Abstract
Mendelian randomization (MR) is a technique that uses genetic variation to address causal questions about how modifiable exposures influence health. For some time-varying phenotypes, genetic effects may have differential importance at different periods in the lifecourse. MR studies often employ conventional instrumental variable (IV) methods designed to estimate average lifetime effects. Recently, several extensions of MR have been proposed to investigate time-varying effects, including structural mean models (SMMs). SMMs exploit IVs through g-estimation and circumvent some of the parametric assumptions required by other MR methods. In this study, we applied g-estimation of SMMs within an MR framework to estimate the period effects of adiposity measured at two life stages, childhood and adulthood, on cardiovascular disease (CVD), type 2 diabetes (T2D), and breast cancer. We found persistent period effects of higher adulthood adiposity on increased risk of CVD and T2D. Higher childhood adiposity had a protective period effect on breast cancer risk. We compared this approach with an inverse variance weighted multivariable MR method, which also uses multiple IVs to assess time-varying effects but relies on a different set of assumptions. We highlight the strengths and limitations of each approach and conclude by emphasizing the importance of underlying methodological assumptions in the application of MR to lifecourse research.