had more. They had additional follow-up and then all hospitalizations as well were reported in that specific group. They were actually looked at more carefully.

But I think your question was were they demographically different. We don't have specific records to that, but there is no reason to think that they were selectively chosen or they selectively chose to be in the substudy.

CHAIRMAN OVERTURF: We can address further questions to industry as we begin to discuss the questions, so I think I will ask the FDA to come back and re-present the questions to us again.

DR. SCHARFSTEIN: Could we ask a couple questions of -- wait?

CHAIRMAN OVERTURF: Let's wait a minute.

DR. ROHAN: Once again, questions for the Committee's consideration. No. 1. Are the available data adequate to support the efficacy of ZOSTAVAX when administered to individuals 50 years of age and older in: (a), preventing herpes zoster, (b), preventing postherpetic neuralgia, preventing postherpetic neuralgia beyond the effect on the prevention of herpes zoster, (c), decreasing the burden of illness, decreasing the burden of illness beyond the effect on the prevention of herpes zoster? If not, what additional information should be provided?

Question No. 2. Are the available data adequate to support the safety of ZOSTAVAX when administered to individuals 50 years of age and older? If not, what additional information should be provided?

And Question 3. Please, identify any other issues that should be addressed, including post-licensure studies. In particular, please, address: (a), the use of the vaccine in persons with co-morbid conditions, for example, those who might typically reside in assisted living residences and nursing homes, (b), use of the vaccine among persons taking chronic immunosuppressive therapies, including corticosteroids, (c), use of the vaccine in certain subsets of the sponsor's proposed age indications, for example, those 70 years of age and older, those 80 years of age and older, (d), duration of immunity and, (e), the sponsor's proposed pharmacovigilance plan.

CHAIRMAN OVERTURF: So we will actually begin the discussion and, at this time, if there are additional questions that the Committee Members want to address to either the FDA or the sponsors, we have a few more minutes to do that. Dr. Markovitz?

MEMBER MARKOVITZ: Yes. I would still like to get the take of the FDA on, if there is someone who can speak to the idea when we addressed the 50 to 60 group in lieu of data, what considerations are there from an Agency point of view, if someone can answer that.

CHAIRMAN OVERTURF: Dr. Baylor?

DR. BAYLOR: Norman Baylor, FDA. We would like you to try to address the questions based on the data presented. We really need the advice based on your interpretation of the data. I don't think you should use logic or gut feelings so, please, use the data.

CHAIRMAN OVERTURF: Yes, Dr. Word?

MEMBER WORD: I just want to go back to the issue with duration again. I know the sponsor showed a slide, but I guess one of the questions I had, you know, if you looked efficacy and they said in the 60 to 69 age range I think it was like 73 percent and maybe it dropped down to 38 percent in those older, so then I'm looking at, well, that's when it was administered at 60.

If they are proposing to administer at 50, do they anticipate that there is going to be a change, that suddenly those numbers are going to drop or is there going to be a point where you think that do I need a booster?

And I guess the other part of the question I have is right now for individuals who are born, I think it's after 1965, the adult immunizations recommend that they all get varicella vaccine if they haven't had it. And so then do you have any information about if they receive varicella and then you want to offer them this, because they are close to that age group, what would you give them?

CHAIRMAN OVERTURF: I'll let the sponsors answer that. My own personal feeling on that issue would be that, obviously, we need data in that regard. I mean, if we -- and, actually, I don't think it's restricted to the population less than 50, because we only have data that we're being presented today that is really, essentially, a four year period.

So the issue, well, this obviously has to be part of the pharmacovigilance issue, is to continue to look at that in order to justify the vaccine. Does the --

DR. SILBER: Okay. I think I heard three questions, so I will take them in turn.

In terms of the expectations of what might happen at the age of 50, again, for the herpes zoster incidence endpoint as we went out over time, we were at roughly 65 percent efficacy and stayed pretty much right there throughout the period of observation. With the younger individuals, younger immune system, we would expect that the durability of the response should be at least as good, let us say as good, in the 50 to 59. So 60 to 69 experience we think will be predictive in that regard.

With respect to the possible need for a booster vaccination, which I think was your second question, which is really a question whether one vaccinates at 50, at 60, at 70 or at any other age, that is not known at this time.

One of the really critical questions that we're answering or hope to answer in the persistence substudy of the Shingles Prevention Study is to take these people out to 10 years, is the target right now, and to determine whether there is any waning of efficacy at any point. Again, after the initial drop, we have not seen that yet.

But should that happen, what could be explored certainly is to define when or if a booster is needed and then, if we use this population basically as a bellwether, then we would be able to assess the potential benefits of booster vaccination ahead of those populations who would be receiving vaccine in the general marketplace later. So we don't know but the persistence substudy, we hope, will give us that answer.

Lastly, with respect to varicella vaccination and what advice -- I think the question was what would the advice be if someone had varicella vaccine?

MEMBER WORD: Right now in the adult recommendations, now they point blank just put it in as a recommendation if you didn't have varicella or if you were born after 1965, that you have to -- to immunize them. So they are in your 50 year-old age range now.

DR. SILBER: Right, yes. And what we know from the VARIVAX experience is that the very large majority of the cases, of the doses of VARIVAX, have been administered in childhood. We do not have -- other than as part of our booster studies, we don't have data that speak to the benefits of someone with prior varicella vaccination, but even in that these were seropositives. And so it will be some time until there will be a body of data really to be able to answer that question.

CHAIRMAN OVERTURF: Do you have a predefined signal for lack of efficacy of the vaccine in those that you continue to follow beyond? At this time, is there any predefined signal or are you simply going to --

DR. SILBER: This is something that will be looked at by annual summaries and other than a signal of a lower bound of efficacy falling below zero, there are no specific criteria at this time.

CHAIRMAN OVERTURF: Dr. Fleming?

DR. FLEMING: Could I have the FDA slide 47 which I will get to in a moment, but at least we can pull it up. My sense is you have shown us after the break now quite an array of additional slides and I am not sure that too much of it is what we hadn't already seen.

You showed us a lot of slides on persistence of effect, which I don't think was something we were challenging as controversial. As your slides point out, in fact, your curves that you presented earlier, S-48 and S-50, were very descriptive of how effect occurred over time and did, in fact, show that it was a larger relative efficacy in the first six months. In fact, what you didn't show is it's probably even larger in the first three months and then after six months, it seems to be fairly constant.

The controversial issues are age, how is that an effect modifier for effect, and is the BOI, which is intended to look at severity-by-duration beyond incidence, telling us something beyond what just the incidence is telling us?

And what your data seem to be showing us as it relates to age is that while we had pooled ages 70 and above before, now when we're looking separately at 70 to 79 and greater than 80, there does seem to be a gradient for PHN by age as well as for BOI. And we also see, as we had already seen before, that at age 60 to 69, the BOI relative efficacy is the same as the HZ relative efficacy.

In terms of the explanation of the BOI, I don't contest what you were showing, but it seems to be entirely consistent with what the FDA already showed in slide 47, which is that there is, in fact, the appearance of this number of people that had very high scores that are more prevalent or predominant in the placebo group.

Although, if we go to the next group that you didn't go to, then there is a bit more of those in the intervention arm, which is part of why a Log-Rank analysis is going to be a little more sensitive than a Wilcoxon analysis. I actually -- oh, go ahead.

DR. ROHAN: I wanted to make a comment at this point. We had talked about the PHN. The duration of follow-up was 182 days and there were equal proportions of the subjects followed in either the vaccine or the treatment arm out to 182 days.

But if you look at other increments in more the mid range, and I don't have those data with me, there is some difference in the proportion of subjects being followed and I think it would be important to look at the pain scores that had been accrued at that point.

Obviously, it's in an exploratory manner, but to look at the pain scores in the subjects that were then lost to follow-up, didn't have complete follow-up, because if we see, for example, half a dozen subjects in the zoster group that don't have that sort of mid range follow-up that had scores in the hundreds and only one in the placebo group or if we see a number of subjects in the placebo group that had zero or very low scores and very few in the zoster group, that would also skew this because the vast majority of the postherpetic neuralgia cases even in the older subjects, they resolve after several weeks and a smaller proportion are carried out --

DR. FLEMING: Right, yes.

DR. ROHAN: -- to 60 or 90 days.

DR. FLEMING: Yes. Could you go to slide, your slide, 68, FDA slide 68? You did answer one of the questions I had asked the sponsor this morning and that is to give us information on hospitalization and HZ-related hospitalization.

DR. ROHAN: This is the sponsor's data though, I will point out.

DR. FLEMING: Okay. But you showed it so I'm asking you about it. It's interesting to me how few of all hospitalizations are HZ-related, so that at least as we look at what might be an anticipated effect on something as significant as hospitalization mediated through a vaccine effect on reducing HZ-related hospitalization, we would expect almost no effect. And, of course, we see almost no effect. Hospitalizations in total are 22 in excess, which is entirely consistent with random variability.

But, basically, what this is telling me is we didn't reduce any HZ-related hospitalizations but, then again, they are so incredibly infrequent I don't suppose it really matters so much.

DR. ROHAN: Well, I think that sort of speaks to two different, I guess, factors. I think, first of all, the question of the subjects that were enrolled and the exclusion criteria, etcetera, I don't know if the burden in very ill subjects that might not be candidate for this vaccine that might not be -- had been enrolled might be a factor, but I also think it also speaks to Dr. Oxman and his colleagues and the care that they have provided, keeping people out of the hospital.

So I think it's the investigators, as well, that deserve -- you know, maybe they are sort of a victim of their own success in that respect, as well, and you might not have seen this low hospitalization rate out in the general community. You might have seen more people hospitalized.

DR. FLEMING: Well, they weren't keeping them out of the hospital, 1,115, 1,137, but at least those people that would have been HZ-related hospitalizations on placebo didn't --

DR. ROHAN: That's what I was talking about. I'm sorry.

DR. FLEMING: So I take more your first point to heart and that is we didn't look at a cohort here where in the placebo arm, there was very much prevalence or incidence of HZ-related hospitalization and, in turn, we didn't decrease it at all either.

DR. ROHAN: I probably wasn't clear, but what I meant was that I think that the health care that was provided in HZ-related disease may have kept subjects out of the hospital, been more effective than you might have seen in a community setting.

