Updated 25 April 2008

This report summarizes research findings, intellectual input, and points of consensus shared by the meeting participants, the invited presenters and attendees of a workshop co-chaired by Marilyn Albert and Ron Petersen held on 19-20 June 2007.

Executive Summary
1. Background and Meeting Objective
2. Overview of Specific Challenges Facing Clinical Trials in AD
3. Measures That Distinguish Populations in Cross-sectional Studies
4. Measures That Detect Progression in Clinical Trials and Longitudinal Studies
5. Limitations of the ADAS-cog, and Its Use as a Basis for New Test Batteries
6. Existing Placebo Data
7. Innovation in Biomarker Development and Cognitive Testing
8. General Considerations for Improving Clinical Trials in AD
9. FDA Expectations for Supporting Data: Q&A with Russell Katz
10. Meeting Outcome: Main Points of Consensus
11. Meeting Outcome: Formation of Working Groups

Executive Summary

Overview
During the past two decades, intensive research in many laboratories has delineated some of the key steps in the pathogenesis of Alzheimer disease (AD). As a result, the AD field is transitioning from today's purely symptomatic therapies to an increasing emphasis on clinical trials of potentially disease-modifying agents. Historically, the first disease-modifying treatments to be approved for a disorder have often been less efficacious than later therapeutics, but these early therapeutics are nonetheless very helpful to patients who previously had no hope of slowing or arresting the disease.

Clinical trials in the near future will primarily be conducted in established AD cases. Though treating the disease earlier in its course may be more beneficial, conducting trials in milder cases may pose challenges—primarily in that owing to slower progression, and less specificity in the diagnosis at earlier stages, it may be more difficult to detect changes in the rate of decline. Thus, to avoid missing a significant clinical benefit arising from a potentially disease-modifying agent in clinical trials, there is an urgent need to identify neuropsychological tests that are sensitive to meaningful clinical change at different points in the course of AD, from early to later disease.

Towards this end, a meeting was held in Washington D.C. on June 19 and 20, 2007, to discuss the sensitivity of currently used tests and to identify areas that need further development. At this meeting, experts in the cognitive assessment of AD reviewed and evaluated psychometric instruments that can sensitively detect and quantify change in mildly affected AD patients. The meeting was the first step in a process intended to achieve a consensus about which tests are most sensitive and useful for detecting a slowing in the decline of cognitive function at different disease stages. The meeting also highlighted a secondary goal, namely to determine the most sensitive and useful tests for detecting a slowing in the decline in clinical functional outcomes, or activities of daily living, in patients early in the course of disease.

Meeting Participants
Marilyn Albert of Johns Hopkins University and Ron Petersen of the Mayo Clinic served as co-chairs of the meeting. Albert and Petersen were joined on the Organizing Committee by Adrian Ivinson of the Harvard Center for Neurodegeneration and Repair, John Morris of Washington University, Dennis Selkoe of Harvard University, and Stacie Weninger of the Fidelity Foundations.

Speakers
Dennis Selkoe, Harvard University; Mary Sano, Mount Sinai School of Medicine; Richard Mohs, Eli Lilly and Company; David Salmon, University of California, San Diego; Michael Grundman, Élan Pharmaceuticals; Robert Wilson, Rush University Medical Center; Ron Petersen, Mayo Clinic; Peter Snyder, University of Connecticut; Marilyn Albert, Johns Hopkins University; Julie Chandler & Lyn Harper-Mozley, Merck Research Laboratories; Roger Lane, Novartis Pharmaceuticals Corporation; and Russell Katz, Food and Drug Administration.

Attendees
Andrew Blackwell, Cambridge Cognition; Heather Dawes, Fidelity Foundations; Michael Gold, GlaxoSmithKline; Robert Goldman, Pfizer; Suzanne Hendrix, Myriad Pharmaceuticals; Kristin Kahle-Wrobleski, Eli Lilly; Marcelle Morrison-Bogorad, NIA; Eve Nichols, Fidelity Foundations; Tony Phelps, NIA; Holly Posner, Eisai; Bill Potter, Merck; Laurie Ryan, NIA; Jeffrey Sevigny, Merck; Nina Silverberg, NIA; Yaakov Stern, Columbia University; Morihiro Sugishita, University of Tokyo; Bill Thies, Alzheimer's Association; Molly Wagster, NIA; Michael Ward, Genetech; Kathleen Welsh-Bohmer, Duke University; Michael Weiner, San Francisco VA Medical Center; Stacie Weninger, Fidelity Foundations; Frost White, Pfizer; Fumio Yamashita, University of Tsukuba.

Discussion of Cognitive Tests
This report details the issues that were raised over the course of this two-day meeting, highlighting next steps necessary to achieve the intended goals. Much of the discussion centered around the Alzheimer Disease Assessment Scale (ADAS), which is the test battery most commonly used today to assess progression of AD. The ADAS was designed as a standardized assessment that could be used to measure the severity of major dysfunction in cognitive and non-cognitive behaviors characteristic of AD sufferers. It has proved to be suitable in a variety of settings (e.g., across cultures and languages). The 21-item ADAS is a performance-based scale that includes 11 items to assess cognitive function, e.g., memory and orientation. This cognitive portion of the test is known as the ADAS-cog.

One of the issues under discussion was whether the ADAS-cog may be better suited to detecting decline in function for patients at some disease stages (for example, moderate AD) than other stages, including mild cognitive impairment (MCI), a condition thought in many cases to precede frank AD. Those patient populations most effectively assessed by the ADAS-cog may be revealed in future analysis. Current interest in this area stems in part from the fact that several recent studies using the ADAS-cog have failed to detect significant decline in the placebo groups of clinical trials—a problem for MCI populations, but in some cases for established AD as well.

The inability of the ADAS-cog to detect decline in some placebo control studies has highlighted the urgent need to develop more sensitive tests, including tests useful at the early stages of disease. Many of the participants suggested that, although the ADAS-cog may be perfectly adequate for moderate AD, the development of a sufficiently sensitive test battery for very mild AD and MCI may best be achieved by combining the most sensitive measures for changes early in the course of disease from the ADAS-cog and other cognitive tests. Given that deficits in declarative memory seem to precede more general cognitive dysfunction in AD, test batteries focusing on this realm of cognition may prove most sensitive at the early stages of disease, though measures of attention or executive function may also prove useful. Analysis of placebo data from the numerous clinical trials conducted by the Alzheimer Disease Cooperative Study as well as by pharmaceutical companies may allow for the identification of specific components of the ADAS-cog that are sensitive at earlier disease stages. These components may then form the basis for a new test battery for trials in MCI or very mild AD.

There are currently many other cognitive tests in use or under development that should be considered when determining the most effective tests for early disease stages. Participants in the meeting discussed various measures, including tests specifically targeting verbal memory, computer-based instruments, and new physiological and cognitive assessment approaches. The meeting participants agreed that there are currently not enough data to allow for a complete evaluation of the usefulness of these newer tests. Moreover, a successful battery for detecting deficits in mild disease my in fact be a combination of tests from current test batteries. Hence, a thorough analysis of the cognitive tests currently in use—particularly regarding their sensitivities to changes at different stages of AD—would appear to be warranted.

