Alzheimer's Advisory Meeting Report

Alzheimer disease (AD) is the most frequent cause of dementia in the elderly, affecting an estimated 10% of Americans over age 65 and half of those older than 85 years. More than four million people in the US already have the disease. And, as the population ages —as it is expected to do over the next several decades— the number of people suffering Alzheimer's is projected to escalate to between 10 and 15 million.⁽¹⁾

This is unfortunate because a diagnosis of AD is one of the grimmest a physician can make. Beginning as a minor memory loss and progressing to confusion, intellectual decline and severe impairment in thinking; AD ultimately steals patients' memories, intellect and dignity -leaving them to linger in a vegetative state for years.⁽²⁾

In light of the coming Alzheimer's "epidemic," it is prudent for pharmaceutical companies to actively investigate novel strategies for disease prevention and control. On a positive note, a decade of intense research has uncovered a great deal about AD pathology, revealing a number of potential targets for therapeutic intervention.⁽³⁾

The treatments resulting from this information are important for economic as well as compassionate reasons. Notably, Alzheimer's is already the third most expensive disease to treat in the US, costing almost $100 billion each year.⁽¹⁾

An Advisory Board Meeting was in December 2002 to give researchers from GlaxoSmithKline a " clinical reality check" from physicians on the Alzheimer's frontlines. To this end, more than a dozen neurologists and psychiatrists were convened in Bal Harbour, Florida for a day's discussion of the the disease.

Thirteen Alzheimer's experts were brought together to:1) candidly discuss the clinical value of available therapies and 2) list unmet needs; 3) compile a wish list of future drugs; and 4) rate clinical trial designs and outcome measures. A brief summary of the discussion is available immediately below and a more complete review follows.

As clinicians, the panel members think they make a difference, but seem frustrated that nothing they do can restore memory or stop disease progression. They agree that acetylcholinesterase inhibitors are effective for patients with standard (i.e., mild to moderate) AD but not effective enough. This may be because, even when prescribed to the appropriate patients, they aren't used long enough or in sufficient dosages. Vitamin E may delay time to clinical worsening. The value of selegiline is as yet determined but adding it to vitamin E doesn't appear to result in additional improvement. Antioxidants, estrogens and anti-inflammatories also need more study. Antipsychotics, while useful, should only be prescribed when absolutely necessary. Memantine, soon to be marketed in the U.S. and in clinical trials with at least one of the panel members, is viewed as incredibly benign agent with some efficacy. As such, panel members though it would probably be added onto other therapy. Notably, drug benefits for patients with severe Alzheimer's or purely vascular dementias haven't been objectively studied. [See Appendix for a list of currently available therapies.]

Not unexpectedly, the inability to detect Alzheimer's early -early enough to intervene successfully-emerged as the group's primary unmet need. Indeed, prevention is considered the "holy grail" of Alzheimer's. AD is a developmental disease, by the time signs and symptoms emerge it's too late to make a major difference. Thus, they stressed the importance of having reliable disease markers, citing growing enthusiasm for serum and CSF measures. A number of neuroimaging techniques also show promise. The group would like to see large-scale screening programs for AD and related dementias.

Ideally, panel members would like cognitive assessment to be a routine part of every physical examination. However, they recognize that this is unrealistic and computer-based testing was suggested as a feasible alternative. Programs designed to test memory, continuous attention and executive function could be self-administered on home computers. Results would yield baseline data and document change over time (enabling early treatment and drug efficacy monitoring). Telephone interviews were offered as an alternative method for cognitive function assessment.

Panel members were eager to get a grasp on the cognitive disruptions of Alzheimer's, asking for drugs that specifically treat this aspect of the disease. They preferentially want medications that restore memory and others that reverse the course of AD. Compounds that work better than placebo are not the goal, drugs that work better than donepezil are. The group would also like therapies safe enough treat large populations of at-risk individuals (the model is widespread antihypertensive use for prevention of cardiovascular disease). They agree that new drugs should be aimed at biological targets, but aren't certain which targets are the most valid for pharmacological intervention.

Clinical testing of AD therapies presents a formidable challenge. For one thing, FDA wants placebo studies and panel members considered it both unethical and impractical to do trials that preclude use of cholinesterase inhibitors. For another, FDA has few valid definitions or standards to evaluate the efficacy of drugs for AD and other dementias.

