The VA Million Veteran Program,Alzheimer's disease,and genetics: Important steps forward but many steps still to go on the march to control dementia and Alzheimer's disease

The VA Million Veteran Program (MVP, https://www.mvp.va.gov) was launched in 2011 and reached a million veterans enrolled in November 2023. The MVP is the nation's largest biorepository of veteran data and has one of the world's most diverse cohorts of any genetic research program. An article by Merritt et al. ¹ analyzing data from the MVP, derives dementia diagnoses in the VA's electronic health record (EHRs) from ICD codes and then relates the diagnoses to other medical factors, including medication prescriptions, and genetic factors, both apolipoprotein E (APOE) and polygenic risk factors. ² This article demonstrates the important relationships that exist between these elements.

With the advent of EHR systems around the country and the world, it could be expected that the available data would provide critical information to understand many medical conditions, particularly dementia. However, such developments in dementia, reviewed by Merritt et al., ¹ have not provided the important advances needed in this field. Further, the medications used for Alzheimer's disease (AD), though they have been approved by the FDA, particularly the cholinesterase inhibitors, used for over 30 years, ³ are not prescribed to AD patients based on rigorous criteria, and the link between diagnosis and medication prescription is weaker than should be expected. And most disturbingly, the genetic relationships of AD have been long known, ⁴ but these critical factors in AD causation have neither been the target of extensive study nor used for improving diagnosis and treatment development, which is clearly possible as is shown by these analyses. And all of these issues also need improvement in clinical venues outside the VA.

The first issue to discuss is the value of EHRs. Moving from paper records to EHRs has been one of the major advances of the computer era. Much of what is done in EHRs is the clinician making additions to a patient's electronic database in a legible format that can be instantly and easily shared across medical systems. Diagnostic codes are entered and available for patient care, billing and research (as utilized by the MVP). However, in most cases the clinical notes are little more than what used to be written on sheets of paper. Only in a few cases are templates used to improve completeness of notation. And the ICD codes, which are diagnostic approximations themselves, are not necessarily provided from a rigorous understanding of the definitions or the underlying conditions. But there are many instances in which electronic forms could be utilized to provide a more standardized and accurate assessment, particularly for the diagnosis of dementia and assessment of the level of severity along the cognitive decline continuum. ⁵ Applying computerized assessments more widely, including cognitive measurements, organized reporting of activities of daily living, and systematized clinical observations ⁵ (see the supplementary material in ⁵ ), would be a great step forward for the dementia and AD fields. Such analyses, including improved accuracy of reports and codes, would greatly enhance studies such as those which the MVP has shown are possible. As noted by Merritt et al., ¹ clinician and coder training should be implemented to improve future algorithmic accuracy and include the use of templates.

The second major issue is the relationship of morbidity, mortality, and dementia onset to age and genetic factors. The Merritt et al. ¹ article does address the effect of age but not at the level of exact years of age, which can substantially clarify many of these relationships. Clearly, after childhood, morbidity and mortality increase with age, but there is little appreciation that this increase is predominantly exponential and affects most living beings, including humans, and all systems of the body. ⁶ The 2020 CDC and Census data indicate that mortality doubles for men approximately every 9 years and for women approximately every 8 years from 30 to 100 years of age. ⁷ However, the rate of doubling of dementia incidence is every 5 years from 50 to 90 years of age, ⁸ showing that the systems of the brain affected by dementia are more vulnerable to aging than most other systems of the body. Further, this age-related vulnerability is highly related to APOE genotype,^7,9,10 as well as race, ethnicity, and population ancestry, ¹¹ though the Merritt et al. ¹ analyses only examine veterans above age 60 with European ancestry. The association of AD incidence with sex is also a major factor, women having about twice the incidence of men, but this disparity is fully accounted for by sex-related mortality: according to 2021 CDC data, women have a life-expectancy of 79 years and men about 73 years, the 6-year difference, with a 5-year doubling time for AD incidence, being more than enough to explain the variation. In studies of complex data which include genotype, these age-related factors must be fully analyzed to understand what the real risks are, particularly when the important directions are efforts to understand predisposing factors and develop strategies for dementia/AD prevention.

