As the renowned physicist Niels Bohr said, “Prediction is very difficult, especially when it’s about the future.” Yet, we can expect risk assessment for AD in upcoming years to differ in important ways from its current state. We can expect significant advances in predictive capabilities as the epidemiology and pathophysiology of AD become better understood. Social epidemiological studies emerged in recent years to identify new behavioral and health history risk factors for AD. As prospective studies accumulate in the literature, risk models for AD may mature to the point where they encompass a wider range of risk factors and perhaps even suffuse into clinical practice. Such models exist for several other common, complex diseases; for example, the online Breast Cancer Risk Assessment Tool based on the Gail model [123
] is available to both patients and professionals via the National Cancer Institute website [203
]. A starting point in developing a similar tool for AD could be a late-life dementia risk index [125
] which accurately categorizes older adults as having low-, moderate- or high-risk of developing dementia within 6 years. The 15-point index includes: age, neuropsychological test performance, body mass index, APOE
ε4 status, MRI results, cardiovascular factors, fine-motor function and alcohol consumption.
Genomic research in AD is also rapidly maturing. Although one might classify the search for single-gene causes of AD via genome-wide association studies as disappointing, we are beginning to see promising results from other strategies including epigenetics, gene–environment studies, and gene–gene interactions. For example, SNPs at CLU
locations have shown significant associations with carriers of APOE
ε4 and risk of disease [17
]. In addition, variants in the gene TOMM40
have been associated with risk for AD among APOE
ε3 carriers; this finding might be explained by TOMM40
being adjacent to and in linkage disequilibrium with APOE
, with their interaction resulting in pathological protein production [24
]. A variable-length sequence repeat polymorphism has been identified on TOMM40
that in a recent study predicted age of onset risk within the next 7 years for older individuals with normal cognition [126
]. A prospective, population-based study is underway that seeks to validate the association of APOE
genotypes and TOMM40
haplotypes with age of onset in AD. This suggests the future possibility of multiplex genetic testing (i.e., testing for a panel of genes rather than a single one) not only to estimate AD risk but also age of onset.
More immediately applicable methods of risk assessment may come from other biomarker strategies used in the preclinical stages of AD. ADNI may ultimately identify biomarkers of clinical utility that reliably distinguish normal aging from MCI from clinical AD [127
]. Considerable literature on factors predicting conversion from MCI to AD indicates that possession of the APOE
ε4 allele is associated with more rapid progression [73
], and several other factors are under investigation, including neuropsychological measures [129
], hippocampal volume as shown by structural MRI [130
], reduced tempoparietal glucose metabolism as shown by PET scan [129
] and levels of Aβ in CSF [132
]. Such biomarkers may become more clinically relevant as MCI continues to emerge as a formal diagnostic category; the revised Diagnostic & Statistical Manual of Mental Disorders (DSM-V) will likely include a category corresponding to MCI [69
Of course, the landscape for any of the aforementioned risk assessment methods would change dramatically should treatment options significantly improve for AD. Given the investments that major research funders (e.g., National Institute on Aging) and leading pharmaceutical companies (over 30 have drugs intended to treat AD in the FDA pipeline) [204
] are making in new treatments for AD, such a development may not be far off. Advancement may occur not only in the development of new treatments, but the targeting of existing ones. Preliminary indications in pharmacogenomics research suggest APOE
might serve as a useful marker when deciding upon and monitoring response to pharmacotherapy for AD. For example, differential treatment response between ε4 carriers and noncarriers has been observed in clinical trials of rosiglitazone and donepezil, respectively [38
]. In the case of rosiglitazone, APOE
allowed the identification of a potentially responsive subset of participants, thus enabling Phase III trials with patient subpopulations stratified by ε4 status [133
]. Although the Phase III trial was negative, pharmacogenomic strategies are likely to continue to be explored in AD treatment and prevention trials [135
]. Broad support of genomic-based approaches, including from research funders and advocacy groups, will be necessary to accelerate discovery of effective prevention strategies in this area.
Although research on its risk factors is rapidly maturing, we still lack a theory of AD causation that integrates known risk factors; we do not know if such factors are interactive, additive or orthogonal in producing AD risk. As a common, complex disease, AD presents challenges not only in terms of risk estimation, but also risk communication. Healthcare professionals are not always skilled in conveying probabilistic risk information to patients or may not have adequate opportunity to do so, given time pressures of clinic visits [100
]. In addition, significant proportions of patients lack even basic health literacy and numeracy skills required to comprehend risk information [136
]. These challenges become compounded given that many patients likely to be interested in and appropriate for risk assessment may be older adults already evidencing mild cognitive difficulties compromising their ability to process and remember risk information or give informed consent. It will therefore be important to develop risk communication procedures that accommodate these difficulties, as well as tools to precisely measure individuals’ decisional capacity regarding AD risk assessment. The MacArthur Capacity Assessment Tool is a validated instrument that could be adapted for use in this context to identify whether cognitive deficits impair decisional abilities to the point where a surrogate should be involved in medical decision making [137
]. Such assessments could also take on legal significance if used by clinicians to determine competency to make medical decisions.
