Distilling causality between physical activity traits and obesity via Mendelian randomization

Abstract

Abstract Background Whether obesity is a cause or consequence of low physical activity levels and more sedentary time has not yet been fully elucidated. Better instrumental variables and a more thorough consideration of potential confounding variables that may influence the causal inference between physical activity and obesity are needed. Methods Leveraging results from our recent genome-wide association study for leisure time moderate-to-vigorous intensity (MV) physical activity and screen time, we here disentangle the causal relationships between physical activity, sedentary behavior, education—defined by years of schooling—and body mass index (BMI), using multiple univariable and multivariable Mendelian Randomization (MR) approaches. Results Univariable MR analyses suggest bidirectional causal effects of physical activity and sedentary behavior with BMI. However, multivariable MR analyses that take years of schooling into account suggest that more MV physical activity causes a lower BMI, and a higher BMI causes more screen time, but not vice versa. In addition, more years of schooling causes higher levels of MV physical activity, less screen time, and lower BMI. Conclusions In conclusion, our results highlight the beneficial effect of education on improved health and suggest that a more physically active lifestyle leads to lower BMI, while sedentary behavior is a consequence of higher BMI.