However, there was very valuable discussion on clinical trials and number of interesting points emerged from the session. The panelist agreed, for example, that drug features should determine a clinical trial design as well as which patient subsets are tested. If an agent impacts on memory, they said, subjects with minimal impairment should be enrolled and improvement or decline monitored. If not, then attention testing might be better because the outcome is broader and easier to achieve.

Longer trials --4 years or so—were considered more valid than the usual 6-month to 1-year investigations. This is especially true if mild cognitive impairment (MCI) is used as a standard measure. When patients are followed long enough, almost half revert to normal without any treatment. However, it was recognized that longer trials are expensive and inefficient because so many patients can be lost to follow up. Small studies are underpowered and also considered misleading. If a drug is safe enough, proof of concept could be established in very large populations -the treated group would show a generally slowed decline in cognition.

Prevention is the major treatment goal but studies of preventative therapies would take considerable time. The group understands that companies don't want to invest too much if the patent runs out before their drug even gets to market. Patent laws need to be changed they said.

As far as testing tools go, the Alzheimer's Disease Assessment Scale—cognitive subscale (ADAS-cog) has established properties known to FDA which is a powerful reason to use it. But this test doesn't measure attention well. Activities of daily living (ADL) measures are very important, especially in the real world of clinical medicine. Quality of life (QOL) is also important but testing methods are very controversial at this point in time. However, studies should also be designed to prepare an agent for the market, and while ADAS-cog has to be done for FDA, behavior and ADL are so important to caregivers, patients and clinicians that they must be part of every trial done.

Last, but certainly not least important, objective measures need to be incorporated into clinical trials. The Alzheimer's Association has an imaging work group developing a consensus statement and FDA is considering neuroimaging use as well. Biological markers are also being developed but most are still in experimental stages.

Alzheimer's was described by Dr. Doody as a "slippery" disease and she discussed the difficulties involved devising practice guidelines. She and other AD experts reviewed more than 3,000 published papers in an effort to make treatment recommendations. Jeffrey Cummings MD noted here that the literature review was rigorous with each paper analyzed and classified by two experts in the field. The group determined that cholinesterase inhibitors --such as tacrine [Cognex, Parke-Davis], donepezil [Aricept, Pfizer], rivastigmine [Exelon, Novartis] and perhaps galantamine [Reminyl, Hoechst Marion Roussel]—do benefit patients with standard AD; however, the average benefit appears small. Vitamin E may delay the time to clinical worsening while the usefulness of selegiline [a selective MAO inhibitor that may provide neuroprotection by scavenging nerve-toxic free radicals] needs further study. Other antioxidants and estrogens as well as anti-inflammatories [such as aspirin and other NSAIDs] require further studies. It was also determined that use of antipsychotic medications, while beneficial, should be minimized.

Regulatory concerns preclude examination of how well therapies work for severe Alzheimer's disease. Nor were therapeutic recommendations made for people purely vascular dementias, because at the time of review there were no double-blind, placebo-controlled studies in this population. Most investigations have been done on patients with mild to moderate AD because, said Dr. Doody, it's the most "convenient group to study."

A big problem with the clinical studies analyzed for the practice parameters was their short-term nature. According to Dr. Doody: "In our center from the first symptoms to death, people usually live about 11 or 12 years, and we're looking at usually three or six months or one year's time worth of change….And all of the data that we're seeing suggests, but doesn't prove, that the benefit continues for a number of years." Another problem with clinical trials, added Ranjan Duara MD, is that study subjects behave differently than average patients, they're a select population. Furthermore, they have very good caregivers and even the individuals on placebo have better outcomes than the Alzheimer's population in general.

As a group though, the panel agree that cholinesterase inhibitors are perceived as less effective than they actually are. This may be because they aren't used in sufficient dosages or long enough and they're prescribed "promiscuously" to patients with "memory complaints of some ill-defined type." More important, the said: Alzheimer's is under diagnosed and "absolutely" under treated. Patients don't stay on their medications for long—an average of four months at most—suggesting that people don't understand the what's involved in "efficacy." The consensus: "…very few people get diagnosed, fewer get treated. And those who get treated aren't being treated properly."

Dr. Doody agreed saying that physicians prescribe more rivastigmine then they should because it makes patients sick and doctors think that's an indication it's working. Other members of the panel said that observation shows no more side effects for rivastigmine than other cholinesterase inhibitors. However, it was noted that most physicians don't perceive much difference between the various agents even though the drugs have different structures, pharmacokinetics and pharmacodynamics. They also agreed, in jest perhaps, that the pharmaceutical companies don't take their "marketing" suggestions seriously which: "…just shows you that advisors have no effect."