MEMBER FARLEY: I would make a comment on that, too. As a clinician, I think that the idea that they were highly educated on the signs and symptoms of zoster and that they were instructed to immediately consult their study physicians and then they were given antivirals within this window, and I think these people were cared for at a higher level at an earlier point on average for sure than the general population.

DR. FLEMING: But what all that would mean is we can, in fact, do something about HZ-related hospitalizations. We don't need this vaccine to do it. We just need the kind of surveillance and quality care that you have referred to.

Either that is the conclusion or the conclusion is we didn't look at a sufficiently high risk group. We excluded a lot of the people that really would have been at risk and we never found out what the efficacy was in that group.

CHAIRMAN OVERTURF: Actually, that addresses one of the subsets of the questions under Question 3, which is obviously the patients. The vaccine use in patients with co-morbid conditions and those taking corticosteroids has really not been answered by this study and clearly is an issue that has to be addressed in any post-licensure procedure, because it's clearly not asked. And it's actually the dilemma that we actually still face somewhat with the varicella vaccine. Yes, Dr. Hetherington?

DR. HETHERINGTON: I just want to come, I think, to closure on some questions we had earlier about dosing and I'm sure you have the information, but I'm not sure that we have finally come to resolution.

And that is, and I will try to ask it very specifically, what are your release specifications going to be for the upper limit, if any, for the potency of this? And I ask the question to try to get some idea of what is the potential range of doses that one might get once this is approved, assuming it's approved? And in that range of potential doses that one would receive, would you expect to see a difference in immune response?

Another way to ask the same question is do you have any dose ranging data on immune response over different plaque-forming unit doses and, in particular, how did you come up with 19,000 as a minimum for your dose?

So there's a number of questions in there, but I think it tries to get to the same sort of understanding about the dose and the immune response.

DR. SILBER: Well, I think there were two main questions, so I will go to the second question first, which was with respect to dose ranging and immune responses.

What we saw in the early studies is that at the low end, basically the varicella, the VARIVAX-type potencies, less than 10,000, that there was not a response. We ended up getting a dose response going up to about that 17,000/19,000 level that I had spoken to with very little dose response beyond that. In our studies conducted since then, there was likewise little dose response over the range once you get into a range above what has been defined as the expiry.

With respect to specifications, the lower specification just is defined by the efficacy study. The upper specification would certainly be no higher than 207,000 plaque-forming units but will be an ongoing discussion between us and the FDA.

DR. HETHERINGTON: All right. So just a two pointed edge on it. The pharmacodynamic response, if I can borrow that term, for immunogenicity, once you hit about 20,000 it's flat going above that. And what you showed in your high/low dose study shows that the relative safety and adverse event rates were similar across the range that you just described.

DR. SILBER: Yes. And perhaps even more importantly than immune response over the range that was studied is the fact that the vaccine efficacy appeared to be relatively flat over the ranges that were studied in the efficacy study.

CHAIRMAN OVERTURF: A somewhat related question. What about the relationship between pre-immunization antibody and adverse events? Was there a relationship between that, particularly local reactions? Actually, you implied that when you commented on the studies when there was a comment made in patients who were 50, the 50 to 59 year-old age group, that there might be a relationship.

DR. SILBER: I'm not sure that we have specific safety tables related to baseline titer, but in three of our studies we had second doses, one at an interval of six weeks, one at an interval of about two years, one at an interval of about eight years.

In all of those the baseline titers were higher than we see in a typical population receiving a first dose, and the second doses had safety profiles in each case that were really the same as was seen with dose one.

CHAIRMAN OVERTURF: Dr. Rowbotham?

DR. ROWBOTHAM: I have a question that I think would be good for Dr. Levin to comment on, and that is that it relates to the hypothesis behind this as a treatment, and it doesn't seem to me that it would be mutually exclusive for younger subjects who have younger immune systems to have primarily a zoster prevention effect, but then in older subjects who have greater immunosenescence to have less of a change in the incidence of zoster, but a change in the natural history of zoster once they get it, such that they might have a change in the burden of postherpetic neuralgia.

DR. LEVIN: So the question is why do we see this difference between the young-old people and the old-old people? Well, I don't know the answer, but the way I look at it is that in the younger people, they have a more vigorous immune response and actually you can show that and, therefore, they have a memory component, a T-cell memory cell response to VZV, that when they reactivate it quickly comes to the fore and the reactivation is often subclinical. You don't see anything. If they do have disease, it will tend to be mild because they have responded so quickly.

In the older individual, there is a delay or they don't mobilize their memory response to reactivation so quickly. The virus takes hold. It does reactivate. You do have some zoster, but then it comes to the fore more quickly than in someone not vaccinated and it limits the disease or attenuates it. And I think that concept fits perfectly with both the efficacy and the immunologic data that we have.

CHAIRMAN OVERTURF: Would any other Member of the Committee like to address any of the specific questions? Dr. Wharton, you had a question.

DR. WHARTON: This is not related to FDA-specific questions, but I have two questions I would like to ask the sponsor.

Was the information collected on the vaccine safety card and in the 48 day telephone call follow-up comparable since they are apparently being used interchangeably as far as safety follow-up is concerned and, specifically, do they both collect information on hospitalization and medical encounters during that 42 day period?

The second question I have has to do with postmarketing surveillance. Once this vaccine is in use in an older population with a high level of co-morbidity, I can anticipate that there will be deaths that occur among recently vaccinated persons and would like to know what plans the sponsor has that will help evaluate such episodes when they occur.

DR. SILBER: Okay. I will address the first question and someone else will get up to address the second question with regard to postmarketing safety.

With respect to the follow-up information, again, the follow-up information for those events that were defined by the typical -- by the ICH and in GCP as serious adverse experiences were collected uniformly, consistently the same way from everybody.

So those questions were asked in the vaccination report card. They were part of the script. They were part of the follow-up, in fact did not have to happen and it was encouraged not to happen at the end of 42 days, but all serious adverse experiences were to be -- were asked to be reported as soon as possible after onset.

So in terms of that sort of safety follow-up, the mechanisms were the same in all types of follow-up, whether diary or otherwise. And I think Dr. Gutsch will get up to talk about the post-licensure.

DR. GUTSCH: Yes. Our post-licensure plans build upon the experience that you have all heard about today with the Zoster Program and upon the experience with the VARIVAX Program using the same active component in over 56 million subjects.

In the placebo-controlled trial for ZOSTAVAX in which no specific adverse experience was identified as being clinically significant for follow-up, we have a great safety profile and we have reasonable power, 97.5 percent power, to detect an adverse experience with a frequency of about 5,500. So this trial gives us a good backdrop going into the postmarketing period.

In addition, we're going to have an additional opportunity to look at the safety in another 17,000 or 18,000 when you combine the Shingles Prevention Study and Protocol 007 in which the placebo recipients are going to be vaccinated.

In addition, we plan to conduct surveillance in the marketplace looking for signals as they are developed and in those instances where some signal might arise from our surveillance systems that are in place, we plan to evaluate those, discuss those with the Agency and, where necessary, adjust the label accordingly.

I think that other than the other things that I mentioned about the identification program to try and get a handle on any AEs and whether they might be related to vaccine and placebo constitutes the package of what is coming up.

And then there's a few other studies that are in place, which we mentioned earlier the Concomitant Use Study and the Bridging Study for a new formulation, which are ongoing. In those studies we are now enrolling subjects 50 to 59 in addition to the older age cohorts so that we'll get additional data in those groups.

And I might add we have also made an effort and are having some success in increasing the minority representation in those studies. So that constitutes the plans that we have.

CHAIRMAN OVERTURF: What is the cohort size over 80 in that placebo group? Do we know, have an estimate, since that was a specific previous question?

DR. SILBER: I think that the number enrolled in the study over 80 was about 2,500 and so it's about 1,250 additional vaccine recipients now through this follow-up.

CHAIRMAN OVERTURF: Dr. Fleming, you had a question.

DR. FLEMING: A question for the FDA. Could I get FDA slide 60, please? And while that is coming up, a quick question for the sponsor. I am pleased that you had a Data Monitoring Committee in place. I am concerned, if not disturbed, that they didn't routinely automatically have unblinded data from the beginning of the trial.

My question is was this a fully independent committee? Were the members of the DMC fully independent of the sponsor?

DR. SILBER: The DSMB met periodically throughout the course of the study.

DR. FLEMING: That's not my question. Just simply, were they independent?

DR. SILBER: I'm sorry? They were completely independent.

DR. FLEMING: Completely independent.

DR. SILBER: In fact, you are here today because of one of the independent people, I think, who served on the committee.

DR. FLEMING: Okay. So all members of the committee were independent?

DR. SILBER: Everybody. Yes, they were all external, independent members.

DR. FLEMING: Then a question for the sponsor relating to slide 60 or, excuse me, a question for the FDA for slide 60. In the FDA presentation, you don't have to show it, in slide 83 you say the completeness of safety ATRS and study termination follow-up is unclear.

And I find myself still struggling to understand the level of completeness that we can be assured has been achieved by the nature of the surveillance, so I think it's slide 60 that you have that I wanted to go to. And I always think of a threshold or a tolerance level for safety in a benefit to risk fashion.

And so what is the benefit here? The benefit appears to be per 1,000 person-years a reduction of 5.7 HZ cases and a little less than one per 1,000 person-years PHN cases, none translating at least in this study into something as significant as reducing hospitalization. So clinically meaningful events, quite infrequent in their occurrence, are being reduced something on the order of 50 percent.

It makes me, from my perspective, believe that understanding safety with great thoroughness is important to make sure that benefit to risk is favorable and, as you have noted here on this slide as well as on the slide that I was quoting from, 83, that were reliant on a fair level of what we might call passive surveillance.

The sponsor again just echoed the rule of 3, i.e., assuming that you look at 19,000 people, we can rule out events that would occur in one in 5,500. Assuming we didn't see any such events, then we can rule out rates at that level under the assumption that were such an event to occur, we would capture it and that subtle, if not not-subtle, assumption is there.