Regulatory Perspective
Russell Katz from the Food and Drug Administration (FDA) provided valuable insights into what might be required by the FDA for the approval of a new drug for AD. Importantly, he confirmed that tests other than the ADAS-cog could be acceptable for use in clinical trials, and that it would definitely be acceptable to use different test batteries for different stages of disease. However, he did raise a caution about the use of tests that were highly sensitive to decline in cognitive function in a specific domain, reiterating that in addition to cognitive data stemming from such tests, drug approval would also require evidence that the drug under investigation had a meaningful clinical effect—that is, a positive effect on patients' daily function. The requirement for functional outcome data is particularly challenging for studies of early disease stages. In MCI, change in functional impairment is difficult to detect even in untreated individuals. Functional measures such as the Disability Assessment for Dementia scale (DAD) and the Alzheimer Disease Cooperative Study Activities of Daily Living (ADCS-ADL) score may not be sensitive enough to detect change in deficits in MCI. This is not surprising, however, since, by definition, MCI subjects should be minimally impaired, if at all. This presents a conundrum for assessing functional changes in the MCI stage of the illness.

Specific labeling for drugs may be possible—e.g., a drug may be labeled to affect MCI, which could be an "approvable" indication. However, Katz emphasized that in such a scenario MCI would need to be able to be diagnosed by trained clinicians in the community—not just those in tertiary care centers—before the FDA could approve such a drug.

Participants agreed that there is an urgent need to develop more sensitive functional tests to detect deficits in global function in very mild disease. Meeting participants suggested that tests that detect "memory-related problems in every day life" may represent a way to assess meaningful functional improvement at stages prior to detectable global functional deficits. The ability to detect change in such measures in the context of clinical trials is an issue that will require attention as the field moves forward.

Next Steps
Based on the discussions at the meeting, three working groups were launched with the following goals:

1. Develop a repository for existing placebo data from clinical trials performed in the academic and pharmaceutical sectors. Once such a database has been developed, the working group will coordinate and oversee the analysis of the placebo data by expert statisticians, who will seek to identify the components of currently used test batteries that are most sensitive to change at different stages of disease, and investigate whether the addition of selective items might improve the sensitivity of these tests.

2. Assess the utility of newly developed measures of cognitive assessment, including computer- and telephone-based tests for evaluating change in performance at different stages of disease.

3. Examine the feasibility of developing a universal scale for assessing change in cognitive performance over the range of MCI and AD. This would incorporate extant and anticipated data from the first two working groups, as well as other sources and additional analysis, in the development of a universal scale.

The meeting participants also highlighted the future need for a fourth working group focusing on the development of sensitive measures of global function in mild disease.

The ultimate goal of these working groups will be the development of a discrete set of standardized test batteries for common types of AD clinical trials. These standard measures, and the analysis that informs their selection, should allow the field to design and conduct effective clinical trials that produce comparable outcome data. The hope is that the findings produced through these continuing efforts will help guide the design of future clinical trials, and help pave the way towards finding effective therapies for AD.

1. Background and Meeting Objective

There is an acute need for the rapid development of approved disease-modifying agents for the treatment of Alzheimer disease (AD) and mild cognitive impairment (MCI), which may be an early manifestation of AD in some circumstances. Successful efforts to develop such therapies will require advances in understanding the basic biology of AD, innovations in drug development, and the creation of more powerful tools for cognitive assessment, particularly in the context of clinical trials.

Certain characteristics of the current clinical environment stand out in shaping the need for improved cognitive assessment techniques. Clinicians are seeing patients earlier and there is increased emphasis on including patients with early AD in clinical trials. Unfortunately, current cognitive tests may not be adequately sensitive at all disease stages, which could account for the observed difficulty in tracking cognitive changes in placebo groups in recent clinical trials. (Another possibility is that cohort effects have interfered with the interpretation of recently obtained placebo data.) Furthermore, innovations in the pharmaceutical realm raise the possibility of testing the effectiveness of drugs at specific disease stages, in particular cognitive domains, or within certain genetic backgrounds, meaning that the demands on cognitive assessment techniques will likely grow increasingly sophisticated.

In this context, and with several potential disease-modifying drugs currently in clinical trials, the need to rapidly improve current approaches to cognitive assessment is clear. Without new approaches, there is the risk that an agent with the capability of at least partially modifying disease will fail to become available because inadequate cognitive testing measures have obscured evidence for its ability to slow global cognitive and behavioral decline.

A central aim of the talks and discussion in the recent meeting was to reach consensus on ways to move forward in identifying the challenges of cognitive testing in AD clinical trials, and to identify next steps for addressing those challenges. The following report includes key observations, general considerations, and points of consensus regarding cognitive assessment in AD clinical trials, and a potential strategy for the development of better tests. This account represents a summary overview of the meeting's discussion on a number of points relevant to the improvement of cognitive assessment in clinical trials, and suggests points of potential consensus. However, owing to the current need for directed analysis to improve our understanding of cognitive testing in AD, the development of a formal consensus report must await forthcoming studies (see below, "Meeting Outcome: Formation of Working Groups").

The following sections touch on main points of the meeting's discussion, but do not in all cases represent a consensus view among researchers.

2. Overview of Specific Challenges Facing Clinical Trials in AD

Specific practical challenges underlie the development of improved cognitive tests for clinical trials. A primary challenge is to overcome the difficulty, encountered in a number of recent clinical trials, in detecting cognitive changes in disease-affected placebo groups. Another challenge is to improve cognitive testing to overcome so-called floor and ceiling effects, which occur when tests are too difficult or too easy to detect changes in cognitive performance at different disease stages.

Important challenges to the improvement of clinical trials more generally include the regulatory need for functional outcome measures (this challenge is particularly strong in the treatment of mild disease); the need for improved measures for trial recruitment and diagnosis; and, as clinical trials increasingly involve international efforts, the need to obtain comparable data from domestic and international studies, in diverse populations

3. Measures That Distinguish Populations in Cross-sectional Studies

In addition to the issues noted above, there is also a need to ensure that tests for clinical trials are relatively stable over the short term (when meaningful change is not likely to be occurring), but sensitive to change as the disease progresses over the longer term. Such considerations may impact the utility of tests for different purposes—e.g., measures that are useful for cross-sectional studies may not necessarily be appropriate for clinical trials; cross-sectional features may not be replicated longitudinally. That said, the ability of a given test to distinguish between individuals at different stages of disease potentially provides some indication of its sensitivity to disease progression and hence its potential utility in longitudinal studies and clinical trials. The extent to which a test can help distinguish clinical populations in cross-sectional studies may also reflect its sensitivity to changes in particular cognitive domains. In addition, analysis of disease progression following cross-sectional cognitive analysis may provide insight into a test's prognostic utility. In the context of clinical trials, sensitive prognostic measures may be useful for enriching subject samples to shorten trial durations or reduce sample sizes, and, potentially, to help establish reliable and valid endpoints in Phase II and III trials.

Prediagnostic Trends
A number of comparative studies of elderly individuals complaining of mild memory loss indicate that tests of verbal memory (e.g., delayed recall, California Verbal Learning Test [CVLT], Logical Memory test) may be most useful for distinguishing between progressors to AD and non-progressors. Such studies suggest that tests of spatial memory, attention, and semantic memory may be less useful in this context. Other findings suggest that decrements in episodic memory can precede the onset of AD by several years or more, and may serve as a marker for the future onset of dementia. However, other studies have indicated that different cognitive measures - such as those assessing abstract verbal reasoning, attention, executive function, or judgment and problem solving, may also be among the best available cognitive predictors of AD. Lack of consensus on this point may be relevant in considering some issues discussed below (e.g., "Domain sensitivities in healthy elders and MCI").