Agents in the immediate AD pipeline were discussed. The first, memantine, a non-competitive blocker of N-methyl-D-asparate receptor channels and a modulator of glutamatergic neurotransmission, is already approved for AD in Europe. The group perceives memantine as an incredibly benign agent with some efficacy and, as such it will probably be added onto other therapy. At least two companies are developing a compound isolated from Chinese herbs -huperzine. According to Dr. Doody, huperzine appears to be a very potent cholinesterase inhibitor in micromolar doses. The last word? A wish for a drug that specifically treats cognition. There are a lot more patients with mild cognitive impairment that could benefit from a medication that could improve memory.

Dr. Dingwall began by quickly reviewing disease characteristics: neuronal loss, neocortical atrophy, senile plaques and neurofibrillary tangles (NFTs).

Beta-amyloid peptides, which are derived from the amyloid precursor protein (APP) by proteolytic cleavage, have a strong tendency to aggregate and are the major component of senile plaques. In contrast, NFTs are intraneuronal structures composed of hyperphosphoryated microtubule-associated protein (MAP) called tau.

The presence of hyperphosphoryated tau is not restricted to AD, occurring in several other dementias such that there is a strong correlation between dementia and the accumulation of the protein. This correlation is, in general, stronger than that between the presence of amyloid plaques and dementia. However, recent data from epidemiological studies and work with transgenic mice indicate that levels of total β-amyloid (Aβ) do correlate with dementia and in vivo Aβ will induce tau hyperphosphorylation. Amyloid is generated through sequential proteolytic cleavage of amyloid precursor protein, a type 1 integral membrane protein present in endoplasmic reticulum, golgi apparatus and plasma membranes. Cleavage by alpha-secretase appears to occur predominately at the cell surface and yields a soluble extracellular fragment which has been reported to have neuroprotective properties. Amyloid peptides are generated by two sequential cleavage events, cleavage at the β-site close to the membrane which is followed by cleavage within the transmembrane domain of the protein by γ-secretase to liberate Aβ peptide. The cleavage by γ-secretase appears to be less precise than cleavage by β-secretase leading to production of 40 and 42 amino acid peptides that differ at their carboxytermini. Because its amino acid sequence is more hydrophobic, Aβ 1-42 has a greater tendency than Aβ 1-40 to aggregate; thus Aβ 1-42 is thought to nucleate the formation of amyloid plaques.

Dr. Dingwall continued by detailing three specific research directions: 5HT6 antagonists as compensatory treatment for AD, β-secretases (BACE1) inhibitors for disease modification and JNK-inhibitors as neuroprotective agents.

The 5HT6 receptor is one of a large family of closely related G-protein coupled receptors (5HT-1 through 5-HT7). Most members of this family activate an intracellular kinase cascade when the extracellular domain is bound by ligand, the exception being 5-HT3 which acts as an ion channel. Dr. Dingwall presented data showing that mRNA for 5HT-6 is preferentially expressed in the brain and reaches very high levels in specific brain regions. This is confirmed by binding of an iodinated antagonist ligand detected by autoradiography. "A number of potent and specific ligands now being investigated show clear cognitive improvement in a number of rodent behavioral paradigms, reversing the amnesic effect of scopolamine administration. These potent and specific antagonists are currently in clinical trials for Alzheimer's disease."

Previous work reveals that both β and γ cleaving enzymes (secretases) are aspartic proteinases making them very attractive intervention targets in dementias.

In the late 1990s, β-secretase (BACE, Asp2, Memapsin) was definitively identified as a novel membrane-bound aspartic proteinase. "This," said Dr. Dingwall, "...gave us an essential tool in our search for specific inhibitors of β-amyloid production. The search for γ-secretase has so far proved more indirect; however, there is sound evidence that it too is a novel transmembrane aspartic proteinase and amenable to chemical inhibition.

"So, how are we going to find β-secretase inhibitors? I don't think it's any secret in the industry that we've run high throughout screens and failed to identify any hits. Part of the reason may be that the active site is extremely large." We've taken an indirect route," said to Dr. Dingwall, "…going after a transition-state peptide-like molecule that inhibits the enzyme. Then we gradually, systematically try to change the characteristics of the amino acids in the inhibitory peptide to make it more drug like. We make it smaller, less polar -able to pass the blood-brain barrier. While this is a significant challenge, it's analogous to the approach taken to develop HIV protease inhibitors and HIV protease is of the same class as β-secretase, an aspartic proteinase. We've come a long way. We have potent and selective compounds. They've been tested in vitro in cells that constitutively produce amyloid peptide and a significant number work well, potently inhibiting production of β-amyloid 1-40 and 1-42. The compounds, he said, are also nontoxic and of a molecular weight enabling brain penetration."