Future ethical, legal & social issues
Research on risk factors and biomarkers for AD are routinely communicated to the lay public via the media, and research participants may become more aware of the potential significance of their own individual research results. Investigators typically do not communicate such results to participants, but some legal, policy, and bioethics scholars have asserted a duty of researchers to disclose research results to participants when this information is of potential clinical and/or personal significance [138
]. Depending on how research policies and legal authorities define this duty, investigators may face the challenging prospect of disclosing individual research results and implications for AD risk.
Another ‘hot topic’ within bioethics is the emergence of DTC genetic risk assessment. The Genetics and Public Policy Center identified 39 companies currently providing such testing, seven of which include AD risk information as part of their services [205
]. Several commentators have raised the following concerns about DTC testing: it relies on an inadequate evidence base for risk estimation; its advertising claims overrate its benefits and minimize its risks; in-person genetics education and counseling is not available to help interpret results and address emotional responses; and testing is not always done in CLIA-approved laboratories. A 2007 statement by the American Society of Human Genetics (ASHG) offered a series of specific policy recommendations for DTC genetic testing and called for the federal government to regulate genetic tests and their marketing more rigorously [143
]. The available literature suggests that individuals are usually able to cope adequately with genetic risk information, even for severe incurable diseases like AD. It should be noted, however, that DTC companies do not follow procedures that may mitigate against adverse psychological outcomes; these include baseline screening of interested individuals for severe depression or anxiety and in-person disclosure of genetic results from GCs.
Should consumers gain greater access to genetic risk information for AD, whether via DTC testing or other means, important policy issues regarding insurance will be raised. LTC insurance is of particular interest given that GINA does not cover it, and AD is a major driver of nursing home costs. APOE
has been shown not only to predict development of AD, but also nursing home placement [144
]. If APOE
testing is utilized by a significant number of consumers to inform their LTC insurance purchasing, then insurers may be within their rights to address this adverse selection by increasing premiums or denying coverage based on APOE
results. Policymakers will then have to decide whether this practice is acceptable and just.
- Alzheimer’s disease (AD) is a major public health concern owing to its increasing prevalence and the growth of the aging population.
- Disclosure of risk for AD is not typically conducted by healthcare professionals but may soon become a more viable practice given future improvements in AD treatment options and understanding of disease etiology.
Risk & protective factors for AD
- Older age, female sex, African–American or Hispanic ethnicity, and lower educational attainment are demographic factors that are associated with increased risk for AD.
- APOE, APP, PSEN1 and PSEN2 are established genetic risk factors for AD. Numerous other susceptibility genes and genetic markers are currently being investigated.
- Vascular and inflammatory factors and diseases may contribute to AD brain pathology.
- Lifestyle factors including exercise, diet, physical and social activity are associated with lower AD risk; while none of these are proven to reduce AD risk, they are often recommended given their potential benefits and low risk of harm.
Methods of risk assessment for AD
- Risk assessment for AD is being developed to improve clinical trials, tailor disease prevention and risk modification strategies and inform later-life planning.
- Possible methods are epidemiology-based risk indices; predictive genetic testing; other biomarkers; and clinical assessment of mild cognitive impairment.
- Predictive genetic testing for AD is now offered by some direct-to-consumer genetic testing companies, but this practice has raised various ethical, legal and policy concerns.
Impact of risk assessment for Huntington’s disease
- The experience of providing genetic testing and counseling for Huntington’s disease (another fatal, incurable neuropsychiatric condition) may provide lessons for AD risk assessment given that extensive research has been conducted on the process and impact of Huntington’s disease testing.
Impact of risk assessment for AD: the REVEAL Study
- The Risk Evaluation and Education for Alzheimer’s Disease (REVEAL) Study, a series of multisite, randomized clinical trials, is the main source of information about the psychological and behavioral impact of AD risk disclosure that incorporates APOE genotype status.
- Findings suggest the following: risk disclosure rarely results in significant, lasting adverse psychological effects; APOE status alters disease risk perceptions; higher risk individuals more frequently report behavior changes such as long-term care insurance purchases and use of vitamin E.
- The development and provision of AD risk information entails numerous challenges but is supported by societal and scientific investment and interest in combating AD.
- Future AD risk assessment involves challenges in generating and disclosing AD risk; these include understanding interactions among risk factors and addressing provider and patient difficulties in communicating and understanding health risk information.
- Should AD risk be disclosed more commonly, this practice will likely raise ethical, legal and social issues including regulation of direct-to-consumer genetic test companies and addressing potential genetic discrimination by long-term care insurance providers.