Trends in MCI and Mild AD
Evidence from past studies in MCI and mild AD suggests that, in the course of disease emergence, declines tend to first appear in episodic memory, followed by semantic (especially verbal) memory, executive function, language, and then finally visuospatial skills, respectively. The progression is such that by AD diagnosis, episodic memory is usually so poor that many traditional cognitive measures are severely hampered by the floor effects noted above.

Such decline is consistent with a number of observations from the Alzheimer's Disease Assessment Scale—cognitive (ADAS-cog), a battery including 11 items that measure performance in memory, orientation, language, and praxis (a 13-item version also measures concentration/distractibility and delayed recall). In comparing baseline ADAS-cog item scores in normal, MCI, and AD subjects, large differences have been seen in memory for normal subjects compared to MCI subjects, while differences in other skills become more pronounced in comparing MCI and AD subjects.

These findings are also in line with an ongoing study in which three-way comparisons of mean baseline scores for normal controls, MCI, and AD, yielded significant differences in the Wechsler Memory Scale, CVLT, Trial-Making Test B, and two computer-based tests, CANTAB graded naming test (confrontation naming) and CANTAB paired associates learning test, with indications that the latter may be particularly strong in distinguishing normal controls from MCI subjects.

Together, findings from cross-sectional studies indicate that the ability to accurately distinguish populations according to disease stage and prognosis might be potentially improved by utilizing test batteries weighted for measures sensitive in certain domains—in episodic and semantic memory (for prediagnostic and MCI groups), and also executive function (for MCI and mild AD groups). Combining such information with imaging or other biomarker data, when available, may prove especially effective. Honing such approaches to improve prognostic power—to benefit both diagnosis and trial recruitment—remains a critical challenge.

4. Measures That Detect Progression in Clinical Trials and Longitudinal Studies

There very likely is no "one-size-fits-all" test battery presently available that effectively detects changes in all stages of AD; given the presently available tests, it would seem unrealistic to expect the same cognitive (and functional) scales to be sensitive in studies of mild and moderate AD, or in "prevention" trials comparing MCI and healthy elders.

ADAS-cog
Studies indicate that the ADAS-cog may not be ideally suited to assessing change at all AD disease stages. For example, some observations suggest that the ADAS-cog may be most useful for tracking changes in longitudinal studies and clinical trials involving moderate AD; the test also may be sensitive for showing short-term differences in mild AD, but it is not clear that it is suitable for use in MCI or late AD.

At least some analyses of the ADAS-cog subscores in early disease indicate that changes first reveal memory problems, but the measures of memory quickly bottom out (the "floor" effect). Later changes are generally seen in language, then in items corresponding to other domains. Thus, before deciding to use the test, it would appear important to consider the range of capacities in the population under study.

Owing to the reliability and sensitivity the ADAS-cog has shown in the past, some researchers feel that, given present options, it remains a good choice for clinical trials—but it is also the case that the ADAS-cog has not always appeared sufficiently sensitive in placebo groups in recent clinical trials, leading to concern about its utility, particularly in milder populations. Comparative analysis of extant placebo data (see below) may clarify the disease stages and patient populations for which ADAS-cog and its components are most powerful in the context of clinical trials.

NTB
The Neuropsychological Test Battery includes tests of immediate and delayed visual and verbal memory, as well as executive function. While currently a non-standard instrument in the AD field, the NTB may have some advantages for mild AD: there is evidence for decline on the NTB in mild AD; it focuses on memory and executive function; it is familiar to psychometricians; and it is comprised of established, validated scales in multiple languages. It also includes multiple tests per domain, which can help with variability problems, offering a statistical benefit by bringing down standard errors of measurement.

The NTB has been used successfully to detect memory improvement (relative to placebo) in the AN1792 clinical trial, and showed consistent rate of change in patients of varying disease severity. Compared to the ADAS-cog, the NTB may in some circumstances be better able to detect change over time in mild patients.

One present limitation of the NTB is that the score is calculated based on the population under study, and hence is not directly comparable to NTB scores from other studies—this lack of standardization may inhibit clinicians from using it. In principle, the scale could be made invariant across populations.

Other Composite Measures
Analyses of recent cohort studies employing composite measures have provided a number of insights. Among these are that measures of global cognition based on two or more tests can usually adequately capture individual differences in cognitive decline; that cognition declines much more slowly in MCI than in AD; and that in assessing cognitive decline in MCI and AD, the ability of cognitive outcomes to accommodate a wide spectrum of ability is probably more important than the domains of ability being assessed. This latter point may represent the main advantage of composite tests. Evidence also indicates that the rate of cognitive decline in AD is strongly related to morbidity and mortality.

A danger for a given composite measure is that it will not be relevant to a given stage of disease. This raises the question of whether, in a population that potentially spans disease stages, one can usefully test subjects on an individual basis. GRE-style adaptive testing, using item-response theory, is a possible way to address this issue, but will require much development. That approach, while potentially useful in broadening sensitivity range, may also provide relatively better information about individuals than groups (or may require larger subject numbers in trials) and may bring up a whole new set of challenges for clinical trials.

Domain Sensitivities in Healthy Elders and MCI
Changes in cognition are dependent upon the stage of the disease process. In healthy elders, testing results indicate that memory performance (e.g., measured by list learning) is a strong predictor of change and appears to show continued deterioration, and that attention and executive function are also sensitive to early change.

In MCI, standard instruments tend to exhibit subtle changes. Nonetheless, in the context of clinical trials involving MCI subjects, measures in various domains—memory, executive function, language, and visuospatial skill—can be useful in detecting drug effects. Tests that have detected change in MCI subjects relative to unaffected individuals in longitudinal studies include the Auditory Verbal Learning Test (for memory), category fluency (for language), and Trails B (for attention/executive function).

In examining MCI, the usefulness of particular instruments may depend on a study's selection criteria, and hence where a population falls in disease progression. Even within the MCI diagnostic class, there is a wide range of performance on instruments depending on where a patient lies on the MCI continuum. At the more advanced end of MCI, memory tests may no longer be the most sensitive because of the degree of memory loss that has already occurred, and language and executive function may likely be more fruitful for testing. At the milder stages of MCI, memory tests may be the most sensitive.

5. Limitations of the ADAS-cog, and Its Use as a Basis for New Test Batteries

A number of people in the field believe that a primary limitation of the ADAS-cog is its apparent lack of power in detecting decline in placebo-treated populations in recent clinical trials. While some meta-analyses have suggested that this is not so, a recently offered estimate is that future AD disease-modification trials relying on the ADAS-cog may require a minimum of 18 months of treatment to account for slowed rates of progression. If this is indeed the case, there is a need for more sensitive tools to enable earlier readout of treatment efficacy, particularly in mild disease.

Furthermore, other analyses of placebo data from multiple trial studies suggest that the ADAS-cog may not be ideal for a short proof-of-concept study with 40-50 patients, and that it would instead be more productive to use more sensitive and precise measures for a small study population, looking at specific cognitive areas.