"Testing the potential drugs in animal models of Alzheimer's is the biggest challenge we have right now," continued Dr. Dingwall. "We're using transgenic mice, bred to overexpress human amyloid precursor proteins. The focus of the program at the moment is to get as many of our potent and selective compounds through these animals. We're doing acute rather than chronic dosing. Scientists at Elan Pharmaceuticals published a study a few years ago where they gave a γ-secretase inhibitor to transgenic animals and demonstrated a C_max at three hours -concentrations that exceed the IC50 value obtained in vitro by a factor of 5. This resulted in a 50% reduction in soluble amyloid. We're doing the same thing with our β-secretase inhibitors." There is a technical challenge though --the inter-animal variation in amyloid accumulation is very large and increases with the age of the animal. Therefore, to power a study with, for example, 6-month old mice we need to process hundreds of animals.

"What we thought we might try as an alternative is to show that we can reduce the age-dependent accumulation of amyloid in the brains of mice bred to over express amyloid." In this study design, young transgenic vs wild-type mice are dosed withβ-secretase inhibitors for a few months then sampled at intervals and amyloid measurements made. Cognitive benefit of the agents is evaluated using the Morris water maze or object recognition tests. The difference between wild-type and transgenic mice becomes evident after six months in the water maze tests. When the submerged platform is removed, the transgenic animals show a significant deficit in their ability to locate the quadrant of the tank where the platformed used to be. We're seeing if we can correct this cognitive degeneration with one or another of our newly developed β-secretase inhibitors.

To summarize the secretase studies, Dr. Dingwall said: "…we've cloned and identified β-secretase and are working on the basic cell biology of the protein and its biochemical properties. Additionally, we have synthesized all of the γ-secretase inhibitors identified in the literature and are using them as tool molecules." Small molecular inhibitors of both β- and γ-secretases are being tested in a battery of transgenic mice models, including a newly bred very aggressive amyloid accumulating animal.

GSK researchers are also actively investigating JNK-3 inhibitors as neuroprotective agents. There are at least two good reasons to do this: 1) strong evidence suggests a role for JNK-induced neuronal apoptosis in AD pathology; and 2) because JNK phosphorylates tau protein and abnormally phosphorylated tau is the main component of neurofibrillary tangles.

Karen Philpott is leading GSK's JNK-3 program. As a first step, her group looked for evidence that this pathway is actually involved in neuronal death. They began by examining the DNA binding protein, c-Jun in cerebellar granule cells during experimentally induced apoptosis. These studies showed that JNK was involved in neuronal apoptosis - c-Jun expression is upregulated and hyperphosphorylates JNK at two key sites. Then they took primary neurons and looked at nuclear fragmentation -a good measure of apoptosis. In one series a pan-JNK inhibitor peptide was tested on cortical neurons and apoptosis was curtailed -that was evidenced by a reduction in amyloid-induced neuronal fragmentation. But which specific JNK -1, 2, or 3—was doing the job? Antisense oligonucleotides were used against each of the three kinases and the oligonucleotide targeted at JNK-3 appeared most protective. Dr. Dingwall noted that in this study they were looking at the number of surviving intact nuclei and not nuclear fragmentation. In summary, Dr. Dingwall reviewed the evidence that inhibition of JNK-3 is neuroprotective against β-amyloid toxicity in a number of models. "The known role of this kinase in tau phosphorylation," he said: "… make it a particularly attractive target for AD."

The therapeutic goals for patients with dementia are fairly simple according to Lon Schneider MD: "We use drugs to relieve suffering and improve functioning." However, care giver interventions are very important and we often prescribe drugs for their benefit rather than the patient's. This, he said, is different from how we need to think about drugs from the regulatory point of view. According to FDA, all drug products in the US have to be "safe for use," and "effective in use." One aspect of this is recognition of therapeutic claim target.