The third major issue is the utilization of age and genetic information in medical diagnosis, particularly AD. The Merritt et al. ¹ article shows the great benefit of genetic information in supporting an AD diagnosis. However, though the value of APOE genotype information has been well documented since 1993,^7,10,12,13 there has been a tremendous reluctance to address or utilize this important information in clinical settings, similar to the general avoidance of the use of genetic information from adults in the medical community. However, genetic information is readily and inexpensively available. Given the clear importance of genetic information and its interaction with age, it is time to: 1) ensure consumer understanding of the significance of genetics; 2) address and limit the potential negative effects of such knowledge; and 3) support the implementation of such information with improved preventive behaviors, clinical care, and research. ¹⁴ Implementation of genetic information utilization could greatly improve clinical diagnosis and care, including improving the algorithms leading to clinical documentation in EHRs and hastening the development of prevention approaches and clinically significant interventions/treatments.

There are several other factors that the MVP can help to understand, particularly for the clinical care of veterans, including military risk factors for AD, military environmental exposures,^15–17 and interactions with traumatic brain injury and post-traumatic stress disorder.^18,19 Such information can further be analyzed with respect to age and genetic factors to determine risk, better decisions regarding prevention, and recognizing clinical conditions that might otherwise be missed.

A broader issue of relevance to the whole population is early recognition of cognitive impairment. While there are numerous reasons to screen for dementia and AD,^20–23 there is progressively more evidence that the Medicare Annual Wellness Visit (AWV), which requires a cognitive assessment (http://www.cms.gov/cognitive), can lead to earlier recognition of dementia and AD,^24,25 which can facilitate earlier implementation of prevention and intervention measures. The VA does not have such a program and the AWV is only partially utilized by Medicare beneficiaries, with only about one-half reporting an AWV and fewer than a third having a structured assessment. ²⁶ And structured cognitive assessments can be conducted quickly and inexpensively online. ²⁷ Nonetheless, the AWV model provides an approach which can maximize early dementia detection and lead to considerable benefits. ²⁸

The most complex issue is “multiple etiology dementia” or “mixed dementia”. The underlying problem, as discussed above, is that with age, all body systems deteriorate and brain systems are even more vulnerable, so that the problem is not just AD pathology, which accounts for a third of dementia cases by itself, is mixed with other causes in an additional third of the cases, and is not involved with the remaining third of the cases. ²⁹ As neuropathology has progressively shown, there are even numerous proteinopathies associated with dementia, and each has a different impact on overall dementia severity and rate of progression, ³⁰ highlighting the need to learn more about brain mechanisms which are vulnerable to pathological factors which lead to dementia.^4,31 And as important as the issue of multiple causation is, even complex algorithms will have difficulty identifying the presence of multiple etiologies bases in EHR data, affecting the sensitivities and specificities of the diagnoses. The issue of complexity is recognized by the NIH, National Plan to Address Alzheimer's Disease (NAPA; see: 2023 Update, links at the end of the text). And with respect to the MVP, the relationship of TBI and PTSD to dementia is of great relevance for Veterans.^19,32

The fundamental issue is that dementia and AD are extremely important problems, they greatly impair the lives of those afflicted and those who care for the patients, they are very costly, and these conditions are getting more common with the improvement of longevity and more prevalent with the decrease of population fertility. Major efforts are needed, particularly as outlined here and in the National Plan to Address Alzheimer's disease. The article by Merritt et al. ¹ shows the importance of analyses of large databases and the use of algorithms and genetics to further the understanding of dementia. But, while the article presents important steps in the right direction, clearly there are many more steps to take to identify dementing illnesses precisely and establish control of dementia and AD.

National Plan to Address Alzheimer's Disease: 2023 Update: https://aspe.hhs.gov/sites/default/files/documents/3c45034aec6cf63414b8ed7351ce7d95/napa-national-plan-2023-update.pdf (accessed 11/16/2024) https://aspe.hhs.gov/collaborations-committees-advisory-groups/napa (accessed 11/16/2024)