Project Summary/Abstract
Dementia is a condition affecting thinking, judgment, memory and other cognitive domains to the degree
that interferes with performing everyday activities. Alzheimer’s disease (AD) is the most common type of
dementia. In 2021, an estimated 6.2 million adults in the US aged ≥65 have dementia, and the number is
projected to be nearly 14 million by 2060. Women a have significantly higher risk of AD than men. About
two thirds of patients with dementia are women. Mild cognitive impairment (MCI), representing the early
stage of the disease, refers to a state in which the patient experiences a decline in short-term memory, or
other cognitive domain, but with no significant impairment in everyday functioning. Though dementia mostly
affects older adults, it is not a part of normal aging. Given that taking prevention at the early stage of MCI
delays the development of the disease. It is of tremendous important to detect MCI during the early pre-
symptomatic stage. Several studies observed that vascular, metabolic disorders, and inflammation are
associated with risk of MCI, AD and AD related dementia (AD/ADRD). However, research gaps remain: (1)
inconsistent findings were observed from the previous studies. (2) Large-scale population-based studies for
AD and dementia risk are limited. (3) Although it is known that the risk of MCI and dementia are associated
with changes in risk exposures, few studies tested the association between time-varying exposures and risk
of outcomes. In the application, we aim at filling these gaps by using a rigorous study design to test the
predictive values of vascular and metabolic disorders, inflammation, as well as genetic factors for the risk of
incident MCI and dementia, and then develop a novel cumulative (combined) prediction index. We have 2
specific aims. Aim 1: To examine the association of vascular, metabolic and inflammatory biomarkers
with risk of incident MCI and dementia in older women. Hypothesis: vascular, metabolic and
inflammatory biomarkers, with time-varying measures significantly predict the risk of incident MCI and
dementia in women aged 65-79, and these associations are modified by APOE gene (ε4 versus the other
alleles). Aim 2: To develop a machine learning (ML)-enabled algorithm to predict individuals who are
at high risk of incident MCI and dementia. Hypothesis: A novel and advanced risk prediction model (e.g.,
a multi-dimensional risk model using ML) that integrates predictive values of multiple risk factors and key
covariates, will enhance the degree of the prediction for the risk of incident MCI and dementia.
The proposed study addresses a significant public health challenge facing an aging population. The
proposed study is innovative, characterized by (1) focusing on sex-specific study in older women; (2)
addressing time-varying risk factors that may have significant predictive effects on the study outcomes. (3)
We will test whether there are potential modification effects of APOE on the association between the study
of exposures and outcomes. The scientifically promising aspects of the study are the fact that it is able to
test the natural development of MCI and dementia using data from an unique large-scale population-based
study, the Women’s Health Initiative Memory Study. Findings of the study will be disseminated through
scientific meetings and research articles published in high impact journals, and add new insights into the
prevention and control of MCI and dementia.