Likely related to the concerns mentioned above is an observation from Rasch analysis indicating a 10- to 12-fold difference in the meaning of a 1-point change in ADAS-cog score, corresponding to where an individual's performance lies on the test's scale—suggesting that the middle of the ADAS-cog scale may be useful, but its "edges" may be difficult to interpret.

In spite of the above limitations, the ADAS-cog may remain a reliable and valid measure of cognitive function for moderately affected AD patients, at least in some circumstances, and modifications can potentially enhance its utility for other populations, including those only mildly affected by disease.

The ADAS-cog possesses a number of advantages that may make it useful as a basis for the development of new test batteries: reliability, brevity, specificity for major AD symptoms, application in a variety of settings, utility in assessment of increasing symptom severity, and low sensitivity (relative to the MMSE) to education levels.

A general approach for modifying the ADAS-cog for use at different disease stages will be to augment it with measures known to be sensitive to specific cognitive domains. For example, to test a hypothesis dealing with loss of memory early in disease, one might ideally choose a standard test employing a weighted battery that retains some items from the ADAS-cog, like word-list learning, but also includes measures such as delayed recall and non-verbal recall. In principle, this general approach would be wholly appropriate from a regulatory standpoint, though down the line it might require discussion of how to best label a drug tested using a weighted battery, if data indicate drug benefit for a restricted set of cognitive domains.

6. Existing Placebo Data

Raw placebo data from previous Alzheimer's Disease Cooperative Study (ADCS) and pharma-sector trials could be highly beneficial to the community if they were collected in a common repository (see Working Group 1, below). Analysis of such data could potentially shed significant new light on central issues in the field, including the sensitivity of different test items to different stages of disease (as discussed in the previous section), and how trial recruitment criteria and other variable factors may affect the ability of various tests to detect appreciable change in untreated placebo groups. Such work may also aid in dealing with ethnic and cultural variations in both national and international trials.

Longitudinal studies involving most of the standard instruments are now published for both AD and MCI (see Further Reading, below). In addition, data collected from a number of leading pharmaceutical companies, including data on mean changes for 6-month and 12-month studies, will soon be published. Such data represent a highly valuable resource for analysis.

In a recent study, comparison of ADCS-ADL (activities of daily living) and ADAS-cog score changes in placebo groups from two large clinical trials with similar patient characteristics (one in 2000, the other in 2005) indicated a 15 percent slower decline in ADCS-ADL scores in the 2005 placebo group compared to the 2000 placebo group, and a 16 percent slower decline in ADAS-cog scores in the 2005 placebo group compared to the 2000 placebo group. Forthcoming analysis of additional placebo data will potentially help illuminate the basis for such findings and other related trends.

7. Innovation in Biomarker Development and Cognitive Testing

As work in the AD field continues, new approaches could potentially contribute to improved cognitive and physiological measures of treatment response in clinical and preclinical AD. Among these are possible biomarkers for memory decline, such as the Late Positive Component (LPC, or P600) of the Evoked Potential Response; cortico-cortical disconnectivity (e.g., evidenced by sensory integration deficits), which has been seen in early AD; and other preclinical signs that may be associated with increased risk of AD, such as cognitive asymmetry. Other areas of ongoing investigation include sensitive cognitive testing paired with pharmacologic challenge, and the assessment of long-term memory consolidation, which has not been measured in tests used in past AD clinical trials. More established approaches, including imaging and CSF biomarker analyses, continue to be studied. While holding potential for the improvement of testing in AD trials, all these tools will require ongoing characterization, especially regarding their ability to detect clinically meaningful change, their functional correlates, and their ability to predict disease progression.

Other areas of innovation include the development and application of electronic testing technologies. Computer-based tests (such as the CANTAB, the CogState Early Phase Battery, the Groton Maze Learning Test, the Computerized Neuropsychological Test Battery, etc.) are being examined for their utility in clinical trials. Generally speaking, there is a place for both technology-based assessment and traditional assessment in AD and MCI clinical trials. In developing standardized tests suited for different disease populations, success will rely on choosing tests that are fit-for-purpose relative to the population under study, with regard to both disease-stage and a given population's acceptance of different technological approaches. Computer-based methods may play a leading role in assessment in coming years, especially in efforts to detect and track very early disease through cohort studies and clinical trials.

A main advantage associated with technology-based measures is their ability to facilitate in-home testing, which is seen as a way to avoid participant drop-out during trials and cohort studies. These approaches include "tester-administered" phone-based cognitive assessment, telephonic assessment with automated presentation and vocal and key-pad response, and computerized assessment for presentation and response capture.

Ultimately, the utility of different technology-based approaches will be determined through evidence obtained from cohort studies and clinical trials. A number of studies utilizing computer-based assessment are underway or in planning stages.

8. General Considerations for Improving Clinical Trials in AD

Trial Populations
Clinical trials deal with unique populations of individuals whose characteristics are shaped by inclusion criteria, cultural background, and numerous other factors—some of which, such as recruitment strategy, may be idiosyncratic to a given clinical trial. This raises the importance of caution in applying conclusions from non-clinical trials data to clinical trial populations, and vice-versa: depending on circumstances, inferences from one study may not translate well to another population. In light of this, an important consideration in designing clinical trials will be to ensure, to the extent possible, that trial population results will be applicable to the general population. Owing to such factors as self-selection among trial populations, the changing characteristics of those populations over time, and the likely prevalence among study participants of "non-responders" (those unresponsive to previous treatments with other drugs), the pharmaceutical sector has struggled with this issue. Age is another important consideration—trial outcomes and drug effects that are salient in younger populations are not necessarily salient in older populations.

Subject Recruitment
Selection variables can be useful for defining a recognizable population, for boosting the likelihood of a given test or test battery to detect an effect, or to show proof-of concept in an early phase trial for a go/no-go decision. Similarly, selection of a study population by stage or symptom can reduce variance, permit a smaller sample, offer tighter control of trial length, permit the use of stage-specific outcomes with well-known effect size, and offer improved control of side effects. However, selection by stage or symptom may require a longer recruitment period or result in loss of generalizability—for example, tests used in an early-phase trial may need to be amended to study effectiveness in a larger population.

Diagnosis is also a challenging issue in subject recruitment. Despite the fact that the disease progresses through a continuum, the diagnosis point for AD varies; in addition, subjects in MCI trials and studies are often identified using variable implementation of criteria.

Clinical Trial Strategies
Clinical trial designs must necessarily balance the expense of large sample sizes with the benefits that large studies confer in detecting slowing rates of disease progression (a recent analysis of placebo data indicates that sufficiently large samples can help overcome some limitations of the ADAS-cog, for example). An additional challenge is the risk of producing a failed trial that may not represent a true negative result.

In addition to such general considerations, some additional factors impacting clinical trials are now relevant for pharmaceutical companies. One such issue is the possibility of a drug improving only one aspect of cognitive decline—given that AD is defined as a multiple-domain cognitive disorder with functional impairments, a single drug may not be able to address all aspects of disease. Such a possibility has implications not only for the analysis of trial data, but also raises questions regarding data requirements for drug approval, and the eventual labeling and marketing of drugs.

