A systematic review of risk factors for vascular dementia using Mendelian randomization: a call for methodologically robust studies to advance knowledge

Vascular cognitive impairment, of which vascular dementia is the most extreme end of the spectrum, is a result of many risk factors (exposures). Mendelian randomization is a genetic epidemiological approach to assess the causal relationship between a genetically predicted exposure with an outcome of interest that has gained traction in recent years in part due to the emergence of biobank-scale datasets. Here, we performed a systematic review of published Mendelian randomization studies to understand the current knowledge of potentially causal genetically predicted exposures on vascular cognitive impairment or vascular dementia risk. No results were obtained for vascular cognitive impairment. 26 of the 53 of the articles (49%) identified in our review for vascular dementia found evidence using Mendelian randomization for a causal relationship between a genetically predicted exposure and vascular dementia. Among the studies that investigated similar exposures, some consistently yielded positive causal results between certain types of genetically predicted exposures (e.g. predicted markers of inflammation or atherosclerotic vascular disease) and vascular dementia. In contrast, the presence of a causal effect between certain exposures (e.g. cholesterol levels) and vascular dementia was inconsistent depending on the study. Only 10 of the 26 articles (38%) suggesting a causal link included a completed STROBE-MR checklist, warranting caution in interpretation of results. From our systematic review, we identified evidence of potential causal effects between genetically predicted exposures, such as atrial fibrillation, major depressive disorder and gut microbiota, on vascular dementia risk, some of which are supported by multiple studies whereas others require replication. We strongly encourage future Mendelian randomization studies to follow and report the appropriate checklist (STROBE-MR) to employ triangulation and to include individuals of diverse genetic ancestries to allow for more robust exposures on vascular dementia risk for follow-up.