Figure 3.

Summarized Mendelian randomization (MR) effect sizes of AD and sedentary behaviors. Summarized MR effect sizes of the causal relationship between AD and (a) television watching time, (b) computer use time, and (c) driving time were estimated using the following approach. The methods used in the analyses included IVW, IVW removing genetic variants with a potentially pleiotropic effect in any phenotype, MR-Egger, weighted median, MR pleiotropy residual sum and outlier (MR-PRESSO), outlier-corrected MR-PRESSO, and weighted mode methods. Odds ratios (ORs) as well as 95% confidence intervals (95% CIs) are presented on the x axis. A two-sided p-value of less than 0.05 was considered statistically significant.