The increasingly important role of testing in international populations also represents a challenge for companies designing clinical trials. Evidence indicates that despite population matches on such measures as the MMSE, or age and education, cultural differences can alter outcomes on cognitive measures and cause significant problems in comparing data within and between multicultural or multinational groups. Analysis of placebo data obtained from such studies may prove essential in the future development of effective and useful tests for international clinical trials. It is also important to recognize that culture and language can influence cognition in poorly understood ways, and that culture-fair testing may not be a wholly realistic goal. An approach to improving comparability in international studies may be to stay away from the direct translation of tests, and to instead collaborate closely with local researchers and clinicians and talk to them about how they would go about testing cognitive performance in a given population.

Finally, another potential hindrance to clinical trial success is that while the U.S. has a large number of study sites with well-trained staff for use in clinical trials, some large trials have revealed that the monitoring practices are not always consistent. In addition, restricting clinical trials to only these locations may prove prohibitively expensive in some cases.

Test Evaluation and Development
An important general consideration in assessing strengths and weaknesses of existing tests, and their utility in different types of studies, is that measures that are useful for diagnosis, cross-sectional studies, cohort studies, and short and long clinical trials are not necessarily the same—for example, a good test for cross-sectional studies is not necessarily a good test for longitudinal studies.

In assessing existing tests and developing new tests and test batteries for use in AD studies and clinical trials, a framework that outlines useful test characteristics might prove constructive. In general, useful tests are rapid, repeatable, and highly reliable. They are not dependent on experts for administration, and are free of significant re-test effects. Such tests encompass a broad range of possible values; exhibit bi-directional sensitivity; and are not culturally, socially, or educationally biased. Tests and their administration should be standardized and employ easily accessible equipment. Ideally, data ought to be normally distributed, with no correlation between variance and mean, and a small coefficient of variation.

A particular set of considerations accompany the development of batteries employed in early-phase trials. To examine proof-of-concept for potential new drugs, studies must be of short duration and include clinical measures that, ideally, meet the following criteria: they target specific cognitive domains affected by AD and have broad domain coverage; have adequate psychometric properties; can be repeated several times; correlate with currently accepted clinical measures; can discriminate across levels of severity; and can predict functional outcomes. Batteries of specific cognitive tests with broad domain coverage may be useful as cognitive biomarkers to enable smaller studies of shorter duration.

To ensure that tests and batteries remain sensitive to changes in specific cognitive domains, some parts of a scale might, in theory, be expanded by making questions easier or more difficult. However, it may not be the case that difficulty equates with sensitivity—the best cognitive test for measuring change may not be the most difficult. Perhaps a more effective way to boost sensitivity would be to assess how a test range matches the ability range of the population under study in a given trial. A new test battery's sensitivity and flexibility might also be built up by enrichment with additional tests, but it should also be noted that cognitive testing itself is a prevalent reason for non-compliance—hence, for at least some trial scenarios, a more extensive battery may not be the best solution.

A related concern is that weighting fit-for-purpose batteries to compensate for floor effects in an early-moving domain such as memory might mean that improvements in that domain would be missed in a clinical trial. Adaptive testing may be one approach for addressing this issue.

Another consideration in longitudinal studies and trials is that the sensitivity of scales in measuring change over time—even over only six months—may potentially be boosted if one were to "smooth out" the data by correcting for baseline values.

Finally, in addition to creating new tests and batteries, it might also prove worthwhile to develop ways to easily compare the sensitivities of different tests—for example, the ADAS-cog vs. the NTB. A number that designates a test or battery's sensitivity—perhaps the mean change per year /SD (i.e., how much of SD is a mean decline per year)—would be useful for comparing data generated by different measures and studies.

Challenges of Functional Assessment
Current FDA guidelines require that data on both cognitive and functional outcome measures be included in data packages supporting approval of new AD drugs. A primary rationale for the requirement is that it provides assurance that a treatment effect observed in the course of a clinical trial will matter for the patient receiving the drug.

The requirement for functional outcome data is a challenge for AD trials in general, but particularly so for tracking change in early disease and MCI—in the latter case, functional impairment is barely evident, if at all, even in untreated individuals. Functional measures such as the Disability Assessment for Dementia scale (DAD) do not appear to be sensitive for distinguishing normal vs. MCI populations. Therefore, there is currently a need to develop global functional scales for MCI and very mildly impaired AD patients.

The regulatory requirement that an effect seen by cognitive measures should be shown to have clinical meaning is likely to remain an important principle, even though it is presently unclear how best to assess functional effects in early disease. That said, the question of the time needed to see functional outcome benefit is generally seen as a fair one to consider now that trials are potentially looking at effects at earlier and earlier disease stages. (One issue this raises is how to demonstrate efficacy for a disease-modifying therapy that does not have a short-term symptomatic effect.) Appropriate drug labeling may eventually help address this issue.

9. FDA Expectations for Supporting Data: Q&A with Russell Katz

Russell Katz, Director of the FDA's Division of Neurology Products, provided valuable insight into the FDA's perspective on the use of cognitive assessment in clinical trials for AD drugs, and the basis for the Agency's current requirements for data packages accompanying successful applications for drug approval. Katz responded to pre-submitted questions in a formal Q&A period.

Q: It is now clear that AD pathology is evident at least a decade before the disease is full blown (e.g., pathology is present in about 25 percent of controls and approximately 50-60 percent of MCI cases). Would a disease-modifying drug need to demonstrate an effect on both the clinical state and pathological state or would the former be sufficient?
A: An effect on a clinical outcome, in an appropriately designed trial, would be sufficient to support a disease-modifying claim. This is true for clear AD as well. An effect on the pathological state might, under certain circumstances, be part of the support, and might, at some point in the future, constitute the sole support for approval (in particular, perhaps in appropriate subjects who are not symptomatic).

The use of a withdrawal maneuver to compare the effects of a putative disease-modifying agent with those of a placebo is considered, at this time, the most compelling evidence of a disease-modifying effect: a structural effect will mean that after withdrawal, the gain is sustained relative to the placebo over time. A symptomatic effect will mean that the gain is not sustained relative to the placebo control group—the performance trend of a treated group after withdrawal returns to the performance level of the placebo group.

Q: Will the agency accept tests other than the ADAS-cog as endpoints in clinical trials for AD and MCI? This question is based on the increasing concern that more sensitive tests are needed than those commonly used to assess efficacy in clinical trials for AD and MCI.
A: Other tests certainly could be acceptable. However, the basis for the desire to identify tests that are more sensitive is unclear. Obviously, the ADAS-cog is sufficiently sensitive to detect small (mean) effects. Indeed, the current requirement to show an effect on a global measure is based on the considerable sensitivity of the ADAS-cog. A primary consideration is that any test or tests chosen should reflect a drug effect that is clinically meaningful for the patient.

Q: If tests other than the ADAS-cog would be accepted by the Agency, would it be correct to say that additional tests would need to have 1) evidence of association with clinical status cross-sectionally, 2) change in relation to clinical status longitudinally, and 3) longitudinal data covering a period of at least six months?
A: This seems reasonable.

Q: If tests other than the ADAS-cog would be accepted by the Agency, could the tests be given either by a computer or paper and pencil?
A: Assuming that the computerized tests are shown to be reliable and valid, computerized testing would be acceptable.