How confident are you? You know the data better than I. You know the system better than I do. How confident are you that this system that has a non-trivial amount of passive surveillance with a fairly low threshold level for safety, given the nature of efficacy, can be reliably capturing events that, if they were occurring, would in fact meaningfully impact benefit to risk assessment?

DR. ROHAN: I guess one of the issues is looking at the vaccine report cards versus the ATRS safety data. The ATRS specifically queried for these things.

The vaccine report card, and this is part of sort of human nature, if you will, as well, specifically asked for local reactions through day 4, specifically queried for temperature through day 21, and then if subjects felt that they were feverish or felt their temperature was abnormal, they could record their temperature and it also allowed them to record unusual or other events, but it didn't specifically ask were you hospitalized, I don't believe.

I don't believe that it asked these specific questions. So I would think that -- I would be concerned that subjects will be focusing on the first four days of local reactions and temperatures for 21 days every day and that what they reported in the vaccine report card, the rates, etcetera, might differ from the data in similar subjects reporting to the ATRS follow-up, and I am not sure.

I, you know, saw the slides that the sponsor put up. This is from the data set that they provided with us and, as I said, the 4,639 are not all clustered between day 51 and 60, but they go out for several years and there are hundreds and hundreds of people in the second and the third year being added into the database and I don't know what to make of that.

The sponsor has told me that there is no window for the day 42 safety follow-up and, as you can see, people are being enrolled before they were vaccinated, which I take day minus 5 in the first couple weeks, and I don't know if that data is the same as data at two years or three years involving the 42 day follow-up period. So that is an issue.

I don't know how many people are actually calling in on the monthly phone calls, how many of the subjects are having data entered by the investigator, the investigator's site, by month. And, obviously, subjects were followed for an average of three years but many were followed for about two. Some were followed out to five, so there is a variety of issues.

DR. FLEMING: And this is just about three quarters of the study, i.e., when you add up all these numbers this is about three quarters. This is about 28,033.

DR. ROHAN: 66 percent of the total population.

DR. FLEMING: Okay.

DR. ROHAN: 55 from the subjects calling.

DR. FLEMING: Plus 11 percent.

DR. ROHAN: 11 from data being entered.

DR. SILBER: Can I clarify?

CHAIRMAN OVERTURF: Yes, please.

DR. SILBER: First to Dr. Fleming's comment. This was active safety follow-up of all subjects through day 42, all subjects in the AE Monitoring Substudy, all subjects in the routine monitoring cohort.

The passive surveillance for vaccine-related serious adverse experiences and deaths is as is done in all studies. It does become passive beyond that point. The 66 percent figure, again, does not include the 16 or 17 percent with vaccination report cards and again --

DR. FLEMING: Plus the gray region, right, your 11 percent?

DR. SILBER: Well, the magenta and the gray were some of these that -- again, as I showed right after the lunch break, about 80 to 90 percent of the magenta and the gray were between day 51 or at least prior to day 60. And the total before 60 was, again, 93 percent across the entire study of both cohorts.

With respect to these calls that may have gone out beyond day 60 or day 90 or the ones that came before day 42, this table is not one per subject. These are all contacts. So if somebody called the ATRS or there was some other contact for an AE on day 6 and then there was another one on day 44, this would show up twice.

DR. FLEMING: That's why this is only 66 and not 75 percent.

DR. ROHAN: Right. And, in fact, many subjects had two or more, up to six additional entries, so some people had seven entries in the day 42 safety data set and this could occur at day 42, a year later, two years later. There are additional entries being put into this data set for a particular subject.

DR. FLEMING: With, approximately, 1,600 people who died. Certainly, that also impacts the nature of safety information we would get from those. Can you comment on that?

DR. ROHAN: Because there was an anticipated relatively high rate of deaths in this particular population, deaths were monitored but narratives and further details weren't necessarily collected. They were collected in the first 42 days with the follow-up of serious adverse events, so we have more confidence, more knowledge about that time period.

But overall deaths and I guess you could also say that deaths that occurred within the day 42 day period might take longer to be reported since the subject themselves had died, you know, that kind of thing.

CHAIRMAN OVERTURF: Dr. Farley?

MEMBER FARLEY: Can I ask a quick follow-up to this? Can you tell us what the study termination follow-up was to be? And I think if I remember your report, it was missing in a high proportion. How important is that? What was that going to provide us and should we be concerned at all about that?

DR. ROHAN: Well, we recently actually have gotten a little bit of clarification on the termination procedures. Subjects were contacted. There was a determination of whether the subject had been immunocompromised during the study or at study termination, whether they had developed herpes zoster or PHN, whether they had died and, if so, there were additional data elements that were included at that point.

MEMBER FARLEY: But was it, in fact, missing in the very high proportion of cases?

DR. ROHAN: We recently became -- I guess were in discussions. I guess it was clarified why those elements were not filled in and that the data resides in a different -- in a column rather than in the row that is left blank, that the actual date is actually in a column that is not called date of last contact.

It's called exam date. So even though the question with when was the subject last contacted is left blank in the majority of the cases, the information is in a column that is termed exam date, but we just were informed of this a couple of days ago.

CHAIRMAN OVERTURF: Yes, Dr. Rowbotham?

DR. ROWBOTHAM: I have a couple of questions. One is related to the issue of vaccinating patients in the 50 to 59 age group. So from the earlier discussion and Dr. Levin's comment, one would expect that in that group you would primarily see an effect on preventing zoster and perhaps even less of an effect on zoster pain or development of postherpetic neuralgia.

And the other thing that came out of the data presented earlier is that if you get zoster, the amount of immune response, the ELISA titers, is much, much greater than what is achieved with the vaccination.

So if you are vaccinating people in the age 50 to 59 category and at this point don't know how long that protection is going to last, especially compared to getting zoster in your 50s when the risk of postherpetic neuralgia is lower, we may not be doing the patients that much of a favor by shifting their zoster episode from the mid 50s to their mid 60s or into their 70s without knowing when would be an appropriate date to give follow-up vaccinations.

The other aspect is that in the younger population, since there is a lower risk of zoster in the first place, the issue of risks and the vaccination becomes more important and that you make it closer to the risk of the vaccination being close to the risk of just having zoster during that decade, and we don't have a lot of information about that.

CHAIRMAN OVERTURF: That actually begins to address the first question, but I would echo that I think there is considerable issues concerning immunization of individuals 50 to 59 without -- and I think it's clear that there are not data that clearly support that. And, although I appreciate what the sponsor has initiated, I think there are problems in trying to make a recommendation for that group.

DR. SCHARFSTEIN: I would like to come back to an issue I raised before and maybe I can just get a yes or no answer to this regarding PHN and both BOI depend upon the quality of the pain data.

Can you assure me that the pain data is of high quality and there is not a lot of missing data, so that we're actually getting proper measures of PHN and BOI?

DR. ROHAN: I believe that there was about 91 percent 182 day follow-up in the PHN cases, so follow-up over that period, but in the intermediate periods, which I don't have that data, there are some differences. Whether they are clinically significant, etcetera, this is obviously going to be exploratory but, you know, it would be, I think, important to look to see if there were more cases in the ZOSTAVAX group with higher AUCs, up to the point where they were missing and at what point they became missing versus the placebo.

DR. SCHARFSTEIN: All right. So to define AUC, you have to have --

DR. ROHAN: A time.

DR. SCHARFSTEIN: -- to be following. Keep a complete follow-up.

DR. ROHAN: Right.

DR. SCHARFSTEIN: So there's probably very few people who have complete follow-up over that time period. I don't know. We haven't seen any of the data. So two of our endpoints critically depend upon the quality. Yes, you have the data? Do you have it?

DR. SILBER: Yes. Actually, I would like to just clarify one other point. Again, in terms of the termination interview, 95 percent of the subjects enrolled in the study completed a termination interview. 4 percent died. Less than 1 percent were lost and so did not have follow-up. Month by month, a very large majority had ongoing follow-up throughout.

So now, if I could turn to slide 1501. I think this speaks to the issue of follow-up. And I think it's important to realize people were not lost to follow-up. What happened was the pain fell below a certain level, so the frequency of visits decreased. At any given time point for a particular visit, about 90 percent were at a visit and the BOI does cover the entire period.

What we see here, again, is that 91 percent completed. Another 5 percent were within a stone's throw of 182 days by having at least 175 days. We're talking about roughly 5 percent who had incomplete follow-up and, among the 33 out of the 950 or so, there were 11 deaths. And, as you can see here, among the individuals, several of them had a healed rash and a score of 1 or lower at the last visit and then there were just a little handful who had no follow-up.

DR. SCHARFSTEIN: So when I see more than 100 -- put that back up. When I see more than 182 days, does that mean the person was around the whole time and reporting at all your visits or does that mean oh, that person only came in twice before 182 days, but I saw him at 190 days?

DR. SILBER: No. What happens is the primary analysis truncated at 182 days. Those who had ongoing pain due to PHN continued to be followed beyond the six months.

DR. SCHARFSTEIN: When I look at the people, the 287 people who had more than 182 days --

DR. SILBER: It may have been 183, 184.

DR. SCHARFSTEIN: I understand, but does that mean that they reported at -- you have to measure the pain, right, a bunch of times. I mean, is it reported every time during that period? Probably not.

DR. SILBER: It was about 80 or 90 percent of the time points, I think, were covered.

DR. SCHARFSTEIN: At each individual time point, right?

DR. FLEMING: That is a key point that Dan is asking. It's not enough just to know that 90 percent had at least an assessment. It's important to know how many people, what fraction of all assessments were, in fact, captured.

DR. CHAN: During the course of the six month follow-up, on the average around 80 to 85 percent of the subjects that have the mandatory visits that they are supposed to come in for the pain measures. And at the last visit, as Jeff just showed you, pretty much over 90 percent have the complete follow-up at the last visit besides those who don't have pain follow-up, about two and four in each group.

DR. FLEMING: 20 percent missing. This is pretty high.

DR. SCHARFSTEIN: I don't think that answers the question. I mean, he said that 80 percent of the people had complete data in every one of the monitored visits up until 182 days?

DR. FLEMING: Can he repeat? I thought you were saying 82 percent of all visits that were to be performed were performed. What? Could you repeat what you are saying?

