Gains in life expectancy and population aging are driving a sharp rise in Alzheimer's Disease and Related Dementias (ADRD), with a predicted global prevalence of 65.7 million people by 2030. In this context, it is crucial to understand what factors can modify ADRD risk and cognitive decline in older ages. Education has been identified as one potential modifier, as higher education is robustly associated with lower ADRD prevalence. However, little is known about how much of this association reflects a causal effect from education to ADRD risk and how much is driven by third factors, such as genetics. In this paper, we use family genetic data to isolate random variation in education and ADRD risk, and to study how much of the association between education and ADRD in late-life is due to a causal effect running from education to ADRD risk. Due to Mendelian inheritance, we can exploit the "genetic lottery" at conception to evaluate the direct (causal) genetic effect of the EA and AD Polygenic Indices (PGIs) on ADRD and educational outcomes. As expected, we find that the EA PGI has a causal effect on education and that the AD PGI has a large causal effect on ADRD diagnosis. However, the AD PGI does not cause educational outcomes and the EA PGI does not cause ADRD outcomes. This is evidence against a causal effect running from education to ADRD diagnosis.
Silvia Helena Barcellos, University of Southern California and NBER
Leandro Carvalho, University of Southern California
Patrick Turley, University of Southern California