Evidence for causal links between known modifiable risk factors and dementia: A systematic review of Mendelian randomization studies

BackgroundWe aimed to systematically review the evidence for causal associations between the known modifiable risk factors and dementia based on Mendelian randomization (MR) studies. MethodFive databases were searched from inception to April 2024 investigating the association between the twelve risk factors identified in the Lancet Commission and dementia. Evaluable analyses were categorised into one of four levels (robust, probable, suggestive, insufficient) based on estimate significance level and concordance of direction of effect between main and sensitivity analyses. Results47 articles were included representing 160 separate analyses (136 unique; 104 evaluable) for ten risk factors and six dementia outcomes. There were no valid analyses for air pollution and traumatic brain injury. Of the unique and evaluable analyses over half (59.3%) were evaluated as providing insufficient evidence of causal links. There was no evidence that genetically predicted liability to hearing loss was associated with dementia and limited genetic evidence for social contact. Evidence for education, obesity, depression, alcohol consumption and physical activity was inconclusive. There was probable evidence that smoking was protective against dementia risk however this may be an artefact of survivor bias. The two risk factors with the strongest genetic evidence for links with dementia were diabetes (probable evidence) and blood pressure (probable and suggestive evidence). ConclusionGenetic evidence for eight of the risk factors examined was insufficient or inconclusive. However, the null findings should be interpreted in the light of the biases inherent to MR studies. The strongest genetic evidence supported a causal link between diabetes and dementia.