DR. CHAN: On average for a given visit, around 80 to 85 percent of the zoster cases came back for their visits. Sometimes, some of these visits are on a weekly schedule. So if they are off by one day, they got slotted into the next schedule which is the next week.

DR. FLEMING: So much less than 80 percent had all visits.

PARTICIPANT: Much less.

DR. FLEMING: Yes.

DR. SCHARFSTEIN: At each visit, you said 85 percent of the people showed up, right? Is that what you said?

DR. CHAN: Right, of all --

DR. SCHARFSTEIN: In order to calculate AUC, you have to have information at all the visits, that for which they are --

DR. CHAN: Say if somebody skip a visit and have to visit on prior on the next --

DR. SCHARFSTEIN: Then you just extrapolate between the two.

DR. CHAN: Exactly.

DR. SCHARFSTEIN: Right.

DR. CHAN: And that is sort of a --

DR. SCHARFSTEIN: So some people you're just extrapolating from one missed visit, some you're extrapolating for two missed visits. Some you are extrapolating for five missed visits. Right?

DR. CHAN: That is the method of calculating the AUC, is really just not all the subject have the pain scores from every day of the visit. So by design, that is the way that AUC was constructed, yes.

CHAIRMAN OVERTURF: Dr. Royal?

MEMBER ROYAL: I have a question about just pain itself. And granted, to just look at pain scores you're leaving out some parameters that are going to be important to a quality of a person's life. But when you compare just the pain scores themselves initially and at the end of follow-up, what do you see when you look at the two groups?

How do they compare? How does the distribution compare? And, specifically, those who are considered to have significant pain, what sort of comparative distribution do you see?

DR. ROHAN: I don't think that the study specifically -- and, again, I think you had asked this before and probably I didn't actually answer the question. Hopefully, I can now. I don't think that the study was designed to look at different gradations of pain. Anyone that had a score -- all scores up to the first 30 days after rash onset were counted.

Scores of 3 and above on the 10 point scale were counted at time points after 30 days, but I don't think that there was any kind of analysis done on people with the highest pain scores. There were many instruments that were administered with quality of life, health care utilization, etcetera, that were monitored during the study though.

So it was fairly extensive as far as the impact of the disease not just in pain. And although a lot of our conversations have focused on the pain and the area under the curve, really the sponsor looked at every imaginable impact in people's life, quality of life, pain medication usage, etcetera.

MEMBER ROYAL: My understanding is that pain scores were collected for every patient at every reporting point during the study. So one should be able to know what the individual scores were, what the median, the range for the group --

DR. ROHAN: We do have that.

MEMBER ROYAL: And you should be able to make those comparisons.

DR. CHAN: Slide No. 39. Dr. Ahnn, could you?

DR. AHNN: Yes, that --

DR. ROHAN: And I presented this earlier, so this gives you an idea of the mean.

DR. AHNN: Yes.

DR. ROHAN: Worst pain at these various time points.

DR. AHNN: We kind of omit the number of subjects who actually take the questionnaire. So, for example, the day 1 in placebo group, there are like 58 patients who answered the IZIQ, the initial questionnaire out of 642. And the day 1, I mean, the day 2, the next day of the rash onset, 158 patient out of 642 HZ cases actually answer either IZIQ or ZDPI, mostly I think IZIQ. And day 3, 242 placebo HZ cases had answered the questionnaire out of 642 HZ cases.

So, you know, I don't think, you know, everybody who developed HZ has same number of questionnaires answered. You know, it's very variable.

DR. SCHARFSTEIN: Some of that is structural, right, because --

DR. AHNN: Yes.

DR. SCHARFSTEIN: Some of that is structural. The question is what is the unstructured level of missingness in the study?

DR. AHNN: You know, the data like 642 HZ cases in the placebo group and all other like, you know, all others are structural zero. But even with those, among those 642 cases, there are still, you know, the area under the curve zero because they didn't develop any pain at all.

So, also, that's the same for the ZOSTAVAX group, too. That's the real zero and mostly others are structural zero like automatic zero in terms of the area under the curve.

CHAIRMAN OVERTURF: Dr. Farley?

MEMBER FARLEY: I wonder if you could clarify again for us the definition that changed in the course of the study that I -- as I recall, it was for postherpetic neuralgia and the time frame, it was earlier. It had been planned to be 30 days, I think, and it was changed to 90 days.

Can you just help us understand why that change was made halfway through and if that is something that we should be thinking harder about?

DR. ROHAN: I guess I would let the sponsor answer the question, but the change was made after the last HZ case was accrued. The study was completed and terminated about six months after the last case was accrued, but the change was made after the last case.

DR. SILBER: The question relates to it was a protocol amendment to. This was actually generated, I think, at the request of the DSMB quite a time before that and Merck and the VA -- and this was based on emerging literature among pain experts and in the medical field that the definition of PHN was, in fact, evolving and that the concept of acute and chronic pain was changing.

And, in fact, there was much debate as to whether the change should be to 90 or 120 days. In fact, two members of the DSMB are part of the literature that has emerged on this. And if we could go to slide 623, please.

So this was something that was discussed amongst us and then in the end submitted to the FDA about the time the last case was accruing, but prior to unblinding of the data. And I had mentioned earlier when I went through the primary PHN analysis that a sensitivity analysis using different time points had, essentially, the same information.

What we have here is that with each successively later time point, the point estimate for efficacy goes incrementally up a little bit. At the same time, there are fewer subjects at the time points and so the lower bound of the confidence interval remains the same. But this is a change that, again was driven by the DSMB and was driven by an evolution in the medical literature and the understanding of pain in the community.

And then if I could just turn to 625 for a moment, I would like to try to get back to Dr. Royal's question. I'm not sure if this quite gets there, but if we take sort of in the theme of levels of pain, this slide shows the different time points.

And, also, if we were to use a cutoff of 2 or a cutoff of 4, and again what we see, as we have seen as a recurring theme, set the bar higher. Use a level of 4 and relative to what we saw with the cutoff of 3 or now the cutoff of 2, the vaccine effect is just a smidgen higher.

CHAIRMAN OVERTURF: Dr. Scharfstein?

DR. ROHAN: I just had one comment. In changing the definition of PHN, the sponsor specified that the point estimate had to be at least 62 percent for this endpoint. And you can see that from the slides that were previously presented, at day 30 and day 60, that endpoint would have failed based on the specified endpoint of at least 62 percent. So it was changed to 90 that if you look at the time course, that's the first point at which it was above 62 percent.

CHAIRMAN OVERTURF: Dr. Scharfstein?

DR. SILBER: If I may clarify. The time point and the point estimate were actually changed in concert and so if the 30 day time point had remained the point estimate observed at 30 and 60, it would have met the original criterion and so --

DR. ROHAN: But the original criteria did not include a point estimate, I believe. It did? It was -- excuse me? 59 percent. So I guess, obviously, what -- the minimum efficacy that is expected depends on when you see it. But, again, it was changed after the last case was accrued.

CHAIRMAN OVERTURF: Dr. Scharfstein, you had a comment.

DR. SCHARFSTEIN: This is a naive question. Is it possible that the effect of this vaccine is not to prevent herpes zoster, but to just prolong its occurrence?

PARTICIPANT: No, sir.

DR. SCHARFSTEIN: Because it shifts the time at which you would get herpes zoster, so we have only got three years of follow-up on each patient.

CHAIRMAN OVERTURF: Actually, that issue has been raised already, I think, and I raised it.

DR. SCHARFSTEIN: Yes.

CHAIRMAN OVERTURF: All right. Thanks.

DR. SCHARFSTEIN: Do you want another answer?

CHAIRMAN OVERTURF: Yes, but I think the sponsor might want to answer that.

DR. SCHARFSTEIN: Well, are you satisfied with the response?

DR. SILBER: Well, although it's certainly reasonable that the vaccine efficacy might wane over time, we have not seen this and this again being a memory response, people are boosting due to endogenous and exogenous exposure to the virus all the time. One would expect that this T-dependent response would come back with a subsequent vaccination.

In fact, booster vaccinations or a two- dose regimen in a short period have shown that the response does, in fact, come back to the level seen after a first dose. So we would anticipate that should the data evolve to demonstrate that there is waning efficacy, that there would be benefits from a subsequent dose.

CHAIRMAN OVERTURF: Yes. I think actually that addresses actually a couple of questions we have not addressed in the 3(c) and (d), which is that I think post-licensure studies have got to include some component of active surveillance or relatively active surveillance to look at this issue, because we really have a four year period of duration right now in any age group.

And it will require, I think, some continued look at this because I think the question you asked is pertinent and relevant to what we are all considering. So I think that we will probably agree, unless somebody disagrees, that some active component or some active subset needs to be continued to be looked at very actively. This may be done in a number of settings.

We talked yesterday about using VSD data to look at this, which would be one active component, and obviously there will be -- it might be actually included in the vero subset, which is the occurrence of herpes zoster following -- it should be reporting of herpes zoster following the receipt of the vaccine ought to be part of the vero subset as well. Yes, Dr. Gellin?

DR. GELLIN: I want to go back in follow-up to a question that Monica started about the medical care of the patients or the subjects in this, and that we heard early on the medical need for this vaccine was because there was -- available therapies had limitations, but built into this was both pain management and antivirals.

Now, I wonder what we have learned about modern day intervention of ready access to these through this trial.

CHAIRMAN OVERTURF: Clearly, it was a benefit of the trial, I think. I think they have made that point, was that enrollment in this trial actually enrolled you in some very good pain management.

DR. SILBER: Obviously, the trial is not designed to look at the treatment of herpes zoster. But when we look at the fraction of individuals who received antivirals, who received anticonvulsant medication such as gabapentin, who received opiates, and when we compare that with large databases that look across a general population, the frequency of use of all of these medications was actually substantially higher than is seen in general medical practice, so again speaks to the level of care across all of the subjects, vaccine and placebo recipients who might have developed zoster.

CHAIRMAN OVERTURF: Dr. Fleming?

DR. FLEMING: I was actually going to wait to make this comment until we were answering the question, but I think our colleagues have raised this issue and it maybe is better to have it open in the discussion.

And I would like to just pursue a little bit further the idea of might we be delaying? And to the credit of this trial, it provides very good data in terms of durability of effects out to three to four years, but this issue of whether we are allowing people to remain at risk to a later point in time is certainly a very relevant one.