The therapeutic claim target for dementia, said Dr. Schneider, be it Alzheimer's or another form, is an entity called mild cognitive impairment. There are a number of MCI trials going on right now. They're largely set up as either survival trials or a combination of time-to-endpoint plus documentation of symptomatic change. What is implicit in these trials is that "MCI is MCI," a condition that can be agreed upon as a therapeutic target. That being said, when you follow patients long enough many will revert to normal without anything being done. Dr. Schneider cited a recent French study showing that within 2-3 years only 6% of the subjects continued to have MCI. Twenty percent or so developed dementia but 40% of these patients reverted to normal. It's clear, said another speaker, that: "…not all MCI is dementia, Alzheimer's disease. No controversy about that." Physicians don't think that every patient with MCI has or will develop AD, but it is a way of identifying those who may develop the disease. That, the speaker said, is the distinction.

Dr. Schneider reminded that panel that the more important goal is prevention; but here we have to work with people who don't have the illness to achieve our targets of therapeutic claim. Few medical tools available at present offer the possibility of pre-clinical dementia detection. However, since many patients with cerebral blood flow abnormalities are known to develop MCI and eventual AD, neuroimaging may be a helpful tool to detect and monitor the evolution of this dementia. Moving on, Dr. Schneider directed attention to the concept of vascular dementia and Alzheimer's disease with cerebral vascular changes as possible targets of therapeutic claim. Pfizer has at least two trials using vascular dementia as a target of therapeutic claim, studying the effects of their drug for six months. There was another study though, a galanthamine trial, where patients with AD and vascular dementia were combined. When the subgroups were examined, patients within each subgroup reacted differently to treatment. Patients with vascular dementia did not deteriorate on placebo, they improved. In contrast the Alzheimer's patients had more typical stabilization with active drug. This showed up on the ADAS-cog (Alzheimer's Disease Assessment Scale -cognitive subscale). The AD patients given placebo deteriorated. The patients were grouped together to get enough data for statistical analysis, but what did that study really prove? In contrast to the galanthamine and Pfizer trials, a memantine study on people with vascular dementia and AD showed both improved on active drug and deteriorated on placebo. This speaks of the heterogeneity of both dementias in outcomes over six months.

"We need to deal with symptomatic treatment vs. disease modifications," continued Dr. Schneider. Can a distinction really be made? If a company decides to do a disease modification study, a difference in cognitive function of the actively treated vs. placebo over time will have to be shown. This can be done in at least two ways: one is by using a milestone endpoint considered to be an index of disease progression, such as dropping of a few points on one of the well-accepted neuropsychological tests or time to nursing home placement. The other is the "randomized start" and "randomized withdrawal" model that Paul Leiber put forth some time ago. Hoechst Marion Roussel tried this with propentofylline. However, convincing evidence of successful treatment or benefit was not shown with this agent [that was proclaimed to have a neuroprotective effect on astrocyte-derived glial cells]. It became obvious that the drug wasn't going to get labeling. Propentofylline was ultimately withdrawn for other more technical reasons as well as lack of efficacy.

John Morris MD said that, in his opinion, patients don't want another drug that's better than placebo. They want a drug that's better than donepezil. "So," he continued, for patients and their caregivers, the comparison is not: "…is the drug better than placebo? It's, is it better than what we have?" There's also the difficulty involved recruiting people to participate in trials that preclude using approved cholinesterase inhibitors. By their very nature, the patients enrolled are very nonrepresentative. "It's not practical -for years I haven't considered it ethical and now I don't think it's practical either." It's not proper to deny patients effective medication. Rachelle Doody MD brought up a slightly different point. She said: "We don't have the methodology to differentiate between a clinically disease-modifying agent and a pathogenically disease-modifying agent. At the end, they're going to look the same." She also urged the group to keep in mind that cholinesterase inhibitors are disease modifying agents, at least they are out to a one-year study. If a five-year study is planned, the investigators are going to have to plan for that and power the studies accordingly.

Dr. Schneider noted that an agent is called "pathogenically disease modifying" when it's effect on disease pathogenesis is understood and "clinically disease modifying" when it's clinical effect is evident but the underlying mechanism(s) by which it exerts that effect are unknown. Peter Whitehouse MD said that patients only care about slowing the progression of their disease -they couldn't care less about the pathogenesis behind that slowing. Jeff Cummings MD remarked that: "…one difference we might expect is that if we really had disease modification, we might see increasing differences between the placebo and treated group over time." This is seen to some degree with the cholinesterase inhibitors, but we're looking for a bigger effect.