Q: If tests other than the ADAS-cog would be accepted by the Agency, would it be acceptable to use different tests for different levels of severity (e.g., AD, MCI, normals)?
A: Yes, as appropriate. Even in patients with very early disease, it seems reasonable to preserve the principle that an effect on a "global" measure (that is, a measure that reflects [and essentially "guarantees"] that the effect seen is clinically meaningful) should also be required.

An effect on a test of a single cognitive domain raises important questions: what claim would this support? That is, would such an effect support a global dementia claim, or only a claim for an effect on that specific cognitive domain? What do we do about patients who appear normal (and who are likely to go on to develop AD) but who have cognitive deficits? Are these patients appropriate for study? Are these patients with AD?

Q: If tests other than the ADAS-cog would be accepted, would the Agency accept a universal cognitive scale that incorporated items with a broad range of difficulty. As envisioned, such a test would have a small number of items that everyone would receive and performance on these items would determine whether easier or harder items would then be given. If such a test could be developed, conceptually it could be considered a "universal cognition scale" and not a scale for AD dementia. Would this be acceptable in principle?
A: In principle, yes. Such a scale could be used to support an AD (or other specific) dementia claim, or even perhaps a "global" cognitive claim. For the latter, an effect across different dementias would be required.

Q: Would the Agency accept as an approvable drug a medication that showed a significant benefit in relation to a placebo for mild AD patients, without an associated change in global functioning? This question is based on the fact that mild AD patients show impairments in daily function that are less dramatic than moderate and severe AD and change may therefore be more difficult to demonstrate on a global scale.
A: We acknowledge the problem, but we are still interested in knowing that a given change in a cognitive measure was clinically meaningful. Absent compelling evidence that a small change in a cognitive measure alone was by itself evidence of clinical meaningfulness, a measure of such meaningfulness would likely still be required.

Q: Does the Agency require that demonstration of a disease-modifying drug in AD patients is necessary prior to testing the drug in studies of milder subjects (e.g., delaying time to progress from MCI to AD or delaying time to progress from normal to MCI?)
A: Clearly, such a demonstration is not necessary. It should be pointed out that a demonstration of delaying time to AD or to MCI does not, in itself, establish a drug as disease-modifying.

Q: Would the Agency accept as an approvable disease-modifying drug a medication that showed a significant benefit on a cognitive measure and a biomarker that demonstrated a reasonably convincing effect?
A: We have said that it is possible that a combination of clinical and biomarker endpoints may be convincing, although perhaps less so (at this time) than a finding on an adequately designed trial utilizing clinical endpoints. It should be noted that such an effect would likely need to be on the typical clinical endpoints (cognitive and global measure), not just on a cognitive endpoint.

Q: Will the Agency consider other study designs in addition to randomized start or withdrawal for examining efficacy in AD? This question is raised because such designs currently require at least 18 months of observation for a Phase 3 trial in AD.
A: Perhaps. The recent "natural history staggered start" design (a single parallel group study in which patients of varying disease severity are randomized, with a goal of showing a differential drug effect in different severity strata) may provide equivalent evidence.

Q: Does the Agency plan to integrate reviews of all approaches to AD therapy under one group with expertise in AD (e.g., drugs, vaccines, genomics, diagnostics)?
A: There is no specific such plan at the moment. However, we are trying to increase communication among the various relevant groups, including staff with expertise in biologics, genomics, and biomarkers.

Q: What forum would be best for continued discussion of new approaches to trial design and outcomes in clinical trials across the range of cognitive function that is the topic of this meeting (e.g., delaying onset of greater levels of impairment in normals or MCI cases and stabilization or improvement in AD)?
A: Workshops (Agency or NIH sponsored). A workshop on designs to demonstrate disease modification in Parkinson disease is currently planned. An Advisory Committee meeting is another possibility.

10. Meeting Outcome: Main Points of Consensus

• Specific practical challenges underlie the development of improved cognitive assessment approaches for clinical trials in AD and MCI; among the most notable is the difficulty of detecting cognitive change in disease-affected placebo groups, and in treated groups, if the compound is not designed to have a symptomatic effect. Another challenge is the need to overcome floor and ceiling effects present for different cognitive domains at different disease stages.
• Owing to the reliability and sensitivity the ADAS-cog has shown in the past, some researchers feel that, given present options, it remains "the way to go" for moderate AD—however, the ADAS-cog has not always proved sufficiently sensitive in placebo groups in recent clinical trials. The ADAS-cog could be used as a basis for improved, fit-for-purpose, standardized batteries.
• It would be highly beneficial to the community to create a repository of placebo data from previous ADCS trials and pharma-sector studies, and to use such data for further analysis and as reference standards.
• There is a place for both technology-based assessment and traditional assessment in AD and MCI clinical trials—success in using both types of assessment will rely on choosing tests that are appropriate to the population under study with regard to both disease stage and a given population's acceptance of different technological approaches.
• It would be premature, at the present time, to identify specific tests that would be the most useful and informative to use at specific stages of disease and in specific types of studies.

11. Meeting Outcome: Formation of Working Groups

A key goal for the field will now be to develop and agree on a discrete set of tests and batteries that are fit-for-purpose and standard for commonly used types of clinical trials in particular patient populations—for example, for proof-of-concept studies with mild AD subjects, or clinical trials with MCI subjects. These standard measures, and the analysis that informs their selection, should allow the field to design and conduct effective trials that produce comparable outcome data, and should lead to a deeper shared understanding of cognitive assessment in AD.

To most effectively work towards this goal, three working groups were formed at the meeting's end. The first group will seek to collect existing, anonymized placebo data from academic and industry sources, prioritize needs for analysis, and oversee analysis of the data by expert statisticians for the purpose of determining which measures of cognitive performance are most effective at evaluating change across the range of severity from MCI through AD, and whether additional items or tasks might increase this effectiveness. The second working group will identify suitably disinterested experts who can make an initial assessment of new instruments, including computer-based tests and telephone tests, to identify which of these scales might be eligible for further evaluation for their utility in different types of clinical trials, and eventual incorporation into standard batteries. The third group will examine the feasibility of developing a universal scale for assessing change in cognitive performance over the range of MCI and AD. They will seek to incorporate extant and anticipated data from the first two working groups, as well as other sources, in the development of a universal scale.

A number of general suggestions were made regarding the working groups. First, that it would be beneficial to advise the community of the formation of the groups, and to solicit suggestions and ideas for the groups might most effectively accomplish their goals. Second, that after the groups determine the scope and budget for their projects, the ADCS coordinating center in San Diego might represent an ideal neutral repository for collected placebo data. And third, that it would be very useful to use the ADNI (Alzheimer's Disease Neuroimaging Initiative) as a model for the efforts of the working groups—to provide clarity and transparency, a place for industry representation, and a structure for communication between the working groups and with the community as a whole.

Working Group 1: Using Placebo Data to Identify Sensitive Measures of Change in Early Disease
Leaders: Ron Petersen (MCI data), Marilyn Albert (AD data)
Upcoming goal: Identify sources of placebo data; prioritize needs for analysis; submit a report outlining the scope and budget for the project proposal; oversee proposed analysis.
Timeframe: Summary planning document prepared and presented to ADCS in fall 2007; Phase 1 (one year) to be carried out in 2008.
Initial group members: Ron Petersen, Marilyn Albert

Working Group 2: Evaluation and Incorporation of New Tests of Cognitive Change
Leaders: Mary Sano and David Salmon
Upcoming goal: Identify non-conflicted expertise and leadership for the evaluation of new tests, including computerized tests and telephone tests. Develop scope and budget for the project.
Timeframe: TBD
Initial group members: Mary Sano, David Salmon, Neil Buckholtz, Laurie Ryan, Holly Posner, Mike Gold, Andrew Blackwell.