The data that we see indicates that there is a substantial immune response that is provided by the vaccine, but roughly in terms of geometric mean titer ratios, twice that that comes from an actual case of herpes zoster. And so the question that I might wonder, is herpes zoster the best approach to protect against a PHN case?

Well, the issue is not if there is, in fact, a risk of a PHN case when you have herpes zoster, but the data that are fascinating that the sponsor has put forward is where you have high levels of risk of herpes zoster relative to risk of PHN is in your 50s.

And if you have 1,000 people and, based on the data, maybe if the sponsor said 300 of them over a 25 to 30 year period would, in fact, be at risk for a case of herpes zoster, during that first decade of the 50s, if you start at age 50 for example, you're accumulating five to six cases per year that you're preventing. That adds up to 50 to 60 cases out of those 300.

Would it have been better for those people to have, in fact, had cases of herpes zoster where they are at, essentially, no risk for PHN, and this is a question specific to starting in your 50s, rather than to allow those or are you better to prevent those cases or allow them to occur when the PHN risk is going to be low in your 50s?

DR. WHARTON: I would point out that in otherwise immunocompetent subjects, once you have a case of herpes zoster, your risk for having a subsequent episode is 5 percent or less based on literature.

DR. FLEMING: Precisely. Therein lies the issue we're discussing.

CHAIRMAN OVERTURF: Any further questions, comments? Dr. Hetherington?

DR. HETHERINGTON: I apologize if this was covered previously, but did you look at the use of pain medications across treatment arms as a potential confounding factor in the pain assessment?

DR. CHAN: So your question is whether we have looked at the pain medication uses as part of the assessment of vaccine efficacy. We did. Obviously, when we look at the zoster endpoint, the pain medication don't come into the picture because all those come after the zoster surveys.

When we look at the supportive analysis in terms of the severity-by-duration of zoster pain among the cases, we did take that into account, and all we found is in general the pain medication uses are very balanced between the two groups and there is no effect on the vaccine effects because of use of the antiviral or pain medications.

DR. SILBER: I would like to get back to Dr. Fleming's point again about the potential for delaying. The evidence that we have is that the vaccine effect is durable and, although people in the 50 to 59 age group do not have PHN at the rate that older individuals do, they have often very severe, acute pain.

200,000 people a year have acute herpes zoster in this age group with severe pain. The rate of complications, other than PHN, is about as high in people 50 to 59, including the ocular and other potentially severe complications and so --

DR. FLEMING: Then why weren't they included in the trial? If it's so obvious that these people are at such considerable risk and potential for benefit, why weren't they in your trial?

DR. SILBER: Well, again, to go back to the original point, that the primary benefit that we would anticipate to see in the younger individuals is from prevention of the episode outright. The scientific information available to Merck and the VA in 1997 when this trial was initiated, in 1992 and 1994 when the protocol was drafted was that the vaccination could not accomplish that.

Further to the point, even if the vaccine at some point wanes and is not durable, that doesn't mean there is no benefit to the individuals. And, again, what we have seen in three different studies with second vaccinations and as we would anticipate since this virus is kept quiescent for many years is that immunologic boosting that could eventually be given with a second dose, if necessary, would biologically plausibly prevent that episode from happening at a later time.

CHAIRMAN OVERTURF: I would agree that it's biologically plausible, but the issue really is why wasn't it studied? If it was part of the original hypothesis, then it should have been studied. And, obviously, you have explained a little bit why it was not and I'm sympathetic with that, and I think the issue is almost more of a public health issue at this point.

This is going to be an issue about how best to control herpes zoster in this population, and I think the question before the Committee to me is do we have data to support this method of control for this public health problem? Dr. Royal?

MEMBER ROYAL: Just going back a minute to potential effects of treatment of individual patients on some of the parameters that you measured.

Is there any reason to think that patients who are treated with an antiviral might have had some differential difference in the frequency with which you isolated your vaccine strain virus versus non-vaccine strain from the lesions themselves? So I believe you found your vaccine strain in two patients, in lesions from two patients, but not in the rest.

Do you think that their being on an antiretroviral would affect that at all?

DR. SILBER: The question refers to the isolation of the VZV in the PCR. I think we may be dealing with two separate issues. In the Shingles Prevention Study, the Oka strain was not seen in any individuals during the efficacy follow-up. All of the cases of zoster that occurred were with wild type. All of the rashes that occurred that had specimens within 42 days were wild type.

In two other trials, one subject each developed -- among those with VZV-like rashes, there was these two individuals who had rashes from whom the PCRs disclosed Oka strain. In one case it was a 92 year-old man who had just a few, some papular lesions 17 days postvaccination, in the other study a 23 year-old female from the VARIVAX study who was seropositive and had some lesions about a week after vaccination. So this was in the immediate postvaccination period.

DR. GUTSCH: One other point to this question is that the samples for PCR were collected before acyclovir was being administered.

CHAIRMAN OVERTURF: Dr. Markovitz?

MEMBER MARKOVITZ: I'm curious how the decision was originally made to only give one dose of the vaccine. It seems like, you know, you obviously have efficacy in certain populations. I'm wondering why a booster given a month later or something wasn't pursued. I know you did that in some of your earlier studies, but I'm curious why that strategy fell by the wayside.

DR. SILBER: A question about the single dose regimen. Again, the studies that had been conducted previously and, in fact, the studies that have been done subsequently have indicated that there was not further immunologic benefit from a second dose, that it got back to where you were with dose one. Now, whether that could translate into some qualitative difference was not studied.

CHAIRMAN OVERTURF: Hearing no further questions or comments from the Committee, I think we'll progress to the main questions and we're instructed to answer these questions as they are asked.

If there are portions of the question that any given Committee Member, when polled, disagreed with, please, state your reasons and provide input to the FDA on what you think needs to be done in order to fully support that particular indication.

So I'm going to start with Dr. Karron. And the first question is "Are the available data adequate to support the efficacy of ZOSTAVAX when administered to persons greater than 50 years of age in preventing herpes zoster, preventing postherpetic neuralgia, preventing postherpetic neuralgia beyond the effect on the prevention of herpes zoster, decreasing the sponsor-defined burden of illness and decreasing the sponsor-defined burden of illness beyond the effect on the prevention of herpes zoster? If not, what additional information should be provided?"

MEMBER KARRON: Herpes zoster is an important cause of morbidity in the elderly and a vaccine that effectively prevented zoster and its complications would make an important contribution to public health.

The sponsor has shown that ZOSTAVAX is effective in decreasing the incidence of zoster, preventing postherpetic neuralgia and decreasing the sponsor-defined burden of illness in individuals who are 60 to 69 and over 70 years of age.

However, as shown in the additional analysis, efficacy against the incidence of zoster is substantially decreased in individuals over 80 on the order of about 18 to 20 percent, though there may be better efficacy against postherpetic neuralgia, burden of illness or prevention of perhaps the most severe pain complications. Though, obviously, the numbers are small and here the confidence intervals overlap zero.

Although the sponsor has asked for an indication for use in individuals over 50 years of age, only 185 individuals in the 50 to 60 year-old age group have been studied and those individuals have been studied for safety.

While it's likely that a vaccine that is efficacious in individuals over 60 would also be efficacious in individuals in that 50 to 60 year-old age group, the question needs to be addressed more completely. Perhaps additional assessments of immunogenicity with a bridging study could be contemplated since the rate of zoster is quite low in the 50 to 60 year-old age group.

An additional important issue that has been touched on by many of the people here today is the issue of duration of protection against zoster. And this is not only a question regarding the need for booster doses to prevent the breakthrough disease, but also importantly the question of whether immunizing the young elderly, those say 50 to 70 years-old, will only delay the time to occurrence of zoster potentially with worse complications in older individuals.

So my conclusions are that the data are not adequate to support efficacy in persons over 50 years of age, though there may be data to support efficacy in a subset of that group.

CHAIRMAN OVERTURF: Dr. Fleming?

DR. FLEMING: Well, I too think this answer requires some specific consideration of groups or subgroups of patients. As the question relates to people in their ages of 50 to 60, there are no data that have been presented to us. And I do believe in principle that labels should reflect what the eligibility criteria and exclusion criteria are in clinical trials. And if people have been systematically excluded in their 50s, it seems logically inconsistent to then judge we can use evidence from that trial to address whether or not efficacy has been established and safety has been established in that group.

It is the case that PHN risk is low below the age of 60. And I think that does, in fact, provide some logic to why those participants weren't included in the trial. And as we were discussing in our open discussion period, there is at least uncertainty about the issue of the prudence of delaying herpes zoster cases in people in their 50s when they are at very low PHN risk to then be at continued risk in ages later in time when PHN risk is much greater.

As it regards to efficacy for preventing herpes zoster in patients over the age of 60, I believe that there are positive efficacy data to establish effects on herpes zoster. As an aside, I would argue as always we should be doing an ITT analysis. The sponsor here did an MITT analysis excluding those cases in the first 30 days, where, in fact, there was evidence of benefit. So as an aside, again we see an instance where start at time zero and count everything that happens, both analyses would have shown essentially the same thing in this case.

The issue though is one of generalizability, as has been pointed out, and we're going to come back to those issues of generalizability. One of the aspects though of generalizability is specifically age. And experience has shown that it's treacherous to look at results by subgroups with the risk of being misled that differences that are uniform may be interpreted or facts that are uniform may be interpreted to be different by subgroups.

However, I do think in this case the evidence for a waning effect or for a lessor effect in older participants is very strong with estimates for herpes zoster on the order of about 64 percent relative efficacy, if you are from 59 to 69, dropping down to 44 in your early 70s, 36 in your late 70s, 20 in your early 80s and about 12 above 85. A monotonic trend in a study of this size that provides very strong indication of an effect that is, in fact, age-specific.

And we see a similar type of evidence for PHN and for BOI. So as we move forward to Part B for preventing PHN, I do believe that there is evidence here in this study for reducing PHN at targeted levels, protocol-specified targeted levels for people who are in their 60s to 80s. But for people who are above 80, the overall PHN efficacy is well below the targeted level. And a similar situation arises with BOI where there is evidence of benefit in those who are in their 60s to 80, but above 80 one again is below targeted levels for efficacy.

