Addressing statistical biases and evaluating sex differences in lifecourse obesity and risk of ADRD in a diverse cohort

PROJECT SUMMARY (ABSTRACT) With an increasingly aging and diversifying US population, the prevalence of Alzheimer’s Disease and related dementias (ADRD) is expected to rise. The expected increase highlights the importance to address modifiable risk factors to decrease the burden of ADRD and lead to better health outcomes. Obesity is a modifiable risk factor that continues to pose a public health concern. Mid-life obesity is considered a risk factor for ADRD, however late-life obesity appears to have either a null or protective association with ADRD. Potential explanations for these mixed findings are selective survival bias, competing risk of mortality, inverse causality, and BMI’s estimation of adiposity. These conflicting results affirm the need for a lifecourse approach to fully delineate the contribution of obesity at different ages as a risk or protective factor on ADRD and cognitive decline. This project will further evaluate the association between lifecourse obesity and ADRD and cognitive change, while addressing selective survival and competing risk biases that frequently impact findings from observational cohort studies on aging. Data will be leveraged from three ongoing well-characterized diverse cohort studies of long-term Kaiser Permanente members: the Kaiser Healthy Aging and Diverse Life Experiences Study (KHANDLE) (n=1712), the Study of Healthy Aging in African Americans (STAR) (n=764), and the LifeAfter90 Study (LA90) (n=999). These harmonized cohorts provide up to 60 years of data with repeated BMI and cognitive measures. BMI will be measured using height and weight collected from health records. Adolescent BMI will be any BMI readings between the years 10-20. Mid-life BMI will be from the ages 40-60 years, and late-life BMI will be from ages 65 years and above. Participants will be grouped into four categories based on their BMI: obese, overweight, normal, and underweight. Cognition will be measured using the Spanish English Neuropsychological Assessment Scale annually for KHANDLE and STAR and twice a year for LA90 participants. To address methodological concerns of selective survival and selection bias, Aims 1,2, and 3, will use inverse probability weighting. Aim 1 will examine different lifecourse pathways to evaluate the risk of adolescent, mid- and late-life obesity on cognitive decline. Aim 2 will address competing risk bias and evaluate the risk of mid- and late-life obesity with ADRD. Aim 3 will determine the mediating impact of lifecourse obesity on the association between gender and ADRD and cognitive decline to better understand the influence of lifecourse obesity on the well characterized gender differences in ADRD and cognitive decline. The results from this proposed study have potential public health impacts by addressing the role of a modifiable risk factor over the lifecourse on ADRD. These results will contribute to a better understanding of the epidemiology of ADRD and cognitive decline in diverse populations.