Working Group 3: Development of a Universal Scale for Cognitive Assessment in MCI/AD
Leader: Holly Posner
Upcoming goal: Identify existing and anticipated data and relevant analysis (including data from Working Groups 1 and 2) that would inform development of a universal scale.
Timeframe: Phase 1 underway in 2008.
Initial group members: All interested parties, including interested representatives from Pharma, should contact Holly Posner.

Meeting Outcome: Upcoming Discussion
A meeting entitled "Analysis of Placebo Data in AD and MCI clinical trials" will be held in Washington, DC, on 21 July 2008, with the aim of devising strategies for utilizing placebo data to improve the power of cognitive tests in drug trials. The meeting is being organized by Marilyn Albert and Ron Petersen.

Other Issues to Consider in the Near Future
A similar discussion regarding functional scales will be needed in the near future, given the importance of functional outcome data in the drug approval process. Cross-cultural issues will also require careful, intensive consideration in the near term, since a large and increasing proportion of clinical trials are being conducted outside the United States. Other issues that will also require further discussion in the future include best approaches for testing healthy elderly; approaches to the diagnosis of MCI and dementia by non-specialist clinicians; the application of cognitive testing principles to other dementias; and the improvement of statistical methodologies applied to cognitive assessment data.

The meeting participants concluded that a follow-up meeting in 2008 or 2009 would be useful to assess the progress of the working groups and other issues related to the improvement of cognitive testing in AD clinical trials.

Further Reading

ADCS Instrument and Trial Papers

Cummings JL, Raman R, Ernstrom K, Salmon D, Ferris SH, Alzheimer's Disease Cooperative Study Group. ADCS Prevention Instrument Project: behavioral measures in primary prevention trials. Alzheimer Dis Assoc Disord. 2006 Oct-Dec;20(4 Suppl 3):S147-51. Abstract

Ferris SH, Aisen PS, Cummings J, Galasko D, Salmon DP, Schneider L, Sano M, Whitehouse PJ, Edland S, Thal LJ, Alzheimer's Disease Cooperative Study Group. ADCS Prevention Instrument Project: overview and initial results. Alzheimer Dis Assoc Disord. 2006 Oct-Dec;20(4 Suppl 3):S109-23. Abstract

Ferris SH, Mackell JA, Mohs R, Schneider LS, Galasko D, Whitehouse PJ, Schmitt FA, Sano M, Thomas RG, Ernesto C, Grundman M, Schafer K, Thal LJ. A multicenter evaluation of new treatment efficacy instruments for Alzheimer's disease clinical trials: overview and general results. The Alzheimer's Disease Cooperative Study. Alzheimer Dis Assoc Disord. 1997;11 Suppl 2:S1-12. Abstract

Galasko D, Bennett DA, Sano M, Marson D, Kaye J, Edland SD, Alzheimer's Disease Cooperative Study. ADCS Prevention Instrument Project: assessment of instrumental activities of daily living for community-dwelling elderly individuals in dementia prevention clinical trials. Alzheimer Dis Assoc Disord. 2006 Oct-Dec;20(4 Suppl 3):S152-69. Abstract

Galasko D, Bennett D, Sano M, Ernesto C, Thomas R, Grundman M, Ferris S. An inventory to assess activities of daily living for clinical trials in Alzheimer's disease. The Alzheimer's Disease Cooperative Study. Alzheimer Dis Assoc Disord. 1997;11 Suppl 2:S33-9. Abstract

Grundman M, Petersen RC, Ferris SH, Thomas RG, Aisen PS, Bennett DA, Foster NL, Jack CR, Galasko DR, Doody R, Kaye J, Sano M, Mohs R, Gauthier S, Kim HT, Jin S, Schultz AN, Schafer K, Mulnard R, van Dyck CH, Mintzer J, Zamrini EY, Cahn-Weiner D, Thal LJ, Alzheimer's Disease Cooperative Study. Mild cognitive impairment can be distinguished from Alzheimer disease and normal aging for clinical trials. Arch Neurol. 2004 Jan;61(1):59-66. Abstract

Koss E, Weiner M, Ernesto C, Cohen-Mansfield J, Ferris SH, Grundman M, Schafer K, Sano M, Thal LJ, Thomas R, Whitehouse PJ. Assessing patterns of agitation in Alzheimer's disease patients with the Cohen-Mansfield Agitation Inventory. The Alzheimer's Disease Cooperative Study. Alzheimer Dis Assoc Disord. 1997;11 Suppl 2:S45-50. Abstract

Mackell JA, Ferris SH, Mohs R, Schneider L, Galasko D, Whitehouse P, Schmitt F, Sano M, Thal LJ. Multicenter evaluation of new instruments for Alzheimer's disease clinical trials: summary of results. The Alzheimer's Disease Cooperative Study. Alzheimer Dis Assoc Disord. 1997;11 Suppl 2:S65-9. Abstract

Mohs RC, Knopman D, Petersen RC, Ferris SH, Ernesto C, Grundman M, Sano M, Bieliauskas L, Geldmacher D, Clark C, Thal LJ. Development of cognitive instruments for use in clinical trials of antidementia drugs: additions to the Alzheimer's Disease Assessment Scale that broaden its scope. The Alzheimer's Disease Cooperative Study. Alzheimer Dis Assoc Disord. 1997;11 Suppl 2:S13-21. Abstract

Patterson MB, Whitehouse PJ, Edland SD, Sami SA, Sano M, Smyth K, Weiner MF, Alzheimer's Disease Cooperative Study Group. ADCS Prevention Instrument Project: quality of life assessment (QOL). Alzheimer Dis Assoc Disord. 2006 Oct-Dec;20(4 Suppl 3):S179-90. Abstract

Patterson MB, Mack JL, Mackell JA, Thomas R, Tariot P, Weiner M, Whitehouse PJ. A longitudinal study of behavioral pathology across five levels of dementia severity in Alzheimer's disease: the CERAD Behavior Rating Scale for Dementia. The Alzheimer's Disease Cooperative Study. Alzheimer Dis Assoc Disord. 1997;11 Suppl 2:S40-4. Abstract

Petersen RC, Thomas RG, Grundman M, Bennett D, Doody R, Ferris S, Galasko D, Jin S, Kaye J, Levey A, Pfeiffer E, Sano M, van Dyck CH, Thal LJ, Alzheimer's Disease Cooperative Study Group. Vitamin E and donepezil for the treatment of mild cognitive impairment. N Engl J Med. 2005 Jun 9;352(23):2379-88. Abstract

Salmon DP, Cummings JL, Jin S, Sano M, Sperling RA, Zamrini E, Petersen RC, Edland SD, Thal LJ, Ferris SH, Alzheimer's Disease Cooperative Study. ADCS Prevention Instrument Project: development of a brief verbal memory test for primary prevention clinical trials. Alzheimer Dis Assoc Disord. 2006 Oct-Dec;20(4 Suppl 3):S139-46. Abstract