Now, key questions are also asked in BNC about how much of the effect goes beyond the prevention of herpes zoster. So specifically, in B, how much of the PHN effect goes beyond prevention of herpes zoster? My own sense about this is again this is an age-specific answer. If you are in your 60s, there is no difference at all.

So the evidence, in fact, would considerably suggest that if you are in your 60s, the effect on PHN is essentially reflecting the effect on herpes zoster. For participants who are in their 70s though and even into their 80s, there is an indication that the effect is exceeding that effect that is simply represented by herpes zoster.

For the similar question as it relates to BOI, I struggle a bit more. Again, it's very clear. If you're in your 60s, there is no evidence that the BOI measure of efficacy exceeds at all what was simply attributable to herpes zoster. There is a suggestion though as with PHN that when you are in your 70s and 80s, there may be some added value, i.e., it's not just incidence, it's severity-by-duration.

But I'm still struggling to understand the BOI. I think the definition is somewhat problematic. The ascertainment of the outcome is not as consistent as one would hope. I do think there is a suggestion in the right hand tail, which would explain why the FDA and sponsor's analyses are so different. So at least, at this point, I'm willing to say like with PHN, there is a suggestion that there might be more than just the herpes zoster effect when you are in your 70s and when you are in your 80s.

I'm going to stop at that point, because you are talking about what additional information. I don't know if you want that answer later, but one thing I have skipped over, because it comes in Question 3, that I think is critical, at least in my answers to A, B and C, is not only does this approval or does this conclusion have to depend on the age, but it certainly is problematic that we have an absence of or very limited information in critical cohorts.

Obviously, nothing in the 50s to 60. In patients with co-morbidities or chronic immunosuppression, we have also no evidence. We have minimal evidence in blacks and Hispanics. And the evidence that we do have in those above age 80 and certainly above age 85 is very concerning in terms of lack of persuasiveness.

So the answers here, I believe, as has already been stated are very dependent on the nature of the baseline characteristics and risk groups of the participants.

CHAIRMAN OVERTURF: Dr. Word?

MEMBER WORD: I don't think I'll be as long as Dr. Fleming. I think he summed it up very nicely. But anyway, I think what the sponsor actually -- the indication that the sponsor is seeking is really in individuals greater than 50. However, they really only provide us with data that examines those and provides evidence for those that are greater than 60 years of age. And that's where I struggle with this.

I mean, we're really based or asked to make a judgment call based on some immunologic data or, you know, as you would say a leap of faith, well, if it works better. We know they are younger, so that they should have a better immunologic response. However, what we are missing are the hard and fast data. So if I stick to what you say, then, you know, some of the questions that I had, it still goes back to the duration of the effect of the vaccine, giving it in this 50 year-old age group.

I don't know about the need for the booster or the effect administering the vaccine that has been brought up by others if you give it earlier to people, what long-term effect will that have. So I guess if I took away the year 50 years and I took it to 60, then the answer would have been yes. But because it stayed at greater than 50, I would have to say my answer would have to be no to all three.

CHAIRMAN OVERTURF: Dr. Scharfstein?

DR. SCHARFSTEIN: I think the sponsor showed that there is a short-term effect of the vaccine on preventing herpes zoster in the 60s and 70s. I'm concerned about the 50 to 59 year-old category as well as the over 80 category. I have serious concerns about the quality of the pain data. It may be fine. I just haven't seen it. And so the endpoints, postherpetic neuralgia and BOI depend critically upon that.

So I would say that I am uncomfortable concluding that the sponsors have shown an effect on preventing postherpetic neuralgia or on BOI. I also have concerns about the analyses that are conditional on the presence of herpes zoster as those populations may or may not be comparable. We saw some data suggest that there were a couple on basepoint characteristics. However, there could be unmeasured confounders that can explain some of these differences.

So again, I'm not comfortable concluding that the sponsor has shown an effect of preventing PHN above and beyond its effect on herpes zoster or its effect on BOI above and beyond its effects on herpes zoster.

CHAIRMAN OVERTURF: Dr. Rowbotham?

DR. ROWBOTHAM: For the great majority of persons who develop and episode of herpes zoster, it's a very severe, but fortunately, relatively short illness. But for those whose pain lingers beyond a month, and even more so for those who still have pain at six months or a year, there is no way to really underestimate the burden of suffering.

This brings up the importance of the right hand tail in the data, in that those patients who have the very high burden of illness scores over the six months after an episode of zoster, those people are very likely to continue to have pain a year or even longer and be really quite severely disabled as a result.

However, it's difficult to answer the three questions here, because of the lack of direct data in the group between the age of 50 and 59. So I can's answer any of the three questions on that. It would be speculative for me to provide a direct answer. For the first question of preventing herpes zoster, the answer is quite clear. That if you are year 60 or greater, there is a very definite effect and that seems to carry on with some reasonable confidence on up into the 70s or perhaps even into the 80s.

With regard to the question of preventing postherpetic neuralgia, there is a semantic difficulty which is that if you don't have herpes zoster, you can't possibly get postherpetic neuralgia, as we usually define it. So to put out an indication for preventing postherpetic neuralgia would encourage patients to try and get vaccinated as soon as they get an episode of zoster in the hopes of preventing postherpetic neuralgia.

And I'm already getting calls from patients asking to be vaccinated even though they have had postherpetic neuralgia for the past 5 or 10 years. So there needs to be clarity as to what exactly the vaccine can do. And what is most clear is that in this age group between 60 and 70, that the vaccine is very effective in preventing herpes zoster.

Now, in the older age group, there is evidence that there is a preventive effect on postherpetic neuralgia beyond the effect of preventing zoster. And there, the labeling language would need to be very careful to try and avoid confusing both patients and clinicians. With regard to the third question of decreasing the sponsor defined burden of illness, the problem there is that the way it was defined also included the preventive effect on herpes zoster.

And so it's difficult to answer that question, because it's really something that should be split out into looking at the burden of illness in those who have developed zoster. And again, the data suggests that especially in the older patients that in the pivotal study that the patients over the age of 70 did have less severe pain, even when their pain persisted. And so there, I think the burden of illness question is very important and it does support that there is an effect on burden of illness in those who are unfortunate enough to develop zoster despite being vaccinated.

The most difficult problem that will come up in the other questions is what to do in the group between 50 and 59. And there we are really hampered by the lack of information on the durability of the vaccination and whether or not patients who are vaccinated at 50 should be revaccinated at some additional time point before they turn 60, when the likelihood of developing postherpetic neuralgia after zoster starts to greatly rise.

CHAIRMAN OVERTURF: Dr. Gellin?

DR. GELLIN: I'll avoid summarizing a lot of the data we heard, but given the question as framed, are there data available to support efficacy of ZOSTAVAX when administered to persons greater than 50? We simply don't have sufficient data in the 50 and above. So for me that makes the answer to all the subparts easy, that there is not the data to support that.

CHAIRMAN OVERTURF: Okay. Dr. Wharton?

DR. WHARTON: I would echo Dr. Gellin's comments regarding the adequacy of available data to support the efficacy in persons 50 years of age and older for herpes zoster, postherpetic neuralgia and burden of illness.

That said, there is good data in the pivotal efficacy trial to support efficacy for prevention of herpes zoster in persons in their 60s and 70s, yet into their 80s, as others have commented, postherpetic neuralgia and burden of illness evaluations in those age groups are very strongly driven in the younger part of that population by reduction of herpes zoster. It does appear on the higher end that there may be independent effect, but it was less definite than one might like.

CHAIRMAN OVERTURF: Dr. Royal?

MEMBER ROYAL: Thank you. I would also like to stick to the question as posed to us. Looking at the data for patients 50 years and over, there is non-uniformity in response and inadequate data for the 50 and 60 year age groups. So for that reason, I feel that the studies do not support efficacy for patients greater than 50 years.

CHAIRMAN OVERTURF: Your industry opinion, Dr. Hetherington?

DR. HETHERINGTON: Well, my comments will parallel pretty much what you've already heard. Just to put it in different words, durability is a relative term and for the older age group, three to four years may put you in the ballpark of something reasonable. But as you get to somebody in the 50 and 60 year-old group, who has 30 to 40 years of life left, then durability of three to four years really doesn't mean much.

And until the question of durability or strategy to deal with any waning immunity in those who might be immunized at a younger older age group is answered, I don't think you could make a recommendation in that 50 to 59 year-old group. The standard of approving therapeutics is still based on data and data for the population that has been studied. And that again is still lacking.

That said, for the subparts, there certainly is data showing this vaccine could be effective in certain age groups for preventing morbidities associated with zoster. Most of the improvements or benefits seem to be in reducing the frequency of actual cases. I confess some indecision about whether the things such as burden of illness or preventing PHN is beyond the effect of prevention of herpes zoster. Nevertheless, I think the bottom line is that there is an overall effect and a potential for this therapeutic.

CHAIRMAN OVERTURF: Dr. Farley?

MEMBER FARLEY: I agree that as posed my answer to Question 1 would be no, that we haven't been presented with adequate data in the 50 and older category. I do think that it's important to acknowledge that they have shown what I think is quite impressive reduction in the incidence of herpes zoster in those 60 and over. And I do believe that that's something that needs to be visited with the idea of whether it has a role currently in terms of the approval process for those for which it was tested.

I believe that the additional data that we all want and would emphasize is in the 50 to 59 age group. Perhaps also in the immunocompromised older elderly, but in the 50 to 59, the emphasis not only on some sort of consistent bridging information, but with a mandate to really look at the issue of the waning immunity and the idea of boosters and data on the boosting effect over time.

So I would vote or answer no to the question as posed, but would prefer to also keep the idea of some consideration for the 60 and older category for consideration of approval.

CHAIRMAN OVERTURF: Yes, Dr. Markovitz?

MEMBER MARKOVITZ: Yes, I would like to echo a few things that were stated, but a few additional things. First of all, I think that, obviously, we cannot say there are any data to support licensing this between age 50 and 60 or 50 and 59. It is unfortunate in the sense that my guess is it will work once the company actually does the studies, but until we have the studies, we can't really comment.