Zaven Khachaturian PhD said another paradigm needs consideration; programs that evaluate drugs -assuming they're safe-- in large populations and see if the treated subjects develop significantly less cognitive impairment over time. Dr. Schneider said this requires very big samples because 40% of MCI patients get better without any intervention. The numbers of subjects needed could present a formidable clinical trial problem. Right now we do 2-month, 6-month or 12-month clinical trials and get our patients either through clinics or advertising. The diagnosis of AD differs somewhat from site to site even though we try to have standardized protocols. The working definition of Alzheimer's is that you must see dementia. This is a consensus, albeit a somewhat arbitrary definition. It's illogical to think that a patient wakes up one day with AD and didn't have it the day before.

What we really need, Dr. Schneider continued, are longer clinical trials, four years for some -and that comes with a high cost. It's not always efficient either. Subjects are completely lost to follow-up…in the end there may be far less than half the original study group left, and probably no information at all on the people who have dropped out. Another problem is that we have a patient population that can be highly manipulated psychosocially. John Morris MD noted here that people who don't really understand AD think there's some sort of diagnostic gold standard. The disease is extraordinarily variable as are the diagnostic criteria used by clinicians.

The last point covered by Dr. Schneider was survival analyses. He said that in most cases they are underpowered. He gave the vitamin E/selegiline trial as an example. If that study had been structured to show risk reduction -we had about a 30% risk reduction with vitamin E or selegiline—we would have designed a trial to include about one-thousand patients. "To do these survival studies," he continued, "…you do need a good amount of patience."

The subject then turned to "proof of concept," and the difficulty establishing it for Alzheimer's. Dr. Schneider said that if a drug is safe, proof of concept could be established in people without illness -in other words, given to large numbers of people and then determine if the treated group develops significantly less dementia than what would be expected. On the other hand, he said, if you have an amyloid buster, proof of concept can't be tested in normals. This type of drug could needs to be tested in severely impaired patients using a surrogate marker to show it's working. The kind of study designed depends on the type of drug being tested.

Another speaker picked up on treatment of large populations -the public health approach successfully used for hypertension. Proof of concept would be the difference in dementia between placebo and treatment groups. It was pointed out that the trial wouldn't show specific efficacy for AD, but just for a general slowing of cognitive decline. Jeffrey Cummings MD suggested using imaging, enrichment designs. These could differentiate people with early AD from others with MCI. "There are," he said: "…a lot of ways to think about imaging that we haven not yet capitalized on." This proved somewhat controversial. Dr. Doody said that "a whole bunch of people"couldn't be in the study just because they "didn't happen to have access to that imaging." Questions would be raised by the regulators. Dr. Schneider moved the discussion to enrollment rates. Recruitment can be quick if people already on Aricept are taken in. However, are they typical patients? Carl Sadowsky MD said it's extremely difficult not to compare whatever we're studying to Aricept, especially in a six-month or longer trial. He used the Plavix study to make his point. "Every patient that came into the office was a candidate" either they were going to be on aspirin or a drug that might be better. There were no recruiting problems in that trial. Another panel member agreed saying: "…from a practical standpoint, most of the MCI patients that I see are already on cholinesterase inhibitors….and I think we have to factor that into our strategies when we design trials." Barry Reisberg MD took another approach in his memantine trial. Cholinesterase inhibitors are not approved for moderately severe and severe disease. "So we felt…we safely could do a trial saying that we're not in a position to prescribe an unapproved medication….we saw it as an opportunity to do what we though might be a purer trial." Dr. Schneider agreed that that is a valid technical position. Ranjan Duara MD does community screening for cognitive impairment -"…these are people who are impaired but didn't realize it or the symptoms were not [severe] enough to mention to their doctors."

Nancy Earl MD asked the panel if there were enough people coming into treatment centers who weren't on drugs. If not, those on drugs could have either placebo or active drug added. "I resume that would be ethical." Or, she said, you could take newly diagnosed patients and give everyone Aricept plus either active drug or placebo. That way there wouldn't be differences in prior drug treatment. Could both these groups be allowed in trials. Dr. Doody said if all were in one trial, they would have to be randomly assigned equally. Separate trials could easily be done. Guidelines may be needed and Dr. Schneider asked the group if they though new ones would evolve in the near future? Peter Whitehouse MD, PhD said he didn't see a burning desire among regulators for a whole bunch of new guidelines. Dr. Schneider agreed that FDA prefers to be as passive as possible saying "show us what you've got and we'll react to it." Unfortunately, FDA sticks to the old dogma that a placebo control is necessary to get a monotherapy indication, said another panel member. But, said Dr. Doody, "…this is a field in which you don't need a monotherapy indication in order to get people to use your drug." And there the discussion ended.