Sano M, Egelko S, Jin S, Cummings J, Clark CM, Pawluczyk S, Thomas RJ, Schittini M, Thal LJ, Alzheimer's Disease Cooperative Study Group. Spanish instrument protocol: new treatment efficacy instruments for Spanish-speaking patients in Alzheimer disease clinical trials. Alzheimer Dis Assoc Disord. 2006 Oct-Dec;20(4):232-41. Abstract

Sano M, Zhu CW, Whitehouse PJ, Edland S, Jin S, Ernstrom K, Thomas RG, Thal LJ, Ferris SH, Alzheimer Disease Cooperative Study Group. ADCS Prevention Instrument Project: pharmacoeconomics: assessing health-related resource use among healthy elderly. Alzheimer Dis Assoc Disord. 2006 Oct-Dec;20(4 Suppl 3):S191-202. Abstract

Sano M, Mackell JA, Ponton M, Ferreira P, Wilson J, Pawluczyk S, Pfeiffer E, Thomas RG, Jin S, Schafer K, Schittini M, Grundman M, Ferris SH, Thal LJ. The Spanish Instrument Protocol: design and implementation of a study to evaluate treatment efficacy Instruments for Spanish-speaking patients with Alzheimer's disease. The Alzheimer's Disease Cooperative Study. Alzheimer Dis Assoc Disord. 1997;11 Suppl 2:S57-64. Abstract

Schmitt FA, Ashford W, Ernesto C, Saxton J, Schneider LS, Clark CM, Ferris SH, Mackell JA, Schafer K, Thal LJ. The severe impairment battery: concurrent validity and the assessment of longitudinal change in Alzheimer's disease. The Alzheimer's Disease Cooperative Study. Alzheimer Dis Assoc Disord. 1997;11 Suppl 2:S51-6. Abstract

Schneider LS, Clark CM, Doody R, Ferris SH, Morris JC, Raman R, Reisberg B, Schmitt FA. ADCS Prevention Instrument Project: ADCS-clinicians' global impression of change scales (ADCS-CGIC), self-rated and study partner-rated versions. Alzheimer Dis Assoc Disord. 2006 Oct-Dec;20(4 Suppl 3):S124-38. Abstract

Schneider LS, Olin JT, Doody RS, Clark CM, Morris JC, Reisberg B, Schmitt FA, Grundman M, Thomas RG, Ferris SH. Validity and reliability of the Alzheimer's Disease Cooperative Study-Clinical Global Impression of Change. The Alzheimer's Disease Cooperative Study. Alzheimer Dis Assoc Disord. 1997;11 Suppl 2:S22-32. Abstract

Walsh SP, Raman R, Jones KB, Aisen PS, Alzheimer's Disease Cooperative Study Group. ADCS Prevention Instrument Project: the Mail-In Cognitive Function Screening Instrument (MCFSI). Alzheimer Dis Assoc Disord. 2006 Oct-Dec;20(4 Suppl 3):S170-8. Abstract

Whitehouse PJ, Rajcan JL, Sami SA, Patterson MB, Smyth KA, Edland SD, George DR. ADCS Prevention Instrument Project: pilot testing of a book club as a psychosocial intervention and recruitment and retention strategy. Alzheimer Dis Assoc Disord. 2006 Oct-Dec;20(4 Suppl 3):S203-8. Abstract

Other Papers Mentioned in Discussion

Albert M, Blacker D, Moss MB, Tanzi R, McArdle JJ. Longitudinal change in cognitive performance among individuals with mild cognitive impairment. Neuropsychology. 2007 Mar;21(2):158-69. Abstract

Bondi MW, Salmon DP, Galasko D, Thomas RG, Thal LJ. Neuropsychological function and apolipoprotein E genotype in the preclinical detection of Alzheimer's disease. Psychol Aging. 1999 Jun;14(2):295-303. Abstract

Cutler NR, Shrotriya RC, Sramek JJ, Veroff AE, Seifert RD, Reich LA, Hironaka DY. The use of the Computerized Neuropsychological Test Battery (CNTB) in an efficacy and safety trial of BMY 21,502 in Alzheimer's disease. Ann N Y Acad Sci. 1993 Sep 24;695:332-6. Abstract

Elias MF, Beiser A, Wolf PA, Au R, White RF, D'Agostino RB. The preclinical phase of alzheimer disease: A 22-year prospective study of the Framingham Cohort. Arch Neurol. 2000 Jun;57(6):808-13. Abstract

Galvin JE, Powlishta KK, Wilkins K, McKeel DW, Xiong C, Grant E, Storandt M, Morris JC. Predictors of preclinical Alzheimer disease and dementia: a clinicopathologic study. Arch Neurol. 2005 May;62(5):758-65. Abstract

Gold M. Study design factors and patient demographics and their effect on the decline of placebo-treated subjects in randomized clinical trials in Alzheimer's disease. J Clin Psychiatry. 2007 Mar;68(3):430-8. Abstract

Lange KL, Bondi MW, Salmon DP, Galasko D, Delis DC, Thomas RG, Thal LJ. Decline in verbal memory during preclinical Alzheimer's disease: examination of the effect of APOE genotype. J Int Neuropsychol Soc. 2002 Nov;8(7):943-55. Abstract

Olichney JM, Morris SK, Ochoa C, Salmon DP, Thal LJ, Kutas M, Iragui VJ. Abnormal verbal event related potentials in mild cognitive impairment and incipient Alzheimer's disease. J Neurol Neurosurg Psychiatry. 2002 Oct;73(4):377-84. Abstract

Rapp MA, Reischies FM. Attention and executive control predict Alzheimer disease in late life: results from the Berlin Aging Study (BASE). Am J Geriatr Psychiatry. 2005 Feb;13(2):134-41. Abstract

Rogers SL, Farlow MR, Doody RS, Mohs R, Friedhoff LT. A 24-week, double-blind, placebo-controlled trial of donepezil in patients with Alzheimer's disease. Donepezil Study Group. Neurology. 1998 Jan;50(1):136-45. Abstract

Saykin AJ, Wishart HA, Rabin LA, Santulli RB, Flashman LA, West JD, McHugh TL, Mamourian AC. Older adults with cognitive complaints show brain atrophy similar to that of amnestic MCI. Neurology. 2006 Sep 12;67(5):834-42. Abstract

Schneider L. [Comparison of placebo data obtained from clinical trials] (in preparation).

Twamley EW, Ropacki SA, Bondi MW. Neuropsychological and neuroimaging changes in preclinical Alzheimer's disease. J Int Neuropsychol Soc. 2006 Sep;12(5):707-35. Abstract

Vellas B, Andrieu S, Cantet C, Dartigues JF, Gauthier S. Long-term changes in ADAS-cog: what is clinically relevant for disease modifying trials in Alzheimer? J Nutr Health Aging. 2007 Jul-Aug;11(4):338-41. Abstract

Veroff AE, Bodick NC, Offen WW, Sramek JJ, Cutler NR. Efficacy of xanomeline in Alzheimer disease: cognitive improvement measured using the Computerized Neuropsychological Test Battery (CNTB). Alzheimer Dis Assoc Disord. 1998 Dec;12(4):304-12. Abstract