And I'm a little reluctant to endorse, at least without a lot of thought, a bridging study. I suspect an efficacy study would really be substantially better. That being said, I like the data for people over 60. I think they are pretty strong data. And I believe that it shows efficacy certainly in preventing herpes zoster.

Now, the issue that people have raised about postherpetic neuralgia and burden of illness, I think that's important in terms of the labeling. But it's my impression that at least clinically if you can actually improve on those parameters by simply preventing zoster, that's still a very important improvement.

So while I think there may be some discussion about how to label this if it does get approved, I think in terms of real life clinical efficacy, I think that preventing zoster and then impacting on those other measures would be fine with me. So I vote, I guess I'm voting no for 50 to 59 and yes for 60 and above, if that's allowed.

CHAIRMAN OVERTURF: Yes, I'm actually splitting the vote on this. I would like to really congratulate the sponsors on what I think was an excellent and a difficult trial. It's a trial because it's one of those -- it's similar to many vaccines that we now are beginning to develop, which really have to deal with long-term consequences that occur long after the vaccine is given.

I suppose that was always true, but with childhood vaccines, we are often dealing with issues like rubella that would occur very shortly after immunization or would have occurred very shortly after immunization and we're somewhat universal and didn't also carry some of the chronic and difficult consequences like zoster does.

That said, actually I think I could move into the hypothetical realm and use what I know about the immunological data in the 50 year-old age group and would have been willing to. I think the biggest concern here is that you're really talking about giving this to a universal large population with what I don't think are adequate safety data yet.

That's perhaps the biggest limiting factor and perhaps probably needs to be the most important prerequisite for post-licensure, if that's going to come. It does also -- and I think another issue is the long-term public health consequences of that vaccine given in that age group and whether that's the best strategy. And I don't think we have enough information.

Plus, we don't have enough -- this sounds to me like it echos an awful lot on what we used to say about when the varicella vaccine was first licensed. We were all -- there was so much concern and still is some concern that we were going to be delaying the problem until a later point in which the severity might be greater. And I think that still is an issue here. Although, obviously, with the varicella vaccine, which the sponsors have also provided, we have eliminated perhaps an awful lot of the wild disease that would have contributed to part of this problem.

So I think to me the data do support the use of the vaccine very clearly in individuals over 60. I think there were strong suggestions that it probably does lower not only the incidence, but probably also somewhat the severity of the disease in individuals over 70. And I think even though at times the data suggests a minimal effect, I think that could have major public health consequences, even with the minimal effect. So I would support the use of the vaccine in individuals over 60, at this point.

We need to proceed to the second question and I think, at this time, what I would suggest we do is at the time I polled the Committee at this point, I would -- if you have additional questions that you want to address under Question 3, I would make that point at this time. If there is any further or last minute clarifications that you think should be brought out by either the FDA or the sponsors, we'll take that at this time as well.

And we ended with Dr. Markovitz and we'll start with him on this one. The second question being "Are the available data adequate to support the safety of ZOSTAVAX when administered to persons greater than 50 years of age? If not, what additional information should be provided?"

MEMBER MARKOVITZ: Well, the simple answer for me would be yes, again, talking about really over 60. Although, there are some pretty decent safety data for over 50. So I guess here we could even say over 50. I am a little concerned about various follow-up issues that have been raised and the statistical issues that have been raised. Although, I think I would probably defer to my more statistically sophisticated colleagues to talk about that more.

So my overall answer is yes, I think the safety data are okay. The second question you raised, Gary, in terms of Question No. 3, "What else do we need?" I think it's obvious we need 50 to 59. We need more data on the more elderly. As I mentioned before, I wonder if really one dose is really the optimal way to proceed with this vaccine or one might be better off with two in the long run.

And then the obvious thing is the people who suffer the most clinically with zoster are obviously people who are immunosuppressed, people on steroids, people with HIV. For these people, this problem is a disaster. Not to downplay the problems with an otherwise healthy person, zoster is an awful disease in those people, too. But I think we clearly need data on the immunocompromised and I don't mean just minor impairments, but truly immunocompromised people.

CHAIRMAN OVERTURF: Dr. Farley?

MEMBER FARLEY: In terms of No. 2, I'm satisfied with the safety data as presented for those 60 and older, of course, not for -- and I'm not for those under 60. Just a couple of comments on No. 3. I'm actually -- I have much less concern about the Subgroup A in that I think these people may get a benefit. It's possible it may be a little less beneficial. These patients may be -- because of their co-morbidities might respond a little less, but they also because of who they are and where they are living, their life span may be shorter than those who were in this study.

So I'm not all that concerned about expanding or generalizing or at least making it available to those in the Subgroup A. I think Subgroup B will need some very careful attention in terms of post-licensure studies that would assure the safety of the use in that group. I think that it would be to our collective benefit for us to really be establishing good monitoring systems for and in an active way and this isn't necessarily all driven by the sponsor, but also by CDC and elsewhere, active surveillance for herpes zoster.

I think it is important from the standpoint that we have now, you know, generation coming along without wild type disease, without native disease, that are vaccine protected, never had chicken pox, where will they go with zoster, and those who would have been boosted by that, the elderly and then introducing this vaccine, all of these things are going to be a complex mixture to be studied and that we need to have a system that accurately assesses it in the best way possible with the best tools.

And let's see, I think I'll close at that.

CHAIRMAN OVERTURF: Dr. Hetherington?

DR. HETHERINGTON: I think I would put a qualified yes on the adequacy of the safety data. There are a couple of issues that I'm still wrestling with and I hope that the FDA will drill down on these as they complete their review. What is the dependence upon recall, patient diaries for the collective safety data? And the second is the use of the subset. While we were told it was comparable to the general population in the study, we weren't shown the data. It wasn't shared. And there may be some subtle differences that may need to be explored a little bit more. And again, I hope that the FDA will take that into account during their review.

The presentation of the safety data, I think, was somewhat limited, but on the top line looked fairly reasonable. For Question No. 3, I'm just going to pick on two issues. One is interaction studies with other vaccines and I believe the sponsor showed that they were planning on doing a study to look at the interaction between flu vaccine and this vaccine. And I think that will be critically important.

The second, I think, would be Part A under 3 and that is the use of vaccines in persons, particularly those who are residing in assisted living situations or nursing homes. While in this population you didn't see any of the vaccine strain appearing, any herpes zoster, perhaps that would not be the case in somebody in more of a debilitated state, somebody who was on some sort of chronic immunosuppressive therapy and in a nursing home setting. There may be the potential that the vaccine strain could be spread cutaneously. So these are the things that I think that the postmarketing pharmacovigilance study would need to address.

CHAIRMAN OVERTURF: Okay. Dr. Royal?

MEMBER ROYAL: I would also say that the data as presented does to a limited extent support the safety of the vaccine and individuals greater than age 50 years. Although the data in the younger age group could be a bit stronger, I do feel that it's good enough, at this point. I would also like to recommend that the sponsor consider looking at a group of patients that can provide information that's more generalizable to the general population by looking at individuals with chronic conditions, not necessarily chronic, immunosuppressive conditions.

I also feel that it would take a more special look at that group. And also to keep in mind the fact that even within the VA population that there is a fair amount of variability in the care that's given, given the fact that many veterans don't use the VA as their only point of care.

CHAIRMAN OVERTURF: Dr. Wharton?

DR. WHARTON: As written, as the question is written, I don't believe you have all the data adequate to support the safety in persons 50 years of age and older. Although, I would give a qualified yes for persons 60 years of age and older. I'm still troubled by the fact that there were information on 7 percent of vaccine recipients were obtained more than 60 days out and I'm concerned about the ability to adequately ascertain safety information with information apparently obtained late.

That said, the information such as it was didn't suggest any safety-related problems. However, as the vaccine is -- assuming the vaccine is licensed and is introduced into general use in the elderly, there will be, I suspect, large numbers of frail elderly people with many co-morbid conditions who will be vaccinated. And it's clear that information is needed on the vaccine used in the more general population of the elderly.

And I remain concerned that safety issues will arise which may have nothing at all to do with the vaccine and maybe have to do with the underlying health status of those persons, that there will need to be a population laboratory so that those questions can be answered in a way that is efficient and can rapidly resolve the issues. And clearly, duration of immunity will need to be addressed as well.

CHAIRMAN OVERTURF: Dr. Gellin?

DR. GELLIN: Again, as written, I don't think there is sufficient data in that 50 to 59 year-old group to answer the question overall and I won't get into the subgroup analysis. Although, I want to comment that I felt that the safety data was otherwise sufficient. On a tangential note, I had by serendipity over the past several years have met many people who have been involved in this study and I would encourage the sponsors to in some way capture the information that happened here today and report back to the volunteers who may read about what happened here today in a different light.

And I think that also speaks to these incredibly large studies and more and more people are involved in these studies. We want to make sure that people continue to want to participate in such studies.

CHAIRMAN OVERTURF: Dr. Rowbotham?

DR. ROWBOTHAM: With regard to this question, I think I would like to just point out that the 009 study had only 185 subjects between the age of 50 and 59. And I don't think that's enough to say that's adequate safety data when the target population in the overall U.S. population is many, many millions.

For patients over the age of 60 though, I think the Shingles Prevention Study, which is really a landmark for those of us who work primarily in the pain area, is adequate to suggest that the safety is quite good, especially given the potential benefits in that age group. Through post-licensure studies, I do have a couple of comments.

I think a very good postmarketing study would be to examine patients who are living in assisted living or nursing homes. And that's a particularly difficult group to manage. If they do develop shingles, their communication abilities may be quite impaired. They may be cognitively impaired. They can't tolerate any of the more aggressive invasive injection procedures, like epidural injections that can be used in younger patients or healthier patients.

They tend to do spectacularly bad on sedating drugs like some of the anticonvulsants and the opiates, so this is a group that's really tailor made for a preventive type treatment like the vaccine. It's important to get this information on patients with chronic immunosuppressive therapies, probably in a progressive approach with those -- starting with those who are the least immunologically impaired and then going on steadily into more and more impaired groups.

The persons with HIV infection actually offer a ready model there, because you can look at the T-cell counts and state the severity of immune system damage in that disorder. So there was an interest in looking in that particular population. You could start with HIV-positive patients who have the least damage to their immune system with a treatment like this before going on to the more severely impaired ones and that certain group that's very high risk for zoster and also has quite a high incidence of postherpetic neuralgia.

