Dementia is an overarching term which describes a group of symptoms that result in long-term decline in cognitive functioning that is significant enough to affect daily function. It is caused by a number of different diseases, the most common of which is Alzheimer's disease. Currently, there are no definitive biomarkers for preclinical or diagnostic use, or which differentiate between underlying disease types. The purpose of this review is to highlight several important areas of research on blood-based biomarkers of dementia, with a specific focus on epigenetic biomarkers. A systematic search of the literature identified 77 studies that compared blood DNA methylation between individuals with dementia and controls and 45 studies that measured microRNA. Very few studies were identified that focused on histone modifications. There were many promising findings from studies in the field of blood-based epigenetic biomarkers of dementia, however, a lack of consistency in study design, technologies, and platforms used for the biomarker measurement, as well as statistical analysis methods, have hampered progress. To date, there are very few findings that have been independently replicated across more than one study, indicating a preponderance of false-positive findings and the field has likely been plagued by positive publication bias. Here, we highlight and discuss several of the limitations of existing studies and provide recommendations for how these could be overcome in future research. A robust framework should be followed to enable development of the most valid and reproducible biomarkers with the strongest clinical utility. Defining a series of biomarkers that may be complimentary to each other could permit a stronger multifactorial biomarker to be developed that would allow for not only accurate dementia diagnosis but preclinical detection.
Keywords: Biomarker; blood; dementia; diagnosis; epigenetics.