NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Williams JW, Plassman BL, Burke J, et al. Preventing Alzheimer's Disease and Cognitive Decline. Rockville (MD): Agency for Healthcare Research and Quality (US); 2010 Apr. (Evidence Reports/Technology Assessments, No. 193.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Preventing Alzheimer's Disease and Cognitive Decline

Preventing Alzheimer's Disease and Cognitive Decline.

Show details


Alzheimer’s disease (AD) is unique in that it may be the only late-life disease that has a long “silent” prodromal phase, no validated biological test for diagnosis, and imprecise measures of correlation between progression of phenotype and progression of pathology. Diagnosis during life is based on the clinical phenotype of symptom progression, which is heterogeneous between individuals. Part of the variation in clinical presentation may be due to the presence of other types of neuropathological changes in the brain in addition to those typically considered to be AD-related. These characteristics make it difficult not only to accurately diagnose AD but also to identify risk or protective factors for the disease; they also make it challenging to implement interventions efficiently and economically. The impact of these challenges is made clear by the fact that few of the putative risk or protective factors covered in this review had sufficient evidence from which to draw firm conclusions about their effect on AD and cognitive decline. But these findings need to be interpreted in the context of the effect size of a treatment or intervention that would make a noticeable difference in the disease burden. Using analytical models, it has been shown that relatively small delays in the onset of AD or the progression of the disease would have a large effect on the prevalence of the disease. A 1-year delay in both onset of AD and progression of AD would decrease the number of prevalent AD cases in 2050 by 9.19 million. This reduction in the number of AD cases is almost entirely due to fewer individuals with late-stage dementia, the point in the disease course when many individuals with AD are institutionalized and when the most care is needed.432

Many of the exposures reviewed in this report likely do not work in isolation in their effect on risk of AD or cognitive decline. Instead, they work in combination with other factors. Thus, the ideal interventions should be multi-dimensional, combining interventions for multiple risk factors and controlling for many other factors. But as noted when discussing the Key Question 6 in Chapter 3, above, few of the exposures reviewed here are appropriate for randomized controlled trials (RCTs). Among those that are not appropriate for intervention trials are exposures that one would want to avoid due to their negative impact on outcomes other than cognition. For example, smoking, diabetes, hypertension, and few years of education are all factors that have deleterious effects on health and lifestyle. Although healthcare interventions may be appropriate for some other potential influential factors (e.g., omega-3 fatty acids, statins, cognitive engagement), many of the other factors may be most appropriately addressed through public policy interventions (e.g., education, designing communities to facilitate physical activity) and public health interventions (educational campaigns on diet). Public education campaigns to change behavior to incorporate or exclude these factors would have relatively less risk (cost) to individuals.

One of the key limiting factors in synthesizing the current literature is the lack of standardization of exposure and outcome measures. Because outcome measures for cognitive decline were not standardized across studies, we limited the use of studies with continuous outcome measures when the conclusions from these studies were consistent with those from studies with categorical outcome measures. This meant that some studies reporting continuous measures were not reported in detail in this report, and the results of these studies were not synthesized quantitatively, but we do not think that this changed the conclusions for any exposure factor. In the future, more standardization at various steps of the research process is needed before all available data can be synthesized. We also acknowledge that standardization can have its weaknesses and can limit innovations that may advance science. The key is to strike a balance between enough uniformity to maximize the use of study results and methods that are novel enough to advance the field to the next level.

Issues related to age are central to the interpretation of all of these results. The incidence of AD increases markedly with age, doubling in rate approximately every 5 years. Due to this, the age distribution of a study sample influences the expected number of AD cases; that is, the older the sample, the greater number of cases of AD expected. For this reason, the age distribution also influences the statistical power present to detect an association between an exposure and AD. Complicating this issue further, the neuropathological evidence available suggests that both typical AD pathology and microvascular changes in the brain become more frequent with age, so the older the sample, the more likely it is that mixed pathologies are present and contribute to the cognitive profile. However, the phenotype of the mixed pathology is often difficult to distinguish from that of AD pathology alone, meaning that the clinical AD group may become more heterogeneous with advancing age. The increasing incidence of AD with age also can affect the interpretation of studies of cognitive decline. The older the sample, the more likely it is that cognitive decline represents prodromal AD, and thus any association with a risk exposure may reflect an association with AD, not just cognitive decline.

Age may also be a central issue in regard to the timing of the exposure. There may be a window of time during which exposures influence risk of AD. For example, obesity in mid-life may be associated with increased risk of AD, while obesity in late life may be associated with reduced risk of disease. The latter finding may be explained by the weight loss often associated with the disease itself. But the point is clear that different exposures may have effects at different times along the life course or the natural history of AD. Ideally the exposure should be measured in different age groups within the same study to control for inter-study variability in measurement, but this may not be realistic given the long period of followup necessary when studying exposures in mid-life. Interventions may also have different effects at different points throughout life or the AD process. Although one might assume that interventions or lifestyle modification should be undertaken as early as possible, there may be other windows during which a given intervention may exert its effect. Careful consideration of the complex relation of exposure, age, and disease will likely be key to understanding the factors that alter risk of AD and cognitive decline.

The present review has some limitations. By excluding small to moderate observational studies and small RCTs, we may have missed some important evidence, particularly for factors with scant data. To evaluate the potential impact of excluding small studies, we coded detailed reasons for exclusion in a subset of citations. Of 549 citations, only three observational studies and two randomized controlled trials were excluded solely for small sample size. Applying these rates to the 6713 citations identified overall from electronic searching, we may have excluded as many as 48 articles for small sample size that otherwise would have met our eligibility criteria. However, small RCTs and systematic reviews based on small RCTs are more prone to bias, including publication bias and failure of randomization. Small observational studies have limited power. For factors where we have large studies already, it is very unlikely that the addition of small studies would change the estimate of effect or conclusions.

The exclusion of RCTs lasting less than 1 year may have missed some studies showing promising short-term results. These would not have been adequate to conclude that the intervention was useful for preventing cognitive decline or AD, but may have provided the impetus to conduct longer trials.

The extant research for a specific factor generally did not include more than a couple of studies using the same cognitive measure for a continuous outcome. Given the variability in outcome measures and the limited resources and time to complete the present project, it was not possible to perform quantitative meta-analyses on studies with continuous outcomes. We acknowledge, however, that quantitative estimates of effect may have been easier to interpret than qualitative syntheses.

The focus of this review was on the association between specific conditions (e.g., diabetes mellitus) and AD or cognitive decline. We did not evaluate the association between AD or cognitive decline and the treatments or interventions for the conditions. These exposures are of potential interest, but were not specified by the planning committee.

We note that this is a difficult literature to search for several reasons, including the wide range of factors assessed, the lack of well-validated search strategies for relevant observational studies, variability in categorizing studies by standard search terms, and variability in the terms used to categorize cognitive decline. For all these reasons, it is possible that relevant studies were overlooked.

Epidemiological studies of complex diseases using observational data often simultaneously evaluate the association between a range of exposures and the outcome of interest, in this case, AD or cognitive decline. These studies do not typically design their analyses specific to one or two factors of interest. We were not able to assess systematically how this approach may influence the association between the factor of interest and the outcome, but we note the issue as one to be considered when interpreting the results.

In summary, previous work on the search for clues to factors that alter the risk of AD and cognitive decline has provided a number of potential leads. These leads now need to be pursued with potentially novel approaches and increasingly rigorous scientific methods to be able to identify a real signal among the numerous factors throughout the life course that may contribute to the complex late-life disorders considered in this report.


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...