The duration of immunity as has been mentioned quite a bit, I think, is the major factor limiting discussion of the utility of this in patients between 50 and 59.

CHAIRMAN OVERTURF: Dr. Scharfstein?

DR. SCHARFSTEIN: For Question 2, I feel that there is data for subjects over 60 years of age to support the safety of the vaccine, although I have some concerns about seeing the data with regard to the comparability of the set of people who were followed in the AE Monitoring Study and the rest of the study population, as well as the uniformity of follow-up for safety information in the study.

In terms of Question No. 3, it seems to me that A, B and D, we didn't have a lot of data on those, so it's hard to comment. In terms of C, in terms of the vaccine greater or equal to 80 years, there wasn't a tremendous amount of data to support. There was some data, but not a tremendous amount of data for over 80. And we haven't seen the pharmacovigilance plan, but I would support active surveillance if the vaccine was approved. And I also have some concerns about generalizability as this was a predominantly white population that was studied.

CHAIRMAN OVERTURF: Dr. Word?

MEMBER WORD: In terms of the available safety data, I think if you were looking at over 60, I think I would be in agreement with everyone else. I think when it comes to the 50 to 60 year-old age group, as has been pointed out, while the study seemed like it was reasonable, you have a very small number in that population.

In terms of additional studies, I think Dr. Hetherington mentioned one already when he talked about co-administration. He mentioned the -- the sponsor mentioned influenza. One of the other things that has been suggested during the age range, that they should have gotten pneumococcal and you may also look at DTaP, even though it's not been formally recommended, I think ultimately it will be for that age group for adults as opposed to just plain TD.

I guess the other question I wasn't sure of was in the pharmacovigilance plan, it talked about identification, some identification program, I'm not quite sure what that is, in terms of being able to identify, I guess, the vaccine versus wild type virus, but I don't know how you are going to disseminate that. And the other thing that just struck me as odd is that you're looking at targeting a group that is over 50 and you have a Pregnancy Registry. So I don't know if that was just a carryover from the other vaccine, but it just seemed a little odd that that was in there.

CHAIRMAN OVERTURF: Dr. Fleming?

DR. FLEMING: Whenever I answer the safety issue, I always view this as an answer in the context of benefit to risk. So I always think of what is acceptable safety based on what is the level of efficacy. With that in mind, just very quickly again, my view is efficacy essentially is established in the 60 to 80 range. We don't have data in the 50 year-olds and in those above 80, HZ incidence was only minimally effected and BOI and PHN levels didn't meet target.

So the efficacy, as I see it, is in 60 to 80 year-olds where in that range throughout there is HZ incidence data that is persuasive. Although, only in the 70s is there evidence of added severity-by- duration beyond the incidence. So with that as background, I'll just very briefly say that the safety data in the 50s, like the efficacy data, is lacking. Although it was interesting to me to look at the sponsor's slide 74 where there was more safety events that were occurring in the 50s than above 50. That was an interesting observation.

So drilling down on the safety in 60 to 80 year-olds, one thing that I noted was the SAE rate is relatively 60 percent higher in this cohort. And, in fact, in the 70s it's relatively 80 percent higher. That translates by my calculation into something on the order of about six SAE events per thousand people. And I put that in contrast with six HZ events prevented and one PHN event prevented, although that is an extension. That will extend over each year in the future.

So my sense about this is that based on what is known in the safety domain, the overall benefit to risk does appear to be favorable in the 60 to 80 year-olds. I remain somewhat uncertain and specifically again refer to the FDA's summary slide 83 saying "Completeness of safety ATRS and study termination follow-up is unclear." I, obviously, believe that it's going to be important for the FDA to be as confident about the completeness of this evidence as possible.

I am suspecting that when that assessment is final, that the overall sense in the 60 to 80 year-olds would be that benefit to risk is favorable. As it relates to Question -- as far as Question No. 3 is concerned and Parts A, B and C, I'll reiterate what I had mentioned before. I don't look at this just as postmarketing. I look at this as premarketing. There are categories of patients here that I would be concerned if they were included in the label.

Those that are 50 to 60, those with co-morbidities on chronic immunosuppressives and, in fact, because of the lack of benefit, those that are above age 80. Therefore, I would hope to see studies done even before marketing that would enlighten us much more clearly about benefit to risk in those categories.

In blacks and Hispanics it is disappointing how limited the evidence is. I would urge the FDA to carefully consider that evidence. I'm stopping short of saying those categories shouldn't be included in the label, but I'm disappointed at the limited amount of information we have there and would want the FDA to look carefully at what information exists.

Regarding duration of immunity, let me just step back and first as an aside make the point that the nature of evidence that has been presented to us, for example, that was on the sponsor slide 62, indicates that ELISA titers and ELISPOT counts are correlated with the level of risk. That is what we, in fact, know.

The sponsor used orally the terminology "they are correlates of protection," and on their slides said it's correlated with efficacy. Those latter two terms convey knowledge of causality. The kind of data that is available doesn't establish causality. It simply establishes a statistical association or a correlation and I have been arguing since 1990 on this Committee.

I would prefer that we state what we know with data such as this, and that is that this evidence is correlated with level of risk rather than calling it correlates of protection or correlations with efficacy, the latter of which suggests causality has been established.

Let me though finish on a positive note. I would like to congratulate the sponsor for conducting a clinical endpoint trial not just an immunogenicity study. Therefore, as it relates to question 3(d), duration of immunity, I think we can trump the answer to the question "Is there evidence of duration of immunity?" by saying the sponsor has established over three to four years of follow-up that there is durable efficacy, and that to me is more impressive than the answer "Is there durability of immunity?"

CHAIRMAN OVERTURF: Dr. Karron?

MEMBER KARRON: So I would say that in terms of data to support the safety of ZOSTAVAX in 50 to 60 year-olds, I don't think there is adequate data. I do think there is adequate data for those over 60.

I do share some of Dr. Wharton's concerns about that missing 7 percent and that number is a bit flexible, but whatever that is, and I would at least encourage the sponsor to get to the FDA some of those demographic data to assure us that those missing individuals are not different from those for whom they were able to get data.

In regard to question 3, I don't think I will make any additional comments, except that I did want to focus on that group over 80 and a comment that Dr. Overturf made at the end of the last question, which is that I think we should not under-estimate the morbidity in that age group, in the very elderly, and that we might want to use a vaccine in that age group that has less efficacy.

I mean, ideally, there would be a vaccine that had sustained efficacy over all age ranges, but that a vaccine even with lower efficacy in that age group might still provide a substantial public health benefit.

CHAIRMAN OVERTURF: I don't think I have anything to add. I think all the questions have been added. I also, based on what I have said previously, don't feel there is enough data to support. Although it suggests and I actually would like to believe that there is safety for the 59 year-old age group, I don't think there is sufficient data to support that.

I have been asked by the FDA to poll the Committee one more time and I don't want a lot of discussion, and I'm going to ask the same question and all I want you to answer is yes or no as to the answer about 60 versus 50 to 59. So I would first ask you again.

I think I have this recorded, but I think they want it on tape. So I would like to ask you first. You can actually ask both questions. I will ask you both questions. The first one is about safety and the second one is about efficacy for the -- specifically, if we rephrased all these questions for the 60, greater than 60 year-old age group. I think I heard one person in support of adequate data for safety for 60, maybe two.

So if I could have you -- I will poll this one more time. Let's go with Dr. Markovitz first. Safety and efficacy for those over 60 years-old?

MEMBER MARKOVITZ: Right. I vote yes, there is efficacy and yes, there are safety data to support licensing this in those 60 and above.

CHAIRMAN OVERTURF: Okay. Dr. Farley?

MEMBER FARLEY: Let me just clarify. Is it efficacy against herpes zoster? Are we keeping it simple?

CHAIRMAN OVERTURF: We're keeping it simple.

DR. FLEMING: But how can we keep it simple? I mean, we have just gone through two hours of clarifying that these aren't yes/no answers.

CHAIRMAN OVERTURF: No, I agree.

MEMBER FARLEY: But it would be if it's herpes zoster for me at least, and my answer for 60 and older for efficacy against herpes zoster, yes, and for safety, yes, given the caveat of making sure the data is fully shared and nothing new comes up.

CHAIRMAN OVERTURF: Dr. Hetherington?

DR. HETHERINGTON: Yes. I would vote yes on both safety and efficacy in the greater than 60 year-old group.

CHAIRMAN OVERTURF: Dr. Royal?

MEMBER ROYAL: I would vote yes to efficacy and safety in the 60 year and above age group.

CHAIRMAN OVERTURF: Dr. Wharton?

DR. WHARTON: Yes on efficacy for prevention of herpes zoster, a qualified yes on safety with the reservations I expressed earlier.

CHAIRMAN OVERTURF: Dr. Gellin?

DR. GELLIN: For zoster, yes, for both efficacy and safety.

CHAIRMAN OVERTURF: Yes?

DR. ROWBOTHAM: Yes on both safety and efficacy.

DR. SCHARFSTEIN: Yes on efficacy for preventing herpes zoster and yes on safety provided -- with the additional caveat.

CHAIRMAN OVERTURF: Okay. Dr. Word?

MEMBER WORD: Yes on both safety and efficacy.

CHAIRMAN OVERTURF: Let's try you, Dr. Fleming.

DR. FLEMING: I think I have nothing to add, i.e., a qualified yes on efficacy and safety with in-depth discussion of the qualifications already being on record.

MEMBER KARRON: Yes on safety, yes on efficacy with the qualifications on safety expressed previously.

CHAIRMAN OVERTURF: Now, I think that was -- and I also would vote yes on safety and efficacy for those over 60. I don't know if I dare ask this, but are there any further comments or discussion that the Committee would like to put forth at this point, any Members? Yes, Dr. Wharton?

DR. WHARTON: Thank you to the FDA staff for all their work, as well as that of the sponsor for putting this on today.

CHAIRMAN OVERTURF: I think with that, we will adjourn the meeting. Thank you very much.

(Whereupon, the meeting was concluded at 4:01 p.m.)