How our emerging understanding of Alzheimer’s disease and dementia risk factors is enhanced using UK Biobank

Abstract

AbstractWe have long known that there are large individual differences in the risk of developing Alzheimer’s disease and other dementia subtypes. Indeed, the hope is that this variation gives us clues that may provide new opportunities for targeted interventions and dementia prevention. Thousands of studies in recent decades provide compelling evidence that there are dozens of modifiable and non‐modifiable dementia risk factors. This provides cause for optimism, though critics stress several limitations that need to be addressed. The thorny issue of possible confounding remains, and the causal status of many dementia risk factors is challenging or even impossible to establish in randomized controlled trials. Most prior studies are modest in size and there is large variation in the associations observed between studies that is difficult to explain. Lastly, the window of opportunity for early intervention is difficult to assess from studies of older adults whose underlying pathologies may already be well established. UK Biobank provides us with a fundamentally different type of resource to help address these limitations and accelerate the pace of our discoveries. So, what’s new? For the first time we can investigate genetic, lifestyle and environmental risk factors for dementia in half a million middle‐aged and older adults. This has been a game changing shift in the size and richness of the data available to us, which in turn has changed the scientific questions that are possible to address. For example, we established for the first time that lifestyle factors do not interact with polygenic risk, suggesting that lifestyle interventions may decrease dementia risk regardless of genetic variation (Lourida, et al., 2019, JAMA). We can now investigate single or multiple risk factors for dementia with greater confidence. For instance, recent studies of frailty and cardiometabolic multimorbidity in combination with genetic, neuroimaging, and other rich data provide practical examples (Ward, et al., 2022, JNNP; Tai, et al., 2022, Lancet Healthy Longevity). Mendelian randomization and causal machine learning are particularly promising approaches when evaluating causality. Future enhancements to UK Biobank will facilitate even more powerful insights into the natural history of dementia, raising new possibilities for precision dementia medicine.