PROJECT SUMMARY/ABSTRACT
The observed decline in the incidence of Alzheimer’s Disease and Related Dementias (ADRD) in the U.S. and
Europe over the past few decades suggests that risk factor modification can prevent or delay ADRD. However,
as the exact causes of this reduction in incidence remain unclear and the number of people affected by ADRD
is large and is expected to grow as the global population ages, there is a critical need to identify which ADRD
risk reduction strategies should be promoted. Although elevated total cholesterol and low-density lipoprotein
cholesterol (LDL-c) in midlife are associated with increased risk of ADRD, it remains unclear whether midlife
statin use is an effective ADRD risk reduction strategy. This is largely due to the timing of the introduction of
statins, which were first approved for clinical use in the late 1980s; the first group of midlife statin users have
only now aged sufficiently to be at substantial risk of ADRD. Thus, the overall objective of this proposal is to
investigate whether midlife statin use reduces risk of dementia or slows cognitive decline. To do so, we
propose to use data from the longitudinal Atherosclerosis Risk in Communities (ARIC) study. ARIC participants
enrolled in midlife in the late 1980s, with the most recent study visit completed in 2018-2019. Given that the
initial focus of ARIC was on cardiovascular health, and the focus shifted to cognitive health and dementia as
the cohort aged, ARIC has robust data on indications for statin use, statin use, and late-life cognition. Thus, the
ARIC data and timing of data collection are perfectly suited to our aims, and we propose to use a propensity
score matching approach to estimate the effect of midlife statin initiation on risk of incident ADRD and cognitive
decline over two decades of follow-up. This approach addresses limitations of the existing literature. Prior
studies have largely focused on late-life statin use or have limited follow-up. These studies would not be
expected to find substantial cognitive benefits because elevated lipids in late-life are not associated with
increased risk of dementia, and studies have not demonstrated near-term cognitive benefits of statin use. To
the contrary, our focus on midlife statin use follows naturally from findings linking elevated midlife blood lipid
levels to increased risk of cognitive decline and dementia, and minimizes concerns about reverse causation.
Our propensity-score matching approach and focus on the effect of treatment in the treated also explicitly
recognizes and is designed to address confounding by indication. Randomized trials of midlife statin use on
late-life cognition are unethical and infeasible due to the required length of follow-up. As such, careful
evaluation of observational data using a causal inference framework to guide analyses – which we propose
here – will be needed to provide evidence for or against a causal effect of midlife statin use on late-life
cognitive health. Ultimately, our results will help public health professionals, clinicians, and advocates
determine whether it is reasonable to promote midlife statin use for lipid management as also maintaining brain
health and reducing risk of ADRD.