Despite much previous research stating that traumatic brain injury (TBI) has been confirmed as a risk factor for dementia and Alzheimer disease (AD), findings from observational studies are mixed and are of low methodological quality. This review aimed to critically evaluate the methodologies used in previous studies. Relevant literature was identified by examining reference lists for previous reviews and primary studies, and searches in MEDLINE, PubMed, Google Scholar, and Research Gate. Sixty-eight identified reports, published between 1982 and August 2018, met inclusion criteria. Common methodological weaknesses included self-reported TBI (62%); poor TBI case definition (55%); low prevalence of TBI in samples (range 0.07-28.7%); reverse causality (86% moderate to high risk of reverse causality); not controlling for important confounding factors. There were also key areas of methodological rigor including use of individual matching for cases and controls (57%); gold standard dementia and AD criteria (53%); symmetrical data collection (65%); large sample sizes (max, 2,794,752); long follow-up periods and controlling of analyses for age (82%). The quality assessment revealed methodological problems with most studies. Overall, only one study was identified as having strong methodological rigor. This critical review identified several key areas of methodological weakness and rigor and should be used as a guideline for improving future research. This can be achieved by using longitudinal prospective cohort designs, with medically confirmed and well characterized TBI sustained sufficient time before the onset of dementia, including appropriate controls and informants, and considering the impacts of known protective and risk factors.
Keywords: Alzheimer disease; dementia; risk factor; traumatic brain injury.