There is a growing consensus in the AD clinical research community of the need to identify individuals at risk for AD or those in whom the neurodegenerative process has already begun [8
]. New diagnostic criteria proposed by three workgroups convened by the National Institute on Aging in partnership with the Alzheimer's Association (NIA-AA) [12
] as well as by an International Working Group (IWG) [16
], recognized the value of incorporating biomarkers of AD and neurodegeneration in research studies but concluded that they have not been sufficiently standardized or validated for use in clinical care. The field is rapidly moving forward, with the goal of using biomarkers to ascertain AD as a diagnosis of inclusion with a level of certainty, rather than a diagnosis to be made after excluding other disorders. However, the consensus at this time is that the science does not support the widespread clinical application of a biomarker-based diagnosis.
The NIA-AA committees distinguished three stages of disease: 1) a preclinical phase, where the pathophysiological processes of AD have begun before any evident signs of cognitive impairment, 2) a prodromal or MCI due to AD phase, where there is evident decline in memory or other cognitive functions, and 3) AD dementia, where cognitive and functional deterioration have reached a threshold at which an individual needs help with activities of daily living.
While substantial data indicate that leading CSF and neuroimaging biomarkers provide clinically useful information for patients with evident cognitive impairment (i.e., MCI and dementia), current longitudinal data are not
sufficient to warrant the use of AD biomarkers in a-
symptomatic individuals [15
]. About one third of older adults with normal cognition exhibit abnormal CSF Aβ levels or Aβ PET scans; however, it is not yet known whether abnormal AD biomarkers in asymptomatic people indicate a disease process that will manifest in MCI and AD at some future point, and if so, when. While ADNI has been extraordinarily successful in identifying, standardizing, and validating biomarkers, the population studied does not necessarily reflect the general population. Thus, in order to establish appropriate biomarker cutoffs for both clinical trials and clinical usage, it is imperative to conduct longitudinal, population-based studies using standardized protocols and multiple biomarkers and cognitive markers. Establishing an international population-based registry of older adults, including geographically and culturally different populations and individuals with co-morbid conditions, is thus a high priority recommendation from this conference [17
]. This registry could take advantage of existing clinical databases such as those managed by the Department of Veteran’s Affairs and the Center for Medicare and Medicaid Services, and could be modeled after other large longitudinal studies such as the Framingham study, the Cardiovascular Health Study, the Health and Retirement Study, and other population based studies of aging (summarized in [18
Using biomarkers to define risk, however, presents several concerns, particularly since a therapy that impacts an AD biomarker is not certain to translate into clinically meaningful endpoints [19
]. The experience of other biomarker-based processes demonstrates the need for caution in disseminating a biomarker into clinical practice without clarifying its value [20
]. Clinicians do not yet have a good way to predict the future of AD, including the speed of progression and to what extent function will be affected.
If an effective disease modifying therapy is reported, research needs to address how it will change AD. In particular, as ongoing and planned research enhances our understanding of the preclinical stages of AD, both how we think about the disease and its treatment will change. What was once a disease defined by clinical signs and symptoms, with treatments designed to reduce these clinical features, will become a disease defined by the risk of developing these signs and symptoms, and, once they occur, the risk of their worsening. The concept of successful treatment (including non-pharmacologic treatment) will then be transformed to mean reducing risk and improving quality of life, and it will be imperative that the prevention of signs and symptoms is shown to be both clinically meaningful and cost effective.
As the value of biomarkers and other risk factors emerge and become clinically useful, patients and society will be better served if we think of biomarkers not as labels of a category such as “preclinical AD,” but instead as one, but not the only, measure of risk for disability as a result of progressive cognitive decline. To make the best use of established and emerging biomarkers and other risk factors for AD, an important goal of the AD research community should be to develop and validate a risk-stratification model for the development of endpoints such as MCI or dementia, in order to guide clinical and care management decisions by patients, families, and health care providers; policymaking; and research to test new interventions. Such a risk-prediction model, which incorporates drugs, clinical trials, and biomarkers [21
], holds substantial promise for improving public health. Moreover, it is possible that certain biomarkers will ultimately identify individuals who will ultimately progress to AD and hence prove critically important for identifying early treatment and intervention strategies as these emerge akin to how a Pap smear is currently used to identify women at risk of uterine cancer. The Economics, Ethics and Health Policy Workgroup drew upon their collective expertise in both neurodegenerative diseases and other diseases of aging, such as cardiovascular disease, osteoporosis, and cancer, to propose how the promise of this risk-based model can be realized.
The risk stratification model should be developed using data from representative population-based samples so as to accurately represent the clinical complexity of typical older adults, and the resulting competing health risks (e.g., heart disease, diabetes, cancer) that may have an important impact on life expectancy and quality of life, and, therefore, the efficacy and value of drug and other interventions to prevent or treat AD. Further, the risk-stratification models should be in the public domain, and should be evaluated and updated by a public body (e.g. HHS) using established standards for the reliable and valid measurement of AD risk factors, including the standardization of AD biomarkers. The ongoing evaluation of risk-stratification models should include assessments of the costs (e.g., unnecessary treatment and treatment side-effects) that result from mis-classification when using the models.
Recommendations aimed at enabling better patient care through early identification
- Establish a national/international population-based registry of older adults, including geographically and culturally different populations and individuals with co-morbid conditions.
- Conduct longitudinal population-based studies to clarify the natural history of AD and other dementias using clinical measures, cognitive tests, and biomarkers.
- Develop best practice standards on when to order biomarker test, what and how to disclose and interpret results.
- Develop a training and certification program for conducting biomarker tests, interpreting data, communicating biomarker data, and quality assurance.
- Develop a risk stratification model that incorporates demographic, genetic, biological, cognitive, and environmental risk factors