Examining the causal relationship between associated risk factors and Alzheimer's disease using Mendelian randomization

BACKGROUND: Alzheimer 's disease (AD) is highly prevalent in the global population, predicted to affect 132 million people by 2050. A recent report by the Lancet Commission has identified 12 modifiable lifestyle risk factors that are associated with an increased risk of AD, including hypertension, education, exercise and tobacco and alcohol consumption. Targeting these risk factors could potentially prevent a third of all dementia cases by allowing people to reach the end of their life dementia free. However, current evidence is primarily obtained from observational studies which does not distinguish between factors that causally contribute to increased AD risk or factors that are correlated with disease risk through reverse causation or mediating variables. METHODS: We examined the genetic relationships between 10 AD-associated risk factors and AD, as well as the genetic relationships across risk factors, using LD Score regression. Next, we used bivariate Mendelian Randomization (MR) to assess the causal relationship between each risk factor and AD. Finally, we used Multivariable MR to model the combined causal relationships of multiple risk factors while accounting for genetic relationships across risk factors. RESULTS: We observed statistically significant genetic correlations between many of the examined risk factors (ranging from -0.88 to 0.75) and between these risk factors and AD (ranging from -0.33 to 0.45). Eight of the 10 examined risk factors were significantly correlated with AD risk. Notably, educational attainment had a significant genetic correlation with all other examined risk factors. The results of the MR analysis will be presented. As an example, preliminary results suggest that although AD and hearing impairment show a positive genetic correlation, hearing impairment does not causally influence AD risk. CONCLUSION: Modifying risk factors will only reduce AD prevalence if the factors are causally linked to disease risk. We will establish patterns of causality for 10 modifiable risk factors. We show that hearing impairment does not causally influence AD risk, suggesting that reducing hearing impairment in the population will not decrease the prevalence of AD.