Statistics from Altmetric.com
The evidence that prevention of Alzheimer's disease is an achievable goal is growing. This meta-analysis summarises the current knowledge and underlines possible gaps to be filled in the near future
The prevalence of dementia in the world is around 50 million and is increasing rapidly. Most projections estimate that by 2050 there will be more than 150 million people with dementia, with a predicted cost of two trillion dollars by 2030.1
The most common dementia is Alzheimer's disease (AD), comprising 60–80% of the total. To date no disease-modifying treatments are available for AD or for any other form of dementia. Therefore, strategies to prevent AD to cope with the huge public health problem are urgently needed. The literature investigating risk factors for AD and other dementias is an essential base on which to build preventive strategies.
The strongest indication that prevention of dementia and AD may be possible is the recently declining incidence shown in many studies, both in Europe and the USA.2 More recently, an analogous trend has been shown in Asia, with a 40% decline in the incidence of dementia and AD over 12 years in South Korea.3
The most likely factors responsible for this trend have been improvement of vascular health with the introduction of effective treatments, especially for hypertension, and in the past decades, improvement of education, generally measured by years of schooling.
In this meta-analysis, Yu et al 4 examined almost 45 000 papers. They analysed 33 145 records of observational longitudinal studies (OLS) and 11 531 for randomised clinical trials (RCTs), to select a final pool of 243 OLS and 143 RCTs, for their analysis.
This meta-analysis by Yu et al 4 identifies 21 evidence-based suggestions that can be used in life-course practices to prevent AD, including 19 rated as strong suggestions, with nine rated as level A evidence. Two-thirds of the risk factors identified in this meta-analysis are vascular risk factors, underlining how vascular health is key in maintaining good brain function, including cognition.
Meta-analyses may reduce some of the distorting effect of the primary studies (sampling errors, especially from small studies), and thus improve cumulative knowledge in the literature. This approach also helps to keep track of the relevant literature, which can be a considerable burden as shown in the present study.5 However meta-analyses do not avoid bias deriving from study design in primary observational studies (selection and measurement bias),6 which may change the direction of the primary measurement of association (the relative risk). Meta-analyses of RCTs are more likely to reflect the real world, adding studies of different quality and settings. RCTs can substantially reduce the risk of confounding because of randomisation but do not prevent specific selection bias, such as restriction at entry or attrition. In RCTs of environmental exposures, the time window and the choice of the population from which the sample is selected are definitely more critical. All sources of methodological heterogeneity should be carefully examined in a meta-analysis.7 However, the inclusion of both OLS and RCTs is a merit of the study by Yu et al, 4 because the examination of studies with different designs adds strength in dealing with the same research question.8
Publication bias is an additional problem when selecting the sources of information for a meta-analysis, by choosing larger studies in the most prominent peer-reviewed journals. Only a rigorous assessment of the quality of the meta-analysis process provides sufficient guarantee against reaching possible false conclusions. In a balance between checking the adequacy of the operations in the primary studies and the presence and relevance of inconsistencies (contaminating or extraneous studies), the authors seem to prefer the first option through a rigorous assessment of quality using recognised metrics.9 Yu et al 4 employed an extensive team (10 graders) to determine the quality of each paper included in the analysis. At the end of this process, eight risk factors (diabetes, orthostatic hypotension, hypertension in midlife, head trauma, stress, depression, midlife obesity, and coronary artery bypass grafting surgery) and three protective factors (cognitive activity, increased body mass index in later life, and education) were rated as moderate-to-high level quality of evidence.
These results strengthen the suggestion that the changing prevalence of risk factors across different populations and over time could have led to the recent decline in incidence of AD and other dementias in many regions.
A major concern about using meta-analysis is that summing up and giving a synthesis may hide possible heterogeneities of risk factors across different populations. This is a topic of investigation in several areas of neurodegenerative diseases research,10 especially in view of the interaction of environmental factors and genetic background with the different prevalence of relevant genes across different populations, ancestries and ethnicities.
In summary, the results of this meta-analysis indicate the relevance of specific risk factors for the onset of dementia and AD. Prevention is the required route for each national health system to cope with the burden of dementias globally. Even after the introduction of high-cost effective disease-modifying treatments, the role of prevention will still be decisive, given the numbers involved.
The results of this study suggest that there is a strong need to pursue more epidemiological studies to obtain new evidence to implement the correct strategies for specific intervention in each population. Meta-analysis is a complex procedure but points out clearly what we know, and where we need to direct new research, considering the urgency for an earlier diagnosis of cognitive decline. The results of the study by Yu et al 4 show that the goal is for research agencies to deliver more funding for epidemiological studies both observational and RCTs on dementia and AD in the real world. This should possibly expand epidemiological research to include the new diagnostic criteria based on biomarkers.11
Contributors None declared.
Funding The author has not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.