Subject selection and identification of optimal clinical outcome measures in dementia trials focused on vascular cognitive impairment and dementia

Abstract

AbstractBackgroundResearch into the understanding and identification of modifiable clinical and lifestyle related factors associated with increased risk of dementia has grown enormously over the past twenty years. The development of the field is now such that we have a general awareness of the main risk factors, and, in many areas, public health messaging is underway to support risk reduction. Given the overlap between the risk factors for dementia and those for other non‐communicable disease, for example, cardiovascular disease and cancer, the potential benefit of risk reduction activities is unequivocal. Despite this, the evidence base for dementia risk reduction is not strong and for several risk factors the clinical trial results have been mixed and inconclusive.MethodTo evaluate reasons for the inconsistency between the epidemiology and the clinical trial evidence we selected key modifiable risk factors that lend themselves to blinded pharmacological intervention and where trial results are available. E.g. antihypertensives, hormone replacement therapy, non‐steroidal anti‐inflammatories, statins and omega 3 supplementation. Experts in the field of dementia risk factor epidemiology, dementia risk reduction trials, and with statistical and clinical, Alzheimer’s and vascular dementia expertise were convened to critically appraise the evidence.ResultExplanations for the discrepancies arise in both the trial and epidemiology evidence. E.g.; selection of trial populations from age or population groups that do not match the groups where the association between the risk factor and dementia outcome is strongest, a lack of understanding of the most appropriate dose of intervention or goal level of risk factor for the best cognitive outcome and the short duration of trials in the context of a disease process with a decades long prodromal phase. Conversely epidemiology that suffers from confounding by indication or that fails to fully account for confounding by key factors such as socioeconomic status, or to take account of the life‐course.ConclusionTo fully understand the potential for risk reduction and to target resources accordingly will require both a new and more sophisticated evaluation of existing data and the delivery of new and better designed trials. Recommendations are made for the next steps.